Casting Design and Performance

  • 67 176 8
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Casting Design and Performance

Materials Park, Ohio 44073-0002 www.asminternational.org

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Copyright # 2009 by ASM InternationalW All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of the copyright owner. First printing, November 2009

Great care is taken in the compilation and production of this book, but it should be made clear that NO WARRANTIES, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE GIVEN IN CONNECTION WITH THIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM’s control, ASM assumes no liability or obligation in connection with any use of this information. No claim of any kind, whether as to products or information in this publication, and whether or not based on negligence, shall be greater in amount than the purchase price of this product or publication in respect of which damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND SOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTING FROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this book shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International. Prepared under the direction of the ASM International Technical Book Committee (2008–2009), Lichun L. Chen, Chair. ASM International staff who worked on this project include Scott Henry, Senior Manager of Product and Service Development; Steven Lampman, Technical Editor; Ann Britton, Editorial Assistant; Bonnie Sanders, Manager of Production; Madrid Tramble, Senior Production Coordinator; and Diane Whitelaw, Production Coordinator. Library of Congress Control Number: 2009935431 ISBN-13: 978-0-87170-724-6 ISBN-10: 0-87170-724-1 SAN: 204-7586

ASM InternationalW Materials Park, OH 44073-0002 www.asminternational.org Printed in the United States of America

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Contents Design Problems Involving Junctions . . . . . . . . . . . . . . . . . . 147 Design Problems Involving Distortion . . . . . . . . . . . . . . . . . 155

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Part I: Casting Design Principles and Practices Casting Design Issues and Practices . . . . . . . . . . . . . . . . . . . . . 1 Casting Design and Processes . . . . . . . . . . . . . . . . . . . . . . . . . 9 Modeling of Casting and Solidification Processing . . . . . . . . . 37 Part II: Process Design Riser Design . . . . . . . . . . . . . . . . . . Gating Design . . . . . . . . . . . . . . . . . Design for Economical Sand Molding Design for Economical Coring . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. 61 . 73 . 81 . 89

Part III: Design and Geometry Casting Design and Geometry . . . . . . . . . . . . Design Problems Involving Thin Sections. . . . Design Problems Involving Uniform Sections . Design Problems Involving Unequal Sections .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

101 121 133 139

Part IV: Casting Performance Corrosion of Cast Irons . . . . . . . . . . . . . . . . . . Corrosion of Cast Carbon and Low-Alloy Steels . Corrosion of Cast Stainless Steels . . . . . . . . . . . Fatigue and Fracture Properties of Cast Irons . . . Fatigue and Fracture Properties of Cast Steels. . . Fatigue and Fracture Properties of Aluminum Alloy Castings . . . . . . . . . . . . . . . . . . . . . . . Friction and Wear of Cast Irons . . . . . . . . . . . . Friction and Wear of Aluminum-Silicon Alloys . Failure Analysis of Castings . . . . . . . . . . . . . . . Inspection of Castings . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

163 171 175 185 199

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

209 219 227 237 247

Appendix: Classification of Casting Defects . . . . . . . . . . . . . . 251 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

iii

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Preface Component geometry is a powerful aspect of casting design in terms of both effective production and the function of a cast part. Many tools have been developed for casting design, and the examples of past designs, even those of years ago, provide an important baseline in producing effective castings. Computer modeling and simulation has greatly facilitated the design process, and the article “Modeling of Casting and Solidification Processing” provides an extensive review of the subject. In addition, the complex aspects of configuration design are detailed in a series of articles in the sections on “Process Design” and “Design and Geometry.” Several of these articles are based on the ASM publication Casting Design Handbook (1962), which has been out-of-print for many years. Nonetheless, the lessons are still relevant today, as the basic fundamentals of geometry, metallurgy, and physics remain unchanged (even within the view of new modern perspectives and the advent of more powerful analytical or numerical tools). It is noted that the distinct sections on “Process Design” and “Design and Geometry” are a somewhat artificial division of topics, because really both process and design are intertwined in complex ways, especially for castings. In a sense, the “design” is like the fulcrum that leverages these two important aspects of castings into effective products. It is hoped that this collection of articles provides a useful reference on casting design. Finally, the performance of cast products is covered in a series of articles in the last section. Ultimately, the performance of product determines its success.

S. Lampman February 2009

v

ASM International is the society for materials

engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials. ASM International, Materials Park, Ohio, USA www.asminternational.org This publication is copyright © ASM International®. All rights reserved. Publication title

Product code

Casting Design and Performance

05263G

To order products from ASM International: Online Visit www.asminternational.org/bookstore Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US) Fax 1-440-338-4634 Mail

Customer Service, ASM International 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA

Email [email protected]

American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, In Europe United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card)

www.ameritech.co.uk Neutrino Inc. In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan Telephone: 81 (0) 424 84 5550 Terms of Use. This publication is being made available in PDF format as a benefit to members and customers of ASM International. You may download and print a copy of this publication for your personal use only. Other use and distribution is prohibited without the express written permission of ASM International. No warranties, express or implied, including, without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this publication. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this publication shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement.

Casting Design and Performance Pages 1–8

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Casting Design Issues and Practices* H.W. Stoll, Northwestern University

DESIGN is the critical first step in the development of cost effective, high quality castings. Designing a successful casting requires an integrated, concurrent engineering approach. It also requires systematic and structured use of sophisticated computer-aided design software and casting analysis and simulation software. In this paper, we discuss these and other issues impacting good casting design. In particular, design decisions that drive casting performance and cost are identified and used as the basis for a proposed holistic approach to casting design. This new design philosophy and methodology is aimed at optimizing both the structural performance and producibility of the casting while minimizing design time and effort. In casting, the part being produced and the tooling used to produce the part interact in complex ways, which effect both the quality and cost of the casting. This suggests that the design of a successful casting requires an integrated approach that considers both functional and process requirements simultaneously. In the traditional setting, however, the design engineer typically first decides the geometry of a casting and then a casting engineer independently develops the mold and process. This decoupled arrangement often results in more costly castings with larger safety factors than necessary. In addition, lead times for casting development using this approach can be excessive. As manufactures seek to reduce weight and cost of products, casting has re-emerged as the manufacturing process of choice in many situations. This is because casting offers the important advantage of being able to produce highly complex functional shapes quickly and easily. Cost is reduced because numerous parts and complex construction and processing typically associated with built up structures and weldments can be replaced by a single cast part. Weight is reduced because material can be distributed to where it is needed and because sections can be thinner since load does not transfer across part interfaces (i.e., through fasteners or welds). To take advantage of these unique capabilities, castings must be properly conceived and designed from the start. This means that design

engineers and casting engineers will need to change the casting design and development process from the historical sequential, decoupled approach to a more integrated and concurrent approach. This new approach must be based on the recognition that the part geometry not only affects the load carrying functionality of the casting, but also the mold construction, mold filling and material solidification processes involved in producing the casting. These processes in turn affect cycle time, casting quality, and material properties such as yield strength, ultimate strength, and fatigue resistance. Casting geometry must therefore be determined based on both functional and processing requirements. Hence, design engineers must become more knowledgeable of the casting process, and casting engineers must have a better understanding of the functional requirements that drive the design. In essence, the new design philosophy must not only facilitate good part and process design, it must also teach it. This paper discusses the issues and practices associated with good casting design. The focus of the paper is on casting design in general, and on sand and permanent mold aluminum casting in particular. We begin by examining the casting design process from a variety of design and processing perspectives. Strategies for casting design process improvement are then proposed which provide the basis for a new casting design philosophy and methodology. Two possible implementations are presented. The first is a structured team approach that is intended as a possible means for quickly improving traditional casting design practice by integrating the casting geometry and process design. The second is a longer-term approach involving the development of casting design guidelines for the design of lightweight, high quality, and high performance structural castings.

Geometry/Material/Process Interactions Carefully planned geometry is the secret to efficient load carrying capacity and acceptable component performance. It is also the secret

*Reprinted from Proceedings from Materials Solutions Conference ’98 on Aluminum Casting Technology

for achieving high quality (i.e., uniform properties, soundness, etc.) and for avoiding costly and time consuming problems associated with pouring and solidification of the alloy. Understanding how casting geometry, the material in both its liquid and solid phases, and the casting process interact provides the insight needed to specify the best casting geometry from a function, form, and fabrication point of view. In the following, we explore several geometry/ material/process interactions that are pivotal to good casting design. Much of this discussion is based on the paper “Cost Effective Casting Design — Developing a Conceptual Framework for Designing Metal Castings” by Michael Gwyn (Ref 1). The reader is referred to this paper for a more in-depth discussion of many of the topics briefly referred to here. Fluid life is the ability of the molten alloy to fill the mold cavity, flow through thin narrow channels to form thin walls and sections, and conform to fine surface detail. In addition to temperature of the molten metal, fluid life also depends on chemical, metallurgical, and surface tension factors. Therefore, the fluid life of each alloy is different. For example, aluminum 356 is considered to have excellent fluid life whereas the fluid life of aluminum 206 is only fair to good. Fluid life determines the minimum wall thickness and maximum length of a thin section. It also determines the fineness of cosmetic detail that is possible. Hence, knowing that an alloy has limited fluid life suggests that the part should feature softer shapes (i.e., generous radii, etc.), larger lettering, finer detail in the bottom portion of the mold, coarser detail in the upper portions of the mold, more taper leading to thin sections, and so forth. Solidification Shrinkage. Shrinkage occurs in three distinct stages: liquid shrinkage, liquid-to-solid shrinkage, and solid shrinkage. Liquid shrinkage is the contraction of the liquid before solidification. Liquid-to-solid shrinkage or solidification shrinkage is the shrinkage that occurs as the liquid’s disconnected atoms and molecules form into the crystals of atoms and chemical compounds that comprise the solid metal. Solid shrinkage is the shrinkage that occurs as the solid metal casting cools to

2 / Casting Design and Performance ambient temperature. Although liquid shrinkage is important to the metal caster, it is not an important design consideration. Solidification shrinkage and solid shrinkage, on the other hand, are extremely important and must be carefully considered during casting design. Different alloys have differing amounts of liquid-to-solid shrinkage (e.g., aluminum 356 has little while aluminum 206 has moderate to large). Most importantly, there are three different types of solidification shrinkage: directional, eutectic, and equiaxed. In alloys such as malleable iron and carbon steel, which solidify directionally, solidification moves along predictable pathways determined by the casting geometry and thermal gradients in the mold. For example, solidification will typically begin at the mold wall and move perpendicularly toward the center of the part. This is called progressive solidification. Solidification will also begin in cooler regions where the mold surface area to metal volume ratio is large and travel toward the hotter regions of the casting. This is called directional solidification. The key is to configure the part geometry so that directional solidification can occur before progressive solidification shuts off the source of liquid metal supply (the riser). Without proper pathway geometry (e.g., risering and tapering), voids or pores due to isolated internal shrinkage can result. In eutectic-type solidification, the liquid metal cools and then solidifies very quickly all over. This behavior minimizes internal shrinkage and the need for risers and makes this type of alloy the most forgiving of the three. Eutectic-type materials that have very little solidification shrinkage like gray iron often require no risering at all. The key geometric concern for eutectic-type solidifying alloys like aluminum 356 which have small but appreciable solidification shrinkage is to ensure that the avenue of liquid metal supply stays open and functioning all the way to final solidification. In addition to solidifying both progressively and directionally from the mold walls, alloys that exhibit an equiaxed solidification behavior also begin to solidify throughout the liquid, forming mushy regions consisting of equiaxed islands of solid. These equiaxed islands can block the avenues of liquid metal supply making these alloys difficult to feed. To offset this tendency, regions that solidify in an equiaxed-type manner should be designed to have small thermal gradients, that is, to be as thermally neutral as possible. Therefore, thermal mass in these regions should be spread out and distributed uniformly throughout the region. This causes the shrinkage to be distributed as microscopic pores throughout the volume of the casting. Although the thought of having microscopic holes in the casting is disturbing, the effect on mechanical properties is greatly minimized by the small size, rounded shape, and uniform dispersion produced by using the proper geometry for this type of alloy. Also, a uniform dispersion of very small pores

is clearly preferable to large, irregular pores concentrated in possibly critical regions of the casting that could result from less appropriate part geometries. Solid shrinkage is often called pattern maker’s shrink because the tooling must be properly sized so that the part will shrink to the desired final size and shape upon cooling. Solid shrinkage is critical for two important reasons. First, the shrinkage must be predicted and then built into the patterns/dies and corebox dimensions. If this is not done correctly, then the tooling will need to be modified iteratively to achieve an acceptable production casting. This adds time and cost to the design cycle and introduces quality risk in the final product. Second, as the casting cools, it may not be able to shrink uniformly because some regions are stiffer than others. This can result in undesirable residual stresses and/or undesirable warpage. Creating geometry’s that make shrinkage predictable and that avoid residual stress and warping is therefore highly desirable. Slag/Dross Formation. Slag is typically composed of liquid nonmetallic compounds (usually fluxed refractories), products of alloying, and products of oxidation in air. Dross refers to nonmetallic compounds produced primarily by the molten metal reacting with air. Some molten metal alloys are much more sensitive to slag/dross formation than others. Castings made from these alloys are much more likely to contain nonmetallic inclusions. In addition to process and quality control techniques, part geometry can be used to dramatically reduce the likelihood of nonmetallic inclusions. For example, for castings made from alloys that have buoyant slag/dross, the probability of having an inclusion in a critical machined surface can be reduced by designing the part so those surfaces are in the lower portion of the mold. Similarly, rigging design (i.e., configuration of sprues, runners, and gates) can be designed to control the amount of oxidation that occurs due to turbulent flow and entrained air. Pouring Temperature. Molds used in the casting process must withstand the extremely high temperature of molten metal. Often, proper casting geometry can help make the mold robust against this thermal abuse. Also, recognizing the undesirable effects of high pouring temperature on casting quality can help. For example, high pouring temperatures can lead to poor as-cast surface finish due to metal penetration into small sand cores. Therefore, when pouring temperature is high, it is often advisable to machine rather than core small holes and other small features. Fluid Flow. Another key geometry/process interaction involves the flow of molten metal into the mold cavity. As mentioned previously, turbulent flow through gates and other channels can effect the amount of oxidation and consequent dross formation that occurs. Another consideration is the force generated by the molten metal as it flows into the mold cavity and the

turbulence of the flow in the cavity since both of these effects can displace cores and erode mold walls, especially sharp edges and high detail features. Steep thermal gradients can also arise due to fluid flow. If the flow separates to pass around cores and other features and the joins together again, weld lines, nonmetallic inclusions, and other flaws can occur due to cooling and oxidation of the flow front. In order to minimize undesirable effects of fluid flow, the casting must be poured slowly. Unfortunately, this gives the molten metal more time to oxidize and increases process cycle time. Undesirable interactions due to fluid flow effects can often be reduced or eliminated by designing the casting geometry and rigging as a system. Heat Transfer Considerations. The geometry must also be selected with an understanding of the heat transfer conditions involved. High pouring temperatures mean that large amounts of heat must be transferred into the mold. If the geometry is such that the heat cannot escape, a hot spot is likely to occur. For example, narrow peninsulas or tight corners of mold material surrounded by molten metal will get hot very quickly and as a result, solidification of the molten metal in these regions will be slower than surrounding regions. This creates the possibility of hot tears or shrinkage pulls because the hotter material will have less tensile strength and is therefor less able to resist internal forces that develop due to solidification and solid shrinkage. Voids can also form because liquid metal supply paths close off before the material in the region of the hot spot is fully solidified. Just the opposite situation occurs when sharp corners or narrow peninsulas of molten metal are surrounded by mold material. In these cases, the molten metal cools and solidifies very quickly. This is generally a desirable situation. However, if cooling is too rapid, it can cause cold cracking due to stressing of the solidified skin or thin region by solidification shrinkage occurring at a slower rate in more massive adjacent regions. Also, difficult to machine or undesirable material properties may result from to rapid cooling of some alloys. Geometry/Alloy Interactions. In most cases, it is the combination of material properties possessed by a particular alloy that determines the most desirable casting geometry. For example, gray iron has a moderately high pouring temperature, excellent fluid life, and very small, eutectic type solidification shrinkage. As a result, accept for the danger of hot spots due to its relatively high pouring temperature, gray iron is very casting friendly. It’s excellent fluid life permits fine detail and thin sections and its low solidification shrinkage provides considerable geometry latitude. Aluminum 356 has a low pouring temperature, excellent fluid life, and more eutectic type than directional type solidification. The low pouring temperature makes it an excellent candidate for precision castings. Also, its excellent

Casting Design Issues and Practices / 3 fluid life permits fine detail and thin walls everywhere. However, although still relatively small, solidification shrinkage is significant enough to warrant consideration especially with respect to risering, section size, and feeding pathways. Also of concern is the sensitivity of aluminum to dross formation. Dross can be a particular problem because the specific gravity of aluminum oxide is close to that of molten aluminum and hence buoyancy does not aid separation. Carbon steel is at the opposite end of the spectrum. Steel has a very high pouring temperature, poor fluid life, and large, directional type solidification shrinkage. This combination of material properties makes steel very casting unfriendly. As a result, careful attention to casting geometry is essential. Because of its poor fluid life and high pouring temperature, fine detail and thin sections are difficult. Most importantly, because of its large solidification shrinkage, feeding of the casting is a great concern. Risers need to be large and the geometry must be carefully designed to ensure proper feeding of the casting. Because of its unfavorable combination of properties, steel and materials like it, require softer shapes (i.e., large radii, rounded shape, large lettering, no sharp detail) compared to casting friendly materials such as gray iron.

Cost Drivers A second way to look at the design of a casting is to understand how design decisions regarding casting geometry drive total cost of the casting. By total cost, we mean the sum of all costs, both the direct and indirect, that result from the design, production, distribution, use, and salvage of the casting over its lifetime. Although all components of total cost are important, we are particularly concerned in this paper with how design decisions drive both the direct cost associated with production of the casting and the cost of designing, building, and proving out the tooling. This cost can be calculated on a per unit basis as follows, CT C0 tcycle Cost ¼ þ Cc þ V Cm þ þ Cs N Y

(Eq 1)

where: CT = total tooling cost ($) N = lifetime number of castings CC = cost of coring ($/unit) V = total casting volume (in3) Cm = alloy cost ($/in3) C0 = casting equipment and labor cost ($/hr) tcycle = total casting lead time (hr) Y = yield (useable castings/N) Cs = cost of secondary processing ($/unit) It is important to note that total tooling cost (CT) includes all cost associated with tooling including the cost of pattern and corebox construction, the cost of producing and inspecting the first article, and the cost of iteratively modifying the tooling to meet specifications. Also, material volume (V)

includes not only the volume of the casting, but also the volume of the risers, runners, and sprues used to feed the casting. Total casting cycle time (tcycle) is given by the following, tcycle ¼ tnp þ tbuild þ tcast þ tcool þ ttrim

(Eq 2)

where: tnp = non-productive time (hr) tbuild = mold build time including core placement (hr) tcast = time to pour the casting (hr) tcool = time to cool to ambient temperature (hr) ttrim = time to remove gates, risers, etc. (hr) Since many castings involve more than one core, the per unit cost of coring (CC) is calculated as, Cc ¼

 nc  X C00 t0cycle 0 V 0 Cm þ 0 YC i i¼1

(Eq 3)

where: nC = number of cores V0 = volume of core material (in3) 0 = cost of core material ($/in3) Cm 0 C0 = core making equipment and labor cost ($/hr) 0 tcycle = core making cycle time (hr) YC0 = yield (useable cores/lifetime number of cores) Similarly, secondary processing may involve more than one process such as machining, heat treating, welding, painting, and plating. In addition, processes such as machining might involve several different operations (e.g., drilling, milling, grinding, etc.). The per unit cost of secondary processing is therefore calculated as Cs ¼

 ns  X C000 t00cycle CT00 þ YS00 i i¼1

(Eq 4)

where: ns = number of secondary processes CT00 = secondary process tooling cost ($/unit) C000 = secondary process equipment and labor cost ($/hr) 00 tcycle = secondary process cycle time (hr) YS00 = secondary process yield (useable castings/N) Equations 1 through 4 are very clear as to what should be done to reduce the per unit production and design cost:     

Design to minimize tooling cost Design to minimize material cost Design to minimize process cycle time Design to maximize yield Minimize the number of cores and secondary processes

By looking at how geometry decisions effect the sources of cost in the above equations, it is possible to make geometry decisions that reduce cost. For example, both tooling cost and

production cost will be reduced by selecting the parting plane early in the design and then creating geometry that minimizes the number of undercuts and other features that must be cored. Locating riser and gate contacts at easy to access areas on the casting will reduce trimming time. Also, locating riser and gate contacts so that they don’t interfere with machining fixture targets will reduce trimming time, the cost of the fixture, and possibly machining cycle time because fixturing will be easier and more consistent. Designing so that critical dimensions do not cross the parting line will decrease build time and increase yield since the positional accuracy required between the cope and drag is reduced.

Shape Optimization Metal casting offers two unique and very desirable design advantages: (1) metal mass can be located exactly where it is needed and (2) complex, three-dimensional geometry is readily created. By properly capitalizing on these advantages, part geometry can be created that minimize both weight and cost of the part. For example, the use of continuously varying section geometry that fully utilizes the material strength while also satisfying deflection requirements is readily achievable. In addition, many parts can be consolidated into one part, thereby eliminating piece part fabrication cost, assembly cost, and all the indirect costs, interfacing information, quality risk, and manufacturing complexity associated with built-up parts and weldments. Shape optimization is the design perspective that seeks to leverage these advantages. This practice has been greatly facilitated by the development of powerful engineering workstations and solid modeling software that significantly enhances the engineer’s ability to visualize complex three-dimensional geometry and to analyze stress levels and deflections of complex three-dimensional shapes.

Rigging System Design The rigging system includes the system of sprues, runners, gates, risers, and chills that channel and control the flow of liquid metal into the mold cavity, feed the casting as it solidifies, and control the heat transfer and rate of solidification in critical regions. Rigging system design specifies the size, dimensions and location of sprues, runners, gates, risers, and chills that comprise the system. In the traditional approach, an expert casting engineer designs the rigging system, usually after the geometry of the casting has been specified. Rigging design decisions typically include selection of the following: orientation of the cast part, parting line, potential sites for chills and chill types, sprue height and location, runner types and configuration, ingate sites, choke area (smallest cross-section area present in the flow system), riser sites and configuration, and pouring rate and temperature.

4 / Casting Design and Performance

Casting Process Simulation In casting process simulation, comprehensive modeling of the intended production process is performed in order to determine the size and shape of sprues, runners, gates, and risers. A variety of simulation software for performing this type of simulation is available. In addition, methodologies have been developed to understand and predict the size and location of process related defects (microporosity, etc.). See for example references [2–6]. Using these methodologies, the rigging system design can be varied in the foundry system simulation to evaluate how defect size and location are to be controlled and/or eliminated. Using computer simulation early in the design process can greatly reduce the amount of guess work involved in specifying cost effective and functionally acceptable casting geometry. Computer based casting process simulation offers two important advantages: (1) design iterations and what-if analysis are much easier to perform and (2) the physics engine underlying the simulation software provides a consistent and predictable science base for casting design. This allows the casting geometry and rigging system to be specified and optimized as a coordinated system. Most importantly, it allows evaluation of the overall design before tools are cut and the design is irreversibly committed to hardware. When used properly, the result is a substantial reduction in design time and tooling iterations. It is extremely important to note however, that the use of casting process simulation software is not, in itself, a viable substitute for early input of experienced tooling and foundry engineers. Rather, it is a very powerful tool that helps leverage and assist the team approach.

Casting Design Improvement Strategies Casting design is an iterative process (Fig. 1). The problem of design is typically formulated in terms of functional requirements and constraints

that must be satisfied. Functional requirements relate to the functions the part must provide while constraints relate to the form (shape, size, surface finish, precision, etc.) and processing (parting line, draft, section thickness, etc.) requirements that constrain the geometry that can be selected. Based on the problem formulation, an initial design is created. This design is then evaluated and modified iteratively until an acceptable design is achieved. Typically, the redesign is guided by the design information, insight, and understanding developed in the evaluation step. To be acceptable, the design must satisfy all functional requirements and constraints. When the traditional approach to casting design is examined, we see that the iterative process characterized by Fig. 1 is essentially repeated at least two and perhaps several times as shown in Fig. 2. First the design engineer goes through the iterative design process to specify the casting geometry. This geometry is then passed on to the casting engineer who repeats the iterative design process to specify the rigging system and apply pattern maker shrink to the casting dimensions. Problems discovered during rigging system design can generate additional iterations if casting geometry changes are required. Additional iterations to the casting geometry and rigging system design may also be required during tool fabrication and preparation for production of the first article. Finally, iterative changes to the tooling and perhaps the casting and rigging system geometry may be necessary to tweak the design to meet production requirements. Excessive design iterations can adversely impact the casting design in two important ways. First, design iterations significantly increase design cost and time. Second, design iterations, especially those performed late in the process, can lead to suboptimal design. The result is a casting that falls short of cost, weight, and performance targets. Such designs give casting a bad reputation and are a disappointment to all concerned. Several strategies for improving the traditional casting design process are possible based on the design perspectives discussed above. These are summarized as follows: 1. Design the casting geometry and casting process as a coordinated system by integrating shape optimization and rigging system design into one concurrent process. Consider geometry, material, and process interactions

Fig. 1

Iterative model of the design process

Fig. 2

Traditional casting design process

2.

3.

4. 5.

and design related cost drivers from the beginning as part of the process. Develop a thorough understanding of all customer needs including downstream processing constraints before beginning the design. Focus on creating an acceptable initial design. By spending the time up-front to create the best possible initial design, a large number of lengthy analyze-redesign iterations are avoided. The evaluation phase should confirm the design rather than create it. Use casting process simulation and other modern computer-aided analysis and inspection methods to quickly optimize the design. Develop a consistent, well defined science base for casting design in the form of casting design guidelines and structured methodologies.

The goal of these strategies is to shorten the design cycle and help ensure that the best possible casting design is created. In the sections that follow, we propose some possible approaches for implementing these strategies.

Structured Team Approach Strategies 1 through 3, and to some extent, strategy 4 can be implemented very easily and quickly by adopting a structured team approach. By a structured team approach, we mean that the casting design is performed using a multi-disciplinary team and a structured design methodology. The goal of the team approach is to have all required product and process knowledge available when the key early design decisions are being made. In a structured design methodology, the overall problem of design is broken down into a series of sequential, easier to perform steps that proceed from the general to the specific. Excessive iteration and long design times are avoided by performing each step in a thorough and disciplined manner. In general, each step in the process can be further subdivided into steps to create a hierarchy of structured methodologies. To illustrate the structured team approach, we propose the simple methodology shown schematically in Fig. 3. This approach recognizes that not all members of the team can be available for designing the casting on a continuous basis. Team meetings are therefore scheduled at which all salient aspects of the design are reviewed and discussed. All members of

Casting Design Issues and Practices / 5 a casting engineer, a tooling design engineer, and perhaps one or more specialists who are familiar with casting process simulation software, finite element analysis, fracture mechanics, non-destructive evaluation (NDE) techniques, and so forth. In addition, it is essential that the end customer, the foundry that is to make the casting, and others concerned with secondary processing that is to be farmed out be properly represented on the team. This not only helps ensure that all customer and processing needs are appropriately considered, it also makes it possible to rapidly negotiate changes to the design specification when necessary, and to quickly assess cost consequences of design decisions. Meeting 1: Clarify the Design Problem. Clarifying the problem consists of developing a general understanding of the cost, performance, and manufacturing goals and constraints of the design. A typical agenda for this meeting might include the following:

Fig. 3

Simple structured team approach

the team must be present at these meetings. The purpose of the meetings is to establish design direction, make key design decisions that require input and consensus from all team members, make sure that all process constraints and requirements are being properly considered, and resolve conflicts and impediments to the proposed design. The outcome of each meeting is a set of action steps to be implemented by individual team members. In this way, all team members are kept informed and participate in the design decision making process. At the same time, the actual detail work of creating the design is delegated to specific team members according to the skills and knowledge required. In the following, we briefly discuss each step of the methodology and imagine how a casting would be designed using this methodology. Step 1: Form Team. This is the pivotal first step in the methodology. Unless the arrangement is formalized in some way, it often is difficult to get effective collaboration between the design and casting engineer early in the casting design process, especially before the casting geometry is defined. By being formally assigned to a team, each individual team member takes personal responsibility for the design from the beginning. This fosters and facilitates the kind of collaborative attitude that is essential for good casting design. All stakeholders who have an interest in the casting should be represented on the team. A typical team might include a design engineer,

1. Review product background. 2. Customer requirements and design objectives. 3. Expected annual production volumes and target costs. 4. Geometry concepts and alternatives. 5. Material and processing options. 6. Foundry and secondary processing locations. 7. Potential geometry/material/process interactions. 8. Develop a preliminary configuration design. 9. Make assignments to team members to create the initial design. As mentioned above, it is essential that all team members be present at each team meeting. For example, although the analyst may not be actively involved with the design before the casting geometry is fully defined, it is extremely important that he or she participate in the early design decisions that lead to the proposed geometry. In this way, the analyst knows all the needs of the design problem and is familiar with the reasoning behind the particular geometry that will eventually be analyzed and optimized. Step 2: Create the Initial Design. The initial design establishes the detail layout of the casting geometry. It includes the configuration and parametric design of the part together with the rigging design required to fill the mold cavity and feed the solidifying casting. Configuration design involves the determination of what features such as walls, holes, ribs, etc. will be present and how these features will be connected to provide the desired form, fit, function, and manufacturability (e.g., parting line, coring, and draft for low tooling cost, short cycle time, and minimal trimming and secondary processing). Parametric design involves the determination of dimensions, tolerances, and exact material specifications needed to meet durability, stiffness, and/or natural frequency targets. As discussed previously, rigging design involves the location and sizing of the sprues, runners, gates, and risers.

As a general approach, a preliminary configuration and rigging design might be proposed in Meeting 1 by the team as a whole. Using this as a starting point, the design engineer and casting engineer would work together to develop the details of the configuration design, seeking input and consensus from various team members as necessary. Once the initial casting and rigging configuration has been firmed up, a preliminary parametric design would be performed. The goal of this task is to quickly determine section dimensions, secondary processing requirements, and material property requirements using simple strength of materials methods or, if necessary, a rough finite element analysis. Once the approximate parametric design is complete, the overall design is evaluated and modified to minimize cost. This is easy and straightforward to perform with a minimum of analysis and iteration because the casting geometry, secondary processing, and rigging system have been conceived and developed as a coordinated system with input from all team members. Hence, all of the information needed to make quality design decisions is readily available. Meeting 2: Refine and Approve the Initial Design. The goal of this meeting is to react as a team to the initial design and to make any adjustments or modifications deemed necessary by general consensus of the team. This is the time when all design and processing issues should be discussed and resolved. If there are significant impediments to the design as proposed, these should be resolved before proceeding to Step 3. A typical agenda for this meeting might include the following: 1. Review the initial design. 2. Identify impediments, potential undesirable interactions and performance and processing concerns. 3. Discuss all design-related costs to ensure that the best casting geometry from a total cost standpoint has been identified. 4. Make assignments to team members to work out solutions to various impediments and schedule a follow on meeting, or 5. Approve the proposed initial design and make assignments to team members to refine and optimize the design. Step 3: Refine and Optimize the Design. Once the team is confident that the initial casting geometry and rigging design is the best solution possible, the effort required to optimize details of the casting geometry and rigging system by computer analysis can be justified. The goal of this step is therefore to computer model the design and iteratively improve it until all aspects have been appropriately optimized. The amount of effort expended on this step will depend on how important it is to optimize the casting. For example, if weight and/or material cost is critical, extensive effort to minimize the amount of material used can be justified. Similarly, if safety is an important issue, comprehensive analysis to

6 / Casting Design and Performance ensure acceptable fatigue life and reliable detection of flaws can be justified. The key to this step is to start with a casting geometry that is close to the optimum. This will minimize the time, analysis effort, and number of iterations required to converge to the optimum design. Meeting 3: Approve the Final Design. The result of Step 3 will be a fully specified casting design including the detailed casting geometry, parting plane, draft, coring, riser locations and sizes, and plumbing system design. In addition, the finished part design including machining, heat treating, and so forth will be fully specified. The purpose of Meeting 3 is to formally review the finished design as a team and approve the design for release to manufacturing. When the structured team approach is performed properly, the final design will almost always be approved. However, if the team decides that the design is not ready to be released, then appropriate action plans for correcting design deficiencies must be developed and implemented. One or more follow-on meetings may then be required before the design is released. Although painful at times, not releasing the part until it is ready helps insure a minimum number of tooling changes and tweaks and, in the long run, is the most cost effective policy. By strictly adhering to this policy, the casting design should proceed quickly and smoothly to first article and production with little or no modification. When this is the case, the team knows that it has done its job well.

facilitate design for damage tolerance, design for manufacture, and design for inspection. 2. Capture and reduce to practice the latest research results quantifying the relationships between casting processes, microstructure characteristics, cyclic (fatigue) properties, and component function for cast aluminum and magnesium alloys. 3. Facilitate the development and implementation of design strategies that seek to leverage part family concepts (i.e., group technology) and to standardize casting geometry, rigging, and tooling features. 4. Help the design engineer make better casting design decisions when input from experienced casting engineers is not readily available. This goal recognizes that many companies may not have deep in-house casting expertise. As shown in Fig. 4, the guidelines would be organized in a manner similar to the structured team approach discussed previously. The guidelines associated with each step in the process are briefly discussed as follows. Step 1: Clarify the Problem. The purpose of these guidelines would be to help the casting design team to properly understand the problem of design by considering all customer needs

including downstream tooling and processing needs. These guidelines might consist of a stepby-step structured methodology for gathering and evaluating needs together with one or more check lists formulated to ensure that all design requirements and constraints are considered. They would also include details regarding design strategy and standardization objectives. Ideally, the design process would not continue until all items have been checked off and appropriately evaluated. Step 2: Create the Initial Design. This step includes the iterative development of the configuration design, parametric design, and rigging design. Guidelines for configuration design would include high level design for light metal casting guidelines and recommended practices. Such guidelines will recognize the important metallurgical characteristics of the foundry alloy and mechanical engineering structural characteristics that drive the geometry of good casting design. These will include the geometry, material, and process interactions discussed earlier as well as best practices for communicating design information to the foundry and design for guidelines pertaining to downstream considerations such as casting inspection, heat treating, and machining. Guidelines for parametric design would include best dimensioning and tolerancing

Casting Design Guidelines Casting design guidelines make needed alloy and process information readily available to the design and casting engineer in a timely fshion and in a form that is easy to understand and apply. Such guidelines typically consist of design rules suggested best practices, structured design methods, check lists look-up tables, graphs and charts, and computerized design tools that can be used by the casting design team to perform each step of the design process. In this section, we present a vision for the development of comprehensive casting design guidelines. The goal of the envisioned guidelines is to facilitate the design of high performance castings that are also easily manufactured in large production quantities for a reasonable cost. Specific development objectives include the following. 1. Provide the sophisticated design information needed to design light metal cast structural component systems for use in safety critical automotive applications. Such castings must be as light as possible and also manufacturable in large production quantities for a reasonable cost. This implies that every section must be optimized to use an absolute minimum amount of material while satisfying durability, stiffness, natural frequency, and other structural and performance targets. The design guidelines must therefore

Fig. 4

Structure of guidelines for light metal casting design

Casting Design Issues and Practices / 7 practices, material properties including those relevant to the geometry, material, and process interactions discussed earlier as well as structural properties needed for proper consideration of damage tolerance, etc. The guidelines would follow existing best practices such as those given in standard texts on automotive component design and machine design (Ref 7 and 8). The major difference would be in the way material properties such as finite life fatigue strength would be estimated. Rather than using the traditional casting factor, the design guidelines would provide charts and property estimation procedures based on recent and developing research results (Fig. 5 and Ref 9 and 10). Rigging design would include the approximate configuration and parametric design of the risers, gates, runners, chills, insulation, etc. to feed the casting, prevent porosity formation, and minimize cycle time. As discussed previously, rigging design plays an extremely important role in the development of a successful casting design. As such, it must be considered early in the design process to achieve the goals outlined above. In addition, many of the computer simulations to be applied in the computer analysis step (see Fig. 4) require that the rigging system be modeled along with the part. The inclusion of as a part of the casting design represents a major departure from traditional casting design practice. Hence, the development of comprehensive guidelines that both teach and ensure good practice is essential. One approach that might be considered is use of computer-aided rigging design tools such as that described in Ref (11). Once an approximate rigging system design has been developed, it becomes possible to consider the casting geometry and rigging system as a coordinated system. This represents an important opportunity to improve the overall initial design quickly with a minimum of analysis and iteration. Development of guidelines for this step might include extensive use of a computer automated rapid prototyping (CARP) environment such as that shown schematically in Fig. 6.

This graph illustrates how material properties might be expressed as a function of section geometry. Developing such graphs will obviously require a focused research effort.

Computer Analysis. This step is necessary to meet the stringent weight, performance, and cost goals described above. It would be performed using computer based analysis and redesign methods to facilitate rapid convergence to the optimum coordinated casting geometry and rigging system design. As discussed in Ref 12 the computer analysis might begin with a casting process evaluation to estimate local material properties and to determine if and where flaws (porosity, etc.) will form. This would be followed by a casting inspectability evaluation to determine the smallest flaw that can be reliably detected in critical using modern NDE inspection techniques. Results of these evaluations would then be used in a damage tolerance evaluation to predict component durability and fatigue characteristics. Using component weight as the objective function and performance, cost, and quality targets as the constraints, these evaluations would be incorporated in an iterative optimization procedure that would modify component and rigging system geometry until an acceptable minimum weight design is achieved. It is expected that the guidelines for this step would consist of a procedure for using standard, commercially available finite element and casting process analysis software together with special computer programs and math based simulation techniques to perform the optimization see Ref 12 and 14.

Summary A broad range of casting design issues and practices has been reviewed. Based on this, a new casting design philosophy and methodology is proposed. The structured team approach is intended as a possible means for quickly improving traditional casting design practice by including early consideration of downstream tooling and processing constraints early in the design process. The development of casting design guidelines represents a longer-term approach for reducing design cycle time and

Fig. 5

Fig. 6

Computer automated rapid prototyping environment (CARP) Source: Ref 13

improving the quality and performance of lightweight structural castings. The methodology used for both approaches is to comprehensively understand the problem of design, quickly define an initial casting geometry and rigging system design that is close to target requirements, and then refine the total design iteratively as a coordinated system using computer based analysis tools. Both approaches are predicted on the underlying assumption that, given appropriate design and processing information and guidance at each step, the design will converge quickly to the best possible result. It is important to note that development of design guidelines such as those suggested in this paper require a substantial research and development undertaking. It is also important to point out that the proposed methodologies represent starting points for casting design process improvement. It is expected that the effectiveness and usefulness of these methods will increase as they are used and improved over time by casting design professionals and the foundry industry.

REFERENCES 1. M.A. Gwyn, “Cost Effective Casting Design,” presented at the AFS 101 Casting Congress, Preprint No. 97–147, American Foundrymen’s Society, Inc., Des Plaines, IL, 1997 2. E. Niyama, et.al., “A Method of Shrinkage Prediction and Its Application to Steel Casting Practice,” 49th International Foundry Congress, 1982 3. K. Tynelius, J. Major, and D. Apelian, “A Parametric Study of Microporosity in the A356 Casting Alloy System,” AFS Transactions, 1993, p 401–413 4. H. Huang, and J. Berry, “Evaluation Criteria Functions to Minimize Microporosity Formation in Long-Freezing Range Alloys,’ AFS Transactions, 1993, p 669–675 5. N. Tsumagari, et.al., “Construction and Application of Solidification Maps for A356 and D357 Aluminum Alloys,” AFS Transactions, 1993, p 335–341 6. J. Huang, “Study of Criteria Function for Porosity Prediction in A356 Castings,” Masters Thesis, Mechanical Engineering, Northwestern University, Dec. 1995 7. Fatigue Design Handbook, 2nd Ed., SAE International. Warrendale, PA, 1988 8. J. Shigley, and C. Mischke, Mechanical Engineering Design, 5th Ed., McGraw-Hill, 1989 9. A.J. Hinkle, J.R. Brockenbrough, and J.T. Burg, “Microstructural Material Models for Fatigue Design of Castings,” SAE Paper 960161, SAE International, Warrendale, PA, 1996 10. J.M. Boileau, J.W. Zindel, and J.E. Allison, “The Effect of Solidification Time on the Mechanical Properties in a Cast A356-T6

8 / Casting Design and Performance Aluminum Alloy,” SAE Paper 970019, SAE International. Warrendale, PA, 1997 11. N. Nanda, et.al., “Feature-Based Design of Gates and Risers in a Casting,” p 75–84, Knowledge-Based Applications in Materials Science and Engineering, Ed. by J.K. McDowell and K.J. Meltsner, The

Minerals, Metals, and Materials Society, 1994 12. J. Conley, B. Moran, and J. Gray, “A New Paradigm for the Design of Safety Critical Castings,” SAE Paper 980455, SAE International, Warrendale, PA, 1998 13. J. Conley, Private communication.

14. N. Palle, S. Singh, S. Mahadeva, R.J. Yang, and J.A. Dantzig, 1997, “An Optimization Based Methodology for Casting Process Design,” presented at the McNU ’97 Joint ASME, ASCE, and SES Summer Meeting, Northwestern University, June 29–July 2, 1997

Casting Design and Performance Pages 9–36

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Casting Design and Processes SHAPE CASTING, as a metal forming process, offers considerable flexibility and wide scope in producing parts. Almost all metals and alloys can be cast with few restrictions on part weight or size, and casting is capable of producing highly reliable, cost-effective components ranging from low-volume, single-part prototype production runs to economies of scale for millions of parts. In terms of component design, metal casting is very flexible in terms of configuration design. Casting permits the formation of streamlined, intricate, integral parts of strength and rigidity that are not obtainable by other methods of fabrication. For example, the flexibility of metal casting, particularly sand molding, may permit the use of difficult design techniques, such as undercuts and curved or reflex contours that are not possible with other high-production processes. Tapered sections with thickened areas for bosses and generous fillets are routine. The inherent design freedom of metal casting allows the designer to combine what would otherwise be several parts of a fabrication into a single piece. This is significant when exact alignment must be held, as in highspeed machinery, machine tool parts, or engine end plates and housings that carry shafts. Nonetheless, there are a number of challenges in the design of castings. To begin, component geometry and properties are closely interrelated in casting design. Most castings involve complex configurations with sections of varying thicknesses, which influences solidification rates and thus properties within various sections of a casting. For example, concave sections, or reentrant angles, solidify more slowly than fins or protrusions, thus affecting the resultant local structure and the properties. In other words, property variations occur within a casting due to differences in cooling rates. These variations are predictable and need to be taken into account during component design. Designers also need to be aware that different casting alloys have different levels of “castability,” meaning that development of reliable casting geometries and properties may be more difficult in some alloys than others. A suitable casting geometry for ductile iron may not be the same as that of cast aluminum bronze depending on key casting parameters such as pouring temperature, liquid-metal fluid life; solidification shrinkage, and formation of inclusions, slag or dross, in the melt. Design engineers thus

need to know how suitable casting geometries vary for the wide variety of casting alloys. Moreover, designers need to know how differences in suitable casting geometries can be utilized to achieve more optimum design configurations in terms of strength and rigidity. Traditionally, for sand-casting analysis, the use of geometric methods has been known as the section modulus approach. The technique uses the well-known Chvorinov’s rule, which is commonly used to compare the solidification times of simple casting shapes. Although originally developed for the solidification of pure metals and alloys solidifying over a very narrow temperature interval, the concept is more broadly applicable and states that the total solidification time of a casting (or casting section) is proportional to the square of the volume-to-area ratio of the casting (or casting section). This rule can be stated mathematically as follows: 

Vc Ac

tf ¼ K 

2 (Eq 1)

where tf is the total solidification time for the casting or casting section, Vc is the volume of the casting or casting section, Ac is the surface area of the casting or casting section, and K is a constant for a given metal and mold combination. For simple shapes, the modulus in Eq 1 can be calculated from the ratio of volume and surface area involved in cooling. However, for complex shapes discretized in a three-dimensional grid, the continuous distribution of modulus can be determined using the concept of distance from the mold. The modulus at each point in the casting is determined by the relation: M¼

2 N P

(Eq 2)

1=di

i¼1

The concept of the section modulus approach has been extended to other casting processes. The modulus approach is an approximate analysis scheme that uses geometry-based considerations to provide valuable insights into the solidification times and, therefore, the propensity for defect formation during solidification.

The next stage is to design the rigging system for the casting, which includes the design of the gate, risers, downsprue, and so forth. This has been based on the rules of thumb of foundry experts and empirical charts. Once the rigging design is established, the solidification behavior can be evaluated. Here, heat, mass, and momentum transfer determine the cooling history of the casting. The factors ultimately influence microstructure evolution and the development of stresses in the casting This article provides a general introduction on casting processes and design. The next chapter describes computer modeling and simulation, which continues to be an important tool in solidification modeling and casting design. The remaining chapters describe in more detail the role of design in meeting specific process or component requirements. Some of these chapters reflect more traditional or conventional solutions from the rules of thumb developed by foundries over years past. This retrospective view of casting design solutions may provide perspective in this age of computer–based engineering and perhaps an appreciation of the continuity in the basic principles of casting as both a science and a processing art.

Casting Methods As one of the oldest manufacturing methods, casting involves pouring molten metal into a mold cavity that is configured to the shapes and dimensions of the finished form. The methods of shape casting can be divided into several broad categories, as illustrated in Fig. 1. The main categories are:  Expendable molds with permanent patterns  Expendable molds with expendable patterns  Metal or permanent mold processes

In the case of an expendable mold, made with bonded sand or other loose granular mold material, the patterns may be permanent as is typical in sand casting or expendable as in lost foam and investment casting. When patterns are permanent, the mold must be separable into two or more parts in order to permit withdrawal of the permanent pattern (Fig. 2). The tapered ends of the pattern permit it to be removed from the sand mold without restriction. Cores are separate

10 / Casting Design and Performance

Fig. 1

Chart of shape casting process

shapes that are placed in the mold to provide castings with contours, cavities, and passages that are not practical or obtainable with molds. Permanent molds must be separable into two or more parts in order to permit withdrawal of the raw casting from the mold or die. With expendable patterns, the limitations of two or more separable parts of the mold is not necessary, the molding media must only surround the expendable pattern and maintain its shape during molten metal pouring and solidification. After the shape has solidified, the sand or other molding media is shaken off and out of part. Casting parts using expendable mold processes with either permanent patterns or expendable patterns is a very versatile molding method that provides tremendous freedom of design in terms of size, shape, and product quality. Table 1 summarizes the differences in the steps of casting a part between the permanent-pattern versus the expendable-pattern methods. In terms of capability, each shape casting process has its niche, depending on the metal or metals being cast, productivity; component size, dimensional tolerances, and configuration details. The choice of pattern materials can affect dimensional tolerances. Tables 2 and 3 briefly compare some of the typical capabilities of shape casting processes. The ratings and data are necessarily typical and approximate, as many material and design factors can influence the economical feasibility of a casting process.

The tables only provide a brief overview only for general approximate comparison. Ferrous metals, steel, gray and ductile iron, account for nearly 75% of all metals cast. On the nonferrous side, high-pressure die casting (HPDC) is the dominant process, largely because it readily accommodates scrap-based secondary alloys, and among metal-mold processes it enjoys the highest productivity because of short cycle times. Success depends, to a large extent, on the skill of the mold designer in understanding the very strong influences of the metal alloy and mold material. Even with the best design, solidification on the mold walls closes passage of melt at a certain distance, and this leads to a minimum allowable section thickness. There are large differences in the fluidity of alloys, and therefore minimum allowable section thickness depends on the alloy being cast. Lower minimum thickness can be allowed with zinc, aluminum, and cast iron than with steel. Streamlined flow is helpful; thus sharp corners are avoided. Thicker sections solidify last and must be fed adequately.

General Principles of Casting Design Regardless of the casting method, good casting designs generally follow the same basic guidelines. The basic objective of good casting

design is to concentrate liquid-solid contraction to the last portion(s) of the casting to solidify. Almost all molten alloys, with some important exceptions, undergo shrinkage during solidification. One exception is cast iron, both gray and ductile, because graphite expands during solidification. This expansion of graphite often, but not always, can compensate for the shrinkage of the iron. During alloy solidification, small nuclei of solid grains or dendrites form and grow as the temperature in the casting falls. As these grains grow, a mush of liquid and solid occurs within the freezing range of the alloy. Because each small dendrite forms a site of shrinkage, the shrinkage allows metal to flow between the dendrites to the location where solidification and shrinkage is taking place. As the dendrites grow, however; the paths for molten-metal flow become smaller, thus restricting complete filling of the cavity. Thus, an overarching principle of good casting design is that casting sections should freeze progressively, so that contraction results in a sound part. There are several ways to achieve the goal of progressive solidification and liquid-solid contraction. One basic technique of metalcasters is to design reservoirs of molten metal risers that feed liquid metals to the last portion of the casting to solidify. The risers supply liquid metal to feed shrinkage that occurs during solidification, so that casting sections freeze

Casting Design and Processes / 11 Table 1 Differences in casting process steps between the permanent pattern and expendable pattern methods of the expendable molding process Expendable mold process Steps in casting a part

Permanent patterns

Design the part and Pick the casting process to be used for the part Design of tooling to make the part Make the tooling Make the patterns for the part Make the part the first time Assemble the pattern sections, negative or positive shape of the part, with its gating system into an expendable mold Cast (create/make) the part

Simultaneous engineering with the product designer is a key to making a quality part. Once designer and caster have reached consensus on the casting process, the design can be finalized to meet functional requirements of both the final part and the chosen foundry process. Design tooling to make the negative Design tooling to make the positive shape shape of the part of the part Permanent pattern tooling to make Expendable pattern tooling to make the the negative shape of the part positive shape of the part Negative shape of the part Positive shape of the part Not applicable If necessary, assemble the positive-shaped patterns into the part to be cast. Although pattern sections may be either a negative or positive shape, they must be assembled into a final expendable mold assembly ready for casting. Introduce molten metal to the part in the selected casting process

and slag/dross forming tendency. However, systems for feeding and filling have evolved largely from the experience and capabilities of individual foundries. New tools and computer simulation have lead to the design of improved filling systems, but filling system designs for castings is not always systematic. Key objectives and principles that should be addressed for gating and risering include (Ref 1 and 2):

 Design runners and gates, if two or more 

 Establish nonturbulent metal flow.  Systematically fill the mold cavity with

metal of minimally degraded quality.

 In conjunction with the selection of an





 Major components of a sand mold (a) pattern assembly for cope and drag sections. (b) Cross section of mold with core

Fig. 2

progressively. The need for risers depends on the alloy being cast and the geometry of the mold cavity. For example, cast iron often needs very little in the way of risers, because graphite expansion during solidification may compensate for the shrinkage of the iron. Adequate filling depends on the configuration of the mold assembly and alloy castability factors such as fluidity, freezing range, and shrinkage during solidification. Thin sections, for example, may require risers, so that all portions of the mold cavity are properly filled with molten metal. Carefully planned geometry also can offset alloy limitations in terms of fluid life, solidification shrinkage, pouring temperature



 







Expendable patterns

appropriate pouring temperature, provide conditions for mold filling consistent with misrun avoidance. Establish thermal gradients within the cavity to promote directional solidification and to enhance riser effectiveness. Design riser size and geometry, and locate risers and riser inlets to minimize the ratio of gross weight to net weight. Minimize to the extent possible the vertical distance the metal must travel from the lowest position of metal entry to the base of the sprue. Taper the sprue or use a sprue geometry other than cylindrical to minimize vortexing and aspiration. Keep the sprue continuously filled during pouring. Avoid abrupt changes in the direction of metal flow; gate and runner passages should be streamlined for minimum induced turbulence at angles or points of divergence in the system. Provide contoured transitions in gate, runner, and infeed cross sections at points of cross-sectional area changes. Employ multiple gates to improve thermal distribution and to reduce metal velocity at entry points. Avoid molten metal impingement on mold surfaces or cores by appropriate gate location.







   

sprues are used, to prevent the turbulence associated with the collision of flow patterns. Design risers to be of sufficient size and effectiveness to compensate for volumetric shrinkage. Riser position, shape, and filling from the gating system relative to the casting cavity are interrelated and critical considerations. In general, risers should be placed to achieve the maximum pressure differential and, when possible, should be open to the mold surface. Blind or enclosed risers must be adequately vented. Observe the principles of directional solidification. The use of chills, riser insulation, and casting design changes may be required. The effects of inadequate gating and riser design can in some cases be corrected only by complete redesign. Provide runner overruns, dross traps, or insystem filtration to avoid the impact of degraded metal on casting quality. Locate the runners in the drag and locate the gates in the cope for horizontal mold orientation. This rule is subject to intelligent variation by the uniqueness of each part. Place the riser cavities in the gate path for maximum effectiveness whenever possible. Never place filters (if used) between riser and cavity. Design the gates so that metal entry occurs near the lowest surface of the casting cavity. Geometrically contour the runners to maintain uniform fluid pressure throughout. Formulas applicable to all gravity casting methods have been developed for this purpose.

General Guidelines of Geometry in Casting Design. In terms of general geometry, good casting designs generally follow some basic guidelines:  Limit drastic changes of thickness to mini-

mize stresses and aid in the casting process.

12 / Casting Design and Performance Table 2 Shape casting processes ratings chart Capabilities Mold method

Applicable metals

Expendable mold with permanent patterns Bonded sand processes Green sand CO2 sand Cold box Hot box Shell Slurry processes Plaster Ceramic Rammed graphite Nonbonded sand processes Vacuum molding Magnetic molding Expendable mold with expendable patterns Lost foam Investment Replicast Centrifugal casting Metal mold Sand mold Permanent mold and semipermanent mold Gravity Static top pour

Tilt pour Low-pressure permanent mold (LPPM) Standard LPPM Vacuum riserless casting/pressure riserless casting Counterpressure casting/pressure-counterpressure casting Countergravity casting High-pressure die casting (HPDC) Conventional HPDC Hot chamber Cold chamber Vacuum die casting Squeeze casting Indirect Direct Semisolid processing Thixocasting (billet) Rheocasting (slurry) Thixomolding (granular magnesium)

Productivity(a)

Disposable cores

Casting size

Dimensional control

All All All All All

5 2 2 2 3

Yes Yes Yes Yes Yes

Unlimited Unlimited Unlimited Unlimited Small to medium

Nonferrous All Reactive

1 1 1

Yes Yes Yes

Unlimited Unlimited Unlimited

Outstanding Outstanding ...

Aluminum Ferrous

1 1

NA NA

Unlimited Unlimited

... ...

Aluminum All All

3 2 1

NA NA NA

Medium to large Small to medium Medium to large

... Outstanding ...

Nonferrous All

3 2

Yes Yes

Small to medium Small to medium

... ...

Nonferrous and metal-matrix composites (MMCs) Nonferrous

3

Yes

Medium to large

...

3

Yes

Medium to large

...

Light metals Aluminum Aluminum All

3 3 3 3

Yes Yes Yes ...

Medium Medium Medium Small

... ... ... ...

Zn and Mg Nonferrous Aluminum

5 5 4

No No

Small Small to medium

Outstanding Outstanding

Al and MMCs Nonferrous and MMCs

4 1

No No

Small to medium Small to medium

... ...

Aluminum Nonferrous Magnesium

4 5 5

No No No

Small to medium Small to medium Small

Outstanding Outstanding Outstanding

... ... ... ... ...

NA, not applicable. (a) Productivity rankings: 1 = low productivity; 5 = high productivity

 Taper sections as liberally as possible

Table 3 Comparison of several casting methods Approximate and depending on the metal



toward risers to help control directional solidification. Do not interpose thin sections between thick sections and access to risers. Avoid extensive horizontal, flat surfaces. Avoid isolated thick sections that create hot spots that are difficult to feed. Critical dimensions should not cross parting lines in molds or injection or core dies.

Parameter

Green and casting

Permanent mold cast

Die casting

CO2-core casting

Investment casting

Relative cost in quantity Relative cost for small number Permissible weight of casting

Low Lowest

Low High

Lowest Highest

Medium-high Medium-high

Highest Medium

 

Up to approx. 900kg (1ton)

45kg (100lb)

40kg (85lb)

g(oz)–45kg (100lb)



Thinnest section castable, mm (in.) Typical dimensional tolerance, mm (in.) (not including parting lines) Relative surface finish

2.5(1/10)

3.0(1/8)

0.8(1/32)

Shell: g(oz)–115kg (250lb) CO2 0.22kg (½ lb)–Mg (tons) 2.5(1/10)

1.5(1/16)

0.3 (0.012)

0.75 (0.03)

0.25 (0.01)

0.25 (0.01)

0.25 (0.01)

Fair to good

Good

Best

Very good

Good

Good

Very good

Shell: good CO2: fair Good

Fair

Fair to good

Fair

Good

Good

Best

Best

Poor

Poorest

Fair

Fair

Designers also should have a good knowledge of the casting characteristics of available alloys, as this influences design approaches to work around limitations of alloy castability. See Chapter 8 “Casting Design and Geometry” in this book. In general, designing for progressive solidification requires tapering walls so that they freeze from one end to the other (Fig. 3a, 3b), thus avoiding situations where two heavy sections are separated by a thin section. This is a poor design because metal must feed one heavy

Relative mechanical properties Relative ease of casting complex design Relative ease of changing design in production

Casting Design and Processes / 13 section through the thin section; when the thin section freezes before the heavy section, the flow path will be cut off and shrinkage may form in the heavy section. Tapered configurations also facilitate removal of patterns from sand molds (Fig. 3c). Junctions also concentrate heat, leading to areas in the casting where heat is retained. These areas solidify more slowly than others, thus having a coarser structure and different properties from other sections, and solidify after the rest of the casting has solidified, so that shrinkage cannot be fed. Minimizing the concentration of heat in junctions, therefore, aids in improving casting properties. Examples of this are shown in Fig. 4. Concave corners concentrate heat, so they freeze later and more slowly than straight sections, while convex corners lose heat faster, and freeze sooner and more quickly than straight sections. In designing a casting, the designer should use properties from test bars that have solidified at the actual cooling rate in that section of the casting. This cooling rate can be determined by instrumenting a casting and measuring the cooling rate in various sections or by simulating its solidification using a commercial solidification simulation program. The hollow spaces in castings are formed by cores, refractory shapes placed in the molds and

around which the casting freezes. These cores are later removed from the casting, usually by thermal or mechanical means. However, each core requires tooling to form it and time to place it in the mold. Casting designs that minimize cores are preferable to minimize costs. Some examples are given in Fig. 5. The designer also must provide surfaces for the attachment of gates and risers. The casting must solidify toward the riser in order to be sound, and gates must be located so that the mold fills from the bottom to the top so that oxide films that form are swept to the top surface of the casting or into risers where they will not affect casting properties. Gate and riser locations must be accessible for easy removal to minimize processing costs. The position of gate and riser contacts may also add costs if they are placed where subsequent machining will be required to remove gate stubs or riser pads in the finished component.

Expendable-Mold Casting Expendable mold processes are widely used in casting. The choice of molding method depends on several factors such as part size and shape, quantity, tooling, and the molten metal being poured into mold. The basic characteristics of foundry molds must address four basic requirements:  Be formable into the desired shape around

some type of pattern

 Be able to hold that shape while molten

metal is introduced

 Be able to break down and become strippable

after the metal solidifies The first requirement is that the mold material must flow or be pliable enough to encapsulate the surface around the shape of the pattern. Patterns can be either expendable; as in lost-foam casting and investment or lost-wax casting; or permanent. Permanent patterns are made from wood, steel, hard plastic, where numerous molds can be duplicated before significant wear would alter the shape of the part being cast. Once the mold material is tightly bonded around the pattern, then the mold must be stripped and removed from that pattern, while not damaging either the permanent pattern material or the mold shape established. The mold shape, which is typically the reverse or mirror image of the part being cast, must then be strong enough to be handled and manipulated, so that it can be combined with other mold shapes such as copes, drags, cores, gatings, etc., that produce the empty cavity of the negative shape of the part being cast. Typically the plumbing system—the sprues, runners, gates—that allows the molten metal to be introduced into the cast part shape are also made from these same molding materials, and becomes part of the mold assembly. After complete assembly of this mold package, the molten metal is introduced in some manner into this mold assembly. The second requirement is the molding material must be able to withstand the erosion action of the molten metal as it flows through the gating system and the part geometry without deteriorating or losing its shape. Any portion of the molding media that may break away could be

 Be able to maintain that shape while the

molten metal solidifies

Redesign with tapered sections. Tapered walls provided progressive solidification in (a) elbow and (b) valve fitting. (c) Taper allows easier removal of pattern from mold.

Fig. 3

Redesign of casting to minimize heat concentration. (a) Design has numerous hot spots (X junctions) that will cause the casting to distort. (b) Improved design using Y junctions.

Fig. 4

Redesign of castings to eliminate cores. (a) Casting redesigned to eliminate outside cores. (b) Simplification of a base plate design to eliminate a core. (c) Redesign of a bracket to eliminate a core and to decrease stress problems.

Fig. 5

14 / Casting Design and Performance washed up in the molten metal, and end up as a defect within the solidified part being cast. The molten metal could also infiltrate the molding media at the mold/metal interface producing a nonhomogenous scab of solidified metal and molding media on the part being cast. The third requirement for molding media is to maintain shape while the molten metal solidifies. This is a critical requirement to ensure that the cast part in the mold accurately reproduces the outline of the pattern used to make the mold. During solidification, the thermal conductivity of the molding medium is also important in meeting the material properties of the metal being cast. Foundries must constantly balance the correct thermal conductivity of the molding media. If heat conducts too quickly from the mold cavity, freezing of the metal before filling the entire mold cavity could result. Conversely, slow heat conduction can result in slow cycle times or improper microstructure of the solidifying metal, which can lead to incorrect material properties for the part being cast. To assist solidification, the mold may include regions with metal chill plates, which are added to assist in quicker solidification in certain sections of cast parts. Most metals can be heat treated after casting to assist in meeting the material properties of any particular part. Of course this adds cost to the product, so any compromises in the heat conductance of the molding media that can be implemented would save this cost. Finally, the fourth basic attribute of the molding media is being able to be removed from the cast part. The expendable mold must be broken away and stripped from the solidified part. More cast parts, especially those with intricate internal passages, are rendered unusable during this shake out step of the casting process than in any other place in the foundry. This is why some of the expendable pattern practices, for example lost-foam and investment casting, may have an advantage over the permanent pattern methods of the expendable molding processes. Mold Materials. Sand is the most prevalent molding medium, and various types of binders are used to bond the sand into useable molds. Besides sand, other materials can meet the aforementioned requirements for molding media. Slurry molding uses a plaster and/or ceramic material to form the shape around either a permanent pattern or an expendable pattern. Graphite molding materials rammed around a pattern have also been utilized. Sand is an abundant raw material, although the sand for foundry use means something much more than a layman’s concept of sand. In casting, sand has numerous properties that need to be carefully controlled and monitored. In terms of sand molding, the basic requirements of the aggregate material include:  Dimensional and thermal stability at eleva-

ted temperatures

 Suitable particle size and shape

 Chemically unreactive with molten metals  Not readily wetted by molten metals  Freedom from volatiles that produce gas

upon heating

 Economical availability  Consistent cleanliness, composition, and pH  Compatibility with binder systems

Many minerals possess some of these features, but few have them all. The most prevalent mold base is silica sand, due to its wide availability and lower cost relative to other types of sands such as zircon, olivine, and chromite. Manufactured ceramics (such as mullite pellets) are also used as mold-base aggregates.

Expendable Molds Produced with Permanent Patterns Various methods are used to fabricate expendable molds from permanent patterns. The methods include:  Molding of sand with a clay-water binder 

 

   

and mechanical compaction green-sand molding Molding of sand with chemically thermal activated binders made from inorganic compounds such as silicates or organic resins Shell molding of sand with a thin resinbonded shell that is baked No-bond vacuum molding of sand, where molding media is held together with a vacuum source Plaster-mold casting Ceramic-mold casting Rammed graphite molding Magnetic or no-bond molding of ferrous shot

Green sand molding is perhaps the most popular molding process. Green sand mold; which are usually never green in color, but black; use natural or synthetic clays with other additives to bond the sand grains together. Green sand molding may be done either with flasks or without flasks. The resin-bonded mold processes called cold-box, hot-box, warm-box, and shell molding were originally based on organic resin binders, although lately even inorganic binders are mixed with the sand and then hardened or cured by chemical or thermal reactions to fixate the shapes. Often these expendable resin bonded processes are used to produce cores that are placed in permanent molds. This combination of expendable cores with permanent molds is referred to as semipermanent molding. This is used extensively in the nonferrous metal casting area. Virtually all sand processes are suitable for casting both ferrous and nonferrous metals, but green sand excels over other sand processes because it is also the most productive. Green sand is by far the most often employed casting process, simply because it is the dominant process for casting ferrous metals. No other sand process can boast hundreds of molds per hour, with the potential for numerous cavities

per mold. Percentage of casting tonnage by molding process is All All All All

Sand Casting Permanent & Semi Permanent Mold Die Casting “other”

75% >5 19 r are also a function of the design parameters being considered, and four equations must be used. For R > t + r and R < T + r, the casting modulus of the junction is: Volume ¼ Surface area 2 2 T þ t þ TR þ tr þ tT þ 0:2146 r2  0:2146 R2 t þ 3T þ 1:5707 r þ 1:5707 R (Eq 9)

For R > T + r and R > t + r, the casting modulus of the junction is: Volume ¼ Surface area 2 2 T þ r þ RT  tT þ tR þ 0:2146 r2 þ 0:2146 R2 t þ T þ 2:8584 R  0:4292 r (Eq 10)

For R < T + r and R < t + r, the casting modulus of the junction is: Volume ¼ Surface area 2 2 T þ t þ Tt þ rt þ rT þ 0:2146 r2  0:2146 R2 3T þ 3t þ 2:8584 r  0:4292 R (Eq 11)

For R > T + r and R < t + r, the casting modulus of the junction is: Volume ¼ Surface area T 2 þ t2 þ Rt þ Tr þ tT þ 0:2146 r2  0:2146 R2 3t þ T þ 1:5707 R þ 1:5707 r (Eq 12)

Although the sequencing curves for R > r shown in Fig. 29(a) and Fig. 29(b) seem complex, the complexity is easily understood by remembering that both the internal and external radii, r and R respectively, are held constant on any given graph. The first unusual feature of graphs of this type is the forbidden zone, that is, the dark area in the lower left corner of the graph. In this area, L-sections using the radii combinations given in the figure caption are impossible, and the internal radius punctures the external radius (Fig. 30a). Another unusual feature of these curves is the crossover point, which in Fig. 29(a) occurs at t = T = 0.25. At this point, the casting moduli of all the components of the L-section (t, T, and J) are equal. This occurs because a uniformly thick plate results from these particular design conditions, as shown in Fig. 30(b). The crossover point always occurs when t = T = (R  r) for L-sections in which R> r. The effects of increasing the radii can be seen by comparing Fig. 29(a) with 29(d).

28 / Casting Design and Performance

Fig. 26

Solidification sequence curves for X-section with three legs equal. (a) R = 0. (b) R = 2. (c) Composite of design curves for various fillet radii. (d) One segment (R = 1) of the composite design curve shown in (c)

External Radius Less Than Internal Radius. If the external radius R of an L-section is designed to be less than the internal radius r, completely different sequencing curves result. The simplified geometry of L-sections in which r > R also considerably simplifies the casting modulus of the resulting J- or L-junctions. Because of the geometry, only the following casting modulus equation is required:

Volume ¼ Surface area T 2 þ t2 þ rT þ rt þ Tt þ 0:2146 r2  0:2146 R2 3T þ 3t þ 3:5707 r  0:4292 R (Eq 13)

Curves for L-sections with two such R < r combinations are shown in Fig. 31. No crossover point occurs, nor does the curve contain

a forbidden zone. The reasons for this can be determined by noting the features of such L-section designs (Fig. 28).

Feeding To compensate for shrinkage during solidification, molten metal is fed into the mold cavity

Casting Design and Processes / 29

Fig. 27 sequences.

Model of L-section for R > r. Compare with Fig. 28. See also Fig. 29 for solidification

Two L-section designs that can occur when R < r. (a) L-section has a higher modulus than the plates when plate thicknesses are nearly equal. (b) Plate thicknesses are very unequal; the modulus of the L-section is less than that of the thick plate.

Fig. 28

from separate reservoirs containing molten metals. These reservoirs are called feeders or risers; although the term riser can refer to a specific type of feeder that is connected to a slot gate such that the metal rises in the reservoir while rising in the mold cavity (Ref 4). The first question to answer is whether feeders or risers are even needed? If possible, the use of feeders should be reduced as much as possible. Just in terms of economics, feeders reduce yield; defined the metals weight going into the foundry divided by the weight of good castings produced. Many small and medium-sized castings do not need to be fed, especially with thinner wall castings. Judicious uses of chills or cooling fins may also reduce the need for feeders. If feeders are required, the next question is what is the required size of the feeder? Guidelines on the sizing and design of feeders are detailed by Campbell (Ref 4) in terms of six basic criteria: 1. Freezing time of the feeder must be as long as or longer than the freezing time of the casting. Freezing time depends on both the casting geometry and the thermal conductivity of the molten metal. 2. Feeder must contain a sufficient amount of metal, for example, volume criterion 3. Freezing time in the junction between the feeder and casting should be greater than the freezing time of both the feeder and casting. If not, under feeding and shrinkage porosity may occur.

Fig. 29

Solidification sequence curves for the L-section shown in Fig. 27 with R > r. (a) r = 0.25, R = 0.5. (b) r = 0.5, R = 1. (c) r = 1.0, R = 2.0. (d) r = 2.0, R = 4.0

4. The occurrence of a proper feeding path depends on geometric factors that should not be overlooked. 5. Sufficient differential pressure must occur for the correct amount and direction of flow. 6. Sufficient pressure is needed at all points in the region of feeding so that cavities do not form and grow These guidelines are detailed by Campbell in Ref 4 and 5. Feeding is also discussed in other chapters of this volume along with following overviews on feeding of flat plates and cylindrical sections. Computer modeling and simulation, as described in Chapter 3, have assisted in the design of improved filling systems. Solidification of Flat Plates. The L-, X-, and T-sections discussed to this point consist of a combination of plates and the junctions themselves. Any discussion of casting design must also consider the solidification of simple flat plates as well as their more complex junctions.

The solidification of metals does not occur as a flat moving front, but rather as a series of protuberances ranging in shape from cone points to very complex, branched shapes called dendrites. The complexity of these shapes is a function of the speed of growth of the front, the cooling rates involved, and the composition of the alloy. For the purposes of design, the nature of these shapes concerns the ability of metal to develop a feeding path through uniform flat plates or cylinders. Because these protuberances in most alloys of interest at commonly found growth rates and thermal gradients usually manifest themselves as dendrite-type shapes, they can provide a significant barrier to the continued flow of metal. This phenomenon can lead to shrinkage defects in castings because the feed metal supply cannot reach into the solidifying front at which the liquid-solid contraction is causing a feed metal demand. In particular, this effect influences casting design in the area of feeding distances.

30 / Casting Design and Performance

L-section designs that fall in the forbidden zones of Fig. 29. (a) Internal radius r punctures external radius R. (b) Crossover point occurs when moduli of all the components are equal—in this case at t = T = 0.25, as shown in Fig. 29(a).

Fig. 30

Example of a design using a solid cylinder (a) and the solidification sequence curves for such a design (b)

Fig. 33

Feeding path design considerations. (a) Circular flat plate with a single riser. (b) Addition of wedge-shaped ribs to ensure proper solidification. (c) Branched ribs to overcome feeding problems at the circumference of the plate

Fig. 32

Solidification sequence curves for L-sections with r > R. (a) r = 0.5, R = 0.25. (b) r = 1, R = 0.5. Compare with Fig. 29.

Fig. 31

In a simple plate having a constant thickness T, a set of solidification fronts will move from each of its surfaces until they meet at its centerline. However, as these fronts move together, their dendritic character begins to cause a blockage of the metal flow from a feeding path, which will result in some degree of shrinkage. This phenomenon has been quantified, and it has been determined that the ability of a riser to feed through a plate is very limited (Ref 6). Often, sound casting production can be achieved only for a few thicknesses along the plate even when chills are used. This is very important because long flat plates will require extensive risering for this reason alone. This limited ability to feed can be overcome by simply reducing the tendency for the front closure to occur as a straight line. This is accomplished by using a plate whose cross section is not uniform in thickness but is continuously changing in thickness, such as the wedge casting discussed earlier. Another way of accomplishing the same objective without changing the geometry of the plate is to alter the thermal shape of the casting through the use of an insulating or chilling

material to change the rate of heat extraction. Still another way of accomplishing feeding through an extensive constant plate thickness is to create a series of feeding paths close to each other. Consider a circular flat plate with a single riser at the center of the plate, as shown in Fig. 32(a). If the feeding distance creates a problem, multiple feeding paths could be considered as a solution. The design of feeding paths involves the development of shape geometries that permit solidification to be directed to a portion of the casting that can be easily fed by a riser. For this example, it is necessary to create a feeding path to the center of the plate where it can be fed by the riser. If the addition of radial ribs presents no problem, wedge-shaped ribs could be added, with the taper increasing as the rib nears the center of the plate (Fig. 32b). Because feeding distance is a problem only for continuous-dimensioned flat plates if the distance between these ribs does not exceed the feeding distance for the plate thickness, the ribs will ensure feeding of the plates to soundness. Because the ribs themselves are tapered to create a path to the riser, they should also be sound or free of shrink. If feeding distance again becomes a problem at the outer circumference of the circular plate, the ribs could be branched to ensure that the feeding distance for the flat plate is maintained (Fig. 32c). If the geometry of the rib causes difficulty with the design of the resulting casting, the same idea could be applied through the use of insulating or chilling materials to create the same effect thermally without the addition of the added metal of the ribs. However, because the use of molding materials

Casting Design and Processes / 31 to create the needed feeding path would no doubt increase the cost and complexity of the solution, it should be avoided if possible. Solidification of Cylinders. A restriction of feeding distances similar to that in plates also occurs in cylinders. Once again, this can be overcome through a change in the thermal shape by developing a cone shape either geometrically or through the use of chilling or insulating materials. Cylinders, however, do create a new condition. The new condition of concern is the junctions produced by the intersection of a cylinder and other cylinders or plates. If the cylinders are hollow, it is not necessary to consider them separately, because these conditions can be visualized as L- or T- sections merely revolved about an axis. However, a separate geometric case develops when the cylinder is solid. Figure 33 shows an example of such a case as well as the resulting design curve. On this curve, the solidification sequence is either D-J-T or T-J-D, and no case exists in which the junction J is the last to freeze. For such a junction, either the cylinder acts as a chill for the junction or the plate tends to act as a chill dependent on the cylinder diameter D and the plate thickness T. Thus, such intersections are straightforward and should cause little problem for designers or casting engineers if the feeding distance relationships and other thermal shape factors for promoting directional design are followed.

Heat Transfer and Transport Phenomena Heat is transferred in any of three modes: conduction, convection, and radiation. One must consider which of these to include in the simulation. Conduction, except for some

Model for an L-section (a) cast using two different molding materials A and B. (b) Solidification sequence model for L-section shown in (a) in which the insulating material is in the A position; r = 0.25, R = 0.5, diffusivity of mold material A = 2, diffusivity of mold material B = 1. (c) Baseline curve for comparison with Fig. 34(b), 35, 36, and 37 in which r = 0.25, R = 0.5, A = 1, and B = 1 where insulating materials are the same for positions A and B. See also Fig. 35 to 37.

approximation methods, is always included. Convection and radiation effects are often approximated as a single convective boundary condition—convection between the casting and mold, and between the mold and surroundings. The thermal resistance between the mold and casting greatly affects results for metallic mold casting simulations, such as high-pressure die-casting. It is also true that one must use an accurate convective heat transfer coefficient between castings and metallic chills. Although literature values are useful, the generalized heat transfer methods must be carefully applied to specific casting designs. Knowledge about fluid flow during the filling of casting is important because it affects heat transfer both during and after filling. Heat transfer by forced convection predominates during the filling stages. The loss of liquid metal superheat in the casting cavity of the mold after the filling transients have died out occurs by buoyancy-generated convection currents. Buoyancygenerated natural convective heat and mass transfer occur before the phase change. Buoyancy-generated convection currents tend to redistribute the melt temperature and composition until solidification begins. In the case of alloy melts, the difference in atomic weight of the constituent metals causes an additional convection pattern. There are several computational techniques to simulate fluid flow during mold filling. Mold-filling simulation, therefore, can be indispensable if a high-quality casting-solidification analysis is desired. The information obtained could help avoid problems of cold shuts, where the melt solidifies before filling a void, or where a molten front of liquid comes in contact with a solidified metal. In general, the transport of heat, mass, and momentum during solidification processing controls such varied phenomena as solute macrosegregation,

Fig. 34

Another solidification curve for the L-section shown in Fig. 18(a) in which a chill has been placed in the A position. In this case, r = 0.25, R = 0.5, A = 2, and B = 1. See also Fig. 34(b) and (c), 20, and 21.

Fig. 35

Another solidification curve for the L-section shown in Fig. 34(a) in which the insulating material has been placed in the B position. r = 0.25, R = 0.5, A = 1, and B = 0.85. See also Fig. 34(b) and (c), 35, and 37.

Fig. 36

32 / Casting Design and Performance distribution of voids and porosity, shrinkage effects, and overall solidification time. These parameters, in turn, result in a variation of the mechanical, thermophysical, and electrical properties of the solidified product. For example, macrosegregation primarily occurs due to the movement of solid phase by convection during solidification. Governing equations for transport of mass, momentum, energy and solute are introduced by Danzig (Ref 7) for a moving solidification front. Differences in density, enthalpy and composition between the liquid and solid are very important for the understanding of solidification processes. A modeling process called scaling is used to understand how material properties and

Another solidification curve for the L-section shown in Fig. 34(a) in which a chill has been placed in the B position. r = 0.25, R = 0.5, A = 1, and B = 2. Note that there is no crossover point in this figure. See also Fig. 34(b) and (c), 35, and 36.

geometry can be analyzed in the context of the hierarchal time and length scales associated with the effects of transport phenomena on important features of cast microstructures (Ref 7).

General Geometric Factors on Heat Transfer In addition to the transport phenomena during mold filling and solidification, heat transfer is also influenced by geometric factors (Ref 5). For example, the designs of plates, cylinders, and T-, X-, and L-sections, as previously described, involve components with mold that extract heat uniformly. However, there are many instances in casting manufacture in which this is not the case. Such situations occur during the normal course of the casting process and in some cases are done intentionally to change the solidification sequence with the use of a core or a chill. Use of Cores and Chills. Cores are often used in casting, but their effect on the solidification sequence of a section depends on the relative heat extractive capability or heat diffusivity differences between the mold and core components. If the heat extractive character of the core components is greater than that of the mold components, the core will tend to act as a chill.

Fig. 37

Solidification sequence curve for T-section shown in Fig. 38 with R = 0.5, A = 1, B = 1, and C = 0.85. Note that in one case (area marked tatb-T-J), ta and tb solidify at the same time. See also Fig. 40 and 41

Model of a T-section cast using three different mold materials. See Fig. 39 to 41 for possible solidification sequences.

Fig. 38

This occurs because heat is removed more efficiently from the portions of the casting in contact with the chill than from those in contact with mold components. More commonly, the diffusivity of the core when made of resin bonded silica sand is lower than that of the green sand mold components, and the core acts as an insulator with respect to the mold components. It must be remembered that the relative heat extractive capacity of the various mold components determines the resulting solidification sequence. The results shown here are independent of the actual mold and cores being considered because the solidification sequence is important in casting design, not the time required for an individual portion of the shape to solidify. Only in a few rare cases, as in continuous casting and slush casting, the manufacture of hollow castings through decantation of unsolidified liquid, is the actual time for solidification important. In these cases, the actual position of the solidification front at a particular time is important. This is not true for most casting design situations, in which only the sequence of events is important because the order affects the formation of a feeding path. However, the actual time required for an individual component to solidify is usually relatively unimportant if its position in the overall solidification sequence can be determined. Because of this comparative unimportance of the actual solidification time of a given component, only the relative ratios of the thermal diffusivities of the two materials involved, rather than the actual diffusivity of each, need to be considered in the design curves. Thus, to understand the major effects on the design of T-, X-, and L-sections, only two cases must be considered—one in which one mold component is insulating with respect to the other, and a second in which one component has a chilling effect compared to the other. By convention,

Fig. 39

Solidification sequence curve for T-section shown in Fig. 38 with R = 0.5, A = 1, B = 0.85, and C = 1. See also Fig. 39 and 41

Fig. 40

Solidification sequence curve for T-section shown in Fig. 38 with R = 0.5, A = 0.85, B = 1, and C = 0.85. See also Fig. 39 and 40

Fig. 41

Casting Design and Processes / 33 the diffusivity ratio is always defined as the ratio of a material divided by the diffusivity of the mold. For example, in a green sand mold, the ratio is the diffusivity of the nongreen sand component and the diffusivity of green sand, and for a permanent mold application, the ratio equals the diffusivity of a given component/diffusivity of die material. If one mold component is insulating with respect to the mold material, the diffusivity ratio would be less than one. A common condition of this type exists for the case in which a resin bonded solid sand core is placed in a green sand mold. The diffusivity ratio would be approximately 0.85 for this case. However, as stated above, the actual materials involved are relatively unimportant provided the same diffusivity ratio exists; the principal concern is the ratio of diffusivities rather than their actual values. L-Sections. Before discussing the design curves for L-sections, it is important to consider the geometric consequences of such placements. An L-section is shown in Fig. 34(a). In such a section, there are two possible positions, A and B, that might be occupied by the insulating material. The basic difference between these two possible placements has to do with the amount of casting surface area exposed to each. For placements of insulation material in the inner corner, the A position, less area at the Jjunction is exposed to the insulation material per unit volume than is exposed along the plates making up the section. This is important because one might expect the placement of insulation material along a junction to slow the solidification of the section. However, when the resulting L-section design curve (Fig. 34b) for this situation is observed, the exact opposite seems to be the case; that is, there are far fewer designs in which the junction solidifies last. This occurs because in position A the material extracts heat differently from the junction than the material along its position B outer area. Because a greater percentage of surface area is contacted by a material in position A along the plates making up the L-section than from the junction itself, the plates are insulated to a greater degree than the junction. This means that insulation material placed in the A position tends to act as a chill from the point of view of the junction with respect to the plates making up the section. This occurs because a greater effect of insulation is experienced by the plates due to the increase in surface area exposed to the insulation material than by the junction. This results in a slowing of solidification of the plates more than its effect on the junction. Note that there are fewer designs in which J solidifies last in Fig. 34(b) than in Fig. 34(c), in which the mold insulating materials are the same for the A and B positions. A similar apparent reversal of the expected effect when a chilling material is placed in the A position of an L-section can be seen in Fig. 35 for a relative diffusivity ratio of 2.0. Once

again, the chill has a greater effect on the plates making up the L-section than on the junction itself. Such chill materials, when placed in the A position, tend to act as insulation materials from the point of view of the L-section. That is, because the plates are chilled to a greater degree than the junction due to the exposed surface area to the chill, chill placement in the A position creates more designs in which the junction tends to be the last portion to solidify, as can be seen in Fig. 35. If material having an insulation character is placed in the B position of an L-section, it contacts a greater amount of surface area to volume in the junction than is contacted by the plates making up the section; the result of this is the reverse of that discussed for materials placed in the A position. This placement is illustrated in Fig. 36. Because the insulation effect is greater in this case on the junction due to the greater surface area contact, it creates more designs in which the L-junction solidifies after the plates. The effect of a chill placed in the B position can be seen in Fig. 37. It is clear from Fig. 34 that, because the chill affects the junction to a greater degree than the plates, the resulting design curves contain more designs in which the junction is the first to solidify. In addition, a radius in the A position placements tends to increase the area in contact with materials in the A position. However, external radii tend to lessen the surface area in contact with a material in the B position. T-Sections. In the case of a T-section, there are three potential positions in which insulating or chilling materials can be placed, namely, positions A, B, and C, as can be seen in Fig. 38. The T-shape causes several interesting effects. The first of these is similar to that observed for the L-sections. Materials placed in either the A or B position have a greater influence on the plates making up the section than on the T-section itself. Similarly, material in the C position has a greater influence on the solidification of the T-junction than on the plates making up the section. These effects are similar to those discussed for the L-sections, but because of the particular geometry involved in a T-section, several others also occur. For example, when the material in the C position is different from that in the A and B positions, the effects on the plates ta and tb are indistinguishable (Fig. 39). If the materials in the A and B positions are different, there will be a difference in the solidification of the two halves of the cross plates ta and tb. Figure 40 illustrates this effect, in which each of the plates ta and tb are shown to provide an independent section on the design curve. This effect is a result of the imbalance of the heat extraction across the intersecting plate labeled T in Fig. 38, and it is evident in the use of chilling as well as insulating materials (Fig. 41). In addition, if A and C or B and C are similar materials, an imbalance still exists across the intersecting plate of the T-section, resulting in

a design curve in which the plates ta and tb separate into individual effects (Fig. 40 and 41). X-Sections. In the case of X-sections, a geometry-related effect exists. When an insulating or chilling material is placed in a single position, it influences two of the plates but not the other two plates. It also has an effect similar to placements on L-sections in their inner corner in that the material has a greater influence on the plates making up the section than on the X-section itself. This means that the effects are similar to those seen in L-sections discussed previously; chills tend to act as insulating materials from the point of view of the X-section. In addition, diagonal placements across the section cause a matched effect on all of the plates making up the section. Therefore, when mold materials of different diffusivities are employed, the results can be predicted, but geometric effects can often produce results that may not be obvious. Thus, to predict the results of a given set of conditions, one must consider the effects very carefully. However, the amount of contact area of each portion of a section in direct contact with a chill or insulator plays a significant role in the eventual solidification behavior.

Structure and Properties Properties of a material depend on its microstructure, which is influenced by the processing and the type of manufacturing method, for example, casting, forging, and machining. In many cases, the structure and properties of a material are assumed to be roughly uniform and isotropic, with some small inherent random scatter in properties. This is not always the case. For example, directionally solidified casting are purposely produced with a distinct orientation of structure for the directionally dependent properties of gas-turbine blades. Fiber-reinforced composites are another example of a nonuniform, anisotropic material. In castings, differences in cooling rate for different sections in a casting section, or from casting to casting result in variations in structure and thus properties. Discontinuities such as porosity or inclusions may form depending on variations in processing conditions. These variations, nonetheless, do have some underlying root causes that can be understood in predicting variations from point to point within a casting. For example, cast iron provides a dramatic example of the differences in properties caused by differences in structure resulting from cooling rate differences. When molten iron solidifies as a solution of carbon and silicon in iron, the carbon can take different forms, depending on its composition and solidification rate and the way the metal has been treated during melting. Chill cast or white iron has properties completely different from gray iron of the same composition, caused solely by the accelerated cooling rate in the chilled iron. White iron freezes so quickly that the carbon combines

34 / Casting Design and Performance with the iron to solidify as cementite (Fe3C), which is hard and brittle. In contrast, gray iron is produced by a slower solidification process, such that the carbon solidifies in the form of soft graphite flakes. Ductile iron has properties significantly different from gray iron; in this case the carbon solidifies as tiny spheres in a steel matrix. Since the spheres of graphite are less effective as stress raisers, compared to the flakes of carbon in gray iron, ductile iron has significant ductility, whereas gray iron does not, even though they both may have nearly identical compositions. In this case, the difference in structure, which produces the property difference, is caused not by cooling rate, but by chemical treatment of the melt, which alters the undercooling at the beginning of solidification, which in turn affects the way the carbon solidifies. It is important here to realize that in castings, similar compositions, in similar geometries, can have very different properties, depending on the way the castings are made. It is critical that the casting designer specify the true design needs. Poor casting design can interfere with the ability of the foundry to use the best techniques to produce reliable castings. Defect-free castings can be produced at a price. The multitude of process variables such as molding mediums, binder, gating and feeding, melting and ladle practice, pouring technique, and heat treatment that must be controlled to produce sound castings requires a thorough understanding of the processes used and strict engineering supervision. The price of a casting often reflects the cost of such efforts. The attention to detail necessary to make good castings can reduce total cost of a manufactured part by significantly reducing machining, repair, and fit problems later in the assembly. Good casting design requires an understanding of how discontinuities affect casting performance and how casting design influences the tendency for discontinuities to occur during the casting process. Proper specification, as opposed to over specification or under specification, is the key to the successful application. Over specification causes needless expenses, and poor casting design can interfere with the ability of the foundry to use the best techniques to produce reliable castings. This can be avoided by understanding the effect of discontinuities on casting performance and the effect of casting design on the tendency for discontinuities to form during the casting process. Important types of casting discontinuities include porosity, inclusions, oxide films, second phases, hot tears, metal penetration, and surface defects. See Appendix: Classification of Casting Defects for illustration of casting defects. The existence of casting discontinuities does not, in and of itself, indicate that casting performance in service will be affected. Equally important are the size, location, and distribution of these discontinuities. Those discontinuities that are small and located near the center of the casting have little effect, while those

located at or near the surface of the casting are usually damaging. Clustered discontinuities and those that occur in a regular array have a greater effect on properties than those that are isolated and randomly distributed. In specifying acceptable levels of discontinuities, such as microporosity and inclusion sizes and distribution, the designer should determine the critical flaw size that will deleteriously affect performance in service. This permits the foundry to design a casting practice that will eliminate such discontinuities at minimum cost.

Discontinuities and Defects Discontinuities are defined as interruptions in the physical continuity of a casting, and the engineering community tends to make specific distinctions between the terms discontinuity and defect. By definition, a defect is a condition that must be removed or corrected. The word defect, therefore, should be carefully used because it implies that a casting is defective and requires corrective measures or rejection. An imperfection or flaw may be a manufacturing defect, in the sense of being an unacceptable condition during an inspection regime, but not a functional defect. Thus, the term casting defect is sometimes used for general imperfections as well, even though the term discontinuity or imperfection is preferred when referring to deviations in less-than-perfect castings. Critical engineering assessment of discontinuities is necessary to define whether an imperfection is a defect or a harmless discontinuity that does not sacrifice the reliability or intended function of the casting. Unfavorable foundry practice can result in various casting imperfections that are detrimental in service and that may contribute to failure. However, many of the common causes of failures of castings occur from one or more aspects of design, materials selection, casting imperfections, faulty processing, improper assembly, or service conditions not initially anticipated. Some casting imperfections have no effect on the function or the service life of cast components, but will give an unsatisfactory appearance or will make further processing, such as machining, more costly. Many such imperfections can be easily corrected by shot-blast cleaning or grinding. Other imperfections or defects that can be more difficult to remove can be acceptable in some locations. Moreover, many castings, especially iron and steel castings, generally lend themselves to being weld repaired, with proper attention to the details of weld repair procedures. Imperfections and defects can be removed and a suitable weld deposit substituted to give a finished product fit for the intended service, although many failures of castings result from improper repair or rework of casting defects. Porosity is common in castings and takes many forms. Pores may be connected to the surface, where they can be detected by dye

penetrant techniques, or they may be wholly internal, where they require radiographic techniques to discover. Porosity and casting costs are minimized in casting designs that emphasize progressive solidification toward a gate or riser, tapered walls, and the avoidance of hot spots. Porosity is not necessarily a defect or cause for rejection; it depends on product requirements of the designer. Macroporosity refers to pores that are large enough to see with the unaided eye on radiographic inspection, while microporosity refers to pores that are not visible without magnification. Both macroporosity and microporosity are caused by the combined action of metal shrinkage and gas evolution during solidification. It has been shown (Ref 8 and 9) that nucleation of pores is difficult in the absence of some sort of substrate, such as a nonmetallic inclusion, a grain refiner, or a second phase particle. This is why numerous investigations have shown that clean castings, those castings that are free from inclusions, have fewer pores than castings that contain inclusions. Microporosity is found not only in castings, but also in heavy section forgings that have not been worked sufficiently to close it up. When the shrinkage and the gas combine to form macroporosity, properties are deleteriously affected. Static properties are reduced at least by the portion of the cross-sectional area that is taken up with the pores, since there is no metal in the pores, there is no metal to support the load there, and the section acts as though its area was reduced. Because the pores may cause a stress concentration in the remaining material (Ref 10 and 11), static properties may be reduced by more than the percentage of cross-sectional area that is caused by the macroporosity. Liquid metal can dissolve much more gas in solution than solid metal. Therefore, when metal solidifies, gas that is present in the liquid is rejected and forms bubbles. A commonly encountered example is that of hydrogen in aluminum. If these bubbles are trapped in the casting when it freezes, the result is a pore. Pores that result from gas may be spherical, indicating that they formed early in solidification when the metal was mostly liquid, or they may be interdendritic in shape, showing that they formed late in solidification, when the liquid that remained was present between dendrites. Microporosity is found between dendrites and, like macroporosity, is caused by the inability of feed metal to reach the interdendritic areas of the casting where shrinkage is occurring and where gas is being evolved. However, because this type of porosity occurs late in solidification, particularly in long-range freezing or mushy-freezing alloys, it is particularly difficult to eliminate. The most effective method is to increase the thermal gradient, often accomplished by increasing the solidification rate, which decreases the length of the mushy zone. This technique may be limited

Casting Design and Processes / 35 by alloy and mold thermal properties, and by casting geometry, that is, the design of the casting. The shape of the micropore is as important as its size, with elongated pores having a greater effect than round pores. Hot Isostatic Pressing. Microporosity can be healed by hot isostatic pressing (HIP). Hot isostatic pressing treatment is relatively inexpensive. If the porosity is relatively large or a macroporosity, HIP will form small dimples on the surface of the casting that may require weld repair. Because HIP is a thermal as well as a pressure treatment, it can alter the microstructure and, therefore, the properties of components if not carefully designed. Hot isostatic pressing will not affect inclusions or oxide folds and, therefore, will not repair castings that have those types of discontinuities. It also has no effect on pores that are connected to the surface. Hot isostatic pressing has made possible the use of alloys that, because of their composition, do not solidify pore free and, without HIP, could not be used for high-integrity applications. The thermal profile of a HIP cycle also is such that the process can eliminate the need for a stress relief or homogenization anneal. By adding time at elevated temperature to the overall thermal history of a casting, microstructural changes such as grain growth and precipitate coarsening can occur. For more information, see the article “Hot Isostatic Pressing of Castings” in Castings, Volume 15 of the ASM Handbook (Ref 12). Inclusions are nonmetallic particles that are found in the casting. They may form during solidification as some elements, notably manganese and sulfur in steel, precipitate from solution in the liquid. More frequently, they are formed before solidification begins. The former are sometimes called indigenous inclusions, and the latter are called exogenous inclusions. Inclusions are ceramic phases; they have little ductility. A crack may form in the inclusion and propagate from the inclusion into the metal, or a crack may form at the interface between the metal and the inclusion. In addition, because the inclusion and the metal have different coefficients of thermal expansion, thermally induced stresses may appear in the metal surrounding the inclusion during solidification (Ref 13). As a result, the inclusion acts as a stress concentration point and reduces dynamic properties. As in the case of microporosity, the size of the inclusion and its location determine its effect (Ref 14, and 15). Small inclusions that are located well within the center of the cross section of the casting have little effect, whereas larger inclusions and those located near the surface of the casting may be particularly detrimental to properties. Inclusions may also be a problem when machining surfaces, causing excessive tool wear and tool breakage. Exogenous inclusions are mostly oxides or mixtures of oxides and are primarily slag or dross particles, which are the oxides that result when the metal reacts with oxygen in the air during melting. These are removed from the

melt before pouring by filtration. Most inclusions found in steel castings arise from the oxidation of metal during the pouring operation (Ref 16). This is known as reoxidation, and takes place when the turbulent flow of the metal in the gating system causes the metal to break up into small droplets, which then react with the oxygen in the air in the gating system or casting cavity to form oxides. Metal casters use computer analysis of gating systems to indicate when reoxidation can be expected in a gating system and to eliminate them. However, casting designs that require molten metal to jet through a section of the casting to fill other sections will recreate these inclusions and should be avoided. Oxide films are similar to inclusions and have been found to reduce casting properties (Ref 17, 18, and 19). These form on the surface of the molten metal as it fills the mold. If this surface film is trapped within the casting instead of being carried into a riser, it is a linear discontinuity and an obvious site for crack initiation. It has been shown (Ref 20, and 21) that elimination of oxide films, in addition to substantially improving static properties, results in a five-fold improvement of fatigue life in axial tension-tension tests. Oxide films are of particular concern in nonferrous castings, although they also must be controlled in steel and stainless steel castings; because of the high carbon content of cast iron, oxide films do not form on that metal. If the film folds over on itself as a result of turbulent flow or waterfalling; when molten metal falls to a lower level in the casting during mold filling; the effects are particularly damaging. Casting design influences how the metal fills the mold, and features of the design that require the metal to fall from one level to another while the mold is filling should be avoided so that waterfalls are eliminated. Oxide films are avoided by filling the casting from the bottom, in a controlled manner, by pumping the metal into the mold using pneumatic or electromagnetic pumps. If the casting is poorly designed, waterfalling will result. An example is given in Fig. 42. Second phases, which form during solidification, may also nucleate cracks if they have the proper size and morphology (Ref 22). Example

are aluminum silicon alloys, where the silicon eutectic is present as large platelets, which nucleate cracks, and along which cracks propagate (Ref 23, and 24). The size of these platelets may be significantly reduced by modifying the alloy with additions of sodium or strontium. However, such additions increase the size of micropores (Ref 25), and for this reason, many foundrymen rely on accelerated solidification of the casting to refine the silicon. As noted above, solidification rates normally increase, and the structure is thus refined, in thin sections. Heavy sections are to be avoided if a fine structure is desired. Generally speaking, however, secondary phases in the structure of castings become important in limiting mechanical behavior of castings only in the absence of nonmetallic inclusions and microporosity (Ref 26). Hot tears form when casting sections are constrained by the mold from shrinking as they cool near the end of solidification. These discontinuities are fairly large and are most often weld repaired. If not repaired, their effect is not readily predictable (Ref 27). While generally they are detrimental to casting properties, under some circumstances they do not affect them. Hot tears are caused by a combination of factors, including alloy type, metal cleanliness, and mold and core hardness. However, poor casting design is the primary cause. Castings should be designed so that solidifying sections are not subjected to tensile forces caused by shrinkage during solidification, as the solidifying alloy has little strength before it solidifies. An example is given in Fig. 43, and an extensive discussion on how to prevent hot tears through casting design is provided in Ref 29. Metal Penetration. Molten metal may penetrate the surface of the mold, forming a rough

Redesign of a casting to eliminate hot tears. Mold restraint coupled with nonuniform freezing of the various sections of this aluminum alloy 356 casting resulted in hot tears. Moving the wall and increasing its thickness corrected the problem. Part dimensions in inches. Source: Ref 28

Fig. 43 Redesign of a casting to avoid waterfalling. (a) In this design, waterfalling results when casting is filled from the bottom. (b) Improved design provides a path for the metal to follow as it fills the mold.

Fig. 42

36 / Casting Design and Performance surface or, in extreme cases, actually becoming intimately mixed with the sand in the mold. In iron castings, this is normally the result of the combination of metallostatic head, the pressure exerted on the molten iron at the bottom of the mold by the weight of the metal on top of it, and the surface tension relationships between the liquid iron and molding materials (Ref 30). In cast iron, it is frequently the result of the expansion of graphite at the end of solidification, forcing liquid metal into the mold if the casting is not properly designed with a tapered wall to promote directional solidification and avoid hot spots. ACKNOWLEDGMENTS

11.

12. 13. 14.

15.

Portions of this article were adapted from:  T. Piwonka, Design for Casting, ASM Hand-

book Volume 20, 1997

16.

 R. Kotschi, Casting Design, Metals Hand-

book, 9th Edition, Casting, 1986

REFERENCES 1. J.Campbell, Casting Practice—Guidelines for Effective Production of Reliable Castings, ASM Handbook, Vol 15, Casting, 2008, p 497 2. J. Campbell, Filling and Feeding System Concepts, ASM Handbook, Vol 15, Casting, 2008, p 506 3. G. Woycik and G. Peters, Low-Pressure Metal Casting, ASM Handbook, Vol 15, Casting, 2008, p 700 4. J. Campbell, Castings, ButterworthHeinemann, 1991 5. J. Campbell, Castings, 2nd Edition, Butterworth-Heinemann, 2003 6. R.A. Flinn, Fundamentals of Metal Casting, Addison Wesley, 1963 7. J. Danzig, Transport Phenomena During Solidification, ASM Handbook, Vol 15, Casting, 2008, p 288 8. E.L. Rooy, Hydrogen: The One-Third Solution, AFS Trans., Vol 101, 1993, p 961 9. N. Roy, A.M. Samuel, and F.H. Samuel, Porosity Formation in Al-9Wt Pct Mg - 3 Wt Pct Cu Alloy Systems: Metallographic Observations, Met. Mater. Trans., Vol 27A, Feb 1996, p 415 10. M.K. Surappa, E. Blank, and J.C. Jaquet, Effect of Macro-porosity on the Strength

17.

18.

19.

20.

21.

22.

and Ductility of Cast Al-7Si-0.3Mg Alloy, Scr. Metall., Vol 20, 1986, p 1281 C.H. Ca´ceres, On the Effect of Macroporosity on the Tensile Properties of the Al7%Si-0.4%Mg Casting Alloy, submitted to Scr. Metall., 1994 S. Mashl, Hot Isostatic Pressing of Castings, ASM Handbook, Vol 15, Casting, 2008, p 408 I.P. Volchok, Non-Metallic Inclusions and the Failure of Ferritic-Pearlitic Cast Steel, Cast Metals, Vol 6 (No. 3), 1993, p 162 P. Heuler, C. Berger, and J. Motz, Fatigue Behaviour of Steel Casting Containing Near-Surface Defects, Fatigue Fract. Eng. Mater. Struct., Vol 16 (No. 1), 1992, p 115 J. Motz et al., Einfluss oberfla¨chener Fehlstellen im Stahlgub auf die Rib einleitung bei Schwingungsbeanspruchung, Geissereiforschung, Vol 43, 1991, p 37 C.E. Bates and C. Wanstall, Clean Steel Castings, in Metalcasting Competitiveness Research, Final Report, DOE/ID/13163-1 (DE95016652), Department of Energy, Aug 1994, p 51 J. Campbell, J. Runyoro, and S.M.A. Boutorabi, Critical Gate Velocities for FilmForming Alloys, AFS Trans., Vol 100, 1992, p 225 N.R. Green and J. Campbell, Proc. Spring 1993 Meeting (Strasbourg, France), European Division Materials Research Society, 4–7 May, 1993 N.R. Green and J. Campbell, The Influence of Oxide Film Filling Defects on the Strength of Al-7Si-Mg Alloy Castings, AFS Trans., Vol 102, 1994, p 341 C. Nyahumwa, N.R. Green, and J. Campbell, “The Effect of Oxide Film Filling Defects on the Fatigue Life Distributions of Al-7Si-Mg Alloy Castings,” presented at the International Symposium on Solidification Science and Processing (Honolulu, HI), Japan Institute of Metals and TMS, Dec 1995 J. Campbell, The Mechanical Strength of Non-Ferrous Castings, Proc. 61st World Foundry Cong. (Beijing), 1995, p 104; available from the American Foundrymen’s Society, Des Plaines, IL K.E. Ho¨ner and J. Gross, Bruch verhalten und mechanische Eigenschafter von AluminiumSilicium-Gublegierungen in unterschiedlichen Behandlungszusta¨nden, Giessereiforschung, Vol 44 (No. 4), 1992, p 146

23. F.T. Lee, J.F. Major, and F.H. Samuel, Effect of Silicon Particles on the Fatigue Crack Growth Characteristics of Al-12Wt Pct - 0.35Wt Pct Mg - (0 to 0.02) Wt Pct Sr Casting Alloys, Met. Mater. Trans., Vol 26A (No. 6), June 1995, p 1553 24. J.F. Major, F.T. Lee, and F.H. Samuel, Fatigue Crack Growth and Fracture Behavior of Al-12 wt% Si-0.35 wt% Mg (0–0.02) % Sr Casting Alloys, Paper 96-027, AFS Trans., Vol 104, 1996 25. D. Argo and J.E. Gruzleski, Porosity in Modified Aluminum Alloy Castings, AFS Trans., Vol 96, 1988, p 65 26. T.L. Reinhart, “The Influence of Microstructure on the Fatigue and Fracture Properties of Aluminum Alloy Castings,” presented at Aeromat ’96 (Dayton, OH), ASM International, June 1996 27. M.J. Couper, A.E. Neeson, and J.R. Griffiths, Casting Defects and the Fatigue Behaviour of an Aluminium Casting Alloy, Fatigue Fract. Eng. Mater. Struct., Vol 13 (No. 3), 1990, p 213 28. Permanent Mold Casting, Forging and Casting, Vol 5, Metals Handbook, 8th ed., American Society for Metals, 1970, p 279 29. A.L. Kearney and J. Raffin, Heat Tear Control Handbook for Aluminum Foundrymen and Casting Designers, American Foundrymen’s Society, 1987 30. D.M. Stefanescu et al., Cast Iron Penetration in Sand Molds—Part I: Physics of Penetration Defects and Penetration Model, Paper 96–206, AFS Trans., Vol 104, 1996

SELECTED REFERENCES  J. Campbell, Castings, 2nd Edition, Butter-

worth-Heinemann, 2003

 M.A. Gwyn, Cost-Effective Casting Design,

Modern Casting, Vol.88, No.5, 1998, pp.32–36.  B. Ravi, Metal Casting—Computer Aided Design and Analysis, Prentice Hall (India), 2005  Hwang Weng-Sing and Kaohsiung, Ed., Modeling of Casting & Solidification Processes, Taiwan, 2004  Kuang-O Yu, Modeling for Casting and Solidification Processing, Marcel Dekker, Inc, 2002

Casting Design and Performance Pages 37–60

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Modeling of Casting and Solidification Processing* Jianzheng Guo and Mark Samonds, ESI US R&D

CASTING AND SOLIDIFICATION PROCESSES are modeled and described in terms of thermodynamics, heat transfer, fluid flow, stress, defect formation, microstructure evolution, and thermophysical and mechanical properties (Ref 1). Simulation technologies are applied extensively in casting industries to understand the effects of alloy chemistry, heat transfer, fluid transport phenomena, and their relationships to microstructure and the formation of defects. Thanks to the rapid progress of computer calculation capability and numerical modeling technologies, casting simulation can quickly answer some critical questions concerning filling, microstructure, defects, properties, and final shape. Modeling can be used as a quality-assurance function and can also help to reduce the time for responding to customer inquiries. It shifts the trial-and-error procedure used on the shop floor to the computer, which does it faster, more easily, with greater transparency, and more economically. Casting simulation software has been available to the foundry industry for over 20 years. Simulation technology has come a long way since the early-to-mid-1980s when the design engineer could only work with two-dimensional models. The early days focused on identifying hot spots in the casting. As the computer-aided design and numerical simulation software packages evolved, foundry engineers could increasingly make quick changes to the feeding design to fix these problems with relative ease. Most of the casting simulation packages in the market today can now handle solidification and fluid flow in the mold. Today, the foundry industry wants to focus on more advanced predictions, such as stress and deformation, microstructure determination, defect formation, and mechanical properties. The primary phenomenon controlling casting is the heat transfer from the metal to the mold. The heat-transfer processes are complex. The cooling rates range from an order of one-tenth to thousands of degrees per second, and the corresponding length scales extend from several meters to a few micrometers. These various cooling rates produce different microstructures

and hence a variety of mechanical properties. Solidification kinetics, including nucleation, growth, and coarsening, are now being investigated extensively. The incorporation of these principles into the more traditional thermal, fluid flow, and stress models enable quantitative predictions of microstructure and engineering properties such as tensile strength and elongation. The coupling of mechanical analysis with thermal analysis enables the prediction of residual stresses and distortion in the castings and molds. These predictions will enable design engineers to evaluate the effects of nonuniform properties and defects on life-cycle performance of components. Different casting processes are used to produce different kinds of casting components. Some of the most common processes are sand casting, investment casting, die casting, permanent mold casting, lost foam casting, centrifugal casting, continuous casting, and direct chill casting. Each casting process has its own features. For example, for pressure die casting machines, it is particularly important to optimize not only the casting quality but also the die behavior with respect to thermal stress and strains and life expectancy. In this chapter, the topic of computational thermodynamics is first reviewed. The calculation of solidification paths for casting alloys is introduced in which back diffusion is included so that the cooling condition can be accounted for. Then, a brief review of the calculation of thermophysical properties is presented. Fundamentals of the modeling of solidification processes are discussed next. The modeling conservation equations are listed. Several commonly used microstructure simulation methods are presented. Ductile iron casting is chosen as an example to demonstrate the ability of microstructure simulation. Defect prediction is one of the main purposes for casting and solidification simulation. The predictions for the major defects of a casting, such as porosity, hot tearing, and macrosegregation, are highlighted. In the end, several industry applications are presented.

*Originally written and subsequently published for Metals Process Simulation, Volume 22B, ASM Handbook.

Computational Thermodynamics Thermodynamic calculations are the foundation for performing basic materials research on the solidification of metals. It is important to have proper phase information for an accurate prediction of casting solidification (Ref 2). The method of phase diagram calculations was started by Van Laar, who computed a large number of prototype binary phase diagrams with different topological features using ideal and regular solution models. Since then, many researchers began to incorporate phase equilibrium data to evaluate the thermodynamic properties of alloys. It was not until the late 1980s that a number of phase diagram calculation software packages became available. Then, thermodynamic calculations for multicomponent systems became feasible, thanks to the rapid progress in the computer industry. Solidification proceeds at various rates for castings. Thus, the microstructure and the composition are not homogeneous throughout the casting. The solidification path determines the solidification behavior of an alloy. For complex multicomponent alloys, the solidification path is very complicated. Hence, the equilibrium of each phase at different temperatures must be calculated. The thermodynamic and the kinetic calculations are the base for the prediction of solidification. The diffusion transport in the solid phase must be solved for each element. This requires knowledge of the diffusion coefficient of the element, the length scale, and the cooling conditions. Thermodynamic modeling has recently become increasingly used to predict the equilibrium and phase relationships in multicomponent alloys, (Ref 3–5). Currently, several packages are able to simulate solidification using the Scheil model and lever rule, such as Thermo-Calc, Pandat, and JMatPro. A modified Scheil model is applied in JMatPro. In their calculation, carbon and nitrogen are treated as completely diffused in the solid, which makes a great improvement, particularly for

38 / Casting Design and Performance iron-base alloy solidification. In reality, there is always a finite back diffusion based on the cooling conditions (Ref 6). It is critical to have an accurate solidification path for casting simulation (Ref 7–9). Obtaining the solidification path is very important for understanding and controlling the solidification process of the alloy. Normally, there are two ways to predict the solidification path. One is the complete equilibrium approach, which can be calculated by the lever rule. The other is the Scheil model, which assumes that the solute diffusion in the solid phase is small enough to be considered negligible and that diffusion in the liquid is extremely fast, fast enough to assume that diffusion is complete. For almost all practical situations, the solidification occurs under nonequilibrium conditions but does not follow the Scheil model. There is finite diffusion in the solid, or back diffusion, which is a function of the cooling rate. Back diffusion plays an important role in the calculation of segregation. There are many numerical and analytical models that attempt to handle such phenomena (Ref 10–16). For most of the models, constant partition coefficients are assumed, which is a good approximation for many alloys. Unfortunately, sometimes the partition coefficients can vary dramatically for some commercial alloys. The partition coefficient of an element in an alloy can change from less than one to greater than one, or vice versa, during solidification (Ref 17). For these cases, the analytical or earlier numerical models are not valid any more. The equilibrium of each phase at different temperatures should be calculated. This can be fulfilled by the coupling with the thermodynamic calculation. Recently, researchers have started to couple thermodynamic calculations with a modified Scheil model including back diffusion (Ref 18). Normally, the liquid is assumed to be completely mixed. The solid concentration is calculated by solving a one-dimensional diffusion equation for each element. The governing equations for conservation of species for multicomponent alloy solidification are (Ref 19): Liquid species conservation: fl

 @f  @Clj  j SDj  j s ¼ Cl  Cij þ Cs  Cij @t @t L

(Ref 20) using the one-dimensional platelike dendrite geometry: L ¼ f6s l where l is the secondary dendrite arm spacing, which is a func

tion of cooling rate, l ¼ a T n . Here, a and n are constants determined by the alloy composi

tion, and T is the cooling rate. The interfacial area concentration S is related to the solid volume fraction and the secondary dendrite arm spacing, S ¼ 2=l. Combining Eq 1 and 2 and then discretizing provides:  SDt j o ÞCi fs Csj  Csj ¼ ðfs þ L SDt j  ðfs þ ÞCs L

where the superscript o refers to the old value of the variable. Hence:

Csj ¼ Cij when o

¼ 0 Scheil model

SDt L

Based on mass conservation, the liquid concentration can be calculated from the solution of the solid concentration profile accordingly. Yan (Ref 18) did an experiment for the solidification of a quaternary Al-6.27Cu-0.22Si0.19Mg alloy with a cooling rate of 0.065 K/s. The microstructure of the solidified samples is dendritic. The calculated fraction of solid versus temperature relationship for this quaternary alloy is shown in Fig. 1. According to the Scheil model, the solidification sequence for this alloy is L ! L+ face-centered cubic (fcc) ! L+fcc+theta ! L+fcc+theta+Al5Cu2Mg8 830

L-->fcc

820

where j is the species index, C is concentration, D is a diffusion coefficient, f is the volume fraction of a phase, t is time, L is a diffusion length, and S is the interfacial area concentration. The subscript i refers to the solid-liquid interface, l is the liquid, and s is the solid. The diffusion length L can be determined using a model proposed by Wang and Beckermann (Ref 19) based on the work of Ohnaka

SDt ! 1Lever rule L

Csj ¼ Csj þ ðCij  Cso Þfs =fs when

(Eq 1)

(Eq 2)

(Eq 4)

Equation 4 is used for calculating the concentration in the solid. Here, it is noticable that Eq 4 can automatically turn into the Scheil model or lever rule if the diffusion is zero or infinity:

Temperature, K

 @f @Csj  j SDj s ¼ Ci  Csj þ @t @t L

j ðfs  fs ÞCsj þ ðfs þ SDt L ÞCi SDt fs þ L o

Csj ¼

Solid species conservation: fs

(Eq 3)

Si6(Q)! L+fcc+theta+Al5Cu2MG8Si6(Q)+Si. Experimentally, there were no Silicon and Al5Cu2Mg8Si6 phases formed according to metallographical examination and electron probe microanalysis. The back-diffusion model indicates that there are only fcc and theta phases formed during solidification for this cooling condition. The results predicted by the back-diffusion mode are in agreement with the experimental quantitative image analysis program. The measured fractions of fcc phase were compared with the calculations from the Scheil model, lever rule, and the current back-diffusion model for three different cooling rates. The comparison is shown in Table 1. The fraction of fcc phase calculated the from the Scheil model is less than the measured values, and the fraction of fcc calculated from the lever rule is higher than that from the experiments for all three cooling rates. The back-diffusion model, which takes into account the cooling rate, gives good agreement with the experiments.

Thermophysical Properties The research on thermophysical properties is a very important part of materials science, particularly at the current time, because such data are a critical input for the simulation of metals processing. Yang and coworkers (Ref 21) investigated the sensitivity of investment casting simulations to the accuracy of thermophysical properties. They found that the temperature prediction and thermal gradient in the liquid are the most sensitive to the accuracy of the input values used for density and thermal conductivity in the solid. Thermal conductivity in the liquid, specific heat, and density have similar levels of influence on solidification time; increasing their values increases the local solidification time. Thermal conductivity in the solid has the opposite effect compared to all the other properties studied. Accurate thermophysical data are difficult to obtain at high temperatures experimentally, especially for reactive alloys such as titanium and nickel-base superalloys. An extensive database for the calculation of thermophysical properties has been developed (Ref 8) that uses the phase fraction information predicted with the Gibb’s free energy minimization routines developed by (Ref 3) and extended by Ref 4.

810

800

Lever rule Back-diffusion model Scheil model

790

L-->fcc+Theta

Table 1 Comparison of experimental and calculated fraction of face-centered cubic phase (volume percent)

L-->fcc+Theta+Q Cooling rate, K/s

780 0.84

Fig. 1

L-->fcc+Theta+Q+(SI) 0.86

0.88

0.90 0.92 0.94 Fraction of solid

0.96

0.98

1.00

Solidification path of a 2219 aluminum alloy

Lever rule 0.065 0.25 0.75 Scheil model

Area scan

Image analysis

Calculation

 96.0 95.8 95.8 

 95.4 95.3 94.7 

96.7 96.0 95.4 94.3 85.4

Modeling of Casting and Solidification Processing / 39 These properties include density, specific heat, enthalpy, latent heat, electrical conductivity and resistivity, thermal conductivity, liquid viscosity, Young’s modulus, and Poisson’s ratio. The thermodynamic calculation is based on the thermodynamic databases from CompuTherm. A simple pairwise mixture model, which is similar to that used to model thermodynamic excess functions in multicomponent alloys, can be used to calculate the properties (Ref 5): X

XX

X

activation model. The former one is mainly based on the monatomic nature. There are some models available, but most of them are still under development and do not meet the technological need. The thermal activation method is applied here to predict the viscosity of alloys. The viscosity, Z, of pure liquid metals follows Andrade’s relationship (Ref 26): ZðT Þ ¼ Zo expðE=RT Þ

(Eq 7)

where P is the phase property, Pi is the property of the pure element in the phase, i is a binary interaction parameter, and xi and xj are the mole fractions of elements i and j in that phase.

where E is the activation energy, and R is the gas constant. Figure 4 shows an example of the calculated liquid viscosity of an IN718 alloy using the mixture model compared with experimental results. Figure 5 shows the comparison between experimental and calculated results for various alloys at different temperatures.

Thermal Conductivity

Density

The thermal conductivity mainly depends on the chemical composition of an alloy. It also depends to a lesser extent on the precipitates, bulk deformation, microstructures, and other factors (Ref 22, 23). These factors can usually be ignored in the calculation of conductivity for commercial alloys. The thermal conductivity of alloys is composed of two components: a lattice component and an electronic component. In well-conducting metals, the thermal conductivity is mainly electronic conductivity. The lattice conductivity is usually very small compared to the electronic one. Hence, only the electronic component is considered here. The thermal conductivity, l, and the electrical resistivity, r, are related according to the Wiedeman-Franz-Lorenz law (Ref 24, 25):

Currently, the casting simulation models have reached the stage where one of the limiting factors in their applicability is the accuracy of the thermophysical data for the materials to be modeled. Among all the thermophysical data, the temperature-dependent density is one of the most

v ðxi  xj Þv

v

(Eq 5)

LT l¼ r

Liquid Viscosity Viscosity is an important property to be considered in dealing with fluid flow behavior. The liquid viscosity is a measure of resistance of the fluid to flow when subjected to an external force. There are two approaches to modeling of complex alloy viscosity. One is a fundamental molecular approach and the other is a thermal

Casting process modeling involves the simulation of mold filling, solidification of the cast metal, microstructure formation, stress analysis on casting and mold, and so on. At the macroscopic scale, these processes are governed by basic equations that describe the conservation of mass,momentum, energy, and species. Heat transfer is perhaps the most important discipline in casting simulation.

200

14

180

12 160 140 Experiment Calculation

120 100

10 8 6

80

4

60

2

40 300

(Eq 6)

where the Lorentz constant, L; ¼ 2:44  1011 W K 2 , and T is the temperature. Based on this model, an example of the calculated thermal conductivity of an A356 alloy is shown in Fig. 2, with experimental results for comparison. The calculation can accurately predict the thermal conductivity variation with temperatures for this alloy in the liquid, solid, and mushy zone. Figure 3 shows the comparison with experimental results from Auburn University for various alloys at different temperatures. The agreement is good in general.

Fundamentals of the Modeling of Solidification Processes

Viscosity, mPas

j>i

Fig. 2 alloy

400

500

600 700 800 Temperature, K

900

Comparison between experimental and calculated thermal conductivity for an A356

Experiment Calculation

0 1400

1000

1500

1600 Temperature, K

1700

1800

Comparison between experimental calculated viscosity for an IN718 alloy

Fig. 4

Thermal conductivity, W/mK

and

Liquid viscosity, cP

200

9 8 7

150

Calculation

i

xi xj

Thermal conductivity, Wm/K

xi Pi þ

Calculations



critical ones for the accurate simulation of solidification microstructure and defect formation, such as shrinkage porosity (Ref 8). A database has been developed containing molar volume and thermal volume coefficients of expansion of liquid, solid-solution elements, and intermetallic phases. This is linked to the thermodynamic calculations mentioned previousley. The densities of the liquid and solid phases of multicomponent systems are calculated by the simple mixture model (Ref 27). Figure 6 shows plots comparing experimental values with calculations for the density of various alloys at different temperatures. Figure 7 shows a comparison between the calculated and experimentally reported density for a CF8M stainless steel alloy.

100

6 5 4 3

50

2 1 0

0 0

Fig. 3 alloys

50

100 Experiments

150

200

Comparison between experimental and calculated thermal conductivity for various

0

1

2

3

4 5 Experiments

6

7

8

9

Comparison between experimental and calculated viscosity for various alloys at different temperatures

Fig. 5

40 / Casting Design and Performance fl is the fraction of liquid, and ui;l is the actual liquid velocity. Momentum equation is:

Density, g/cm3 9 8

Calculation

7

r

6 5 4

  @ui @ui @ @ui m þ ruj þ pdij m ¼ rgi  ui K @t @xj @xj @xj (Eq 10)

where dij is the Kronecker delta, p is pressure, gi is gravitational acceleration, and K is permeability.

3 2

Continuity equation is:

1 1

2

Fig. 6

3

4

5 6 Experiments

7

8

Comparison between experimental calculated density for various alloys

9

and

8.0 7.8

@r @ðrui Þ þ ¼0 @t @xi

(Eq 11)

In order to solve the previous equation, proper initial conditions and boundary conditions are needed. Basic initial conditions include temperature, velocity, pressure, and so on:

This is a simple version of radiation in which it is assumed that there is only one ambient temperature present. Thus, the view factor is equal to one. For investment casting modeling, the view factor must be calculated carefully. For that case, a view-factor radiation model can be used. View-Factor Radiation Model. A net flux model can be used for more complex view-factor radiation. Rather than tracking the reflected radiant energy from surface to surface, an overall energy balance for each participating surface is considered. At a particular surface i, the radiant energy being received is denoted qin;i . The outgoing flux is qout;i . The net radiative heat flux is the difference of these two: qnet;i ¼ qout;i  qin;i

(Eq 16)

Using the diffuse, gray-body approximation, the outgoing radiant energy can be expressed as:

Density, g/cm3

7.6 7.4

T ðx; 0Þ ¼ To ðxÞ

qout;i ¼ sei Ti4 þ ð1  ei Þqin;i

uðx; 0Þ ¼ uo ðxÞ

The first term in the previous equation represents the radiant energy, which comes from direct emission. The second term is the portion of the incoming radiant energy, which is being reflected by surface i. The incoming radiant energy is a combination of the outgoing radiant energy from all participating surfaces being intercepted by surface i. Specifically, the view factor Fij is the fraction of the radiant energy leaving surface j that impinges on surface i. Thus:

7.0

Calculation Experiment

6.8

vðx; 0Þ ¼ vo ðxÞ

6.6 6.4 6.2

T ðx; 0Þ ¼ wo ðxÞ

6.0 0

Fig. 7

200

400

600 800 1000 1200 1400 1600 Temperature, °C

pðx; 0Þ ¼ po ðxÞ

Comparison between experimental and calculated density for a CF8M stainless steel alloy

The solidification process depends on heat transfer from the part to the mold and from the mold to the environment. The heat can be transferred by conduction, convection, and/or radiation. Conduction refers to the heat transfer that occurs as a result of molecular interaction. Convection refers to the heat transfer that results from the movement of a liquid, such as liquid metal or air. Radiation refers to the heat transfer of electromagnetic energy between surfaces that do not require an intervening medium. Radiation is very important for investment casting processes, for instance. Basic mathematical formulations that govern the solidification process can be summarized as follows (Ref 28): Energy equation is: r

(Eq 17)

7.2

@H @H þ rui  rðkrT Þ  qðxÞ ¼ 0 @t @xi

(Eq 8)

where r is density, k is thermal conductivity; H is enthalpy, a function of temperature that encompasses the effects of specific and latent heat; and ðT HðT Þ ¼ Cp dT þ L½1  fs ðT Þ

(Eq 9)

0

where L is latent heat, fs is the fraction of solid, ui ¼ fl ui;l is the component superficial velocity,

Boundary Conditions. There are many kinds of boundary conditions. Some of the most common ones are listed as follows. The fixed value or Dirichlet boundary condition is: Y ðxÞ ¼ Yd ðxÞfðtÞ on 1

(Eq 12)

where Y ðxÞ can be temperature, velocity, or pressure; 1 is some subset of the total boundary; Yd ðxÞ is specified variable vector; and f(t), the time function. The specified heat flux boundary condition is: krT  n_ ¼ qn fðtÞ on 2

(Eq 13)

where qn is the specified heat flux, n_ is a unit vector normal to the surface, and 2 is some subset of the total boundary . The convective heat flux boundary condition is: krT  n_ ¼ qn fðtÞhfðtÞgðT Þ½T  Ta ; on 2 (Eq 14)

where h is the convection coefficient, g(T) is the temperature function, and Ta is the ambient or media temperature that could be a function of time. The radiation heat-boundary condition is: krT  n_ ¼ segðT Þ½T  4

Ta4 ;

(Eq 15)

on 2 where s is the Stefan-Boltzmann constant, and e is emissivity.

qin;i ¼

N X

Fij qout;j

(Eq 18)

j¼1

where N is the total number of surfaces participating in the radiation model, and the view factors are calculated from the following integral: Fij ¼

1 Ai

ð ð

cos yj cos yi dAi dAj r2

(Eq 19)

Aj Ai

where Ai is the area of surface i, yi is the polar angle between the normal of surface i and the line between i and j, and r is the magnitude of the vector between surface i and j. Then, the vector of radiosities qout;i can be solved by: N  X j¼1

Ai ei Ai dij  Ai Fij qout;j ¼ sT 4 ð1  ei Þ ð1  ei Þ i (Eq 20)

Hence, the net radiant flux is obtained by:  qnet;i ¼

ei t sTi  qout;i 1  ei

(Eq 21)

This heat flux then appears as a boundary condition for the heat conduction analysis.

Modeling of Casting and Solidification Processing / 41 Based on the previous equations, the basic heat-transfer and fluid flow problems can be solved with proper initial and boundary conditions.

Microstructure Simulation The purpose of casting solidification micromodeling is to predict the microstructure of a casting, such as grain size. Understanding the casting solidification process and microstructure formation will greatly facilitate casting design and quality control. Computer modeling also provides the basis for computer-aided manufacturing and product life-cycle management by being able to predict mechanical properties, quality, and useful life (Ref 29–34). There are several ways to simulate the microstructure formation during alloy solidification, such as the deterministic method or the stochastic method. For the deterministic method, the density of grains that have nucleated in the bulk liquid at a given moment during solidification is a deterministic function, for example, a function of undercooling. The stochastic method is a probabilistic means to predict the nucleation and growth of the grain, including the stochastic distribution of nucleation locations, the stochastic selection of the grain orientation, and so on. The stochastic method includes cellular automata and the phase-field method.

Deterministic Micromodeling Modeling of solidification processes and microstructural features has benefited from the introduction of averaged conservation equations and the coupling of these equations with microscopic models of solidification. When conservation equations are averaged over the liquid and solid phases, the interfacial continuity condition automatically vanishes and average entities (e.g., mean temperature or solute concentration) appear. Rappaz et al. (Ref 35, 36) proposed a model using averaging methods to predict the growth of equiaxed grains under isothermal conditions. The nucleation is based on the model proposed by Thevoz and co-workers (Ref 37), which is illustrated in Fig. 8. At a given undercooling, the grain density, n, is calculated by the integral

dn d (DT )

of the nucleation site distribution from zero undercooling to the current undercooling. The density of the grain nuclei is: nmax nðT ðtÞÞ ¼ pffiffiffiffiffiffi 2  Td ! TððtÞ ðT ðtÞ  Tn Þ2 exp  dðT ðtÞÞ 2Ts2



(Eq 22)

where nmax is the maximum grain nuclei density, Td is the standard deviation undercooling, and Tn is the average undercooling. Rappaz and Boettinger (Ref 29) studied the growth of an equiaxed multicomponent dendrite. In their study, for each element, the supersaturation is cl;j  co;j ¼ IvðPej Þ cl;j ð1  kj Þ

(Eq 23)

where J = 1, n is the solute element, cl;j is the tip liquid concentration, co;j is the nominal concentration, and kj is the partition coefficient. The Peclet number is defined as: Pej ¼

Rv 2Dj

where Dj is the diffusion coefficient. The Ivanstsov function is defined as IvðPeÞ ¼ Pe  expðPeÞ  E1 ðPeÞ, where E1 ðPeÞ is the first exponential integral. Assuming growth at the marginal stability limit, the dendrite radius is calculated by: R¼P n j¼1

22  co;j ð1kj Þ mj Pej 1ð1k j ÞIvðPej Þ

(Eq 24)

where G is the Gibbs-Thomson coefficient. Hence, the tip velocity is: v ¼ D1 Pe1

2 R

(Eq 25)

and the tip liquid concentration will be: cl;j ¼

co;j 1  ð1  kj ÞIvðPej Þ

(Eq 26)

Assume the liquid concentration in the interdendritic region is uniform. The solute profiles in the extradendritic liquid region can be obtained from an approximate model (Ref 29):

DTd DT

Jj ¼ Dj 

n nmax



cl;j  co;j dj =2

 (Eq 27)

where the solute layer thickness is: DTn

Fig. 8

Nucleation model

DT

dj ¼

2Dj v

So the solute balance will become:

n X

(Eq 29)

mj Jj ¼ 0

j¼1

0

j ¼

n dfs X dT mj ðkj  1Þcl;j þ ðfg  fs Þ dt dt j¼1

(Eq 28)

where fg is the envelope volume divided by the final grain volume. Please refer to Ref 29 for details about the derivation of the previous equations. The secondary dendrite arm spacing is calculated by: l2 ¼ 5:5ðMtf Þ1=3

(Eq 30)

Where M¼P n j¼1



 mj ð1  kj Þðce;j  co;j Þ=Dj 0

1 m ð1  k Þc =D j j e;j j Bj¼1 C B C lnB P C @n A mj ð1  kj Þco;j =Dj n P

(Eq 31)

j¼1

Based on the previous equations, the solidification of a multicomponent casting can be predicted. The averaged microstructure, such as grain size and secondary dendrite arm spacing, can be calculated based on the chemistry and cooling conditions. Some examples using such deterministic micromodeling are provided in a later section.

Cellular Automaton Models Cellular automaton (CA) models are algorithms that describe the discrete spatial and/or temporal evolution of complex systems by applying deterministic or probabilistic transformation rules to the sites of a lattice. In a CA model, the simulated domain is divided into a grid of cells, and each cell works as a small, independent automaton. Variables and state indices are attributed to each cell, and a neighborhood configuration is also associated with it. The time is divided into finite steps. At a given time step, each cell automaton checks the variables and state indices of itself and its neighbors at the last time step and then decides the updated results at the present step according to the predefined transition rules. By iterating this operation with each time step, the evolution of the variables and state indices of the whole system is obtained. The CA model is usually coupled with a finite-element (FE) heat flow solver, such as the CAFE model developed by Rappaz and colleagues. The CA algorithm can be used to simulate nucleation and growth of grains. This model can be used to predict cellular-to-equiaxed transition (CET) in alloys. Several models have been developed over the years for the prediction of microstructure formation in casting (Ref 38–42). One example is shown in Fig. 9.

42 / Casting Design and Performance Cellular automaton-finite element models are particularly well suited to tracking the development of a columnar dendritic front in an undercooled liquid at the scale of the casting (Ref 40,41). Although these models do not directly describe the complicated nature of the solid/liquid interface that defines the dendritic microstructure, the crystallographic orientation of the grains as well as the effect of the fluid flow can be accounted for to calculate the undercooling of the mushy-zone growth front. Two- and three-dimensional CAFE models were successfully applied to predict features such as the columnar-to-equiaxed transition observed in aluminum-silicon alloys (Ref 40), the selection of a single grain and its crystallographic orientation due to the competition among columnar grains taking place while directionally solidifying a superalloy into a pigtail shape (Ref 41), as well as the effect of the fluid flow on the fiber texture selected during columnar growth (Ref 42). Coupling with macrosegregation has been developed (Ref 43), thus providing an advanced CAFE model to account for structure formation compared to purely macroscopic models developed previously, for example, Ref [44 to 46]. While both structure and segregation were predicted (Ref 43), comparison with experimental observation concerning structure formation was limited due to the lack of detailed data (Ref 47). Comparison is thus mainly conducted with the as-cast state. Wang et al. (Ref 48) investigated the effect of the direction of the temperature gradient on grain growth. As shown in Fig. 10(b), the temperature gradient was inclined at 45 relative to the macroscopic solidification direction, and the magnitude of the gradient was 12 K/mm. It clearly shows that the direction of temperature gradient can affect both the macro- and microscale dendritic structures, as well as the maximum undercooling. While the CA methods produce realisticlooking dendritic growth patterns and have resulted in much insight into the CET, some questions remain regarding their accuracy. Independence of the results on the numerical grid size is rarely demonstrated. Furthermore, the CA techniques often rely on relatively arbitrary rules for incorporating the effects of crystallographic orientation while propagating the solid-liquid interface. It is now well accepted that dendritic growth of crystalline materials depends very sensitively on the surface energy anisotropy (Ref 49, 50).

Three-dimensional view of the final grain structure calculated in the weak coupling mode for a directionally solidified turbine blade. The pole figures are displayed for various cross sections perpendicular to the main blade axis. Source: Ref 41

Fig. 9

Phase-Field Model An alternative technique for investigating microstructure formation during solidification is the phase-field method. Phase-field models were first developed for simulating equiaxed growth under isothermal conditions (Ref 51, 52). A desirable extension of the model was to study the effect of heat flow due to the release of latent heat. A simplified approach was proposed in which the temperature was assumed

(a) Predicted dendritic structure and (b) solutal adjusted undercooling distribution under thermal conditions of 45 inclined isotherms with respect to the growth direction moving at a constant velocity of 150 mm/s. Source: Ref 48

Fig. 10

Modeling of Casting and Solidification Processing / 43 to remain spatially uniform at each instant, and a global cooling rate was imposed with consideration of the heat extraction rate and increase of the fraction of solid (Ref 53). The attempt to model nonisothermal dendritic solidification of a binary alloy was made by Loginova et al. by solving both the solute and heat diffusion equations and considering the release of latent heat as well (Ref 54). Besides equiaxed growth in the supersaturated liquid, the phase-field model was also applied to the simulation of directional solidification, under well-defined thermal conditions (Ref 55, 56). The phase-field model has also been used to simulate the competitive growth between grains with different misorientations with respect to the thermal gradient (Ref 57). Further development in phasefield models includes the extension into three dimensions (Ref 58, 59) and multicomponent systems (Ref 60, 61). Usually, a regular grid composed of square elements is used in the phase-field models (Ref 52, 53), but an unstructured mesh composed of triangular elements has also been used, which enables the phase-field method to be applicable in a domain with complex geometry shape and also in a large scale. From a physical point of view, the phase-field method requires knowledge of the physical nature of the liquid-solid interface. However, little is known about its true structure. Using Lennard-Jones potentials, molecular dynamics simulations of the transition in atomic positions across an interface have suggested that the interface width extends over several atomic dimensions (Ref 62). At present, it is difficult to obtain usable simulations of dendritic growth with interface thickness in this range due to the limitations of computational resources. Thus, the interface width will be a parameter that affects the results of the phase-field method. It should be realized that in the limit as the interface thickness approaches zero, the phase-field equations converge to the sharp interface formulation (Ref 63, 64). In contrast to CA models, which adopt a pseudo-front-tracking technique, phase-field models express the solid/liquid interface as a transitional layer that usually spreads over several cells. The diffusion equation for heat and solute can be solved without tracking the phase interface using a phase-field variable and a corresponding governing equation to describe the state in a material as a function of position and time. This method has been used extensively to predict dendritic, eutectic, and peritectic growth in alloys, and solute trapping during rapid solidification. (Ref 65). The interface between liquid and solid can be described by a smooth but highly localized change of a variable between fixed values such as 0 and 1 to represent solid and liquid phases. The problem of applying boundary conditions at an interface whose location is an unknown can be avoided. Phase-field models have recently become very popular for the simulation of microstructure evolution during solidification processes (Ref

66–70). While these models address the evolution of a solid-liquid interface using only one phase-field parameter, interaction of more than two phases or grains, and consequently the occurrence of triple junctions, needed to be included into the multiphase-field approach (Ref 71–74). The deterministic model is capable of tracking the evolution of the macroscale or average variables, for example, average temperature and the total fraction of solid, but it cannot simulate the structure of grains. The CA models can simulate the macroscale and mesoscale

grain structures, but it has difficulty in resolving the microstructure. The phase-field method can well reproduce the microstructure of dendritic grains (Fig. 11, 12). However, with the current computational power, phase-field models can only work well on a very small scale (up to hundreds of Micrometers). The typical scale of laboratory experiments is 1 cm, and the scale of an industry problem can be up to 1 m. Both of them are beyond the capability of the phase-field method. In the industry, for larger castings, deterministic micromodeling still is a main player.

Predicted results for two-dimensional dendritic solidification of a binary alloy into an undercooled melt with coupled heat and solute diffusion for Le = 50 at tD/d0 2 = 3500. The upper- and lower-right quadrants show the dimensionless concentration U and temperature fields, respectively; the left quadrants both show concentration c/c0 fields, with different scales used in the upper and lower quadrants to better visualize the concentration variations in the solid and liquid, respectively. Source: Ref 74

Fig. 11

Fig. 12

Melting of dendritic structure and formation of fragments when temperature is increased from the growth temperature of 1574 K shown in (a) to 1589 K shown at later times in (b–d). Source: Ref 65

44 / Casting Design and Performance Micromodeling Applications in the Industry Here, the focus is on deterministic modeling due to its wide application in the casting industry. Thermodynamic calculations are coupled with the macroscale thermal and fluid flow calculations (Ref 75). Ductile iron is chosen as an example to demonstrate the capability of deterministic micromodeling because of its complex solidification behavior. Ductile irons are still important engineering materials due to their high strength and toughness and relatively low price. In the foundries, ductile irons suffer from shrinkage porosity formation during solidification, which is detrimental to the mechanical properties. To minimize porosity formation, large risers are normally used in the design, which reduces porosity level sometimes but leads to a low yield. Due to the complex solidification behavior of ductile irons and their extreme sensitivity to the process, it is very difficult to optimize the casting design to ensure the soundness of castings. To better understand the shrinkage behavior of ductile iron during solidification, a micromodel was developed to simulate the microstructure formation. The density change during solidification and the room-temperature mechanical properties can be calculated based on the microstructure. The simulation has been compared with the experimental results and found to be in good agreement. Cast iron remains the most important casting material, with over 70% of the total world tonnage now (Ref 76). Based on the shape of graphite, cast iron can be lamellar (flake) or spheroidal (nodular). In the last 40 years, many papers have been published on the modeling of ductile iron solidification. It started with the computational modeling by analytical heat transport and calculation of transformation kinetics (Ref 77–83). The computer model can calculate the cooling curve with an analytical method together with the kinetics calculation of carbon diffusion through the gamma-phase shell. In 1985, Su et al. coupled heat transfer and a solidification kinetics model first, using the finite-difference method (Ref 84). After that, many papers have been published on micromodeling of ductile iron solidification (Ref 85–95). The carbon diffusioncontrolled growth through the gamma shell was modeled. In those models, the nodule count, graphite radius, and austenite shell radius were calculated. Reference 96 used the internal state variable approach to model the multiple phase changes occurring during solidification and subsequent cooling of near-eutectic ductile cast iron. In this simulation, the effects on the microstructure evolution at various stages of the process by the alloy composition, graphite nucleation potential, and thermal progress were illuminated. The heat flow, fading effect, graphite/ austenite eutectic transformation, ledeburite eutectic transformation, graphite growth in austenite regime, and the eutectoid transformation were all modeled. A comprehensive micromodel

is developed that can give accurate microstructure information as well as the mechanical properties, such as yield strength, tensile strength, and hardness. The density of austenite, ferrite, pearlite, graphite, liquid, and ledeburite are all calculated. The prediction has been compared with the experimental results and found to be in good agreement (Ref 75). Nucleation Model. Here, Oldfield’s nucleation model is applied. In this mode, bulk heterogeneous nucleation occurs at foreign sites that are already present within the melt or intentionally added to the melt by inoculation: No ¼ AðT Þn

(Eq 34)

where rG ,rg , and rl are the densities of graphite, austenite, and liquid, respectively, and the calculation can be found in the next section; RG , Rg , andRl are the radii of graphite, austenite, and the final grain, respectively; and mav is the average mass of the grain. Assuming complete mixing of solute in liquid, the overall solute balance is written as:

(Eq 32)

where A is the nucleation constant, No is the nucleation number per unit volume, T is the undercooling, and n is another constant that depends on the effectiveness of inoculation. Fading Effect. Fading is the phenomenon whereby the effectiveness of inoculation diminishes as the time between inoculation and casting increases. It is believed that the nucleation of graphite occurs on small, nonmetallic inclusions that are entrapped in the liquid after inoculation (Ref 93). The small particles will grow with time. The particle diameter can be calculated by: d ¼ ðd3o þ ktÞ1=3

4 4 4 rG R3G þ rg ðR3g  R3G Þ þ rl ðR3l  R3g Þ 3 3 3 ¼ mav

(Eq 33)

where d is the particle diameter with time, do is the particle diameter at the beginning of the inoculation, and k is a kinetic constant. Graphite/Austenite Eutectic Transformation. The eutectic growth process in ductile iron is a divorced growth of austenite and graphite, which do not grow concomitantly. At the beginning of the liquid/solid transformation, graphite nodules nucleate in the liquid and grow in the liquid to a small extent. The formation of graphite nodules and their limited growth in liquid depletes the carbon in the melt locally in the vicinity of the nodules. This facilitates the nucleation of austenite around the nodules, forming a shell. Further growth of these nodules is by diffusion of carbon from the melt through the austenite shell. Once the austenite shell is formed around each nodule, the diffusion equation for carbon through the austenitic shell is solved in one-dimensional spherical coordinates. The boundary conditions are known from the phase diagram because thermodynamic equilibrium is maintained locally. Conservation of mass and solute is maintained in each grain. Because of the density variation resulting from the growth of austenite and graphite, the expansion/contraction of the grain is taken into account by allowing the final grain size to vary. Toward the end of solidification, the grains impinge on each other. This is taken into consideration by using the Johnson-Mehl approximation. Using a spherical coordinate system, a mass balance is written as:

rG  1 

4 3 R þ 3 G

R ða

rg cðr; tÞ4r2 dr Rc

4 þ rl cl ðR3l  R3g Þ ¼ cav 3

ðEq 35Þ

Differentiation of the previous two equations and the use of Fick’s law in spherical coordinates lead to two equations for graphite and austenite growth rates, following some manipulation. Ledeburite Eutectic Transformation. When the temperature reaches below the metastable eutectic temperature, the metastable phase forms. The metastable cementite eutectic is also called ledeburite, in which small islands of austenite are dispersed in the carbide phase. It has both direct and indirect effects on the properties of ductile iron castings. Increasing the volume percent of the hard, brittle carbide results in an increase in the yield strength but a reduction in the tensile strength. Following the assumptions from Onsoien (Ref 96, 97) that the graphite/ austenite nodule distribution is approximated by that of a close-packed face-centered space lattice, and that the ledeburite eutectic appears in an intermediate position, the total number of ledeburite nucleation sites would be the same for graphite/austenite nodules. The grain is assumed to be spherical. Hence, the growth of the ledeburite can be calculated as: dRLE ¼ 30:0  106 ðT Þn dt

(Eq 36)

Thus, the fraction of ledeburite can be written as: 4 fLE ¼ NR3LE 3

(Eq 37)

Eutectoid Transformation. The eutectoid reaction leads to the decomposition of austenite into ferrite and graphite for the case of the stable eutectoid and to pearlite for the metastable eutectoid transformation. Usually, the metastable eutectoid temperature is lower than the stable eutectoid temperature. Slower cooling rates result in more stable eutectoid structure. Following solidification, the solubility of carbon in austenite decreases with the drop in temperature until the stable eutectoid temperature is reached.

Modeling of Casting and Solidification Processing / 45 decreases the tensile strength. The yield strength and hardness continuously decrease as the cooling rate decreases. The yield strength is very high on the left part because of the formation of carbide. On the other hand, the carbide decreases the tensile strength. The results are shown in Fig. 16. Experimental Validations. A series of experiments was performed to validate the micromodel (Ref 99). The three-part cast iron foundry mold containing the gating system is shown in Fig. 17. The casting is GGG60 ductile iron. The pouring temperature is 1400  C, the initial die

temperature is 165  C, and the initial sand temperature is 20  C. To establish the structure of the casting and the morphology of graphite, specimens were taken, as shown in Fig. 18. The specimens were then ground, polished, and etched for structure evaluation. It can be seen in the pictures of the microstructure that graphite was segregated in the form of spheroids. Because of the rapid cooling, a large amount of the metastable phase, ledeburite, was formed in the corner. The ledeburite phase reduces gradually as the cooling rate decreases. In the center of the casting, no ledeburite phase was

9

110 100

8

Solidification time

7

Nodule count

Solidification time, s

80 6

70 60

5

50

4

40

3

30 2

20

1

10 0

0 0

2 4 6 8 Distance from the cooling end X, cm

10

Solidification time and nodule count at various distances from the chill

1.0

9

0.9

8

0.8

7

0.7

6 Fraction of ledelurite

0.6

5

Fraction of ferrite

0.5

Fraction of pearlite

0.4

4

Elongation

3

0.3 0.2

2

0.1

1

0.0

0 0

Fig. 14

1

2

3 4 5 6 7 8 Distance from the cooling end X, cm

Phase fractions and elongation of the casting at various distances from the chill

9

10

Elongation, %

Fig. 13

Nodule count, 107/cm3

90

Phase fraction

The rejected carbon migrates toward graphite nodules, which are the carbon sinks. This results in carbon-depleted regions in austenite around the graphite nodules. This provides favorable sites for ferrites to nucleate, which grow as a shell around the graphite nodules. If the complete transformation of austenite is not achieved when the metastable temperature is reached, pearlite forms and grows in competition with ferrite. The ultimate goal of process modeling is to predict the final mechanical properties. The mechanical properties (hardness, tensile strength, yield strength, and elongation) of ductile iron castings are a function of composition and microstructure. The graphite shape, graphite structure, graphite amount, carbide content, and matrix structure (pearlite, ferrite) all affect the mechanical properties of ductile iron castings. As for the matrix structure, the increasing of pearlite increases the strength and hardness but reduces the elongation (Ref 98). To show the capability of this model, a simulated ductile iron casting with a simple geometry was investigated. The dimension of the casting is 10 by 10 by 200 cm. On the left face, it is cooled by contact with a constanttemperature media (15  C) at a heat-transfer coefficient of 500 W/m2K. All the other faces are adiabatic. The initial melt temperature is 1400  C. According to the boundary condition, the left side cools faster than the right side. Figure 13 shows the solidification time for different distances from the cooling end. At the very left end, the solidification time is less than 1 s. On the other hand, the solidification at 10 cm from the cooling end is more than 100 s. Because of the different cooling, the nodule count varies and is shown in the same figure. The metastable phase forms when the cooling is too fast. Figure 14 shows the volume fraction of different phases at room temperature. On the very left end, there is approximately 90% volume fraction of ledeburite phase. It reduces gradually from left to right until approximately 3 cm from the chill end. There is no ledeburite phase after 3 cm. At the same time, as cooling decreases, the volume fraction of ferrite increases and that of pearlite decreases. Ledeburite is a very hard, brittle phase. The pearlite phase is harder than ferrite. Hence, the ductility increases as the cooling rate decreases. From the micromodeling, the calculated grain and graphite radius at different distances from the chill are shown in Fig. 15. Faster cooling results in smaller grain and graphite sizes. The ratio of the radius of graphite and austenite increases as cooling decreases but reaches a constant value of approximately 0.44, even though the radius of graphite and austenite still continue to increase. This constant ratio is determined by the initial carbon content. It can determine the expansion level during solidification. Based on the microstructure, the mechanical properties can be calculated. As mentioned previously, carbide increases the yield strength but

46 / Casting Design and Performance 0.03

0.50 0.48 0.46 0.44

Radius, mm

0.02

0.40 0.38 0.01

Porosity

0.36

Grain radius

0.34

Graphite radius Rg/Ra

0.32

0.00

0.30 0

Fig. 15

Rg/Ra

0.42

2 4 6 8 Distance from the cooling end X, cm

10

Grain and graphite size of the casting at various distances from the chill

650

400 380

600 360 340

500 450

Tensile strength

320

Yield strength

300

Hardness

280

Hardness, HB

Strength, MPa

550

260

400

240 350 220 300

200 0

Fig. 16

1

2 3 4 5 6 7 8 Distance from the cooling end X, cm

9

10

Mechanical properties of the casting at various distances from the chill

found. The radius of the black balls, graphite, increases as cooling decreases. The structure of the metal is formed by pearlite and ferrite. Figure 19 shows the volume fraction of metastable phase (top) and volume fraction of ferrite (bottom). It is difficult to measure the yield strength of the sample at different locations because the strength could change dramatically based on the microstructure variation. On the other hand, hardness is an excellent indicator of strength and relatively easy to measure. Figure 20 shows the hardness measurement points on the sample. Table 2 shows the

the microstructure. There are many kinds of casting defects. Those defects are dependent on the chemistry of alloys, casting design, and casting processes. Defects can be related to thermodynamics, fluid flow, thermal, and/or stress. For most cases, all of those phenomena are correlated. It is necessary to consider every aspect to prevent a defect forming. Here, some common casting defects in foundries are discussed. They are porosity, hot tearing, and macrosegregation.

comparison between the measurement and prediction results of the hardness at different locations. It can be concluded that the prediction matches the experiments very well.

Defect Prediction Defects reduce the performance and increase the cost of castings. It is critical to understand the mechanism of defects and microstructure on the performance so that an effective tool can be developed to prevent defects and control

Porosity formed in castings leads to a decrease in the mechanical properties (Ref 100–106]. This porosity may be a combined result of solidification shrinkage and gas evolution. They can occur simultaneously when conditions are such that both may exist in a solidifying casting. One of the most effective ways to minimize porosity defects is to design a feeding system using porosity prediction modeling. In such a way, the model can determine the location of microporosity so that the feeding system can be redesigned. This process is repeated until microporosity is minimized and not likely to appear in the critical areas of the castings. There are many models that can predict the shrinkage porosity from the pressure drop during interdendritic fluid flow and gas evolution (Ref 104–111). The model of Felicelli et al. can predict the pressure and redistribution of gas and the region of possible formation of porosity by solving the transport of gas solutes (Ref 112). A comprehensive model should calculate the shrinkage porosity, gas porosity, and pore size. As for many casting defects observed in solidification processes, the mushy zone is the source of microporosity. The basic mechanism of microporosity formation is pressure drop due to shrinkage and gas segregation in the liquid (Ref 101, 102, 104). The liquid densities of many alloys are lower than that of the solid phase. Hence, solidification shrinkage happens due to the metal contraction during the phase change. The dynamic pressure within the liquid decreases because of the contraction and sometimes cannot be compensated by the metallostatic pressure associated with the height of the liquid metal. The decrease of pressure lowers the solubility of gas dissolved in the liquid. If the liquid becomes supersaturated, then bubbles can precipitate (Ref 113). Most liquid metals can dissolve some amount of gas. The solubility of gases in the solid phase is usually much smaller than that in the liquid phase. Normally, the rejected gases during solidification do not have enough time to escape from the mushy zone into the ambient air. Being trapped within the interdendritic liquid, the gas can supersaturate the liquid and eventually precipitate under the form of pores if nucleation conditions are met. The formation of bubbles requires overcoming the surface tension (Ref 114). Homogeneous

Modeling of Casting and Solidification Processing / 47 Metal die

Specimen Sand mold

Metal base Gating system

Fig. 17

Experiment setup

to shaped castings was done by Kubo and Pehlke (Ref 114). A more accurate fluid flow model was presented by Combeau et al. (Ref 105). In their study, the interdendritic flow for three-dimensional simulations of mold filling was included but without considering microporosity. Among all these models, no one calculated the diffusion and convection of gas. Felicelli et al. (Ref 112) calculated the redistribution of gas during solidification but did not predict how much porosity forms and the sizes of the pores. To predict microporosity defects in casting processes accurately, the following factors that contribute to microporosity formation should be considered: macroscopic heat transfer, interdendritic fluid flow, gas redistribution by diffusion and convection, microstructure evolution, and microporosity growth. The conservation equation beside the equations mentioned early on is gas conservation. Gas conservation: @r c @ð1  fl Þr þ rui rcl ¼ ð1  kÞcl þ rðfl rrcl Þ @t @t (Eq 38)

where c ¼ fl cl þ ð1  fl Þcs ,cs ¼ kcl ;cl is the gas concentration in the liquid;cs is the gas concentration in the solid; and k is the partition coefficient. In addition to the liquid and solid fraction, which are calculated from the energy equation, the dendrite spacing is needed to estimate pore curvature and permeability in the mushy zone. The pore radius or curvature is taken to be proportional to the dendrite cell spacing through the relationship: 1 r ¼ fl d 2

where d is the secondary dendrite arm spacing. Gas Porosity Evolution. Pores will form in a solidifying alloy when the equilibrium partial pressure of gas within the liquid exceeds the local pressure in the mushy zone by an amount necessary to overcome surface tension. Hence, gas porosity develops when:

Fig. 18

Microstructure of ductile iron at indicated points

nucleation is very difficult. In castings, nucleation of pores can be expected to occur primarily on heterogeneous nucleation sites, such as the solid-liquid interface and inclusions (Ref 114, 115). Generally speaking, there are two ways to predict the level of microporosity in castings. One is a parametric method derived from first principles by using a feeding resistance criteria function combined with macroscopic heat flow calculations (Ref 116–119). Parametric models are easy to apply to shaped castings and have been mainly directed at the prediction of centerline shrinkage. Another approach is a direct simulation method (Ref 104–108, 112, 114).

Pg > Pa þ Pm þ Pd þ Pd

Fig. 19

Simulation results of fraction of metastable phase (top) and fraction of ferrite (bottom)

They usually derive governing equations based on a set of simplifying assumptions and solve the resulting equations numerically. By combining the cellular automata technique, some models can not only predict the percentage porosity but also the size, shape, and distribution of the pores (Ref 109–111). There has been some research that attempted to understand the physics of microporosity formation, too (Ref 103). The earliest work to directly predict microporosity distribution that was general and applicable

(Eq 39)

where Pg is the Sievert pressure, Pa is the ambient pressure,Pm is the metallostatic pressure, Pd is the pressure drop due to the friction within the interdendritic liquid, Pd ¼ 2s=r is the surface tension, s is the surface tension, and r is the pore radius. The maximum dissolved gas (hydrogen or nitrogen) in the liquid, gl , or solid, gs , and the gas pressure are related through Sievert’s law: gl ¼

Kg 1=2 P gs ¼ kgl fg g

(Eq 40)

where Kg and fg are the equilibrium constant and activity coefficient, respectively, for hydrogen. The activity coefficient for hydrogen in

48 / Casting Design and Performance aluminum alloy is estimated (Ref 112, 113) as, for example: ln fH ¼

N X

ajH C j

þ

j¼1

N X

bjH ðC j Þ2

(Eq 41)

j¼1

where ajH and bjH are interaction coefficients, and C j is the concentration of solute element j. The detailed coefficients can be found in Ref 112 and 113. The equilibrium constant is calculated as: ln KH ¼ 3:039 

6198:47 T

(Eq 42)

The volume fraction of gas porosity is: cl fl þ cs ð1  fl Þ ¼ gl fl þ gs ð1  fl  fv Þ þ 

Pg fv T (Eq 43)

where fv is the volume fraction of gas porosity, and  is the gas conversion factor. If no pore has formed yet, then: cl fl þ cs ð1  fl Þ ¼ gl fl þ gs ð1  fl Þ

(Eq 44)

Shrinkage Porosity. If the pressure drops below the cavitation pressure, it is assumed that liquid feeding ceases and that the solidification shrinkage in that computational cell is compensated only by pore growth. In general, cavitation pressure is very small. When the liquid pressure drops below the cavitation pressure, the porosity is determined such that it compensates for the entire solidification shrinkage within the current time step: fvnþ1 ¼ fvn þ

rnþ1  rn rnþ1 =ð1  fvn Þ

Macrosegregation Modeling and simulation of macrosegregation during solidification has experienced explosive growth since the pioneering studies of Flemings and co-workers in the mid-1960s (Ref 120–122). Beckerman did a comprehensive review of recent macrosegregation models and their application to relevant casting industries (Ref 123). There are numerous factors that can cause macrosegregation during casting solidification processes. Those include thermal- and soluteinduced buoyancy, forced flow, solid movement, and so on. Macrosegregation models have been applied extensively in the casting industry, such as steel ingot castings, continuous and direct chill castings, nickel-base superalloy single-crystal castings (freckle simulation), and shape castings as well. Freckles have been the subject of intense research efforts for approximately 30 years due to their importance as a defect in alloy casting (Ref 124, 125). They represent a major problem in directionally solidified superalloys used in the manufacture of turbine blades (Ref 124– 128). Upward directional solidification provides an effective means of producing a columnar

microstructure with all the grain boundaries parallel to the longitudinal direction of the casting. In conjunction with a grain selector or a preoriented seed at the bottom of the casting, directional solidification is used to make entire castings that are dendritic single crystals. During such solidification, the melt inside the mushy zone can become gravitationally unstable due to preferential rejection of light alloy elements (for a partition coefficient less than unity) into the melt. Since the mass diffusivity of the liquid is much lower than its heat diffusivity, the segregated melt retains its composition as it flows upward and causes delayed growth and localized remelting of the solid network in the mush. Ultimately, a pencil-shaped vertical channel, devoid of solid, forms in the mushy zone through which low-density, highly segregated liquid flows upward as a plume or solutal finger into the superheated melt region above the mushy zone. This flow is continually fed by segregated melt flowing inside the mushy zone radially toward the channel. At the lateral boundaries of the channel, dendrite arms can become detached from the main trunk, and those fragments that remain in the channel are later observed as freckle chains. The complex convection phenomena occurring during freckle formation represent a formidable challenge for casting simulation

(Eq 45)

Experimental Validation. In this example, a set of castings with different initial hydrogen content using an iron chill plate was simulated and compared with the experiment for an A319 casting. The geometry and mesh are shown in Fig. 21. The casting is 132 mm in height, 220 mm in length, and of varying thickness. Wedges are cut horizontally at 35 mm from the bottom end and with a thickness of 12 mm. The initial pouring temperature is 750  C. Initial hydrogen contents are 0.108, 0.152, and 0.184. The experiments and simulation results are taken at different distances from the chill end. The comparison of the value of percentage porosity against local solidification time and hydrogen content between simulation and experiment is shown in Fig. 22. It shows that increasing solidification time and hydrogen content increases considerably the percentage of porosity. Numerical simulation results give excellent agreement with the measurements of percentage of porosity. The results also show that local solidification time and initial hydrogen content are very important factors influencing the formation of porosity.

Ingate

Fig. 20

Sample for hardness measurement

Table 2 Comparison between measured and predicted hardness Location

1A B 2 3 4

Dimension x, mm

4 10 50 50 50

Dimension y, mm

4 7 4 10 48

Measurement, HB

368 313 249 236 209

Simulated, HB

371 320 255 245 203

Modeling of Casting and Solidification Processing / 49 (Ref 123, 129–131). In 1991, Felicelli et al. simulated channel formation in directional solidification of Lead-tin alloys in two dimensions (Ref 132). Since then, numerous studies have been performed to simulate and predict freckling in upward directional solidification (Ref 133–142). Neilson and Incropera performed the first three-dimensional simulations of channel formation in 1993 (Ref 136). However, the coarseness of the mesh caused a serious lack of resolution and inaccuracies. Threedimensional simulations have also been performed by Poirier, Felicelli, and co-workers for both binary and multicomponent alloys (Ref 141–143). Freckle formation can be simulated with a commercial package, such as ProCAST. Figure 23 illustrates the freckle formation for a Pb10%Sn binary alloy directionally solidified in a simple geometry. A temperature gradient is initially imposed to simulate a directional solidification system. Cooling is achieved by lowering the temperatures of the upper and lower walls of the cavity at a constant rate, such that the overall temperature gradient is maintained over the height of the cavity. The lateral walls of the cavity are taken as adiabatic.

Hot Tearing Hot tearing is one of the most serious defects encountered in castings. Many studies have revealed that this phenomenon occurs in the late stage of solidification when the fraction of solid is close to one. The formation and propagation of the hot tearing have been found to be directly affected by the cooling history, the chemical composition and mechanical properties of the alloy, as well as the geometry of the casting. Various theories have been proposed in the literature on the mechanisms of hot tearing formation. Detailed reviews on the theories and

experimental observations of the formation and evolution of hot tearing can be found in Ref 144 and 145 and the references therein. Most of the existing hot tearing theories are based on the development of strain, strain rate, or stress in the semisolid state of the casting. For strain-based theory, the premise is that hot tearing will occur when the accumulated strain exceeds the ductility (Ref 146–148). The strain-rate-based theories suggest that hot tearing may form when the strain rate, or strainrate-related pressure, reaches a critical limit during solidification (Ref 149, 150). The stress-based criteria, on the other hand, assumes that hot tearing will start if the induced stress in the semisolid exceeds some critical value (Ref 151, 152). Although these theories were proposed independently as distinct theories, they can, indeed, be considered as somewhat related due to the relationship between strain, strain rate, and stress. It is such a relationship that motivates the development of a hot tearing indicator which uses the accumulated plastic strain as an indication of the susceptibility of hot tearing. This considers the evolution of strain, strain rate, and stress in the last stage of solidification. A Gurson type of constitutive model, which describes the progressive microrupture in the ductile and porous solid, is adopted to characterize the material behavior in the semisolid state. The proposed hot tearing indicator, while verified specifically for magnesium alloys, has much wider application. To reliably predict the formation and evolution of hot tearing in casting by numerical simulations, it is critical to have accurate thermophysical and mechanical properties, especially in the mushy zone. It is also essential that the solidification path of the alloy be accurately described. The prediction of the thermophysical and mechanical properties has recently become possible by using the

knowledge of the microstructure, phase fractions, and defects present in a metallic part (Ref 100). The solidification path can be obtained with the help of thermodynamic calculations of phase stability at given temperatures and compositions. A comprehensive multicomponent alloy solidification model, coupled with a Gibbs free-energy minimization engine and thermodynamic databases, has been developed to facilitate such calculations (Ref 8). With the integration of a back-diffusion model in the calculation, solidification conditions, such as cooling rate, can also be taken into account. Hot Tearing Indicator. A constitutive model used to describe the material behavior in the semisolid state is the Gurson model (Ref 153–155), which was originally developed for studying the progressive microrupture through nucleation and growth of microvoids in the material of ductile and porous solid. When the material is considered as elasticplastic, the yield condition in the Gurson model is of the form: fðs; x; T; ep ; Gu Þ ¼ F ðsÞ  Gu ðs; ep ; fv Þkðep ; T Þ ¼ 0 (Eq 46) 1=2

where F ðsÞ ¼ ð3ðs  xÞ : ðs  xÞ=2Þ is the Mises stress in terms of the deviatoric stress s ¼ s  ðtrsÞI=3. k represents the plastic flow stress due to isotropic hardening, and x denotes back stress due to kinematic hardening. The accumulated effective plastic strain is written as: ep ¼

ð t pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð2=3Þ_ep : e_ p dt

(Eq 47)

0

with e_ p ¼ g_

@f @s

(Eq 48)

and g_ being the plastic flow parameter. The Gurson coefficient, Gu, is defined as:   trðsÞ Gu ¼ 2f  q1 cosh þ f1 þ ðq1 f  Þ2 g 2k

Chill

Casting

(Eq 49)

in which, q1 is a material constant and: Percentage of porosity, %

132 mm

2.5

Porosity, %

2.0 1.5 H=0.108ppm H=0.108ppm H=0.152ppm H=0.152ppm H=0.184ppm H=0.184ppm

1.0 0.5 0.0

60 mm

Fig. 21

Geometry and mesh

150 mm

0

100

70 mm

Fig. 22

200

300 400 500 600 700 Local solidification time, s

800

900

Comparison between experiment (symbols) and calculation (lines)

50 / Casting Design and Performance

Fig. 23

Macrosegregation for directional solidification of a Pb-10%Sn binary alloy. (a) Final tin composition after solidification. (b) Cut-off view

f_nucleation

f  ¼ fv

for fv fc fu  fc f  ¼ fc þ ðfv  fc Þ for fv > fc fF  fc

(Eq 51)

with the rate of void growth defined as: p f_growth ¼ ð1  f  Þtrð_   e Þ  3f q1 trðsÞ  _ f Þ sin h ¼ gð1 2k k

(Eq 53)

where

Here, fu = 1/q1, fc is the critical void volume fraction, and fF is the failure void volume fraction. Their values should be different for different materials. In the calculation for an indicator for hot tearing, constants are used. Here, q1 = 1.5, fc = 0.15, and fF = 0.25, as used in (Ref 154). The Gurson coefficient characterizes the rapid loss of material strength due to the growth of void volume fraction, fv . When fv = fF, then f  = fu = 1/ q1 , are Gu = 0, for zero stress; that is, the stresscarrying capacity of the material vanishes. The evolution of the void volume fraction is described by the nucleation of the new void and the growth of the existing void: f_v ¼ f_nucleation þ f_growth

¼ e_ ht

(Eq 50)

(Eq 52)

In this study, the nucleation of the void is assumed to be strain controlled and is written as:

eht ¼

ð t pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð2=3Þ e_ p : e_ p dt

tc t ts

(Eq 54)

tc

is defined as the hot tearing indicator. tc represents time at coherency temperature, and ts denotes time at solidus temperature. It is observed that the hot tearing indicator is in fact the accumulated plastic strain in the semisolid region, and it corresponds to the void nucleation. Therefore, it should provide a good indication for the susceptibility of hot tearing during solidification. The value of the hot tearing indicator is determined by finite-element analysis (Ref 156). For materials described by a viscoplastic or creep model, yield condition does not exist. The function f defined in Eq 46 can be used as a potential for the inelastic flow, so that the inelastic part of the strain rate can still be given in the form of Eq 48. Experiment Validation. Cao et al. performed some experiments to study the hot tearing formation during solidification of binary Mg-Al and ternary Mg-Al-Ca alloys in a steel mold (Ref 157–160), which is shown in Fig. 24. A hot cracking susceptibility was

introduced, which is a function of maximum crack width, crack length factor, and the crack location. It was found that it is easier to have a crack at the sprue end than at the ball end. It is less likely to have a crack in the middle of the rod. Also, a longer rod is easier to crack. Figure 25 shows the simulated results of a hot tearing indicator for a Mg-2%Al alloy casting. The computed hot tearing indicator agrees very well with the experiments. Figure 26 shows the experimental results of hot tearing at the sprue end of the rods for three different alloys. The calculated hot tearing indicators are shown in Fig. 27 accordingly. It can be seen that the hot tearing is less severe as the aluminum content increases from 2% to 4% and then to 8% at the same location for the same casting with the same casting conditions. Again, the simulated hot tearing indicators agree well with the observations. The susceptibility rises sharply from pure magnesium, reaches its maximum at Mg-1%Al, and decreases gradually with further increase in the aluminum content. The hot tearing indicator is calculated at the end of the sprue for the longest rod with a different alloy composition. For comparison, the hot tearing indicators as well as a crack susceptibility coefficient (CSC), which is defined as the temperature difference between the fraction of solid

Modeling of Casting and Solidification Processing / 51

Fig. 24

Steel mold for constrained rod casting. Source: Ref 157

at 0.9 and at the end of solidification, are shown in Fig. 28. Same as the experiment, the susceptibility of hot tearing rises sharply from pure Magnesium, reaches its maximum at Mg-1% Al, and decreases gradually with further increase in the aluminum content. Similarly, different ternary magnesium alloys have different hot tearing susceptibility. The calculated CSC by this model for different alloys is shown in Fig. 29. The experimental hot tearing susceptibility (Ref 158) is included in the same figure for comparison. The addition of calcium to Magnesium-aluminum alloys can reduce the temperature range between fraction of solid at 0.9 and end of solidification, which is shown in Fig. 30; hence, susceptibility to hot tearing formation will decrease as well. It indicates that the current hot tearing indicator can predict hot tearing trends very well. The alloy chemistry, casting geometry, and cooling conditions all contribute to the formation of hot tearing, and they are included in this model directly or indirectly.

Examples of Modeling Applied in Casting Industries

Fig. 25

Hot tearing indictor for a Mg-2%Al alloy casting

Fig. 26

Close-up views of hot tearing (cracks) in the bottom rods near the sprue. (a) Mg-2%Al. (b) Mg-4%Al. (c) Mg8%Al

As mentioned earlier, radiation is very important for modeling of investment casting. In this example, radiation is discussed extensively.

High- or Low-Pressure Die Casting To demonstrate the application of modeling in the casting industry, a couple of examples are presented next. First, a die casting process is modeled, while focusing on the stress analysis of the casting and die and the gap formation between mold and part. Next, an example of modeling of an investment casting is presented.

For such casting processes, metallic molds are used, occasionally with cooling or heating channels. It is called permanent mold casting if the filling is by gravity. Otherwise, it is called die casting, for which a pressure is applied to provide rapid filling. These processes are

mainly focused on high-volume components such as automobile parts. Fluid flow and stress analysis for both casting and mold are important to other casting processes. Because the molten metal is introduced into the mold by gravity, permanent mold casting flow analysis is similar to that of sand casting. However, back pressure or trapped gas cannot be ignored, since a metal mold is not permeable, unlike that of a sand mold. The flow analysis can be complicated because of the inclusion of the effect of back pressure in mold

52 / Casting Design and Performance 200

HTI CSC

150

0.1

100

50

0.0

0 0

Fig. 28 (CSC)

20

Crack susceptibility coefficient

105

CSC

100

HTC

15

95 10 90 5 85

80

0 Mg-4Al Mg-4Al- Mg-4Al- Mg-4Al- Mg-4Al- Mg-5Al- Mg-6Al0.5Ca 1.5Ca 2.5Ca 3.5Ca 2.5Ca 2.5Ca

Fig. 29

Comparison between hot tearing susceptibility (HTC) and crack susceptibility coefficient (CSC)

1.00 Mg-4Al Mg-4Al-0.5Ca Mg-4Al-1.5Ca Mg-4Al-2.5Ca Mg-4Al-3.5Ca Mg-5Al-2.5Ca Mg-6Al-2.5Ca

Fraction of solid

0.98

0.96

0.94

0.92

0.90 420

440

460

480

500

Temperature,°C

Fig. 30

1

2 3 4 5 6 Aluminum content, wt%

7

8

Comparison between hot tearing indicator (HTI) and crack susceptibility coefficient

Hot tearing indicator in the bottom rods near the sprue. (a) 2% Al. (b) 4% Al. (c) 8% Al

Hot tearing susceptibility

Fig. 27

CSC (Ts0.9-Ts (K))

Hot tearing indicator

0.2

Solidification paths for some Magnesium alloys

520

540

560

filling. If a relatively low pressure is applied to the sealed furnace, it is called low-pressure casting process. The low pressure pushes the molten metal to fill the mold cavity slowly. For low-pressure casting filling analysis, the boundary conditions are pressure, which is a function of time instead of an inflow velocity. This is to simulate the furnace pressure controlling flow in the low-pressure casting presses. In permanent mold casting, the molds are used repeatedly. Hence, the molds develop a nonuniform temperature distribution during the initial cycles of the casting process that approaches a periodic quasi-steady-state condition. For a cyclic analysis, all casting parameters, such as the liquid pouring, dwell time, open time, and spraying conditions, must be considered for the calculation before the solidification analysis. As the modern foundry continues to evolve in implementing new technology, process modeling must also advance to meet the next hurdles facing foundry engineers. It is critical that these new hurdles — stress and deformation in the casting and mold, micro- and gas porosity, as-cast mechanical properties — be accurately predicted and corrected. Beyond simply identifying shrink porosity and fill problems, numerical tools have been developed to predict defects at a microstructural level that can be used effectively by the foundry engineer. Knowing the answers to these questions early in the manufacturing process affords significant time and cost-savings. The main goal of the casting process is to approach and achieve net shape. Large deformation or distortion in the part requires more rework, such as hot pressing, even after a heat treatment operation. To aid in the understanding of why parts deform or develop residual stresses, simulation software can now predict thermally induced stresses, which include the effects of the strength or constraint of the mold. Simulating the strength of molding components has been proven as a necessary input in determining the amount of distortion in the casting, because the main prevention against significant warpage is the mold itself.

Modeling of Casting and Solidification Processing / 53

ð ð t  dud s  gradðduÞd  b  dud    s ð xðuÞgðuÞn  dud ¼ 0 ðEq 55Þ þ

ð

c

Here, a frictionless contact is considered for simplicity. In the previous equation, O represents the geometry of casting and all the mold parts, and G represents all the contact interfaces between all parts. The body forces and surface tensions are denoted by b and t, respectively. The augmented penalty function is given by x. Thermal contact between parts is considered by adjusting the interface heat-transfer coefficient with respect to either the air gap width or the contact pressure, as computed by the mechanical contact algorithm. When the gap width is greater than zero, the adjusted heattransfer coefficient has the form: heff ¼

1 h0

1 1 þ ðhair þh rad Þ

(Eq 56)

where h0 is the initial value of the heat-transfer coefficient, hair is the conductivity of air divided by the gap width (if a vacuum is used, this term equates to zero), and hrad is the radiation heat-transfer coefficient. If the contact pressure is greater than zero, the effective heat transfer is increased linearly with that pressure up to a maximum value.

used to make jewelry. The investment casting process does not require elaborate or expensive tooling and has the ability to produce several casting in one pour. The various stages include creating the wax pattern, growing the investment shell, dewaxing, and pouring the casting. The investment casting process uses ceramic shells as molds. The ceramic shells can be preheated to very high temperatures, up to and above the liquidus temperature. Investment castings are mostly used for aerospace and medical implant applications. Investment castings are commonly made from nickel-base superalloys, titanium alloys, aluminum alloys, cobalt-base alloys, and steels. The most important application for investment casting is in the aerospace industry. Nickel-base alloys are used for jet engine structural castings as well as turbine airfoils. Structural castings normally have an equiaxed grain structure, whereas turbine airfoils have equiaxed grain, columnar grain, and single-crystal types of structure. The unique features of the investment casting process include high-temperature ceramic molds, frequently casting in a vacuum, a furnace withdrawal process for directional solidification, strict requirements for microstructure, and sometimes applying centrifugal force for mold filling. Hence, it is critical to apply some special treatments for modeling of the investment casting process. The mold of investment castings, called a shell, is one of the unusual features of the process. The outer contour of a shell mold generally follows the shape of the casting, unlike most other molds. Since the shape of the shell mold is determined by the procedure of repeatedly dipping the wax part into a ceramic slurry, followed by drying, it is somewhat freeform. It is thus difficult to generate a computer-aided design representation of the shell geometry. Hence, it is necessary to use specialized meshing techniques to create a model of the shell. Some commercial packages can generate the shell automatically based on the investment process, such that the thickness of the shell will be different at different geometrical locations, such as convex sides, concave sides, flat surfaces, or some corners.

When the casting is ejected from the die, the mechanical contact is no longer applied to the casting/die interfaces. Care must then be taken to apply an appropriate displacement constraint to prevent solid body movement. In the first example (Fig. 31), a simple Tshaped casting of A356 in an H13 mold is simulated. The effective interface heat-transfer coefficient at two different points on the casting is plotted. The top curve is from a point experiencing increasing contact pressure as the casting contracts. The middle curve is from a point where a gap is opening up between casting and mold, assuming the presence of air. The bottom curve is from that same point but assuming a vacuum. The large variation in the coefficient illustrates the importance of accounting for local conditions. In addition, this example illustrates the value of the reverse coupling of the mechanical deformations with the energy solution. This effect can be seen in Fig. 31 on the right, where the heat flux contours are plotted. The heat flux is greatest where the contact pressure is highest. Postejection springback and further relaxation after mold removal can also be tracked, completing the necessary cycle to determine final part shape (Fig. 32). Finally, heat treatment processes may further define the final net shape. By appropriately simulating the thermal loading on the part during heat treatment, any additional shape change may be captured, either independent of the casting process or as a continuance of the deformation and residual-stress evolution formed from casting.

Investment Casting The investment casting or lost wax process, as it is commonly called, is one of the oldest known manufacturing processes. The ability to produce near-net shapes leading to minimized machine cost makes the investment casting process one of the most attractive casting processes, especially for making exotic castings with expensive alloys. This process can be used to make complex shapes, from aircraft jet engine components to small, intricate castings 1.40000E+03 1.20000E+03 2 Effective heatflux, W/m . K

Coupling the stress analysis with the thermal and fluid calculation gives a clearer understanding of the physical phenomenon. By comparing the mushy-zone location and the evolution of stresses while cooling, problems such as hot tearing and cracking can be clearly indicated. Hot tearing and cracking can occur when the stresses in the local region go beyond the yield stress, thus requiring a tight coupling of the thermal and stress calculation. Another typical phenomenon involves the gap formed between casting and mold as the casting solidifies and shrinks. When this gap develops, there can be a significant reduction in heat transfer from the casting to the mold components where contact is lost (Ref 101). Conversely, there may be locations where, due to the shrinkage of the casting onto the core, an increasing contact pressure will increase the heat-transfer rate from casting to core. By accurately tracking the heat transfer, a better indication of surface shrink can be achieved. A multibody mechanical contact algorithm is employed to compute the contact and gap formation between the castings and die parts. Contact between different die parts is also considered. An automatic penalty number adjustment technique is implemented in the contact algorithm. Such a technique greatly enhances the stability and robustness of the contact computation algorithm. The variational form of the equilibrium equation with mechanical contact at any time, t, is written as (Ref 161):

Gap closed 1.00000E+03 8.00000E+02 6.00000E+02 4.00000E+02

Gap air 2.00000E+02

Gap vacuum

0.00000E+00 0

200

400

600

800

1000

Time (s)

Fig. 31

The T-bar example shows the heat flux increasing in locations of higher contact pressure and decreasing where gaps develop.

54 / Casting Design and Performance

Fig. 32

Stresses in the component (a) before and (b) after ejection

Another unique feature for investment casting has to do with solving the heat-transfer problems. In general, the model should handle heat conduction in the core and the mold, convection and radiation across the metal and mold interface, and radiation and convection at the mold outer interface. If the alloys are cast in a vacuum environment, such as for most nickeland titanium-base alloy castings, radiation heat transfer is the only method for heat loss from the mold surface to the furnace. For a radiation calculation, proper view-factor calculations are very important. Sometimes, the view factors can change, for example, during the withdrawal process for direct chill and single-crystal castings. To change the cooling rate for some investment casting processes, some insulation materials such as kaowool can be used to wrap some specific areas of the shell. For such situations, this can be represented in the model by changing the shell surface emissivity or heattransfer coefficient. The most significant mode of heat transfer in an investment casting process is through radiation. It is critical that this be fully understood and accounted for when planning the process parameters for any casting. One must

understand the radiation effects during the metal pouring and cooling cycle, because self-radiation and external casting conditions will affect cooling rates and solidification patterns. Additionally, prepouring soaking determines the shell temperatures before the pour, and if this is not taken into account accurately, it will adversely affect the desired outcome. This section delves into radiation and the downstream effects in the casting of certain process scenarios, such as heat loss in the mold before pouring, self-radiation affects on solidification, cooling of the shell and pouring castings in ambient conditions versus in an enclosed chamber (i.e., a can), and the effects of kaowool and other insulating media. These scenarios are demonstrated using computer analyses with the ProCAST casting simulation software by ESI. For a simulation program to model the physics of the process, the computer program must accurately mimic the various stages, from wax injection to the actual pouring of the molten metal and the solidification process. In addition to the mold filling and solidification analysis, advanced computer simulation programs such as ProCAST can also evaluate the thermally

induced stress in the casting and underlying grain structure and mechanical properties of the final cast part. Accurately modeling the trapped air and shrink porosity is critical to determine the validity of an existing gating system and feeder locations. Defects are usually observed in the last areas to fill, especially if the permeability of the shell is not appropriate and if the last liquid area during solidification is in the casting and not properly fed. In badly designed gating systems, the liquid metal can prematurely solidify, leading to cold laps and other such defects. To accurately model the physics, it is essential that the proper heat extraction from the casting to the investment shell is calculated. There are three modes of heat transfer: conduction, convection, and radiation. In the investment casting process, the radiative heat transfer is dominant over the other modes of heat transfer. Due to the elevated temperature of the mold relative to its surroundings and the nature of the system setup, radiative heat transfer is the dominant mode of heat transfer. Since each location on the mold surface sees a different view of its environment, each will have a different rate of heat loss. For example, faces around the outer perimeter of the mold are exposed mostly to the environment, while the view space of those facing the inside of the mold see mostly other hot regions of the casting. As a result, radiative view-factor calculations are required to accurately account for the heat-exchange variations across the geometry. The initial mold temperature and heat loss to the environment prior to casting dictate the thermal profiles that exist at the time of pour. This creates a thermal gradient from the inside to the outside of the shell, depending on the self-radiating effects of the whole assembly. Most analysis packages on the market today assume basic radiation heat-transfer calculations, making single-body assumptions that do not account for multiple bodies or shapes. Calculating view factors allows for multiple bodies and the shapes of those bodies to be involved in the radiation heat transfer to give a much more precise and accurate calculation. By not including multibody radiation, part-to-part radiation effects or self-radiation effects would not be seen. As a result, regardless of the casting orientation on the tree, each casting would heat or cool exactly the same — even if it was a casting on the end of the runner or a part that was fully surrounded by other hot parts. Planning setup design for investment casting involves visualizing the “invisible” radiation heat-transfer effects. Therefore, to further understand radiation effects, analysis tools that are able to calculate the view-factor radiation and shadowing effects are used to present various common investment casting scenarios. A simple example given subsequently shows four test bar castings, one with the effects of selfradiation and the other without. Figure 33(a), which does not have the effects of self-radiation, shows all four test bars losing the same amount

Modeling of Casting and Solidification Processing / 55 of heat at the given time. Even though the radiative heat effects are considered, there is no reflective heat back from the hot surfaces; thus, all the test bars show exactly the same temperature profile, regardless of their location in the model. Figure 33(b), on the other hand, shows the effects of self-radiation on each specimen. As expected, the interior test samples are hotter than those on the outside, because they receive reflected heat from the test bars on either side of them. The outside bars only receive heat from one side. This temperature gradient effect is also reflected in the solidification rate and porosity location and size in both the test configurations. To prevent freezing of the metal or cold flow during filling, it is desirable to pour the casting as soon as possible when the mold is removed from the furnace. This is especially so for thin-shell molds or very thin parts. For “small” castings, where the mold is typically handmoved from the furnace to the pouring bed, it is typical to have 10 to 20 s elapse before the metal is poured. Even this small amount of time can cause a significant reduction in temperature on shell faces that are open to the environment, especially if a thin-shell mold or highly conductive shell material is used. Figures 34(a) and (b) show the effects of shell cooling between the time it takes the shell to be extracted from the furnace to the beginning of metal entering the mold cavity. In the casting shown, a 25 to 30 s time delay can cause certain regions of the casting to drop 300 to 400 . Note how the shell on the end has cooled much more rapidly than the interior regions. With large shells or ones that must be moved from a furnace into a vacuum chamber, the elapsed time can be much more, and thus, the difference between internal shell temperatures and external temperatures may be quite large — potentially resulting in unexpected filling issues or patterns. In many investment castings, when the casting is poured, the shell glows, indicating the large amount of heat inside. During and after mold filling, solidification starts to take place. The energy content of the metal continually decreases by heat conduction with the cooler mold and by radiation from the mold surface. Exterior faces radiate this heat freely and allow for a relatively high rate of cooling. However, internal surfaces or locations where there is mostly part-to-part radiation are, in essence, insulating each other by radiating heat onto itself. It is quite common to have internal parts or parts involved in a high amount of radiation have a solidification pattern that is much different than parts that may be on the end of a row or more exposed to the ambient conditions. Therefore, it may be beneficial to design casting setups such that all of the parts experience the same heat transfer (that is, radiation) effects. By having the same cooling pattern, any changes to the design or rigging of the part will apply evenly to each part. Otherwise, the casting engineer may be chasing a defect that occurs in one part position that does not occur in another.

The case study shown in Fig. 35(a) and (b) demonstrates an in-line gating design. The initial setup shows nine castings stacked next to each other, with a common top-fed runner design. A coupled fluid and solidification analysis was run to understand the problems associated with molten metal flow and porosity. A quick review of the computer analysis run in ProCAST gave a very good indication on the flow pattern of the mold. The molten metal seems to fill most of the central casting, because the sprue is located right above it, then the other castings fill from the center outward. Since the central casting is already filled before the other castings, the temperature in the center casting is initially lower than the others. As the filling progresses to fill the entire mold cavity, the shell starts to become heated, thus reflecting heat to the outside. The central casting remains

much hotter than the outer casting region due to the self-radiation effects, as explained in the earlier section. A quick look at the solidification plots shows similar results. The outside castings solidify quicker than the inside ones, even though the filling sequence was reverse, where the hot metal filled the outside casting last. Figure 35(b) shows the porosity plot for the in-line gating design. The varying degree and size of porosity validates the fact that even though each part is exactly the same, the location on the tree greatly influences the solidification rate and porosity. To minimize the erratic flow and solidification pattern, a circular gating design is introduced. A ring gate on the top fed the castings from a central sprue. Due to the symmetric nature of the gating design, the flow pattern for the entire casting is very similar. Also, the

Fig. 33

Four test bar castings. (a) Without self-radiation. (b) With effects of self-radiation

Fig. 34

(a) Cooldown inside. (b) Cooldown outside

56 / Casting Design and Performance onto the part. With a casting chamber, the enclosure emissivity is controlled with specific chamber wall materials, and the temperature can be controlled with heaters or cooling systems placed on or in the chamber.

Conclusions

Fig. 35

Fig. 36

Modeling of casting and solidification has been used extensively in foundries to solve routine production problems. Casting defects such as those related to filling, solidification, stress, and microstructure can be predicted with confidence, thanks to comprehensive models and the ability to compute thermophysical and mechanical properties of multicomponent alloys. As always, better understanding and accurate material properties, including mold materials, lead to improved predictive capabilities. Further development efforts should emphasize the enhancement of the accuracy of predicting and eliminating various casting defects. Coupling heat treatment simulation with casting simulation can then predict the final mechanical properties of the part in service.

Temperature plot. (b) Porosity plot

Temperature plot with local insulation

fill rate for each part on the tree is the same, so the effects seen in the in-line design, where the central casting filled much earlier than the others, is not observed. Also, due to the circular ring gate design, the radiation view factor that each part “sees” is the same for each part, leading to a similar behavior in radiative and reflective heat from each other. An additional benefit of using a ring gate was that this design accommodated 12 castings on the tree instead of 9 in the in-line design. The filling pattern improved with the ring gates design, but the solidification pattern was similar to the one observed in the

in-line design. The porosity magnitude was reduced by approximately 30% compared to the worst porosity observed in some of the inline design castings. The casting engineer has a few variables with which to adjust and optimize when considering the rigging design. The preceding paragraphs focused on the configuration of the tree—how to place the various parts in reference to each other. The other parameters involved in radiation are also controllable by the engineer: the emissivity off of the shell faces, the temperature surrounding the shell (ambient or controlled enclosure temperature), and the emissivity of the casting environment (open-air cooling versus cooling in some chamber). Going to an extreme case, radiation can be removed completely by the depth at which the shell is buried in a sand bed. Figure 36 displays the cooling pattern on the shell for the in-line casting design when local insulation is applied on the shell. By putting a ½ or 1 in thick kaowool insulation on the critical regions that require insulation to enhance the feeding of the casting, the temperature in any area of the shell can be effectively controlled by controlling the radiation heat-transfer loss through that region. To aid a more balanced and even cooling of the exterior faces of the shell compared to the interior shell locations, adding a can or other enclosing structure can provide this effect. Technologically advanced chambers, such as those used in singlecrystal casting, are even engineered to force certain solidification behaviors and resultant microstructures by controlling the amount of radiation to and from the casting. With a can, the temperature is not controlled; however, the can itself holds in heat by reflecting some of the heat back

REFERENCES 1. K.-O. Yu, Modeling for Casting and Solidification Processing, Marcel Dekker, Inc., 2002 2. Y.A. Chang, S. Chen, F. Zhang, X. Yan, F. Xie, R. Schmid-Fetzer, and W.A. Oates, Phase Diagram Calculation: Past, Present and Future, Progress in Materials Science, Vol 49, 2004, p 313–345 3. H.L. Lukas, J. Weiss, and E.Th. Henig, Strategies for the Calculation of Phase Diagrams, CALPHAD, 1982, Vol 6 (No. 3), p 229–251 4. U.R. Kattner, The Thermodynamic Modeling of Multicomponent Phase Equilibria, JOM, 1997, Vol 49, p 14–19 5. N. Saunders and A.P. Miodownik, CALPHAD: Calculation of Phase Diagrams, A Comprehensive Guide, Elsevier Science Ltd., New York, NY, 1998, p 299–411 6. N. Saunders, Z. Guo, X. Li, A.P. Miodownik, and J.-Ph. Schille, Using JMatPro to Model Materials Properties and Behavior, JOM, 2003, Dec., p 65 7. J. Guo and M.T. Samonds, Property Prediction with Coupled Macro-Micromodeling and Computational Thermodynamics, H. Weng-Sing, Ed., Modeling of Casting & Solidification Processes, Taiwan, 2004, p 157–164 8. J. Guo and M. Samonds, Thermophysical Property Prediction Coupled Computational Thermodynamics with Back Diffusion Consideration, J. Phase Equilibrium and Diffusion, Volume 28, Number 1/ February 2007, pp 58–63 9. J. Guo and M. Samonds, The Thermophysical and Mechanical Property Prediction

Modeling of Casting and Solidification Processing / 57

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22. 23.

of Multicomponent Alloys by the Coupling of Macro-Micromodeling and Computational Thermodynamics, Modeling of Casting, Welding and Advanced Solidification Processes XI, 2006, Opio, France H.D. Brody and M.C. Flemings, Solute Redistribution during Dendrite Solidification, Trans. Met. Soc. AIME, 1966, Vol 236, p 615 T.W. Clyne and W. Kurz, Solute Redistribution during Solidification with Rapid Solid State Diffusion, Metall. Trans., 1981, Vol 12A, p 965 I. Ohnaka, Mathematical Analysis of Solute Redistribution during Solidification with Diffusion in Solid Phase, Trans. ISIJ, 1986, Vol 26, p 1045 C.Y. Wang and C. Beckermann, Unified Solute Diffusion Model for Columnar and Equiaxed Dendritic Alloy Solidification, Mater. Sci. Eng., 1993, Vol 171, p 199 L. Nastac and D.M. Stefanescu, An Analytical Model for Solute Redistribution during Solidification of Planar, Columnar and Equiaxed Morphology, Metall. Trans., 1993, Vol 24, p 2107 V.R. Voller and C. Beckermann, A Unified Model of Microsegregation and Coarsening, Metall. Trans., 1999, Vol 30, p 2183 V.R. Voller, On a General Back-diffusion Parameter, J. Cryst. Growth, 2001, Vol 226, p 562–568 M.C. Schneider, J.P. Gu, C. Beckermann, W.J. Boettinger, and U.R. Kattner, Modeling of Micro- and Macrosegregation and Freckle Formation in Single-Crystal Nickel-Base Superalloy Directional Solidification, Met. Mat. Trans. A, 1997, Vol 28A, p 1517–1531 X. Yan, “Thermodynamic and Solidification Modeling Coupled with Experimental Investigation of the Multicomponent Aluminum Alloys,” Ph.D. Thesis, University of Wisconsin-Madison, Madison, WI, 2001 C.Y. Wang and C. Beckermann, A Multiphase Solute Diffusion Model for Dendritic Alloy Solidification, Met. Trans. A, 1993, Vol 24A, p 2787–2802 I. Ohnaka, Mathematical Analysis of Solute Redistribution during Solidification with Diffusion in Solid Phase, Trans. ISIJ, 1986, Vol 26, p 1045–1052 X.L. Yang, P.D. Lee, R.F. Brooks, and R. Wunderlich, The Sensitivity of Investment Casting Simulations to the Accuracy of Thermophysical Property Values, Superalloys, The Minerals, Metals & Materials Society, 2004 B. Alkan, R. Karabulut, and B. Unal, Electrical Resistivity of Liquid Metals and Alloys, Acta Phys. Pol. A, 2002, Vol 102, p 385–400 P.L. Rossiter, The Electrical Resistivity of Metals and Alloys, Cambridge University Press, New York, NY, 1987, p 137–272

24. A. Rudajevova, M. Stanek, and P. Lukac, Determination of Thermal Diffusivity and Thermal Conductivity of Mg-Al Alloys, Mat. Sci. Eng., A, 2003, Vol A341, p 152–157 25. P.G. Klemens and R.K. Williams, Thermal Conductivity of Metals and Alloys, Int. Mat. Rev., 1986, Vol 31 (No. 5), p 197–215 26. D. Sichen, J. Bygden, and S. Seetharaman, A Model for Estimation of Viscosities of Complex Metallic and Ionic Melts, Met. Mat. Trans. B, 1994, Vol 25B, p 519–525, August 27. N. Saunders, X. Li, A.P. Miodownik, and J.P. Schille, Modeling of the Thermophysical and Physical Properties for Solidification of Al-Alloys, Light Metals, 2003, p 999–1004 28. ProCAST user’s manual and technical references, ESI, 2008 29. M. Rappaz and W.J. Boettinger, On Dendritic Solidification of Multicomponent Alloys with Unequal Liquid Diffusion Coefficients, Acta Mater., Vol. 47 (No. 11), pp. 3205–3219, 1999 30. W.J. Boettinger, S.R. Coriell, A.L. Greer, A. Karma, W. Kurz, M. Rappaz, and R. Trivedi, Solidification Microstructures: Recent Developments, Future Directions, Acta Mater., Vol 48, 2000, p 43–70 31. J. Guo and M. Samonds, Microstructures and Mechanical Properties Prediction for Ti Based Alloys, Journal of Materials Engineering and Performance, May 17, 2007 32. J. Guo and M. Samonds, Modeling of Microstructures and Mechanical Properties of a + b Ti Based Alloys, Materials Science and Technology, Volume 3, 2005, p 41–50 33. A. Kermanpur, M. Mehrara, N. Varahram, and P. Davami, Improvement of Grain Structure and Mechanical Properties of a Land Based Gas Turbine Blade Directionally Solidified with Liquid Metal Cooling Process, Materials Science and Technology, Vol 24 (No. 1), January, 2008, p 100–106 34. I. Steinbach, C. Beckermann, B. Kauerauf, Q. Li, and J. Guo, Three-Dimensional Modeling of Equiaxed Dendritic Growth on a Mesoscopic Scale, Acta Materialia, Vol. 47, 1999, pp. 971–982 35. M. Rappaz and P. H. Thevoz, Solute Diffusion Model for Equiaxed Dendritic Growth, Acta Metall., Vol 35, p 1487–1497, 1987 36. M. Rappaz and P. H. Thevoz, Solute Diffusion Model for Equiaxed Dendritic Growth: Analytical Solution, Acta Metall., Vol 35, p 2929–2933, 1987 37. P. Thevoz, J.L. Desbiolles, and M. Rappaz, Modeling of Equiaxed Microstructure Formation in Casting, Metall. Trans. A, 1989, Vol 20, p 311–322 38. M. Rappaz, Modelling of Microstructure Formation in Solidification Processes, Int. Mater. Rev., Vol 34, 1989, p 93

39. C.Y. Wang and C. Beckermann, Equiaxed Dendritic Solidification with Convection: Part I. Multiscale/Multiphase Modeling, Metall. Mater. Trans., Vol 27A, 1996, p 2754–2764 40. Ch.-A. Gandin and M. Rappaz, Coupled Finite Element-Cellular Automation Model for the Prediction of Dendritic Grain Structures in Solidification Processes, Acta Metall. Mater., Vol 42, 1994, p 2233–2246 41. Ch.-A. Gandin, J.-L. Desbiolles, M. Rappaz, and Ph. Thevoz, Three-Dimensional Cellular Automaton-Finite Element Model for the Prediction of Solidification Grain Structures, Metall. Mater. Trans., Vol 30A, 1999 p 3153–3165 42. H. Takatani, Ch.-A. Gandin, and M. Rappaz, EBSD Characterization and Modelling of Columnar Dendritic Grains Growing in the Presence of Fluid Flow, Acta Mater., Vol 48, 2000, p 675–688 43. G. Guillemot, Ch.-A. Gandin, and H. Combeau, Modeling of Macrosegregation and Solidification Grain Structures with a Coupled Cellular Automaton-Finite Element Model, Solidification Processes and Microstructures—A Symposium in Honor of Wilfried Kurz, T.M.S., Warrendale, PA, 2004, pp. 157–163 44. N. Ahmad, H. Combeau, J.-L. Desbiolles, T. Jalanti, G. Lesoult, J. Rappaz, M. Rappaz, and C. Stomp, Numerical Simulation of Macrosegregation: A Comparison Between Finite Volume Method and Finite Element Method Predictions and a Confrontation with Experiments, Mater. Trans., Vol 29A, 1998, p 617–630 45. J.-L. Desbiolles, Ph. Thevoz, and M. Rappaz, Micro-/Macrosegregation Modeling in Casting: A Fully Coupled 3D Model, Modeling of Casting, Welding and Advanced Solidification Processes X, T. M.S., Warrendale, PA, 2003, pp. 245–252 46. M. Bellet, V. D. Fachinotti, S. Gouttebroze, W. Liu, and H. Combeau, “A 3DFEM Model Solving Thermomechanics and Macrosegregation in Binary Alloys Solidification, Solidification Processes and Microstructures—A Symposium in Honor of Wilfried Kurz, T.M.S., Warrendale, PA, 2004, pp. 41–46 47. D.J. Hebditch and J.D. Hunt, Observations of Ingot Macrosegregation on Model Systems, Metall. Trans., Vol 5, 1974, p 1557 48. W. Wang, A. Kermanpur, P.D. Lee, and M. Mclean, Simulation of Dendritic Growth in the Platform Region of Single Crystal Superalloy Turbine Blades, Journal of Materials Science, Vol 8, 2003, p 4385 – 4391 49. J.S. Langer, Existence of Needle Crystals in Local Models of Solidification, Phys. Rev. A, Vol 33, 1986, p 435 50. D.A. Kessler H. Levine, Velocity Selection in Dendritic Growth, Phys. Rev. B, 1986, Vol 33, p 7867

58 / Casting Design and Performance 51. R.J. Braun, G.B. Mcfadden, and S.R. Coriell, Morphological Instability in Phase-Field Models of Solidification, Phys. Rev. E, Vol 49, p 4336–4352, 1994 52. J.A. Warren and W.J. Boettinger, Prediction of Dendritic Growth and Microsegregation Patterns in a Binary Alloy Using the Phase-Field Method, Acta Mater., Vol 43, p 689–703, 1995 53. W.J. Boettinger and J.A. Warren, The Phase-Field Method: Simulation of Alloy Dendritic Solidification during Recalescence, Metall. Mater. Trans. A, Vol 27, p 657–669, 1996 54. I. Loginova, G. Amberg, and J. Agren, Phase-Field Simulation of Nonisothermal Binary Alloy Solidification, Acta Mater., Vol 49, p 573–581, 2001 55. W.J. Boettinger and J.A. Warren, Simulation of the Cell to Plane Front Transition during Directional Solidification at High Velocity, J. Cryst. Growth, Vol 200, p 583–591, 1999 56. H.-J. Diepers, D. Ma, and I. Steinbach, History Effects during the Selection of Primary Dendrite Spacing. Comparison of Phase-Field Simulations with Experimental Observations, J. Cryst. Growth, Vol 237–239, p 149–153, 2002 57. J. Tiaden and U. Grafe, A Phase-Field Model for Diffusion and Curvature Controlled Phase Transformations in Steels, Proceedings of the International Conference on Solid-Solid Phase Transformations ’99, M. Koiwa, K. Otsuka, and T. Miyazaki, Eds., 1999 58. A. Karma and W.J. Rappel, Phase-Field Method for Computationally Efficient Modeling of Solidification with Arbitrary Interface Kinetics, Phys. Rev. E, Vol 53, p 3017–3020, 1996 59. I. Steinbach, B. Kauerauf, C. Beckermann, J. Guo, and Q. Li, Three Dimensional Modeling of Equiaxed Dendritic Growth on a Mesoscopic Scale, Proceedings of the Eighth International Conference on Modeling of Casting and Welding Processes, B. G. Thomas and C. Beckermann, Eds. San Diego, California, p 565–572, 1998 60. I. Steinbach, F. Pezzolla, B. Nestler, M. Seebelberg, R. Prieler, G.J. Schmitz, and J.L.L. Rezende, A Phase Field Concept for Multiphase Systems, Physica D, Vol 94, p 135–147, 1996 61. J. Tiaden, B. Nestler, H.J. Diepers, and I. Steinbach, The Multiphase-Field Model with an Integrated Concept for Modelling Solute Diffusion, Physica D, Vol 115, p 73–86, 1998 62. J.Q. Broughton, A. Bonissent, and F.F. Abraham, The Fcc (111) and (100) Crystal-Melt Interfaces: A Comparison by Molecular Dynamics Simulation, J. Chem. Phys., Vol 74, p 4029, 1981 63. A.A. Wheeler, W.J. Boettinger, and G. B. Mcfadden, Phase-Field Model for

64. 65.

66.

67.

68. 69. 70.

71.

72.

73.

74.

75.

76.

77. 78.

Isothermal Phase Transitions in Binary Alloys, Phys. Rev. E, Vol 45, p 7424– 9439, 1992 G. Caginalp and W. Xie, Phase-Field and Sharp-Interface Alloy Models, Phys. Rev. E, Vol 48, p 1897–1910, 1993 W.J. Boettinger, J.A. Warren, C. Beckermann, and A. Karma, Phase-Field Simulation of Solidification, Annu. Rev. Mater. Res., 2002, Vol 32, p 163–194 W.J. Boettinger, A.A. Wheeler, B.T. Murray, G.B. McFadden, and R. Kobayashi, Calculation of Alloy Solidification Morphologies Using the Phase-Field Method, Modeling of Casting, Welding and Advanced Solidification Processes, 1993, p 79 A.A. Wheeler, W.J. Boettinger, G.B. McFadden, Phase-Field Model for Isothermal Phase Transitions in Binary Alloys, Phys. Rev. A, 1992, Vol 45, p 7424 G. Caginalp and W. Xie, Phase-Field and Sharp-Interface Alloy Models, Phys. Rev. E, 1993, Vol 48, p 1897 S.G. Kim, W.T. Kim, and T. Suzuki, Phase-Field Model for Binary Alloys, Phys. Rev. E, 1999, Vol 60, p 7186 A. Karma, Y.H. Lee, and M. Plapp, Three-Dimensional Dendrite-Tip Morphology at Low Undercooling, Phys. Rev. E, 2000, Vol 61, p 3996 I. Steinbach, F. Pezzolla, B. Nestler, M. Seelberg, R. Prieler, G.J. Schmitz, et al., Phase Field Concept for Multiphase Systems, Physica D, 1996, Vol 94, p 135–147 J. Tiaden, B. Nestler, H.J. Diepers, and I. Steinbach, Multiphase-Field Model with an Integrated Concept for Modelling Solute Diffusion, Physica D, 1998, Vol 115, p 73 I. Steinbach, and F. Pezzolla, A Generalized Field Method for Multiphase Transformations Using Interface Fields, Physica D, 1999, Vol 134, p 385 J.C. Ramirez, C. Beckermann, A. Karma, and H.-J. Diepers, Phase-Field Modeling of Binary Alloy Solidification with Coupled Heat and Solute Diffusion, Physical Review E, Vol 69 (No. 5 2), May, 2004, p 051607-1 to 051607-16 J. Guo, “Modeling and Experimental Validation of Cast Iron Solidification,” 5th Decennial International Conference on Solidification Processing SP07, July 23– 25, 2007, University of Sheffield, U.K D.M. Stefanescu, Solidification and Modeling of Cast Iron — A Short History of the Defining Moments, Materials Science and Engineering A, Vol 413–414, 2005, pp. 322–333 W. Oldfield, A Qualitative Approach to Casting Solidification, ASM Trans, Vol 59, 1966, p 945 D.M. Stefanescu and S. Trufinescu, Kinetics of Solidification of Gray Iron, Z. Metall. Vol 9, 1974, p 610

79. H. Fredriksson and L. Svensson, Computer Simulation of Structure Formation and Segregation during the Solidification of Cast Iron, The Physical Metallurgy of Cast Iron, H. Fredriksson and M. Hillert, Eds., Elsevier, 1985, p 273–284 80. H. Fredriksson and L. Svensson, Simulation of Grey Cast-Iron Solidification in a Shaped Casting, Solidification Processing of Eutectic Alloys, D.M. Stefanescu, G.J. Abbaschian, and R.J. Bayuzick, Eds., The Metallurgical Soc., Warrendale, PA, 1988, p 153–162 81. D.M. Stefanescu and C. Kanetkar, Computer Modeling of the Solidification of Eutectic Alloys: The Case of Cast Iron, Computer Simulation of Microstructural Evolution, D.J. Srolovitz, Ed., The Metallurgical Soc., Warrendale, PA, 1985, p 171–188 82. J. Lacaze, M. Castro, C. Selig, and G. Lesoult, Solidification of Spheroidal Graphite Cast Irons, Modeling of Casting, Welding and Advanced Solidification Processes, V.M. Rappaz, Ed., The Metallurgical Soc., Warrendale, PA, 1991, p 473–478 83. E. Fras, W. Kapturkiewicz, and A.A. Burbielko, Micro-Macro Modeling of Casting Solidification Controlled by Transient Diffusion and Undercooling, Modeling of Casting, Welding and Advanced Solidification Processes VII, M. Croos and J. Campbell, Eds., The Metallurgical Soc., Warrendale, PA, 1995, p 679–686 84. K.C. Su, I. Ohnaka, I. Yamauchhi, and T. Fukusako, Modelling of Solidified Structure of Castings, The Physical Metallurgy of Cast Iron, H. Fredriksson and M. Hillert, Eds., Elsevier, 1985, p 181–189 85. S. Chang, D. Shangguan, and D.M. Stefanescu, Prediction of Microstructural Evolution in SG Cast Iron from Solidification to Room Temperature, Metall. Trans., Vol 22A, 1991, p 915 86. D.M. Stefanescu and C.S. Kanetkar, Computer Modeling of the Solidification of Eutectic Alloys: Comparison of Various Models for Eutectic Growth of Cast Iron, AFS Trans., 1987, pp. 139–144 87. L. Nastac and D.M. Stefanescu, Prediction of Gray-to-White Transition in Cast Iron by Solidification Modeling, AFS Trans., Vol 103, 1995, pp. 329–337 88. K.M. Pedersen, J.H. Hattel, and N. Tiedje, Numerical Modelling of Thin-Walled Hypereutectic Ductile Cast Iron Parts, Acta Materialia, Vol 54 (No. 19), November, 2006, p 5103–5114 89. G. Lesoult, M. Castro, and J. Lacaze, Solidification of Spheroidal Graphite Cast Irons — I. Physical Modelling, Acta Mater., Vol. 46 (No. 3), pp. 983–995, 1998 90. G. Lesoult, M. Castro, and J. Lacaze, Solidification of Spheroidal Graphite Cast Irons — II. Numerical Simulation,

Modeling of Casting and Solidification Processing / 59

91.

92.

93.

94.

95.

96.

97.

98. 99. 100.

101. 102.

103.

104.

Acta Mater., Vol. 46 (No. 3), pp. 997–1010, 1998 F.J. Bradley, A Stereological Formulation for the Source Term in Micromodels of Equiaxed Eutectic Solidification, Metall. Trans. B, Vol 24B, p 539, 1993 R. Vijayaraghavan and F.J. Bradley, Micro-Model for Eutectoid Phase Transformations in As-Cast Ductile Iron, Scripta Materialia, Vol. 41 (No. 11), pp.1247–1253, 1999 D. Venugopalan, Prediction of Matrix Microstructure in Ductile Iron, Transactions of the American Foundrymen’s Society, 1990, p 465 L. Nastac and D.M. Stefanescu, Prediction of Gray-to-White Transition in Cast Iron by Solidification Modeling, AFS Trans., 1995, Vol. 103, pp. 329–337 Q. Chen, E.W. Langer, and P.N. Hansen, Comparative Study on Kinetic Models and Their Effects on Volume Change during Eutectic Reaction of S.G. Cast Iron, Scand. J. Metall., 1995, Vol. 24, pp. 48– 62 M.I. Onsoien, O. Grong, O. Gundersen, and T. Skaland, Process Model for the Microstructure Evolution in Ductile Cast Iron: Part I. The Model, Metallurgical and Materials Transactions A, Vol. 30A, 1999, pp. 1053–1068 M.I. Onsoien, O. Grong, O. Gundersen, and T. Skaland, Process Model for the Microstructure Evolution in Ductile Cast Iron: Part II. Applications of the Model, Metallurgical and Materials Transactions A, Vol. 30A, 1999, pp. 1069–1079 J.R. Davis, ASM Specialty Handbook: Cast Irons, ASM International, Materials Park, OH V. Krutisˇ and J. Roucˇka, “Shrinkage of Graphitic Cast Irons,” ESI Group Internal Report J. Guo and M. Samonds, Microporosity Simulations in Multicomponent Alloy Castings, Modeling of Casting, Welding and Advanced Solidification Processes X, D.M. Stefanescu et al., Ed., Sandestin, FL, 2003, pp 303–310 J. Campbell, Castings, Butterworth-Heinemann, Oxford, United Kingdom, 1991 A.V. Kuznetsov and K. Vafai, Development and Investigation of Three-Phase Model of the Mushy Zone for Analysis of Porosity Formation in Solidifying Castings, International Journal of Heat and Mass Transfer, Vol 38 (No. 14), Sept, 1995, p 2557–2567 G.K. Sigworth and C. Wang, Mechanisms of Porosity Formation during Solidification: A Theoretical Analysis, Metallurgical Transactions B, Vol. 24B, April, 1993, pp. 349–364 D.R. Poirier, K. Yeum, and A.L. Maples, A Thermodynamic Prediction for Microporosity Formation in Aluminum-Rich

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115. 116.

117.

Al-Cu Alloys, Metallurgical Transactions A, Vol 18A, Nov. 1987, pp. 1979–1987 H. Combeau, D. Carpentier, J. Lacaze, and G. Lesoult, Modelling of Microporosity Formation in Aluminum Alloys Castings, Materials Science and Engineering, Vol A173, 1993, pp. 155–159 P.K. Sung, D.R. Poirier, S.D. Felicelli, E.J. Poirier, and A. Ahmed, Simulations of Microporosity in IN718 Equiaxed Investment Castings, Journal of Crystal Growth, Vol 226, 2001, pp. 363–377 J.D. Zhu and I. Ohnaka, Modeling of Microporosity Formation in A356 Aluminum Alloy Casting, Modeling of Casting, Welding and Advanced Solidification Processes V, 1991, pp. 435–442 S. Sabau and S. Viswanathan, Microporosity Prediction in Aluminum Alloy Castings, Metallurgical and Materials Transactions B, Volume 33B, April 2002, pp. 243–255 D. See, R.C. Atweed, and P.D. Lee, A Comparison of Three Modeling Approaches for the Prediction of Microporosity in Aluminum-Silicon Alloys, Journal of Materials Sciene, Vol 36, 2001, pp. 3423–3435 R.C. Atwood and P.D. Lee, A ThreePhase Model of Hydrogen Pore Formation during the Equiaxed Dendritic Solidification of Aluminum-Silicon Alloys, Metallurgical and Materials Transactions B, Vol. 33B, April, 2002, pp. 209–221 P.D. Lee, R.C. Atwood, R.J. Dashwood, and H. Nagaumi, Modeling of Porosity Formation in Direct Chill Cast Aluminum-Magnesium Alloys, Materials Science and Engineering, Vol A328, 2002, pp. 213–222 S.D. Felicelli, D.R. Poirier, and P.K. Sung, Model for Prediction of Pressure and Redistribution of Gas-Forming Elements in Multicomponent Casting Alloys, Metallurgical and Materials Transactions B, Vol 31B, Dec. 2000, pp. 1283–1292 P.N. Anyalebechi, Analysis and Thermodynamic Prediction of Hydrogen Solution in Solid and Liquid Multicomponent Aluminum Alloys, Light Metals, 1998, pp. 827–842 K. Kubo and R.D. Pehlke, Mathematical Modeling of Porosity Formation in Solidification, Metallurgical Transactions B, Vol 16B, June 1985, pp. 359–366 B. Chalmers, Principles of Solidification, John Wiley & Sons, Inc, 1964 E. Niyama, T. Uchida, M. Morikama, and S. Saito, Predicting Shrinkage in Large Steel Castings from Temperature Gradient Calculations, AFS Int. Cast Met. J., 1981, Vol. 6, pp. 16–22 Y.W. Lee, E. Chang, and C.F. Chieu, Modeling of Feeding Behavior of Solidifying Al-7Si-0.3Mg Alloy Plate Casting, Metall. Trans. B, 1990, Vol. 21B, pp. 715–722

118. V.K. Suri, A.J. Paul, N. El-Kaddah, and J.T. Berry, Determination of Correlation Factors for Prediction of Shrinkage in Castings — Part I: Prediction of Microporosity in Castings; A Generalized Criterion (AFS Research), AFS Transactions, Vol. 103, 1995, pp. 861–867 119. R.P. Taylor, H.T. Berry, and R.A. Overfelt, Parallel Derivation and Comparison of Feeding-Resistance Porosity Criteria Functions for Castings, HTD-Vol. 323, National Heat Transfer Conference Vol. 1, ASME, 1996, pp. 69–77 120. M.C. Flemings and G.E. Nereo, Macrosegregation: Part I, Transactions of the Metallurgical Society of AIME, 1967, Vol 239, p 1449–1461 121. M.C. Flemings, R. Mehrabian, and G.E. Nereo, Macrosegregation, Part II, Transactions of the Metallurgical Society of AIME, Vol. 242, Jan., 1968, pp. 41–49 122. M.C. Flemings and G.E. Nereo, Macrosegregation, Part III, Transactions of the Metallurgical Society of AIME, Vol. 242, Jan., 1968, pp. 50–56 123. C. Beckermann, Modeling of Macrosegregation: Application and Future Needs, International Materials Reviews, 2002, Vol. 47 (No. 5), pp 243–261 124. A.F. Giamei and B.H. Kear, On the Nature of Freckles in Nickel Base Superalloys, Metall. Trans., Vol. 1, pp. 2185– 2192, 1970 125. S.M. Copley, A.F. Giamei, S.M. Johnson, and M.F. Hornbecker, The Origin of Freckles in Unidirectionally Solidified Castings, Metall. Trans., Vol. 1, pp. 2193–2204, 1970 126. A. Hellawell, J.R. Sarazin, and R.S. Steube, Channel Convection in Partly Solidified Systems, Phil. Trans. R. Soc. Lond. A, Vol. 345, pp. 507–544, 1993 127. M.G. Worster, Convection in Mushy Layers, Ann. Rev. Fluid Mech., Vol. 29, pp. 91–122, 1997 128. T.M. Pollock and W.H. Murphy, The Breakdown of Single-Crystal Solidification in High Refractory Nickel-Base Alloys, Metall. Mater. Trans. A, Vol. 27A, pp. 1081–1094, 1996 129. J. Guo, “Three-Dimensional Finite Element Simulation of Transport Phenomena in Alloy Solidification,” The University of Iowa, Iowa City, Iowa, 2000 130. J. Guo, and C. Beckermann, ThreeDimensional Simulation of Freckle Formation during Binary Alloy Solidification: Effect of Mesh Spacing, Numerical Heat Transfer, Part A, Vol. 44, pp. 559– 576, 2003 131. M.C. Schneider, J.P. Gu, C. Beckermann, W.J. Boettinger, and U.R. Kattner, Modeling of Micro- and Macrosegregation and Freckle Formation in Single-Crystal Nickel-Base Superalloy Directional Solidification, Metallurgical and Materials

60 / Casting Design and Performance

132.

133.

134.

135.

136.

137.

138.

139.

140.

Transactions A: Physical Metallurgy and Materials Science, Vol 28A (No. 7), July 1997, p 1517–1531 S.D. Felicelli, J.C. Heinrich, and D.R. Poirier, Simulation of Freckles during Vertical Solidification of Binary Alloys, Metall. Trans. B, Vol. 22B, pp. 847–859, 1991 S.D. Felicelli, D.R. Poirier, A.F. Giamei, and J.C. Heinrich, Simulating Convection and Macrosegregation in Superalloys, JOM, Vol 49 (No. 3), March 1997, p 21–25 P.J. Prescott and F.P. Incropera, Numerical Simulation of a Solidifying Pb-Sn Alloy: The Effects of Cooling Rate on Thermosolutal Convection and Macrosegregation, Metall. Trans. B, Vol. 22B, pp. 529–540, 1991 C.S. Magirl and F.P. Incropera, Flow and Morphological Conditions Associated with Unidirectional Solidification of Aqueous Ammonium Chloride, ASME J. Heat Transfer, Vol. 115, pp. 1036–1043, 1993 D.G. Neilson and F.P. Incropera, Effect of Rotation on Fluid Motion and Channel Formation during Unidirectional Solidification of a Binary Alloy, Int. J. Heat Mass Transfer, Vol. 36, pp. 489–505, 1993 D.G. Neilson and F.P. Incropera, ThreeDimensional Considerations of Unidirectional Solidification in a Binary Liquid, Numer. Heat Transfer A, Vol. 23, pp. 1–20, 1993 H.-W. Huang, J.C. Heinrich, and D.R. Poirier, Numerical Anomalies in Simulating Directional Solidification of Binary Alloys, Numer. Heat Transfer A, Vol. 29, pp. 639–644, 1996 M.C. Schneider, J.P. Gu, C. Beckermann, W.J. Boettinger, and U.R. Kattner, Modeling of Micro- and Macrosegregation and Freckle Formation in Single-Crystal Nickel-Base Superalloy Directional Solidification, Metall. Mater. Trans. A, Vol. 28A, pp. 1517–1531, 1997 S.D. Felicelli, D.R. Poirier, and J.C. Heinrich, Macrosegregation Patterns in

141.

142.

143.

144. 145.

146.

147.

148.

149.

150.

Multicomponent Ni-Base Alloys, J. Crystal Growth, Vol. 177, pp. 145–161, 1997 S.D. Felicelli, J.C. Heinrich, and D.R. Poirier, Finite Element Analysis of Directional Solidification of Multicomponent Alloys, Int. J. Numer. Meth. Fluids, Vol. 27, pp. 207–227, 1998 S.D. Felicelli, J.C. Heinrich, and D.R. Porier, Three-Dimensional Simulations of Freckles in Binary Alloys, J. Crystal Growth, Vol. 191, pp. 879–888, 1998 S.D. Felicelli, D.R. Poirier, and J.C. Heinrich, Modeling Freckle Formation in Three Dimensions during Solidification of Multicomponent Alloys, Metall. Mater. Trans. B, Vol. 29B, pp. 847–855, 1998 J. Guo and J.Z. Zhu, Hot Tearing Prediction for Magnesium Alloy Castings, in preparation for Material Science Forum J. Guo and J.Z. Zhu, “Hot Tearing Indicator for Multi-Component Alloys Solidification,” 5th Decennial International Conference on Solidification Processing SP07, July 23–25, 2007, University of Sheffield, U.K A. Stangeland, A. Mo, M. M’Hamdi, D. Viano, and C. Davidson, Thermal Strain in the Mushy Zone Related to Hot Tearing, Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science, Vol 37 (No. 13), March, 2006, p 705–714 D.G. Eskin, Suyitno, and L. Katgerman, Mechanical Properties in the Semi-Solid State and Hot Tearing of Aluminium Alloys, Progress in Materials Science, Vol 49, 2004, p 629–711 C. Monroe and C. Beckermann, Development of a Hot Tear Indicator for Steel Castings, Materials Science and Engineering A, Vol 413–414, Dec 15, 2005, p 30–36 B. Magnin, L. Maenner, L. Katgermann, and S. Engler, Ductility and Rheology of an Al-4.5% Cu Alloy from Room Temperature to Coherency Temperature, Mater. Sci. Forum, 1996, Vol 1209, p 217–222 J. Zhang and R.F. Singer, Hot Tearing of Nickel-Based Superalloys during

151. 152.

153.

154.

155.

156.

157. 158. 159.

160.

161.

Directional Solidification, Acta Materialia, Vol 50 (No. 7), Apr 19, 2002, p 1869–1879 M. Rappaz, J.M. Drezet, and M. Gremaud, New Hot-Tearing Criterion, Metall. Mater. Trans A, 1999, Vol 30A, pp 449 D.G. Eskin, Suyitno, and L. Katgerman, Mechanical Properties in the Semi-Solid State and Hot Tearing of Aluminium Alloys, Progress in Materials Science, Vol 49 (No. 5), 2004, p 629–711 Z.L. Zhang, C. Thaulow, and J. Odegard, Complete Gurson Model Approach for Ductile Fracture, Engineering Fracture Mechanics, Vol 67 (No. 2), Sept 2000, p 155–168 J. Langlais and J.E. Gruzleski, Novel Approach to Assessing the Hot Tearing Susceptibility of Aluminium Alloys, Mater. Sci. Forum, 2000, Vol 167, p 331–337 A.L. Gurson, Continuum Theory of Ductile Rupture by Void Nucleation and Growth: Part 1 — Yield Criteria and Flow Rules for Porous Ductile Media, Journal of Engineering Materials and Technology, Transactions of the ASME, Vol 99, Ser H (No. 1), Jan, 1977, p 2–15 V. Tvergaard and A. Needleman, Analysis of the Cup-Cone Fracture in a Round Tensile Bar, Acta Metall., 1984, Vol 32, p 157 A. Needleman and V. Tvergaard, Analysis of Ductile Rupture Modes at a Crack Tip, J. Mech. Phys. Solids, 1987, Vol 35, p 151 O.C. Zienkiewicz, R.L. Taylor, and J.Z. Zhu, The Finite Element Method: Its Basis and Fundamentals, 2005, Elsevier G. Cao, S. Kou, and Y.A. Chang, Hot Cracking Susceptibility of Binary Mg-Al alloys, Magnesium Technology 2006, A. A. Luo, Ed., TMS, 2006, p 57–61 G. Cao and S. Kou, Hot Tearing of Ternary Mg-Al-Ca Alloy Castings, Met. and Mat. Trans. A, Vol 37A, Dec. 2006, pp 3647–3663 J.C. Simo and T.A. Laursen, An Augmented Lagrangian Treatment of Contact Problems Involving Friction, Comput. & Structures, Vol. 42, pp 97–116, 1992

Casting Design and Performance Pages 61–72

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Riser Design RISER DESIGN, or risering, deals with the development of suitable reservoirs of feed metal in addition to the desired casting shape so that undesirable shrinkage cavities in the casting are eliminated or moved to locations where they are acceptable for the intended application of the casting. When metals solidify and cool to form a casting, they generally undergo three distinct stages of volume contraction, or shrinkage. Exceptions to this shrinkage behavior of some graphitic cast irons are noted later in this article. These stages, shown schematically in Fig. 1, are:  Liquid shrinkage: The liquid metal loses

volume as it gives up superheat and cools to its solidification temperature.  Solidification shrinkage: The metal freezes, changing from a liquid to a higher-density solid. For pure metals, this contraction will occur at a single temperature, but for noneutectic alloys, it will take place over some temperature range or freezing interval.  Solid shrinkage: The solid casting cools from its solidification temperature to room temperature.

Visible signs of shrinkage-induced casting defects include internal shrinkage voids, surface deformation or dishing, and surface puncture. These defects will vary with different alloys; for example, internal shrinkage may be more dispersed, or alloys with strong skinforming behavior may not exhibit surface deformation. To eliminate these defects in the casting, a riser is added to accommodate the liquid shrinkage and to supply liquid feed metal to compensate for the solidification shrinkage within the casting (Fig. 3). Therefore, the shrinkage in the riser/casting system is concentrated in the riser, which will then be removed from the finished casting. As illustrated in Fig. 3, the riser may be larger than the casting it feeds, because it must supply feed metal for as long as the casting is solidifying. Various methods are used to reduce the size of the required riser, including chilling the casting, that is, reducing its solidification

time or insulating the riser, that is, extending its solidification time.

Optimum Riser Design The role of the methods engineer in designing risers can be stated simply as making sure that risers will provide the feed metal:  In the right amount  At the right place  At the right time

To this list can be added several other considerations:  The riser/casting junction should be designed

to minimize riser removal costs.

 The number and size of risers should be

minimized to increase mold yield and to reduce production costs.

Solid shrinkage, also called patternmaker’s shrinkage, is accommodated by making the pattern and, therefore; the mold cavity, somewhat larger than the desired dimensions of the final casting. Liquid shrinkage and solidification shrinkage are the concern of risering practice. In the absence of risers, a casting would otherwise solidify as shown in Fig. 2.

Schematic of the shrinkage of low-carbon steel. The contribution of each one of the three distinct stages of volume contraction is shown: liquid shrinkage, solidification shrinkage, and solid contraction. Source: Ref 1

Fig. 1

Fig. 2

Schematic of sequence of solidification shrinkage in an iron cube. (a) Initial liquid metal. (b) Solid skin and formation of shrinkage void. (c) Internal shrinkage. (d) Internal shrinkage plus dishing. (e) Surface puncture

62 / Casting Design and Performance  Riser placement must be chosen so as not to

exaggerate potential problems in a particular casting design; for example, tendencies toward hot tearing or distortion. In practice, these considerations are often in conflict, and the final riser design and pattern layout represent a compromise.

Fig. 3

Feed Metal Volume In addition to the rise volume required to satisfy liquid and solidification shrinkage in the casting, the riser itself will be solidifying, so the total shrinkage requirement will be for the riser/casting combination. The total feeding requirement will depend on the specific alloy,

Methods of controlling shrinkage in an iron cube to reduce riser size. (a) Open-top riser. (b) Open-top riser plus chill. (c) Small open-top riser plus chill. (d) Insulated riser. (e) Insulated riser plus chill

the amount of superheat, the casting geometry, and the molding medium. Liquid shrinkage depends on the alloy and the amount of superheat. Liquid shrinkage for carbon steels is generally in the range of 1.6 to 1.8%/100  C (0.9 to 1.0%/100  F) superheat. For graphitic cast irons, liquid shrinkage has been reported in the range of 0.68 to 1.8%/ 100  C (0.38 to 1.0%/100  F) (Ref 2–5). Solidification Shrinkage. Table 1 indicates that solidification shrinkage varies considerably according to the alloy melted and that, within the graphitic cast irons, expansion may occur. This phenomenon is often ascribed to the precipitation of the less dense graphite phase overcoming the contraction associated with the solidification of austenite. Theoretical calculations indicate that such density differences cannot account for the higher reported expansion percentages (Ref 2, 5). Practice shows that, with proper control of metallurgical and mold conditions, expansion phenomena can be used to reduce greatly or eliminate risers; with the liquid shrinkage accommodated by the gating system instead of the riser system (Ref 5). Mold Dilation. Mold wall movement after a mold cavity has been filled with liquid metal can enlarge the casting and thus increase the feed metal requirements. Such mold dilation is a function of the molding medium, the mold filling temperature, and the alloy. With gray and ductile irons, mold dilation may result partially from expansion pressures within the solidifying casting generated by the precipitation of graphite. In soft green sand molds, such mold dilation may produce an additional 15% feed metal requirement above that needed to satisfy the calculated liquid and solidification shrinkages (Ref 7). In copper-base alloys, it has been suggested that an additional 1% volumetric shrinkage should be expected as a result of mold cavity expansion in green sand molds (Ref 8). Casting Geometry. The shape of a casting will affect the size of the riser needed to meet its feed requirements for the obvious reason that the longer the casting takes to solidify, the longer the riser must maintain a reservoir of liquid metal. For rangy, thin-section castings,

Table 1 Solidification contraction for various cast metals Metal

Carbon steel 1% carbon steel White iron Gray iron Ductile iron Copper Cu-30Zn Cu-10Al Aluminum Al-4.5Cu Al-12Si Magnesium Zinc Source: Ref 6

Percentage of volumetric solidification contraction

2.5–3 4 4–5.5 Varies from 1.6 contraction to 2.5 expansion Varies from 2.7 contraction to 4.5 expansion 4.9 4.5 4 6.6 6.3 3.8 4.2 6.5

Table 2

Minimum volume requirements for risers on steel castings Minimum riser volume/casting volume (Vr/Vc), % Insulated risers

Type of casting

Very chunky (cubes, etc.): dimensions in ratio 1:1.33:2 Chunky: dimensions in ratio 1:24 Average: dimensions in ratio 1:3:9 Fairly rangy: dimensions in ratio 1:10:10 Rangy: dimensions in ratio 1:15:30 Very rangy: dimensions in ratio 1:>15:>30 Source: Ref 9

Sand risers

H/D = 1:1

H/D = 2:1

H/D = 1:1

H/D = 2:1

32

40

140

198

26

32

106

140

19

22

58

75

14

16

30

38

9

10

13

15

8

8

11

13

Riser Design / 63 where solidification will be rapid, feed metal requirements may be smaller than what would ordinarily be calculated. This is because a portion of the liquid and solidification shrinkages will be fed by liquid metal entering the mold from the gating system. Table 2 indicates the effect of differences in casting geometry on minimum riser volume requirements for steel castings.

Riser Location To determine the correct riser locations, the designer should make use of the concept of directional solidification. If shrinkage cavities in the casting are to be avoided, solidification should proceed directionally from those parts of the casting farthest from the riser, through

Fig. 4

Directional and progressive solidification in a casting equipped with a riser. Source: Ref 10

Schematic of mode of freezing in pure metals. Crystallization begins at the mold wall and advances into the casting interior on a plane solidification front. Source: Adapted from Ref 11

the intermediate portions of the casting, and finally into the riser itself, where the final solidification will occur. Shrinkage at each step of solidification is thus fed by liquid feed metal being drawn out of the riser. The ability to achieve such directional solidification will depend on:  The alloy and its mode of solidification  The mold medium  The casting design

Two distinct types of castings must be considered: castings with uniform wall thickness and castings with wall sections of varying thickness. Progressive and Directional Solidification. Figure 4 illustrates the interplay of progressive and directional solidification in a casting. With the mold cavity filled, solidification will generally proceed from the mold wall, where a skin of solid metal will form. As heat is lost to the mold, that skin will grow progressively inward. Two conditions serve to change the rate of this growth. At the casting edge, where the greater surface area allows more rapid transfer of heat to the mold, the solidification rate will be faster. At the riser, where the mass of the riser provides more heat, and where heat transfer to the mold is reduced at the internal angle of the riser/casting junction, the rate of skin formation will be reduced. This combination of edge effect, or end effect, and riser effect provides directional solidification.

If the wedge-shaped pattern of the solidification front begun at the casting edge can be maintained, a channel of liquid feed metal should be available throughout its progress toward the riser. If, however, the parallel advancing walls progressively solidifying in the intermediate zone begin to meet, movement of liquid feed metal will be restricted, and centerline shrinkage will result. Solidification Mode. The ability to promote and sustain directional solidification will depend greatly on the manner in which an alloy solidifies. Alloys can be classified into three types based on their freezing ranges:  C (110  C (>200  F)

 Short: liquidus-to-solidus interval 3T) with a cooling edge is the sum of the riser effect and edge effect. Several key points are shown:  The contribution from the edge effect is

generally greater than that from riser effect.

 In the absence of cooling edges, FD between

risers is dramatically reduced.

 If the maximum FD in a section is exceeded,

the edge effect will give a sound edge to its usual length, but centerline shrinkage may extend for some variable distance into the area that would ordinarily be expected to be sound because of riser effect.

Fig. 8

Schematic of intermediate mode of freezing in alloys having a moderate freezing range

Fig. 9

Forms of shrinkage porosity in the sand castings of alloys that freeze in a pasty manner

Figure 13 illustrates the same relationships in steel bars (width equals T). When compared with Fig. 12, Fig. 13 also highlights the fact that bar-shaped sections will have shorter FDs than platelike sections of the same thickness. Figure 14 shows the use of chills to extend FD. When applied to the edge of a casting section, the chill will withdraw heat rapidly, enhancing the development of directional solidification away from the edge. This will add to the length of the zone that will be sound due to end effect. In addition, if a chill is placed between risers in a casting section where there is no natural cooling edge, it can be used to establish an artificial end effect. In this way, the distance between risers can be dramatically increased, thus reducing the number of risers needed to ensure a sound casting. Such a use of chills is illustrated for a steel flange (Fig. 15). The first attempt (Fig. 15a) at subdividing this casting into feedable sections with riser placement based on the absence of end effect (except at the periphery of the flange) requires eight risers (two on the hub and six on the flange). The overlapping feeding zones of the riser (based on riser effect) cover the feeding requirements of most of the flange, but there still remain unfed regions (in which centerline shrinkage would occur) that would probably require the addition of at least one more riser on the flange. The second subdivision of the casting (Fig. 15b), using chills to establish artificial end effects, reduces the number of risers needed to only five (one on the hub and four on the flange). Such an application of chills, in addition to ensuring a sound casting, can provide economic advantages by simplifying molding and patternmaking procedures, increasing casting yield, and reducing riser removal costs. It should be reemphasized that FDs in sections of uniform thickness depend on alloy characteristics and section configuration (that is, whether barlike or platelike) to determine how far directional solidification can be sustained. If the walls of the progressively solidifying intermediate sections begin to come together, disrupting and constricting the feeding

Fig. 10

Shrinkage cavities produced by skin formation in alloys

Riser Design / 65

Feeding distance relationships in steel plates (section width greater than 3T, where T – thickness). Source: Ref 13

Fig. 12

Schematic of the effect of mold and metal variables on progressive solidification. (a) Effect of mold conductivity on solidifying metal. (b) Effect of liquidus-to- solidus range of solidifying metal. (c) Effect of conductivity on solidifying metal. (d) Effect of temperature level on solidification. Source: Ref 12

Fig. 11

channels through which feed metal can move, centerline shrinkage will occur. Once the FD of a section has been exceeded, the development of centerline shrinkage will not be overcome by increasing the size of the riser. Feeding Distance. It should also be noted that most of the data on FD is for steel. A variety

of nomograms and tables have been widely used for decades (Ref 13–16). These served the industry well, but in the 1990s it was felt these FDs were overly conservative. After researching low-alloy plate casting throughout North America, analyzing foundry results, and testing simulation software, the Steel Founders’ Society of

Feeding distance relationships in steel bars (section width equal to thickness, T ). Source: Ref 13

Fig. 13

66 / Casting Design and Performance America (SFSA) presented FD rules (Ref 17). They address the section configuration (bar or plate) by expressing riser and end zone effects as a function of width to thickness, W/T. Multiplier’s are provided for differences in alloy composition, mold materials, pouring superheats, and level of casting soundness desired. For most other metals, such precise data are not available, so their FDs are often characterized by their similarity (or lack of similarity) to carbon steel. One method of making this comparison is by the calculation of centerline feeding resistance (Ref 6). This measurement indicates that some alloys (for example, Monel) will have FDs very much like those of carbon steel, and the methods engineer can use FD tables and nomograms established for the latter. Some alloys, such as 18-8 steel, 12% Cr steel, 99.8% Cu, and 60-40 brass, will have greater FDs in similar casting sections, so a multiplier can be applied to steel-base rules. Multipliers for high-alloy steel castings have been developed for steel alloys CF-8M, CA-15, HH, HK, and HP (Ref 18). In other alloys, such as 88-10-2 bronze, 85-55-5 bronze, Al- 8Mg, and Al-4.5Cu, the centerline feeding resistance is so great that FD is virtually nonexistent unless severe chilling is used. Graphitic Cast Irons. In these materials, the crystallization of the low-density graphite as flakes or nodules should promote self-feeding behavior in the solidifying casting and allow infinite FDs as long as the early liquid feed

demand of the casting is satisfied by the gating system or by a riser. Very large gray and ductile iron castings are often produced to meet radiographic or ultrasonic soundness standards with minimal risering. The key to such practice is a rigid molding medium to minimize mold wall movement.

Castings with Sections of Varying Thicknesses. Most commercially produced castings consist of sections of varying thickness and configuration. Thicker, more slowly solidifying sections are separated from each other by thinner, rapidly solidifying sections. The heavier sections will then act as risers, supplying the feed metal demands of the lighter sections. The selection of a risering method changes from a problem of riser spacing to one of riser placement so that each of the late- solidifying sections has its feed requirements met. Therefore, the methods engineer must divide a casting into sections requiring risers by determining the feeding paths by which solidification will move directionally from early- to late- solidifying sections (Ref 19). These feed paths can often be manipulated by proper application of chilling or insulating materials to minimize risering needs. Several methods of risering isolated heavy sections are shown in Fig. 16, using a casting with two heavy sections joined by a thinner connection. In Fig. 16(a), with no risers, shrinkage develops in the two heavy sections. When an adequate riser is applied to one side (Fig. 16b), shrinkage remains unfed in the other heavy section, or hot spot, because the connecting section freezes first. A simple solution is to use risers on both sides (Fig. 16c). Two feed paths are thus established, running from the center section outward to the two risers. Two alternative methods are shown by which a single feed path can be generated. In Fig. 16(d), a chill is applied to the isolated section to reduce its solidification time below that of the connection. In Fig. 16(e), the solidification time of the connection is extended by applying an insulating or exothermic pad to the casting walls.

Duration of Liquid Feed Metal Availability

Use of chills to reduce the number of risers (T) on a steel flange casting. (a) Side and top view of the casting illustrating locations of the eight risers used when the workpiece is divided into feeding areas without considering end effects. (b) Top view of identical casting showing locations of five risers used when the workpiece is divided into feeding areas in which riser effect and end effect considerations are accounted for through the use of chills. Source: Ref 14

Fig. 15

Fig. 14

Effect of chills on feeding distance relationships in steel bars. Source: Ref 13

A variety of methods have been devised to calculate the riser size needed to ensure that liquid feed metal will be available for as long as the solidifying casting requires. Several of the most commonly used methods are discussed briefly. Shape Factor Method. Drawing on the theoretical work of Caine (Ref 20), researchers at the U.S. Naval Research Laboratory (NRL) devised a method to determine riser size by calculating a shape factor by adding the length and width of a casting section and dividing this sum by the section thickness (Ref 21). The NRL method is illustrated in Fig. 17 with the example of a 508 mm (20 in.) square plate 50 mm (2 in.) thick. According to the earlier discussion of FD, this casting (at least in steel) should be able to be made free of centerline shrinkage with a single top riser in the center of the casting. Figure 12 shows that the FD of this 50 mm (2 in.) plate, with end effect, will be 229 mm (9 in.) from the edge of the riser. If the proposed central riser is at least 50 mm (2 in.) in diameter, the FD requirements should be met.

Riser Design / 67

Procedure for determining a minimum riser size using the shape factor method. (a) The shape factor, derived from the dimensions of the casting, is used to determine the minimum ratio of riser volume to casting volume. (b) The same dimensions used to obtain the shape factor provide the data for casting volume, Vc, from which riser height and riser diameter can be determined. Dimensions given in inches. Source: Ref 21

Fig. 17

Risering of isolated heavy sections joined by a thinner section to minimize shrinkage and number of risers. (a) Workpiece with no risers. (b) Riser added to one side. (c) Risers located on both ends. (d) Chill applied to one end and riser to other end. (e) Riser used on one end and insulator or exothermic pad on opposite end

Fig. 16

Figure 17 shows the calculation of the shape factor (SF) and volume of the casting. In Fig. 17(a), one reads up from the SF (20) on the x-axis to intersect the NRL-developed curve. From this point, one reads across to the y-axis to find a minimum riser volume, Vr, needed to feed the casting volume, Vc. Figure 17(a) shows that, for this example, the required Vr/Vc ratio is 0.25; that is, the riser must be designed to contain a minimum of 25% as much metal as is required by the casting itself. The SFSA guideline (Ref 17) makes use of the data (Ref 21) to express the volume ratio SF relation similar to that shown in Fig. 17(a).

The actual curve is slightly different, being plotted from the equation (Ref 22):

in a riser as it simultaneously feeds metal into the casting and solidifies inward from its walls (Fig. 18) and simplifies it to the configuration of a cylindrical pipe (Ref 23). The size of this pipe depends on the weight of the casting section and the shrinkage percentage (both liquid and solidification shrinkages) of the casting alloy. Simple nomograms and tables are available to determine pipe sizes for a given alloy and casting weight and for a variety of ratios for the height of the pipe, Hp, divided by the diameter of the pipe, Dp. As shown in Fig. 19, the actual riser size is determined by surrounding the pipe by a riser wall, W, that is expected to solidify at the same rate as the walls of the casting. The calculation of the required wall thickness will thus vary with casting configuration:

Vr =Vc ¼ 2:51SF0:74

 For platelike castings, W = casting thickness.  For cube-shaped castings, W = 35% of the

Formation of (a) the conical type of shrinkage cavity due to (b) the accumulation of solidified layers on the outer walls of the riser

Fig. 18

(Eq 1)

The volume ratio can be calculated directly from this equation. To find the appropriate riser size, simple nomograms like the one shown in Fig. 17(b) are used; risers of various height-to-diameter ratios can be found to satisfy the calculated volume requirement. As shown in Fig. 17(b), most such nomograms assume a riser H/D ratio of approximately 1 to 1. Geometric Method. Originally devised for determining side risers for malleable iron castings, the geometric method takes the typical conical shape of the shrinkage cavity formed

edge length of the cube.

 For round or square bars, W = twice the bar

diameter or thickness. The final riser diameter, Dr, thus equals Dp + 2W. The overall riser height, Hr, is determined by adding together:  The pipe height, Hp  A middle section, Hm, above the riser con-

nection with the casting (neck) to provide a safety margin  A base section, Hb, including the depth of the riser neck

68 / Casting Design and Performance This concept was developed by Wlodawer for practical riser calculations by eliminating the need to calculate actual solidification times in favor of simply determining the relative solidification times of casting sections and risers (Ref 14). Chvorinov’s rule is thus simplified to: t

Vc Ac

(Eq 3)

The volume-to-area ratio of the casting is termed the casting modulus, Mc: Mc ¼

Fig. 19

Fig. 20

Cross section of a piping side riser designed by the geometric method

Example of a design system for tapered tall risers. Source: Ref 24

Casting yield is usually maximized by choosing a ratio for Hp/Dp of 2.5 to 1, generally resulting in a riser with an H/D ratio (above the neck) of approximately 1 to 1. Design rules for the geometric method have been further developed to design tapered risers, as shown in Fig. 20. This design not only improves casting yield but also promotes the development of definite piping behavior in the riser, ensuring that metal will feed out from riser to casting. The modulus method is based on the concept that the freezing time of a casting or a casting section can be approximated by using Chvorinov’s rule:  t ¼ k2

Vc Ac

2 ¼ k2 Mc2

(Eq 2)

where t is the freezing time of the casting, Vc is the volume of the casting, Ac is the surface area of the casting, and k is the constant governed by metal and mold properties.

Vc Ac

(Eq 4)

The freezing times of risers and castings are proportional to their respective moduli, and if the modulus of the riser, Mr, is sufficiently greater than the modulus of the casting, Mc, good feeding will be obtained. In steel, if Mr = 1.2 Mc, feeding will be satisfactory. For other skin-forming alloys, including many aluminum- and copper-base alloys, the Mr/Mc ratio of 1.2 to 1 is appropriate. With gray and ductile irons, depending on the carbon equivalent, the required Mr/Mc ratio can range from 0.8:1 to 1.2:1 because the riser may not be required to supply feed metal throughout the entire solidification time of the casting. Wlodawer simplified the modulus method by showing that many casting sections can be reduced to simple geometric shapes for which Mc can be readily found without elaborate calculations of actual surface areas and volumes (for example, for a plate section, Mc = half the plate thickness). Wlodawer further simplified the method by providing simple conversion charts, such as that shown in Fig. 21. This chart allows easy calculation of risers of various shapes if the necessary modulus is known. One note of caution must be kept in mind when using the modulus approach to riser design. This method can recommend risers that are too small on very rangy, thin-section castings. In such cases, the thermal requirements of the casting would indicate very small risers, but the demand for feed metal volume can still be substantial. For such volumecontrolled castings, the riser size arrived at through modulus calculations must be verified against riser volume charts such as that given in Table 2. Computerized Methods. Over the past 15 years, a virtual revolution has taken place regarding computerized modeling methods. Programs can be grouped into two very general categories. The first category includes empirical programs to assist with riser sizing (Ref 25–29). These programs are largely based on practical experience and run nearly instantaneously on the most basic computer platforms. The basis of the calculations is generally one or more of the manual calculation methods discussed previously coupled with measured performance data from foundry trials. These methods



usually contain algorithms to assist in calculating casting weights, section modulus, and FD. In addition, riser calculations generally require simple inputs of such factors as section weight and shape, shrinkage percentage, section modulus, mold medium, ingate location, and desired riser shape for the program to provide a variety of riser alternatives. Quite often, the riser recommendation from these types of programs represents the starting point for a more detailed analysis using the second category of programs. The second category comprises programs that simulate the entire casting process using firstprinciples physics equations. Casting filling, solidification, heat transfer, residual stress/ strain, heat treatment, and other important casting processes are modeled using complicated, detailed algorithms based on the underlying physics equations. As an example, the casting filling process for programs in this category would use the Navier-Stokes equations to predict the filling characteristics of the gating system and mold cavity. Computational power for simple personal computers has increased substantially over the past few years, which has allowed the use of these types of programs to become routine. In many foundries today (2008), every new part is modeled using casting process simulation, and the gating and risering are optimized before the casting is ever produced on the shop floor. The reference on feeding and risering (Ref 18) provides an extensive set of simulations for high-alloy steel casting.

Feeding Systems Used in Riser Design Feeding systems are widely used by the foundry industry to increase casting soundness and to reduce the cost of casting manufacture. Feeding systems reduce the rate of heat transfer from the riser to the molding medium and to the atmosphere. In riser design, three types of feeding systems are commonly used:  Riser sleeves or panels are used to insulate

the riser sidewall or riser top from the mold.

 Topping compounds are used to insulate the

top of open risers from the atmosphere.

 Breaker cores are used between the riser and

the casting to facilitate the removal of the riser from the casting; the use of these materials is illustrated in Fig. 22. Metal is transferred from the riser by three mechanisms: gravity, atmospheric pressure (or pressure applied by other means, as in die casting), and capillary action. For most casting processes, atmospheric pressure is the most important.

Feeding System Advantages Feeding systems delay solidification and the formation of a skin of solid metal in the riser.

Riser Design / 69

Schematic showing the application of feeding systems for (a) a closed-top riser and (b) an open-top riser

Fig. 22

Feeding systems, by reducing the rate of heat transfer between the riser and the mold, increase the effective modulus of a riser when compared with the geometric modulus of that riser. For a given riser modulus, a riser incorporating feeding systems will be smaller than an uninsulated riser. Less liquid metal is needed to produce a given casting, and the cost of producing that casting is reduced (Ref 30). Reducing riser size may allow a given casting to be produced in a smaller mold or may allow the production of more castings in a given mold, both resulting in lower cost. The value of feeding systems in delaying riser solidification is indicated in Table 3, in which the solidification time of a 102 mm (4 in.) diameter, 102 mm (4 in.) high riser is calculated in three alloys with various combinations of insulation on the riser sidewall and top. Also given is the percentage of total heat lost from the solidifying riser through radiation from the riser top. This highlights the fact that hot topping to reduce radiative heat loss is significantly more important in steel casting than in aluminum casting.

Fig. 21

Riser configurations and their characteristic values (Mr, Vr, D, H, and so on). Source: Ref 14

Thermal Properties of Materials This promotes atmospheric puncture of the riser, assisting in the transfer of metal from the riser to the casting. Alloys, specifically those that have a wide freezing range, benefit the most from feeding systems.

Feeding systems increase the temperature gradient between riser and casting. As a result, directional solidification from casting to riser is promoted; that is, a stronger feeding path is established.

Riser sleeves and topping compounds are classified by thermal properties as exothermic, insulating, and exothermic-insulating. The generalized thermal properties of each of these classifications are shown in Fig. 23.

70 / Casting Design and Performance Exothermic feeding systems are based on the oxidation of aluminum to produce heat. These feeding systems tend to be of relatively high density, and the matrix has thermal properties similar to those of a sand mold after the exothermic reaction has subsided. They display an initial chilling effect on the molten metal and rely on a strong exothermic reaction to remelt any metal that may have solidified. Exothermic materials tend to be used on small and intermediate-sized risers and are not generally recommended for large risers having a long solidification time. Exothermic materials also have the greatest likelihood of metal contamination. Metal-producing topping compounds are a special class of exothermic material, and they have a composition based on the thermite reaction. These materials are sometimes used on large risers. Contamination must be carefully monitored; the metal produced by the thermite reaction should not be allowed to feed into the casting. Insulating feeding systems are formulated from refractory materials to have a low density. They exhibit a very low initial chill of the molten metal and rely on their low density to minimize heat loss from the riser. Insulating feeding systems tend to be used on small to intermediate-sized risers and alloys having lower pouring temperatures. They are not

generally recommended for large risers, because these low-density materials thermally degrade when exposed to high pressure and high temperature for extended periods of time. Insulating feeding systems are least likely to cause problems with metal contamination. Exothermic-insulating feeding systems consist of exothermic materials surrounded by a highly refractory, insulating matrix. These materials are the most versatile, having a low initial chill, extended exothermic reaction, and good insulation after the exothermic reaction has subsided. Exothermic-insulating materials are used for the entire range of riser sizes. Their propensity for metal contamination lies between insulating and exothermic materials.

Riser insulation in the form of a sleeve and hot topping may be regarded as effectively decreasing the surface area of the riser. The effect of the sidewall (sleeve) insulation may be represented by a factor x and that of a hot topping by a factor y, both relative to sand (where x = 1 for sand). The factors x and y have been termed apparent surface alteration factors (ASAF). The effective modulus Mr of a cylindrical riser incorporating feeding systems on both the sidewall and top is given by:

Feeding systems are formulated to minimize the possibility of metal contamination. Depending on the formulation of the feeding system and its method of application, certain elements may be absorbed by the metal in the riser. Carbon, silicon, aluminum, oxygen, nitrogen, and sulfur may be absorbed. The widespread use of feeding systems in riser design is an indication that contamination is not normally a problem, but the possibility should not be overlooked.

Mr ¼

Heat loss from a riser as a function of time for the major classifications of feeding systems, with sand included as the basis for comparison

Table 3 Alloy cast

Steel Copper Aluminum Source: Ref 31

(Eq 7)

This should be compared with the modulus of a similar 1 to 1 sand-lined riser, which is 0.2D. The ASAF values for feeding systems are generally available from manufacturers of these products. Although the size of a riser incorporating feeding systems can be readily calculated using the approach described previously, computer programs are now widely available to speed the process of designing risers—both conventional sand-lined risers and risers incorporating feeding systems.

The thermal properties of feeding systems can be readily incorporated into the modulus method of riser calculation discussed previously (Ref 32). Cylindrical risers are the most common because this shape has a high modulus for a given volume of metal and because this shape is easy to mold. The modulus of a cylindrical riser, Mr, is given by:

Fig. 23

D2 D2 ¼ 4Dð0:65Þ þ Dð0:7Þ 3:3D

¼ 0:303D

Factors in Riser Size

Riser Necks and Breaker Cores

2

Vr R H RH ¼ ¼ Ar 2RH þ R2 2H þ R DH ¼ 4H þ D

(Eq 6)

The ASAF values of insulating and exothermic feeding system materials vary and generally range from 0.50 to 0.90. The smaller the ASAF value, the more efficient the insulation. For example, if x = 0.65 and y = 0.7, then for a cylindrical riser with a height-to-diameter ratio of 1 to 1 (where H = D):

Metal Contamination

Mr ¼

DH 4Hx þ Dy

Mr ¼

To promote directional solidification from the casting into the riser, the modulus of the riser neck, Mn, should be intermediate to the moduli of the casting and riser, Mc and Mr (Ref 14). The general rule given previously for the required riser modulus was Mr = 1.2 Mc.

(Eq 5)

where R is the riser radius, D is the riser diameter, and H is the riser height.

Effect of risering systems on the solidification times of various alloys Solidification time, min

Radiation loss through top, %

Sand riser/open end

Sleeved riser/open top

Sand riser/insulated top

Sleeved riser/insulated top

42 26 8

5 8.2 12.3

7.5 15.1 31.1

13.4 14.0 14.3

43.0 45.0 45.6

Riser Design / 71

Fig. 24

General design rules for riser necks used in iron casting applications (side view and top view, respectively). (a) General type of side riser. (b) Side riser for plate casting. (c) Top round riser. Source: Ref 33

The general rule for riser neck design—at least for skin-forming alloys—is Mn = 1.1 Mc. Once again, the graphitic cast irons are an exception because the graphite expansion phase makes it unnecessary for the riser neck to stay open for feed metal transfer throughout the entire solidification time of the casting. Depending on the chemistry of the casting, riser necks for gray and ductile irons can have moduli in the range of 0.67 to 1.1 times the casting modulus (Ref 5). General design rules for riser necks for iron castings are widely available (Fig. 24). For the economical removal of risers, breaker cores can be used between the riser and the casting, as shown in Fig. 22. Breaker cores are typically made of bonded sand or fused ceramic materials. For risers lined with sleeves, the breaker core thickness is generally 10% of the riser diameter, and the breaker core opening is generally in the range of 40 to 50% of the riser diameter. By keeping the mass of the breaker core small, the breaker core quickly reaches the temperature of the surrounding metal and does not appreciably affect the solidification of the riser.

Optimum Riser and Neck Configurations Applying the required riser to a specific casting may generate problems. For example, the necessary riser or neck size may not easily fit the casting configuration. An important problem may be created simply by attaching the riser, because the new riser/casting configuration will have its own solidification pattern, sometimes with unintended results. This is illustrated in Fig. 25, in which the attachment of a side riser at the center of the rim of a gear blank casting generates a hot spot inside the casting, where shrinkage would be expected to occur after the neck freezes. This problem can be avoided either by moving the attachment of the side riser or by using a top riser on the rim. The top riser has the added benefit of improving casting yield. ACKNOWLEDGMENT We thank Richard Williams and Tony Midea, Foseco, Inc., for their assistance in

Effect of risers and riser contacts on solidification wave fronts. Last points to solidify indicated by +. (a) Gear blank casting. (b) The side riser attached directly at rim center generates a hot spot inside the casting itself. (c) and (d) The addition of a thin section between the workpiece and the riser overcomes the problem of hot spots inside the casting. Dimensions in inches

Fig. 25

reviewing this article and revising the section “Computerized Methods.” The remainder of the article, with the exception of material related to new references, is substantially as written by L.A. Plutshack and A.L. Suschil in “Riser Design,” Casting, Vol 15, ASM Handbook, 1988. REFERENCES 1. H.F. Taylor, M.C. Flemings, and J. Wulff, Foundry Engineering, John Wiley & Sons, 1959 2. C.E. Bates and B. Patterson, Volumetric Changes Occurring During Freezing of Hypereutectic Ductile Irons, Trans. AFS, Vol 87, 1979, p 323–334

72 / Casting Design and Performance 3. B.P. Winter, T.R. Ostrom, D.H. Hartman, P.K. Trojan, and R.D. Pehlke, Mold Dilation and Volumetric Shrinkage of White, Gray, and Ductile Cast Irons, Trans. AFS, Vol 92, 1984, p 551–560 4. R.W. Heine, The Fe-C-Si Solidification Diagram for Cast Irons, Trans. AFS, Vol 94, 1986 5. S.I. Karsay, Ductile Iron III: Gating and Risering, QIT—Fer Et Titane Inc., 1981 6. R.A. Flinn, Fundamentals of Metal Casting, Addison-Wesley, 1963 7. J. Briggs, Risering Gray and Ductile Iron, 90% Yield Seminar: Gating and Risering of Iron Castings, Foseco, Inc., 1983 8. B.P. Winter, R.D. Pehlke, and P.K. Trojan, Volumetric Shrinkage and Gap Formation During Solidification of Copper-Base Alloys, Trans. AFS, Vol 91, 1983, p 81–88 9. R.W. Ruddle, Influence of Feeding Practice on Energy Conservation and Production Economics in Steel Foundries, Br. Foundryman, Sept 1978, p 197–222 10. J.L. Francis and P.G.A. Pardoe, The Feeding of Iron Castings, Applied Science in the Casting of Metals, K. Strauss, Ed., Pergamon Press, 1970 11. R.W. Ruddle, Solidification of Copper Alloys, Trans. AFS, Vol 68, 1960, p 685–690 12. R.W. Heine, C.R. Loper, Jr., and P.C. Rosenthal, Principles of Metal Casting, McGraw-Hill, 1967 13. W.S. Pellini, Factors Which Determine Riser Adequacy and Feeding Range, Trans. AFS, Vol 61, 1953, p 61–80

14. R. Wlodawer, Directional Solidification of Steel Castings, Pergamon Press, 1966 15. H.F. Bishop and W.S. Pellini, The Contribution of Riser and Chill-Edge Effects to Soundness of Cast Steel Plates, Trans. AFS, Vol 58, 1950, p 185–197 16. E.T. Myskowski, H.F. Bishop, and W.S. Pellini, Feeding Range of Joined Sections, Trans. AFS, Vol 61, 1953, p 302–308 17. Feeding and Risering Guidelines for Steel Castings, Steel Founders’ Society of America, 2001 18. O.U. Shouzhu, K.D. Carlson, and C. Beckermann, Feeding and Risering of HighAlloy Steel Castings, Metall. Trans. B, Vol 36 (No. 1), Feb 2005, p 97–116 19. R.W. Heine, Feeding Paths for Risering Castings, Trans. AFS, Vol 76, 1968, p 463–469 20. J.B. Caine, A Theoretical Approach to the Problem of Dimensioning Risers, Trans. AFS, Vol 56, 1948, p 492–501 21. E.T. Myskowski, H.F. Bishop, and W.S. Pellini, A Simplified Method of Determining Riser Dimensions, Trans. AFS, Vol 63, 1955, p 271–281 22. W.D. Spiegelberg, “Computation of Solidification Gradients in Cast Steel Sections,” Ph.D. thesis, Department of Metallurgy and Materials Science, Case Western Reserve University, 1970 23. R.W. Heine, Riser Design for Mold Dilation, Mod. Cast., Feb 1965 24. R.W. Heine, Design Method for Tapered Riser Feeding of Ductile Iron Castings in

25.

26. 27.

28.

29.

30. 31. 32. 33.

Green Sand, Trans. AFS, Vol 90, 1982, p 147–158 R.W. Ruddle and A.L. Suschil, Riser Sizing by Microcomputer, Modeling of Casting and Welding Processes II, 1983 Engineering Foundation Conference, Conference Proceedings, The Metallurgical Society, 1983, p 403–419 C.F. Corbett, Methoding of Castings Using a Microcomputer, Br. Foundryman, Aug 1983, p 67–78 G.L. Moffat, FEEDERCALC—The Development of a Micro-Computer Program to Determine Optimum Feeder Sizes for Ductile Iron Castings, Foundry Pract., No. 210, Jan 1985, p 3–8 J.E. Pickin, The Development of a Computer Assisted Feeding System, Proceedings of 1981 Annual Conference, Steel Castings Research and Trade Association, 1981 W.T. Adams and K.W. Murphy, Optimum Full Contact Top Risers to Avoid Severs Under Riser Chemical Segregation in Steel Casting, Trans. AFS, Vol 88, 1980, p 389–404 R.W. Ruddle, Profit Through Steel Risering Technology, Trans. AFS, Vol 87, 1979, p 423–432 C.M. Adams, Jr., and H.F. Taylor, Fundamentals of Riser Behavior, Trans. AFS, Vol 61, 1953, p 686 R.W. Ruddle, “Risering of Steel Castings,” Foseco, Inc., 1979 J.F. Wallace and E.B. Evans, Risering of Gray Iron Castings, Trans. AFS, Vol 66, 1958, p 49

Casting Design and Performance Pages 73–80

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Gating Design A GATING SYSTEM is the conduit network through which liquid metal enters a mold and flows to fill the mold cavity, where the metal can then solidify to form the desired casting shape. The basic components of a simple gating system for a horizontally parted mold are shown in Fig. 1. A pouring cup or a pouring basin provides an opening for the introduction of metal from a pouring device. A sprue carries the liquid metal down to join one or more runners, which distribute the metal throughout the mold until it can enter the casting cavity through ingates.

Design Variables Methods used to promote any of the desirable design considerations discussed subsequently often conflict with another desired effect. For example, attempts to fill a mold rapidly can result in metal velocities that promote mold erosion. As a result, any gating system will generally be a compromise among conflicting design considerations, with the relative importance of the consideration being determined by the specific casting and its molding and pouring conditions. Rapid mold filling can be important for several reasons. Especially with thin-section

Fig. 1

Basic components of a simple gating system for a horizontally parted mold

castings, heat loss from the liquid metal during mold filling may result in premature freezing, producing surface defects (for example, cold laps) or incompletely filled sections (misruns). Superheating of the molten metal will increase fluidity and retard freezing, but excessive superheat can increase problems of gas pickup by the molten metal and exaggerate the thermal degradation of the mold medium as well as increase costs. In addition, the mold filling time should be kept shorter than the mold producing time of the molding equipment to maximize productivity. Minimizing Turbulence. Turbulent filling and flow in the gating system and mold cavity can increase mechanical and thermal attack on the mold. More important, turbulence may produce casting defects by promoting the entrainment of gases into the flowing metal. These gases may by themselves become defects (for example, bubbles), or they may produce dross or inclusions by reacting with the liquid metal. Turbulent flow increases the surface area of the liquid metal exposed to air within the gating system. The susceptibility of different casting alloys to oxidation varies considerably. For those alloys that are highly sensitive to oxidation, such as aluminum alloys; magnesium alloys; and silicon, aluminum, and manganese bronzes, turbulence can generate extensive oxide films that will be churned into the flowing metal, often causing unacceptable defects. Avoiding Mold and Core Erosion. High flow velocity or improperly directed flow against a mold or core surface may produce defective castings by eroding the mold surface (thus enlarging the mold cavity) and by entraining the dislodged particles of the mold to produce inclusions in the casting. Removing Slag, Dross, and Inclusions. This factor includes materials that may be introduced from outside the mold (for example, furnace slags and ladle refractories) and those that may be generated inside the system. Methods can be incorporated into the gating system to trap such particles (for example, filters) or to allow them time to float out of the metal stream before entering the mold cavity. Promoting Favorable Thermal Gradients. Because the last metal to enter the mold cavity will generally be the hottest, it is usually desirable to introduce metal in those parts of the casting that would already be expected to be

the last to solidify. One obvious method of accomplishing this is to direct the metal flow from the gating system into a riser, from which it then enters the mold cavity. Because the riser is designed to be the last part of the riser/casting system to solidify, such a gating arrangement will help promote directional solidification from the casting to the riser. If the gating system cannot be designed to promote desirable thermal gradients, it should at least be designed so that it will not produce unfavorable gradients. This will often involve introducing metal into the mold cavity through multiple ingates so that no one location becomes a hot spot. Maximizing Yield. A variety of unrecoverable costs reside with the metal that will fill the gating system and risers. These components must then be removed from the casting and generally returned for remelt, where their value is downgraded to that of scrap. Production costs can be significantly reduced by minimizing the amount of metal contained in the gating system. The production capacity of a foundry can also be enhanced by increasing the percentage of salable castings that can be produced from a given volume of melted metal. Economical Gating Removal. Costs associated with the cleaning and finishing of castings can be reduced if the number and size of ingate connections with the casting can be minimized. Again, it may be advantageous to introduce metal into the mold cavity through a riser, because the riser neck can also serve as an ingate. Avoiding casting distortion is especially important with rangy, thin-wall castings, in which uneven distribution of heat as the mold cavity is filled may produce undesirable solidification patterns that cause the casting to warp. In addition, the contraction of the gating system as it solidifies can pull on sections of the solidifying casting, resulting in hot tearing or distortion. Compatibility with Existing Molding/ Pouring Methods. Modern high-production molding machines and automated pouring systems often severely limit the flexibility allowed in locating and shaping the pouring cup and sprue for introducing metal into the mold. They also generally place definite limits on the rate at which metal can be poured. Controlled Flow Conditions. A steady flow rate of metal in the gating system should be

74 / Casting Design and Performance established as soon as possible during mold filling, and the conditions of flow should be predictably consistent from one mold to the next.

Principles of Fluid Flow Proper design of an optimized gating system will be made easier by the application of several fundamental principles of fluid flow. Numerous software packages employ these design principles. Chief among these fluid flow principles are Bernoulli’s theorem, the law of continuity, and the effect of momentum.

Bernoulli’s Theorem This basic law of hydraulics relates the pressure, velocity, and elevation along a line of flow in a way that can be applied to gating systems. The theorem states that, at any point in a full system, the sum of the potential energy, kinetic energy, pressure energy, and frictional energy of a flowing liquid is equal to a constant. The theorem can be expressed as: wZ þ wP v þ

wV 2 þ wF ¼ K 2g

(Eq 1)

where w is the total weight of the flowing liquid, Z is the height of the liquid, P is the static pressure in liquid, v is the specific volume of liquid, g is the acceleration due to gravity (9.807 m/s2, or 386.4 in./s2), V is the velocity, F is the friction loss per unit weight, and K is a constant. If Eq 1 is divided by w, all the terms reduce to dimensions of length and will represent:    

Potential head Z Pressure head, Pv Velocity head V2/2g Frictional loss of head F

Equation 1 allows prediction of the effect of the several variables at different points in the gating system, although several conditions inherent in foundry gating systems complicate and modify its strict application. For example:  Equation 1 is for full systems, and at least at

the start of pouring, a gating system is empty. This indicates that a gating system should be designed to establish as quickly as possible the flow conditions of a full system.  Equation 1 assumes an impermeable wall around the flowing metal. In sand foundry practice, the permeability of the mold medium can introduce problems, for example, air aspiration in the flowing liquid.  Additional energy losses due to turbulence or to friction (for example, because of changes in the direction of flow) must be accounted for.

Fig. 2

Schematic illustrating the application of Bernoulli’s theorem to a gating system. Source: Ref 1

 Heat loss from the liquid metal is not consid-

ered, which will set a limit on the time over which flow can be maintained. Also, solidifying metal on the walls of the gating system components will alter their design while flow continues. Bernoulli’s theorem (Eq 1) is illustrated in Fig. 2, and several practical interpretations can be derived. The potential energy is obviously at a maximum at the highest point in the system, that is, the top of the pouring basin. As metal flows from the basin down the sprue, potential energy changes to kinetic energy as the stream increases in velocity because of gravity. As the sprue fills, a pressure head is developed. Once flow in a filled system is established, the potential and frictional heads become virtually constant, so conditions within the gating system are determined by the interplay of the remaining factors. The velocity is high where the pressure is low, and vice versa.

The Law of Continuity This law states that, for a system with impermeable walls and filled with an incompressible fluid, the rate of flow will be the same at all points in the system. This can be expressed as: Q ¼ A1 v1 ¼ A2 v2

(Eq 2)

where Q is the rate of flow, A is the cross-sectional area of the stream, v is the velocity of the stream, and the subscripts 1 and 2 designate two different locations in the system. Again, the permeability of sand molds can complicate the strict application of this law, introducing potential problems into the casting process.

One practical implication of the law of continuity is shown in Fig. 3, which illustrates the flow of metal from a pouring basin. Due to gravitational acceleration, velocity increases as the stream falls, so the cross-sectional area of the stream must decrease proportionately to maintain the balance of the flow rate. The result is the tapered shape typical of a free-falling stream shown in Fig. 3(a). If the same flow is directed down a straightsided sprue (Fig. 3b), the falling stream will create a low-pressure area as it pulls away from the sprue walls and will probably aspirate air. In addition, the flow will tend to be uneven and turbulent, especially when the stream reaches the base of the sprue. The tapered sprue shown in Fig. 3(c) is designed to conform to the natural form of the flowing stream and therefore reduces turbulence and the possibility of air aspiration. It also tends to fill quickly, establishing the pressure head characteristic of the full-flow conditions required by Eq 1. Many types of high-production molding units do not readily accommodate tapered sprues, so the gating system designer will try to approximate the effect of a tapered sprue by placing a restriction, or choke, at or near the base of the sprue to force the falling stream to back up into the sprue (Fig. 4).

Momentum Effects Newton’s first law states that a body in motion will continue to move in a given direction until some force is exerted on it to change its direction. Reynold’s Numbers and Types of Flow. The flow of liquids can be characterized by the Reynold’s number:

Gating Design / 75

Schematic showing the advantages of a tapered sprue over a straight-sided sprue. (a) Natural flow of a free-falling liquid. (b) Air aspiration induced by liquid flow in a straight-sided sprue. (c) Liquid flow in a tapered sprue

Fig. 3

Schematic showing the formation of lowpressure areas due to abrupt changes in the cross section of a flow channel. (a) Sudden enlargement of the channel. (b) Sudden reduction of the channel

Fig. 6

Choke mechanisms incorporated into straightsided sprues to approximate liquid flow in tapered sprues. (a) Choke core. (b) Runner choke

Fig. 4

NR ¼

vdr m

(Eq 3)

where NR is the Reynold’s number, v is the velocity of the liquid, d is the diameter of the liquid channel, r is the density of the liquid, and m is the viscosity of the liquid. As shown in Fig. 5, if the Reynold’s number is less than 2000, the flow is characterized as laminar, with the molecules of the liquid tending to move in straight lines without turbulence (Fig. 5a). For NR between 2000 and 20,000, some mixing and turbulence will occur (Fig. 5b), but a relatively undisturbed boundary layer will be maintained on the surface of the stream. This type of turbulent flow, common in most foundry gating systems, can be considered relatively harmless so long as the surface is not

Reynold’s number, NR, and its relationship to flow characterization. (a) NR < 2000, laminar flow. (b) 2000 NR < 20,000, turbulent flow. (c) NR 20,000, severe turbulent flow

Fig. 5

ruptured, thus avoiding air entrainment in the flowing stream. With NR of 20,000 or greater, flow will be severely turbulent (Fig. 5c). This will lead to rupturing of the stream surface, with the strong likelihood of air entrainment and dross formation as the flowing metal reacts with gases. Abrupt Changes in Flow Channel Cross Section. As shown in Fig. 6, low-pressure zones—with a resulting tendency toward air entrainment—can be created as the metal stream pulls away from the mold wall. With a sudden enlargement of the channel (Fig. 6a), momentum effects will carry the stream forward and create low-pressure zones at the enlargement. With a sudden reduction in the channel (Fig. 6b), the law of continuity shows that the stream velocity must increase rapidly. This spurting flow will create a low-pressure

zone directly after the constriction. The problems illustrated in Fig. 6 can be minimized by making gradual changes in the flow channel cross section; abrupt changes should be avoided. Abrupt Changes in Flow Direction. As shown in Fig. 7, sudden changes in the direction of flow can produce low-pressure zones, as described previously. Problems of air entrainment can be minimized by making the change in direction more gradual. Abrupt changes in flow direction, in addition to increasing the chances of metal damage, will increase the frictional losses during flow. As shown in Fig. 8, a system with high frictional losses will require a greater pressure head to maintain a given flow velocity. Using a Runner Extension. Use of a runner extension beyond the last ingate is illustrated in Fig. 1. The first metal entering the gating system will generally be the most heavily damaged by contact with the mold medium and with air as it flows. To avoid having this metal enter the casting cavity, momentum effects can be used to carry it past the ingates and into the runner extension. The ingates will then fill with the cleaner, less damaged metal that follows the initial molten metal stream. Equalizing flow through ingates by decreasing the runner cross section after the ingate is illustrated in Fig. 9; this is done for systems with multiple ingates. As noted earlier, in the filled system shown, potential and frictional energies become constants, so they can be dropped from consideration in Eq 1 to show the interaction of pressure and velocity effects. At the first ingate, velocity is high as momentum effects carry the flowing stream past the gate. At the second ingate, velocity decreases in the runner as it reaches the end,

76 / Casting Design and Performance

Schematic illustrating fluid flow around right-angle and curved bends in a gating system. (a) Turbulence resulting from a sharp corner. (b) Metal damage resulting from a sharp corner. (c) Streamlined corner that minimizes turbulence and metal damage

Fig. 7

Effect of pressure head and change in gate design on the velocity of metal flow. A, 90 bend; B, r/d = 1; C, r/d = 6; D, multiple 90 bends. The variables r and d are the radius of curvature and the diameter of the runner, respectively. Source: Ref 2

Fig. 8

causing higher pressure and resulting in higher flow through the gate. By stepping down the runner after the first ingate, metal velocities and pressures at the two ingates can be equalized. This effect can be achieved by gradually tapering the runner to a smaller cross section along its length, but patternmaking limitations usually make it simpler to incorporate actual steps in the runner.

Design Considerations In applying fluid flow principles to the design of a specific gating system, several design decisions must be made before the actual sizes of the various components can be calculated.

Runner and Ingate. Figures 1 and 2 show gating systems with ingates coming off the top of the runner and then into the casting. This arrangement of cope ingates and drag runners is common and has the advantages that the runner will be full before metal enters the ingates. This establishes the full-flow conditions discussed earlier in this article. A full runner will reduce turbulence and will help to allow any low-density inclusions in the flowing stream to float out and attach themselves to the mold wall. A system of cope runners and drag ingates (or ingates coming off the base of a cope runner) is also common and has strong proponents (Ref 3). The basis of this design is that momentum effects will carry the first metal past the ingates, and if the runner can be quickly filled

(at least above the level of the ingates), clean metal will flow from the bottom of the runner, while inclusions carried along in the metal stream will float above the ingates. A common element of this system is that the total cross-sectional area of the ingates should be smaller than the cross-sectional area of the runner. Such a pressurized system is intended to force the metal to back up at the ingates and rapidly fill the runner, although complete filling of a cope runner will often depend on at least partial filling of the casting. During this period of incomplete filling, turbulence and the potential for air entrainment and dross generation are increased. Techniques for prepriming and countergravity flow are discussed in the article “Filling and Feeding System Concepts” in Casting, Volume 15 of the ASM Handbook. Pressurized versus Unpressurized Systems. The difference between these two systems is in the choice of the location of the flow-controlling constriction, or choke, that will determine the ultimate flow rate for the gating system. This decision involves the determination of a desired gating ratio, that is, the relative cross-sectional areas of the sprue, runner, and gates. This ratio, numerically expressed in the order sprue:runner:gate, defines whether a gating system is increasing in area (unpressurized) or constricting (pressurized). Common unpressurized gating ratios are 1:2:2, 1:2:4, and 1:4:4. A typical pressurized gating ratio is 4:8:3. Both types of systems are widely used. The unpressurized system has the advantage of reducing metal velocity in the gating system as it approaches and enters the casting. Lower velocities help encourage laminar (or less turbulent) flow, so unpressurized systems are recommended for alloys that are highly sensitive to oxide and dross formation. Pressurized systems generally have the advantage of reduced size and weight for a given casting, thus increasing mold yield. The single greatest disadvantage of pressurized systems is that, by design, stream velocities are highest at the gates just as the metal enters the casting. This increases the potential for mold or core erosion and places a premium on proper location of ingates to minimize such damage. Vertical versus Horizontal Gating Systems. This decision may be imposed by the orientation of the mold parting line. Figure 1 shows conventional, horizontally parted molds with the gating system most conveniently arranged along the parting line. Vertically parted molds impose a vertical placement of gating components, but the design considerations for these components are often the same as those for horizontal systems, for example, the desirability of tapered sprues and runner extensions. In addition, the need for stepped-down sprues to equalize flow through multiple ingates is, if anything, more critical than in a horizontal system. This element of a properly designed vertical gating system is illustrated in Fig. 10.

Gating Design / 77

Properly designed bottom gate that ensures smooth filling of the casting while producing minimal turbulence

Fig. 11

Applying Bernoulli’s theorem to flow from a runner at two ingates for a filled system and comparing velocity and pressure at the ingates for two runner configurations. (a) Same runner cross section at both ingates. (b) Stepped runner providing two different runner cross sections at each ingate. Source: Ref 1

Fig. 9

system necessary to deliver the metal with the minimum flow rate required, which in turn determines the cross-sectional area of the choke. Once that is calculated, the rest of the components are easily calculated, moving downstream from the choke in an unpressurized system or upstream from the choke (that is, the gates) in a pressurized system.

Ceramic Filters in Gating Design Ceramic filters are extensively used in the foundry industry to improve casting cleanliness and to reduce the cost of casting manufacture. Incorporated into the gating system, ceramic filters remove slag, dross, and other nonmetallic particles from the metal stream before the metal enters the mold cavity. Most casting alloys are subject to the presence of particles that can deleteriously affect the physical properties and appearance of the casting. These particles commonly include:  Oxides formed during melting, metal trans-

fer, and pouring

Fig. 10

Comparison of flow patterns in two vertical gating systems. (a) Poorly designed system. (b) Properly designed system using a tapered runner that equalizes flow through the ingates

 Refractory particles from the furnace and

ladle

 Refractory particles present in the gating

One form of vertical gating common in both vertically and horizontally parted molds is called bottom gating and countergravity filling. Elements of a properly designed bottom gating system are shown in Fig. 11. This method has the particular advantage of introducing metal into the casting cavity at its lowest point, thus helping to ensure smooth filling of the casting with minimal turbulence. Numerous examples of calculations for gating system design are

available in the References cited at the end of this article and are not covered in detail here. Also see the article “Filling and Feeding System Concepts” in Casting, Volume 15 of the ASM Handbook. Optimum mold filling times are determined by such factors as metal type, casting weight, and typical casting section thickness. Once the optimum time is established, principles of fluid flow are used to determine the size of the

system or dislodged from the mold or cores during pouring  Reaction products from metallurgical operations  Undissolved metallic or nonmetallic particles made as additions to the molten metal for metallurgical modifications These particles, or inclusions, act as discontinuities in the metal matrix of a casting and can have a variety of adverse effects:

78 / Casting Design and Performance  Large inclusions can reduce mechanical prop    

erties such as tensile strength and elongation. Fatigue life can be reduced (Ref 4–6). Machining can become more difficult, and the rate of tool wear can increase. Surface finish can deteriorate in appearance. Lack of pressure tightness can occur. Subsequent surface treatments such as anodizing or ceramic coating can be adversely affected.

The conventional approach taken to remove inclusions from the metal stream is through gating system design. Using this approach, gating systems are designed to encourage the separation of particles from the metal due to the density difference between the metal and the inclusion (Table 1). Both pressurized and unpressurized gating systems are used for this purpose. To be effective, the gating system must have sufficient length to allow the lower-density particles sufficient time to float and adhere to the mold surface before they can enter the casting cavity. In practice, this approach does not always provide castings of adequate quality, and casting yield is often reduced, especially with unpressurized systems. Filter Advantages. Ceramic filters, when correctly applied, can be relied on to trap particles before they can enter the casting cavity. This feature can result in the following advantages:

   

Machine stock allowances can be reduced. Casting physical properties can be increased. Casting surface finish can be improved. Reliability of the casting process can be increased.

designed to incorporate a ceramic filter are shown in Fig. 13. By relying on a filter to remove particles from the metal stream, the gating system can be designed with the following features and advantages:

The use of a ceramic filter in a gating system of conventional design can be effective in reducing inclusion-related defects, but a gating system designed specifically to incorporate ceramic filters will be more effective. Typical gating systems for horizontally parted molds

 The system can be unpressurized, resulting

Two crankshafts produced using a ceramic foam filter positioned vertically in the drag just downstream of the sprue. Casting yield is 91%.

Fig. 12

in reduced metal velocity as metal enters the casting cavity.  Runners can be reduced in size—both length and cross-sectional area—thus increasing casting yield.  Runners are in the drag, resulting in more rapid, complete filling of the runner and reduced opportunity for metal oxidation. Filter Types. Ceramic filters are available in a wide variety of materials and in many different forms. Mullite, alumina, cordierite, silica, zirconia, and silicon carbide are all commonly used. The most common forms are open-weave cloth, reticulated foam, extruded cellular shapes, pressed shapes, and perforated sheets (Fig. 14). Ceramic foams and cellular ceramics have been shown to be the most effective for inclusion removal and are the most widely used. Strainer cores were widely used prior to the advent of ceramic filters, and it is important to note the distinction between these two devices and their applications.

 Inclusion-related scrap and inclusion-related

rectification can be reduced.

 Gating system size can be reduced and cast-

ing yield increased (Fig. 12).

 Casting machinability can be improved.  Machine tool life can be increased.

Table 1 Densities of metals and metal oxides commonly used in casting alloys Casting alloy

Density, g/cm3

Aluminum alloys Al Al2O3 3Al2O3  2SiO2

2.41 3.96 3.15

Magnesium alloys Mg MgO

1.57 3.58

Copper alloys Cu CuO ZnO SnO BeO

8.00 6.00 5.61 6.45 3.01

Iron alloys Cast iron Low-carbon steel 2% C steel FeO Fe2O3 Fe3O4 FeSiO4 MnO Cr2O3 SiO2

6.97 7.81 6.93 5.70 5.24 5.18 4.34 5.45 5.21 2.65

Fig. 13

Gating system designs for optimizing the effectiveness of ceramic filters in horizontally parted molds having sprue:filter:runner:ingate cross-sectional area ratios of (a) 1:3–6:1.1:1.2 and (b) 1.2:3–6:1.0:1.1

Gating Design / 79

Several common filtration and flow modification devices (from left to right): strainer core, extruded ceramic filter, ceramic foam filter, mica screen, and woven fabric screen. The two types of ceramic filters are by far the most widely used.

Fig. 14

Cross section through a runner bar containing a ceramic foam filter. Particle capture occurs at the front face (top) and inside the filter. The alloy is aluminum-base LM25.

Fig. 15

 Filter type must be correct for the application.  Gating system design must provide mini-

mum metal turbulence downstream from the filter and in the casting cavity.  Gating system size must be kept to a minimum.

Common methods of filter placement in horizontally parted molds. (a) Parallel to parting line. (b) Between 0 and 90 to parting line. (c) 90 to parting line. Arrows indicate the direction of metal flow.

Fig. 16

Strainer (choke) cores are designed to function as a restriction or choke in the gating system. The open frontal area—the ratio of the area available for metal passage to total part cross-sectional area—is in the range of 20 to 45%. Strainer cores are designed into a gating system to control the rate of mold filling and can aid rapid filling of the gating system. This action can assist in the flotation and separation of large particles from the molten metal stream, but strainer cores are not intended to remove particles. Their large, widely spaced holes split the metal flow into several separate metal streams, often resulting in aspiration and subsequent oxide formation downstream. Ceramic filters, on the other hand, are designed to remove inclusions from molten metal. Filtering occurs by two mechanisms: physical screening (or sieving) and chemical

Common methods of filter placement in vertically parted molds. (a) Filter located in pouring basin. (b) Filter located inside the mold. Arrows indicate the direction of metal flow.

Fig. 17

attraction (Fig. 15). When properly designed into a gating system, filters do not act as a significant restriction to metal flow. The open frontal area of most ceramic filters is in the range of 60 to 85%, and the flow downstream of the filter is much less turbulent than with strainer cores. Use of Ceramic Filters. Design of the optimum gating system for ceramic filter use incorporates the following principles:  Filter placement must be easily accomplished.  Mold filling time must be constant and not

affected by the presence of the filter.

Filter Placement. The location and position of a ceramic filter are influenced by the molding method, pattern layout, and any metallurgical processes performed inside the mold cavity, such as the nodularizing and inoculating additions made in certain iron castings. The molding method is particularly important in determining filter position. In molding processes that use an expendable pattern, filters are most easily incorporated into the pouring basin. In horizontally parted molds, filters are commonly positioned as shown in Fig. 16. Filters should not be placed at the base of the sprue, because this increases the possibility of filter breakage and reduces filter effectiveness. In vertically parted molds, filters are commonly positioned as shown in Fig. 17. Although the pouring basin location is often used, filters are more effective when located farther down the gating system. When metallurgical additions are made in the mold cavity, filters must be located downstream, as shown in Fig. 18. Filter Area/Choke Area Ratio. The size and number of ceramic filters required are determined by the rate of mold filling required and the volume of metal to be filtered. As filtration occurs, individual cells in a filter become

80 / Casting Design and Performance

Plot of flow rate versus time showing the behavior of a ceramic filter during the course of pouring until complete blockage occurs

Fig. 19

Fig. 18

Filter placement when a metallurgical operation (that is, magnesium treatment or inoculation of iron) occurs in the mold

blocked, and the rate at which the filter can pass metal is reduced. This phenomenon is illustrated in Fig. 19. Filter size is determined such that the filter operates in the range termed normal flow in Fig. 19. In general, the necessary ratio of filter area to choke area is in the range of 2:1 to 4:1 at the start of pouring. If the filter is greatly undersized, complete filter blockage can occur, and incomplete mold filling will result.

REFERENCES 1. J.F. Wallace and E.B. Evans, Principles of Gating, Foundry, Vol 87, Oct 1959 2. J.G. Sylvia, Cast Metals Technology, Addison-Wesley, 1972 3. S.I. Karsay, Ductile Iron III: Gating and Risering, QIT—Fer et Titane, Inc., 1981 4. P.R. Khan, W.M. Su, H.S. Kim, J.W. Kang, and J.F. Wallace, Flow of Ductile Iron Through Ceramic Filters and the Effects on the Dross and Fatigue Properties, Trans. AFS, Vol 95, 1987, p 106–116

5. W. Simmons, The Filtering of Molten Metal to Improve Productivity, Yield, Quality and Properties, Proceedings of the 52nd International Foundry Congress, 1985 6. W. Simmons and H.A. Bowes, Efficient Filtration Improves Casting Quality, Foundry Pract., Vol 209, 1984 SELECTED REFERENCES  R.A. Flinn, Fundamentals of Metal Casting,

Addison-Wesley, 1963

 L.F. Porter and P.C. Rosenthal, Fluidity Test-

ing of Gray Cast Irons, Trans. AFS, Vol 60, 1952  J.M. Svoboda, Basic Principles of Gating and Risering, American Foundrymen’s Society, 1973  H.F. Taylor, M.C. Flemings, and J. Wulff, Foundry Engineering, John Wiley & Sons, 1959

Casting Design and Performance Pages 81–87

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design for Economical Sand Molding BASIC PRINCIPLES of sand molding are shown in Fig. 1. The cylindrical shape of this casting provides natural clearance or draft along the length of the casting, as seen in Fig. 1. The tapered ends of the pattern permit it to be removed from the sand without restriction. The sand core is positioned in the core print cavities provided in the molds by the ends of the pattern (the core prints). For processes other than sand casting, these molding principles are equally valid. In die casting and investment casting, metal dies and cores are used. In permanent mold casting, metal molds are used with either expendable or metal cores; with expendable cores the process is referred to as semi-permanent molding. When four external ribs are added to the tubular shape (Fig. 2), molding practice remains unchanged. However, when six ribs are added (Fig. 3), either cores or loose pieces must be used to avoid locking the pattern in the mold. This principle is applicable to all shapes, regardless of their complexity.

General Design Factors Parting Lines. The location of a parting line may be dictated either by the shape of the casting or by special requirements such as those illustrated in Fig. 4. Parting the casting as shown in Fig. 4(a) to (e) makes it possible to cast faces that are flat and parallel. The parting lines shown in Fig. 4(f) and (g) necessitate that the faces be tapered in order to provide draft. Other special requirements are illustrated in Fig. 4(a) to (f). For example, the method of parting in Fig. 4(a) provides normal draft in the hole and on the sides of the casting; the hole and the sides are concentric. Any flash formed at the parting line can be easily removed. Figure 4(b) provides parallel sides and normal draft in the hole and on the sides of the casting. It is more difficult, however, to maintain concentricity between the hole and the sides, because of the possible misalignment of the mold halves after the pattern has been withdrawn.

In Fig. 4(c), hole taper is reduced by 50%. Only a minimum of material need be machined from the hole to provide perfect straightness, a possible advantage when difficult-to-machine materials are involved. As in Fig. 4(b), however misalignment of the mold halves after the pattern has been withdrawn is a potential disadvantage. Figure 4(d) is similar to 4(c), except that taper remains the same in the hole and is reduced by 50% on the outer walls. In Fig. 4(e), taper is reduced by 50% both in the hole and on the outer walls. Metal requirements in casting and machining are minimized to a greater degree than in any of the other designs. Figure 4(f) requires taper on two sides, and 4(g) provides sides that are parallel. These seven examples illustrate the adaptability of parting line location to casting design requirements. For the bell-shaped casting of Fig. 5 locating the parting line at the base of the bell, as in Fig. 5(a), would eliminate parting line reflection from the body of the casting. However, because the core cannot be vented at the top, trapped gases might cause defects in the casting

Fig. 1

A simple casting that illustrates the principle of molding a pattern and withdrawing it from a sand mold

Fig. 2

Adding four ribs to a tubular casting introduces no problems in removal of pattern from mold. Molding practice is the same as for the cylindrical shape of Fig. 1.

82 / Casting Design and Performance

Fig. 3

Adding six ribs to a tubular casting necessitates core prints on the pattern, to permit withdrawal of pattern from mold. Cores are necessary to complete the mold.

Fig. 4

Seven different parting lines are possible in producing this simple casting. Each parting line has a different influence on the casting. (See text for discussion.)

Parting the casting as in (a) eliminates any parting line reflection from the sand mold on the casting, and circularity would be good. A riser at the isolated boss could eliminate porosity if it were a problem. Gas trapped by the metal could cause defects. Parting the casting as shown in (b) would eliminate any gas problems but would adversely influence stability of diametral dimensions. The parting line would appear as a seam along the length of the casting.

Fig. 5

metal. If this shape were to be cast in some metals, a second riser might be required at the top of the casting, in order to assure sound metal. Placing the casting on its side so that the parting line is at right angles to the base, as in Fig. 5(b), would permit adequate venting of the core, provide an improved means of gating, and eliminate the need for a second riser; a parting line seam is unavoidable, but it can be easily removed. Figure 6 provides an example of how the shape of a cross section of a casting can dictate location of the parting line. Figure 6(a) shows a

Fig. 6

Redesign to eliminate the need for a large core and to permit relocation of the parting line

casting as originally designed, with a cross section having an I-shape. This shape can be produced only by the use of a core, and the parting line must be located as shown in Fig. 6(a). When the casting is redesigned to have an X-shaped (or a T-shaped) cross section, the need for a core is eliminated and the parting line can be relocated, as shown in Fig. 6(b). This revision appreciably reduces production cost. A stepped parting line results in the formation of fins that may be difficult to remove. It should be avoided in casting design whenever possible. Figure 7 shows a bearing housing in the original design of which, Fig. 7(a), a

stepped parting line was used to allow for simple coring of the hole. To provide a straight parting line, this casting could be produced without the cored hole, Fig. 7(b). The hole could be drilled later, increasing the cost of machining but saving the cost of coring. The relative economy of coring and not coring can be determined only by analysis of the total cost of finished parts produced by each method. Several examples are given in the next chapter. Location of Radii. The casting sections shown in Fig. 8 illustrate how a minor design concession serves to avoid possible mismatch and simplifies removal of fins at the parting line. The original design, Fig. 8(a), was modified to eliminate radii and thus enable the parting line to be located at the top surface of the casting, as shown in Fig. 8(b). A similar concession applied to coring is shown in Fig. 9. Here the possibility of core shift may be a problem, but it can be avoided by eliminating the radius at the end of the core. If such a radius is required, it can be provided easily by machining. A more complicated design involving radii is shown in Fig. 10. Figure 10(a) represents the more difficult approach to the problem of coring the flanged perimeter. Here the core forms both the slot and the radii on the inner faces of the flange. The major disadvantage of this design is increased core cost. Figure 10(b) shows how the same flange can be designed without radii, at a lower cost. A simple modification, providing one radius only, is shown in Fig. 10(c). Bosses and Undercuts. Frequently, it is necessary to locate a boss some distance from the parting line. The section shown in Fig. 11(a) illustrates the positioning of a boss well below a flange whose upper surface serves as a parting

Design for Economical Sand Molding / 83

Fig. 7

Eliminating the cored hole in this sand casting permitted a simple flat parting plane. Previously, a stepped parting plane was required, to permit removal of the core print from the mold.

A radius where the flat face joins the edges of the casting would require parting line as shown in (a). Seams or mismatch may result. By eliminating the radius, the parting line can be located as in (b).

Fig. 8

line. In this design, a core is required, to permit removal of the pattern from the mold. In producing the casting as shown, accurate positioning of the core is difficult, and any shifting of the core results in surface irregularities. A less complicated design, shown in Fig. 11(b), extends the boss to the flange, eliminating the undercut and the need for a core. A design in which the boss has been extended downward to the parting line to satisfy molding requirements is shown in Fig. 11 (c) and (d). The boss can be extended beyond the parting line and provided with a radius to improve appearance, as shown in Fig. 11(d). (Another method would be to form the boss by means of a “loose piece” on the pattern, the loose piece being drawn sideways from the mold after the pattern has been withdrawn vertically.) Figures 12 and 13 deal with the elimination of undercuts in achieving simpler and less expensive designs. As originally designed, the undercuts shown in Fig. 12(a) were needed to provide clearance for other components in the assembly. An improved design without undercuts is illustrated in Fig. 12(b); however, metal is added and the weight of the casting is increased. The design of a wheel hub, Fig. 13(a), required a ring-shaped core to form the exterior of the hub, because of the undercut adjacent to the flange. Eight ribs were needed to strengthen the part and assure its satisfactory performance. In the improved design of Fig. 13(b) the hub section is tapered, eliminating the undercut and the need for a ring core. The strength and rigidity of the casting are improved enough so that the supporting ribs are no longer required. Rib Locations. The housing shown in Fig. 14 can be molded in two different ways at an appreciable difference in cost, depending principally on a design simplification involving

A radius at the junction of a cored hole and a sand casting face requires a shaped core, as in (a). Mismatch could result. Elimination of radius, as in (b), simplifies the core and removes the possibility of mismatch as a result of core shift.

Fig. 9

the number and location of ribs. The original design, Fig. 14(a), specified three ribs (one of which is the tubular shape) projecting radially from the center of the housing and spaced 120 apart. With this arrangement, it was most economical to locate the parting line at the face of the small flange, Fig. 14(b), in order to guarantee stability of the two internal cores. The method of coring is shown in Fig. 14(c). Core box details for the entire casting are presented in Fig. 14(d) to (i). A study of these details indicates the complexity of the coring arrangement. In a redesign, Fig. 14(j), the parting line was located at the centerline of the housing, and the boss was placed in the cope half of the mold at right angles to the parting line. Four ribs (of which one is the tubular shape) spaced 90 apart replaced the three-rib arrangement of the original design. The twelve bolting bosses were revised to eliminate an undercut and permit straight withdrawal from the mold. Core box details for the redesigned casting are shown in Fig. 14(k) and (l). The core box for core Z, Fig. 14(j), is not shown but is similar to core box X, Fig. 14(k). The pattern and rigging for the revised casting are shown in Fig. 14(m). In the original design, the heavy sections of the casting were not accessible for risering. In the new design, risers were easily located where they were needed. Also, the revised design permits the use of an alternate sprue.

Molding and Coring Details The sand casting shown in Fig. 15 was designed solely to demonstrate alternative design solutions to molding and coring problems. Although there are other ways to produce this casting, the methods described here are intended to indicate the important influence of design on the cost of a casting and to

(a) A radius on the inside edges of the two flanges would increase the cost of the core, because of the requirement for a loose piece in the core box. (b) Eliminating the radius simplifies the core. (c) One radius can be incorporated without complicating the core of this hypothetical casting. (See Chapter “Designing for Economical Coring” for an explanation of the term “drier”.)

Fig. 10

emphasize the relation of design to molding and coring practice. The pattern for this casting, as originally designed, is shown in Fig. 15(a). The pattern is made in five sections and requires a special four-part flask. Although the end sections of the flask (for mold sections 1 and 4) may be made to any height that provides enough sand to retain the metal within the mold cavity, the height of the center sections must correspond to the height of the sections of the pattern involved. The molding sequence is: 1. Pattern sections 2 and 3, Fig. 15(a), are assembled, flask section 2 is set in place, and mold section 2 is produced. 2. Pattern sections 4 and 5 are placed in position on pattern section 3. Flask section 3 is set in place, and mold section 3 is produced. 3. An end section of the flask is set in place on flask section 3, and mold section 4 is produced. 4. The three sections of the mold that have been produced are turned over as a unit, exposing the mold surface at parting line A and pattern section 2. 5. Pattern section 1 is set in place on pattern section 2, and the remaining end flask section is set in position. Mold section 1 is produced. At this point, the four flask sections are in place. The mold is complete, but the patterns have not been removed. The steps that follow concern removal of the pattern.

84 / Casting Design and Performance

Fig. 11

An undercut created by an isolated boss on the side of a sand casting requires either a core as in (a) or continuation to a flange as in (b). Absence of a flange as in (c) requires continuation of the boss to the parting line as in (d).

Fig. 12

(a) Undercuts required cores as shown, to permit withdrawal of the pattern from the sand mold. (b) Redesign eliminated the undercuts and the need for cores.

Fig. 13

(a) A reduced diameter adjacent to the flange of this sand cast wheel hub necessitated a core. In addition, eight ribs were required, to provide the desired strength. (b) Redesign as shown eliminated the ring core and the ribs, and a stronger casting was produced.

6. Mold section 1 is removed and turned over to expose pattern section 1. Pattern section 1 is then removed from the mold. 7. Pattern section 2 is removed from mold section 2. Mold section 2 is removed and turned over. Pattern section 3 is withdrawn from the mold. 8. Pattern section 4 is withdrawn from the mold. Mold section 3 is then removed and turned over. Pattern section 5 is withdrawn from the mold. In the foregoing description of the molding sequence, details of gating and risering have been omitted, in the interests of simplicity. Because two of the outside diameters of the casting are smaller than three of the inside diameters, a one-piece core cannot be placed in the mold. A three-piece core, Fig. 15(b), is required. The preparation of each section of the threepiece core entails as much work as would the making a single large core for the entire casting (if such were possible). To understand the details of preparing a core section, the reader

may begin by considering the cross section of the core box used for core 2, which is shown in Fig. 15(c). To produce the required core shape, both a loose ring and a loose center plug are needed. The sand used to fill the core box must be introduced in the limited space separating the ring from the plug. Thus, packing the sand in the core is difficult and timeconsuming. After packing the sand and removing the ring and plug, it is not possible to set the core down on its upper face without sagging of the core. As shown in Fig. 15(c), this upper face or contact area consists of a relatively narrow ring beyond which a portion of the core extends. If only a few castings were to be produced, it would be feasible to prevent core sag by filling the cavities formed by the ring and plug with bedding sand. For production quantities, however, a core drier, Fig. 15(d), is required. The core drier provides for greater stability when the core is removed from the box. An individual drier is required for each core placed in the baking oven.

An alternate method for making core 2 is shown in Fig. 15(e). Here, the core is produced in two halves, each having a large flat face that makes the use of core driers unnecessary. The large core box opening also simplifies packing of the sand. After baking, the two halves are pasted together to form the complete core. The possible disadvantages of this method are the chance of misaligning the halves in assembly and the effect of parting on the accuracy of dimensions measured across the parting line. Cores 1 and 3 are prepared in a manner similar to that used for core 2. Core 1 is placed in mold section 1, and the remaining cores and mold sections are assembled alternately as shown in Fig. 15(b). Although sectional cores and molds provide a high degree of flexibility in forming complicated shapes, they have certain inherent disadvantages. Among these is the possibility of movement, which would cause dimensional variation in the casting. Also, at every junction of the loose pieces in the pattern or core box, and at joints between mold and core sections, there is a possibility

Design for Economical Sand Molding / 85

Housing producible (with permissible modification) by either of two molding methods with different parting line locations. For the casting as originally designed (a), parting line location is shown in (b), and method of coring in (c); core box details are shown in (d) to (i). Redesign of casting is shown in (j); note relocation of parting line. Core box details for redesign are shown in (k) and (l); pattern and rigging, in (m).

Fig. 14

86 / Casting Design and Performance

A hypothetical casting illustrating core and mold problems if placement of cores is not considered in the design of a casting. The required patterns and the necessary mold sections are shown in (a). The core and related mold sections are shown in (b). Views (c) and (d) show core box and drier to produce core 2 in one piece. View (e) shows core 2 produced in two pieces. Revision as in (f) would simplify mold and cores. Revision as in (g) would simplify still more the core necessary to produce the outside shape. (See text for discussion.)

Fig. 15

of forming seams or fins, which would have to be removed at extra cost. The first of two revisions of the original design of this casting is shown in Fig. 15(f). This redesign provides a maximum core diameter smaller than the minimum mold diameter through which the core must pass. The contour of the outer midsection of the casting requires the use of loose pieces in the core box forming core 1. This design eliminates sectional cores and permits the use of a conventional copeand-drag mold. A further design improvement, Fig. 15(g), simplifies the core box producing core 1, and results in a cleaner casting with less dimensional variation.

Conventional molding of a circular sand casting is shown in (a). A specification requiring the large face to be smooth necessitated placing this face in the drag half of the mold, as shown in (b), to obtain maximum smoothness of surface.

Fig. 16

Design for Economical Sand Molding / 87 Draft refers to the amount of taper given to the sides of a pattern to enable it to be withdrawn easily from the mold. For this reason, draft requirements are intimately related to casting design. A simple example demonstrating the influence of a casting requirement on the direction of draft is given in Fig. 16.

Conventionally, this design would be molded as shown in Fig. 16(a). However, when the largest plane surface of the casting is required to be straight and smooth, as cast, the casting is molded as shown in Fig. 16(b), to avoid risering into the plane surface and the possibility of inclusions floating to this surface during

pouring. This reverses the draft direction on the sides of the casting (or moves the casting into the cope). In this example, the importance of draft direction increases with an increase in the over-all dimensions of the casting. To some extent, this problem parallels that of parting line location discussed with Fig. 4.

Casting Design and Performance Pages 89–99

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design for Economical Coring CORES are separate shapes, of sand, metal or plaster, that are placed in the mold to provide castings with contours, cavities and passages not otherwise practical or physically obtainable by the mold. In general, the principles that apply to molding a pattern apply also to molding a sand core. The core must be freed from the core box by moving the box away from the core in a direction generally perpendicular to the face of the core box. Loose pieces that have been placed in the core box to assist in producing the shape of the core will remain with the core sand when the box is removed from the core, and must be removed from the core by a movement in the appropriate direction.

General Principles General principles of coremaking are illustrated in Fig. 1 by four basic examples that depict, in simplified form, most of the problems entailed in molding a core. At the top of each illustration, the core box is shown in section after ram-up; in the lower part, the core box has been inverted onto a plate upon which the core rests preparatory to being baked. Figure 1(a) shows a core that is semicylindrical — one of the most common of core shapes. No problem is presented in removing this core

Fig. 1

Basic principles of core molding.

from the core box; natural draft is provided by the shape of the core, and no obstructions exist. Figure 1(b) shows two bosses formed by the core box. Boss A is simple to produce and adds nothing to the cost of producing the core, although it does add a small amount to the cost of the core box. Boss B, however, creates an undercut that requires a loose piece, as shown, to permit the core to be freed without destroying the desired shape. This adds to the cost of the core box. It also adds to the cost of producing the core, and may increase the dimensional variations in the location of the boss. Figure 1(c) illustrates a core box into which two more conditions are incorporated that require loose pieces. Loose piece A permits the side of the core to be produced without draft. Bosses or ribs could be appended to this face without adding to molding problems, provided they could be drafted to permit easy withdrawal of the loose piece. Loose piece B forms a cavity in the core that assists in producing a rib on the casting. The overhang of the core box shape would lock the core in the box unless a loose piece were provided as shown. Figure 1(d) illustrates the additional problem that arises when the loose piece at the face of the core box projects excessively into the core. When the core box is inverted and the loose piece removed, excessive overhang of the core could cause the sand to sag and lose its

dimensional accuracy. To prevent this, support must be provided for the core. In the core box shown, the loose piece is removed and replaced with bedding sand before the core box is inverted. (Bedding sand is a special sand that does not adhere readily to the core and so can be removed cleanly after the core has been baked.) Core Driers. In the core box shown in Fig. 1 (d), a core drier could be the loose piece. (A core drier is a contoured metal form or “cradle” that remains with and supports an irregularly shaped core while it is being baked.) However, since a drier is required for each core being baked, the expense of producing all the necessary driers adds substantially to core cost. Although supporting an overhanging or irregular section of a core with metal core driers provides greater accuracy and production speed, the use of bedding sand is usually more economical. The tolerance specified will influence the choice. Advantages. Although cores increase cost and tolerance requirements, they enable the foundryman to cast intricate internal shapes not producible by any other process. Examples are curved passages of various kinds and large interior cavities connected to the outside surface by relatively small openings. The interior passages need not have uniform shape or cross section. The internal cavities can be of almost

90 / Casting Design and Performance any shape dictated by the requirements of design. Cores also allow the casting designer to eliminate superfluous metal, thus reducing the weight of a casting and the cost of subsequent machining operations. Cores may be used to produce draft-free surfaces, to simplify parting lines, or to eliminate the use of loose pieces in the pattern. Cores may also provide an additional benefit, in that patterns that would be too fragile to withstand normal handling will, if made with cores, be braced by the core prints, thus permitting shapes and cross sections not otherwise practicable. The part sketched in Fig. 2 was easily producible as a magnesium alloy sand casting. The twist at its center section, which was readily achievable by the use of cores, would pose a problem in the production of this part by any process other than casting. The intricately cored shell mold casting shown in Fig. 3, a manual control valve body

A skewed wall readily producible as a casting but not economical by any other manufacturing process

Fig. 2

An intricately cored value body cast in a shell mold. This shape was economically producible only by the casting process. Cored passages between individual chambers are not visible in the above views.

Fig. 3

Fig. 4

for an automatic transmission, is another part that would be economically impractical to produce by any process except casting. Core Size Limitations. Figure 4 presents recommended relations between core length and core diameter. These maximums were established by one foundry and, although other foundries may be more or less restrictive, they can serve as guides in the design of cored holes. Recommended maximum lengths for horizontal sand cores supported at both ends are given in Fig. 4(a), for steel, aluminum, and magnesium castings. Such cores often bend upward because of the buoyant force of the molten metal. Steel causes more bending than does aluminum because, being heavier, it exerts a greater upward pressure on the core. Also, the higher temperature at which steel is cast is more likely to break down the binder that holds the core sand together. One foundry producing a casting with a horizontally supported core reported an upward bending of 1/32 in. in a tubular casting with a 1½-in. OD, 1-in. ID, and an over-all length of 10 in. Occasionally, one of these cores bent 1/16 in. Cantilever-supported cores, Fig. 4(b), are even more likely to bend, because they are supported at only one end and are less able to resist the buoyant force of the molten metal. A larger tolerance is needed on dimensions at the unsupported end of the core, because of the necessity for a small amount of side clearance between the core and the mold at the opposite end. This clearance permits a displacement of the core when the molten metal enters the mold. The displacement is amplified as the core extends into the casting, and has a pronounced influence on dimensional discrepancies. A blind hole whose length is approximately equal to its diameter usually offers no cleaning problems, because excessive heat concentration is unlikely in the core. As the length of the blind hole increases in relation to its diameter, the possibility of burnt-in sand increases, making cleaning progressively more difficult. It is for these reasons that blind cored holes should be kept to minimum length; as indicated

in Fig. 4, the maximum length recommended is less than for cores supported at both ends. For cantilever-supported cores, the absolute minimum that must be allowed for coreprint lengths is 1/3 L, as shown in Fig. 4(b). This minimum length is suggested for use only in dry sand molds and in shell molds, where the fit is accurate enough to hold the core firmly in place. In green sand molds, enough mass must be provided in the core print section to balance the core so that it stays in place while the mold is being closed. The designer must keep in mind the requirements for core print lengths and mass when it is necessary to locate a part of the casting close to the opening of a cored hole. Vertical cylindrical cores supported at both ends are not distorted by the buoyant action of the metal. As long as they are free to expand, they remain straighter than horizontal cores. With such vertical cores, however, it may be difficult to close the mold without either disturbing the core location or shaving sand from the mold. This sand may fall down into the mold and appear as a casting defect. Figure 4(c) shows maximum recommended lengths. Although vertical cores supported only at the bottom are not subject to distortion by the buoyant effect of the casting metal, they can float out of position if they are not anchored firmly. Furthermore, such cores vent less efficiently than horizontal cores or cores with a core print into the cope. Another method of anchoring such a core in place is illustrated here in Fig. 5(a). The printed section of the core extends past the cavity and is held firmly in place by the cope half of the mold. This method requires the use of more core sand than a ram-up core and is more expensive, but it is also more accurate. The insertion and removal of ram-up cores abrade the pattern and permit variation in core location from casting to casting. Because of the difficulties usually present in molding and coring vertical blind holes, the maximum length of the cores should be reduced to about half the values given in Fig. 4(c).

A comparison of maximum recommended core length versus core diameter for baked sand cores positioned in the mold (a) at the parting line and firmly anchored at both ends, (b) at the parting line but anchored at one end only, and (c) vertical to the parting line and anchored at both ends

Design for Economical Coring / 91 Vertical cores supported from the top, as in Fig. 5(b), are not subject to distortion from the buoyant effect of the molten metal, and they are prevented from floating by the core print, which is held in place by the cope half of the mold. Such cores are easily vented, because core gas ordinarily travels upward to escape. For vertical cores supported at the top only, recommended maximum lengths with reference to diameters are the same as those given in Fig. 4(b) for cantilever-supported cores. Green Sand Versus Dry Sand Cores. Some cavities can be cored by the use of “green sand cores” (cores formed from the molding sand and generally an integral part of the pattern and mold). Because these cores are formed when the mold is made around the pattern, they usually add no expense to the production of a casting, and save the cost of producing and placing in the mold equivalent dry sand cores. Although green sand cores are less expensive than dry sand cores, they are at a disadvantage in several respects. Because they are less rigid than dry sand cores, they are less resistant to movement or deformation under the static pressure of the molten casting metal. This is an important consideration when wall thicknesses must be held to close tolerances. Dry sand cores can provide smoother casting surfaces, and so would be preferable when surface finish is a consideration. Furthermore, green sand cores usually require more draft than would be normal for the same length of draw on an outside surface of the pattern. This is a precaution to

prevent the green sand core from breaking away from the rest of the mold when the pattern is withdrawn. Figure 6 shows a thrust bearing housing, sand cast of malleable iron, that was successfully redesigned to employ green sand cores to form the entire inside diameter, and thereby to lower production costs substantially. As originally designed, Fig. 6(a), this casting required a baked sand core and one of green sand. It had a deep circular lightener pocket on one face and a circular oil passage opening to its outside diameter. The lightener pocket was too deep to be made in green sand with normal processing, so a ram-up core was utilized. This core also formed a portion of the inside diameter of the casting, as shown. (A “ram-up” core is one which is placed in the pattern before the flask is filled with sand, so that as the mold is rammed-up the core becomes an integral part of the mold.) A ring core was required also, to form the circular oil passage. Since the pocket was merely a lightener and the function of the oil passage was noncritical, no precise dimensions or tolerances were involved. A redesign, shown in Fig. 6(b), combined the oil passage and the lightener, was no heavier, and required only one baked sand core. The inside diameter of the casting is now completely formed using green sand cores, half in the drag and half in the cope. They meet at one of the steps and are produced as the pattern is molded.

(a) Core float was prevented by a large coreprinted section clamped between the mold halves. Venting of the core would be difficult. (b) By inverting the pattern, better venting of core gas was possible. A large core print was still necessary to position the core securely and to prevent shifting.

Fig. 5

Fig. 6

A redesign combined two cores into one, resulting in substantial cost savings.

Designing for the Use of Sand Cores Cored holes should be designed as simply as the intended function of the casting permits. An effort should be made to eliminate projecting and overhanging parts, pockets and recesses, and to use simple and smoothly contoured core shapes. The casting design should take into consideration the requirements for draft necessary for successful removal of the core from the core box. Although the incorporation of loose pieces permits cores with straight sides, overhangs and undercuts, these add to dimensional discrepancies and increase the cost of producing the core. Bosses should be extended to the parting line if this can be done without adding impermissibly to the weight of the part. When this is a questionable procedure, the undesirability of additional weight should be balanced against the desirability of savings in the casting cost through the design simplification. Of course, if bosses can be located on the parting line, both cost and weight are held to a minimum. Core Fragility. It is important that the smallest dimensions of the core be large enough to facilitate production of the core. Thin or small-diameter cores are extremely fragile and are a cause of both core scrap and casting scrap. The armament stores pylon casting shown in Fig. 7, which was produced from 4330 steel, presented a problem because of core fragility. View (a) shows the as-cast configuration of the part. View (b) is the machined casting cut in half. In this casting as originally designed, holes A and B, formed by cores 1 and 2, respectively, were all ½ in. in diameter. Additional support for core 1 was provided by the core print for the 1-in.-diameter core at its left end Core 2 was supported solely by the four (two on each side) ½-in.-diameter core prints. This design was unsatisfactory because, with normal foundry handing, the ½-in.-diameter core sections were too readily broken off the main bodies of the cores. In addition, it was difficult to remove the core sand through the small holes in the casting. The casting was redesigned. Holes A were enlarged to be 3/4 in. in diameter. These four holes (two on each side of the casting), plus the 1-in.-diameter core at the left end, now provided adequate support for core 1. In core 2, holes B were enlarged to 5/8 in. in diameter. Hole C, 3/4 in. in diameter, was added for additional support. Increasing the size of holes A and B not only eliminated the problem of core breakage but, also increased the dimensional accuracy provided by the cores, through more positive placement and anchorage during the pouring and solidification of the metal. Venting of the cores was also improved, and removal of core sand was made easier. Thin Core Sections. The separation of heavy masses of metal with thin core sections should

92 / Casting Design and Performance

Fig. 7

Small appended core sections were broken off the main core in handling. Increasing the size of the cored holes reduced breakage, improved core venting, and facilitated removal of core sand after casting this steel part.

Fig. 8

By eliminating a thin core between two heavy sections of this AZ63 magnesium alloy casting, a problem of heat retention was solved. The casting was slotted by a subsequent machining operation: (See next page for discussion.)

be avoided. If too much metal is adjacent to a core or surrounds it, the core will be unable to dissipate the heat from the molten metal. This will result in hot spots in the casting during cooling and solidification. Such hot spots inevitably give rise to shrinkage or porosity, which can cause rejection of the castings. An additional problem is the inability of these thin cores to vent core gas adequately, which can result in blow-holes. Also, metal penetration may cause difficulty in core removal. The designer should attempt to forestall these problems by increasing the cored area or reducing the metal walls that surround it, or both. Often, the thickness of metal wall sections can be decreased, without sacrificing strength, by the use of ribs. The presence of a thin core section between two heavier masses of metal presented a problem in production of the turbine door fitting shown in Fig. 8. Made from magnesium alloy AZ63, this sand casting, as originally designed, had two isolated walls, each 0.38 in. thick, separated from each other by a core section 0.23 in. thick. When the casting was poured, this core section was too thin to dissipate the heat from the surrounding molten metal. A hot spot developed, which resulted in a heavy shrinkage defect on each inside surface of the bosses. Rejection rate was 90%.

This aluminum plaster mold casting is an example where thin cores became excessively hot when surrounded by molten metal, causing shrinkage porosity.

Fig. 9

The casting was redesigned to eliminate the core section between the bosses, so that the metal section was cast solid in this region. Although this now required a milling operation to remove the added metal, rejections dropped to 3%. Another casting with production problems caused by cored areas being-too thin with respect to surrounding metal masses is shown in Fig. 9. When this casting (an aluminum alloy impeller, cast in a plaster mold) was poured, the inability of the thin cores to dissipate heat adequately from the molten metal caused hot spots that produced shrinkage porosity.

Well-proportioned cored passages. Sand core was able to withstand normal foundry handling, provided stable positioning in the mold, and was easily vented.

Fig. 10

Figure 10. shows a sand cast fitting with cored passages exemplifying good design for any process that uses expendable cores, because:

Design for Economical Coring / 93 1. The cores are large enough in cross section to enable them to support themselves without sagging and also to withstand normal foundry handling without breaking. 2. The cores have sufficient mass, in relation to that of the surrounding metal, to dissipate heat rapidly enough to prevent hot spots that can cause shrinkage defects. 3. The cores can be easily and adequately vented, to eliminate any problems arising from generation of core gas. 4. The presence of a core print at each of the three areas where the core emerges from the casting assures that the core can be accurately positioned and that it can be firmly anchored to remain in place while the molten metal is filling the cavity. 5. The straight-through shapes of the cored sections permit easy removal of core sand after the casting has cooled. Thin-Wall Casting Sections Formed by Cores. The minimum thickness of walls formed entirely by cores is governed by the same considerations that control the minimum thickness of walls formed by the mold, as detailed in Chapter 3. Most important is that thin sections be given access to an adequate supply of molten metal, to prevent cold shuts or misruns. This molten metal can be supplied either from adjacent heavier sections or from risers. The thin-wall magnesium sand casting illustrated in Fig. 11 had a center web formed by two adjacent cores. The necessity for core prints to anchor the two cores made it difficult to gate and riser this casting properly. Consequently, misruns and cold shuts, particularly in the thin center web, occurred frequently when castings were made in production. In producing the wave guide power divider shown in Fig. 12, an aluminum alloy plaster mold casting, the 0.040-in. vertical sections,

Six additional cored holes (three on each side) were incorporated to provide the core support necessary to prevent core distortion when molten metal filled the cavity. These added holes were later tapped and plugged.

formed by cores, were difficult to fill completely with casting metal. The metal cooled so quickly when poured into the mold at normal casting temperature that it lost the fluidity necessary to fill the thin sections. Increasing the temperature of the metal caused shrinkage of the heavy sections and excessive formation of gas. Whenever possible, extremes in section size, as in this example, should be avoided, preferably by increasing the thin sections.

Support for Sand Cores The fact that certain complicated shapes can be produced economically only by the casting process may require the design of cored sections difficult to support. It is not always possible to provide as many or as large core prints as necessary to support the core properly. And, although a core may appear to be rigid, if supports for it are too far apart it can warp, sag or float when surrounded by the molten casting metal. This is particularly true with a long core of relatively small cross section. Inadequate support was a problem with the core necessary to form the inside configuration of the oil slinger shown in Fig. 13, sand cast in aluminum alloy. The original design had six cores, three on each side, spaced as shown. After casting, the 5/16-in. wall thickness varied (in some instances, to as little as 3/32 in.), and

rejection rate was high. It was necessary to add six more holes, three on each side, as shown, to provide adequate support for the core. These additional holes were later tapped and plugged, to permit the casting to function as originally intended. When cores are produced in halves, baked and pasted together, any distortion will prevent the faces from meeting firmly. The gap between the cores will cause fins to be present in the cored passage unless special processing is given the cores. In the pilot’s steering wheel shown in Fig. 14 (a magnesium sand casting), a large fin on the inside passage wall was a hazard to the electrical wires that ran through this passage to

A wave guide power divider casting difficult to produce because 0.040-in. sections, formed by cores, froze prematurely. Increasing the temperature of the metal caused excessive gas to form in areas where the cores were in contact with the heavier sections. Increasing the thin sections would be a practical solution in this aluminum plaster mold casting.

Fig. 12 Fig. 11 (See text.)

Thin center section, formed by two cores, was not readily accessible to gating and risering.

Fig. 13

A heavy fin occurred on the inside of this pilot’s wheel (a magnesium sand casting) as a result of a gap between the two core halves that formed the inner shape. Special processing of the core was necessary to eliminate this condition and prevent damage to electrical wires that were installed in this cored passage.

Fig. 14

94 / Casting Design and Performance the control switches at the tips of the wheels. The fins are shown in the cutaway sections in the illustration. It was necessary to specify maximum allowable fins, and to add appropriate inspection procedures. Chaplets. It is sometimes permissible to provide the necessary support for cores by the use of chaplets, which are metal spacers placed in the mold to support a core or to space it a fixed distance from the mold wall. Chaplets (illustrated in Fig. 15) are usually made of the same metal as the casting in which they are to be used. In theory, the chaplet fuses with the casting, to form a continuous wall. In practice, this does not always happen, and poor fusion may create a weak spot in the wall that will leak if subjected to pressure. Fusion between aluminum or magnesium chaplets and castings is even more uncertain than in steel, because of the lower heat capacity of aluminum and magnesium and the fact that an oxide film is usually present on the surface of aluminum and magnesium chaplets. A long, slender core supported at one end only, such as the core shown in Fig. 16(a), requires support to prevent it from sagging or floating and thus causing distortion. Chaplets would provide this support and would also help the molder to position the core properly. Two cored openings, in the same positions as the chaplets, would provide support for the core (from their core prints) and would also allow for better venting, Fig. 16(b). The cored holes would be tapped and plugged later. The enlarged end of the core print, Fig. 16(a), positions the core axially. The addition of the small core prints, Fig. 16(b), eliminates the need for this enlarged section.

During the past several years, however, considerable progress has been made in the development of techniques employing metal tubing for these cores. These give the designer greater freedom in the inclusion and arrangement of passages, and permit highly complex design. Three different techniques are currently available: 1. The older method of drilled passages plus complex, bulky external tubing. When feasible, this may still be the most economical method. 2. Lined passages produced by pouring aluminum or magnesium around tubing which remains in the casting. 3. The most recent method, which involves the introduction of unlined passages into sand and permanent mold castings. The unlined passages (3, above) require cores that can be dissolved or otherwise removed from the casting. These cores are usually made of steel, copper, glass or some siliceous refractory material. For practical purposes, the unsupported length of such cores depends not only on the diameter but also on the design of the core. Large-diameter cores support themselves over a longer span than do cores of smaller diameter, because increasing the diameter increases the rigidity of the core. Best design practice allows for at least a ¼-in. wall of cast metal around a passageway. In some castings a 3/16-in. wall may be adequate, but using a wall of less than 3 /16 in. invites trouble.

Cores in Permanent Mold Castings Coring of Tortuous Passages For reasons of function, many castings are required to have tortuous internal passages of small diameter. Sand cores for these passages are difficult to produce. Ramming of the sand in the core box, in conformity with the outlines of the intended core, is often impossible to do properly. Because of the shape of the cores, they must be supported by core driers during baking. Also, such cores are easily broken in normal foundry handling. The designer of castings has been greatly restricted as to the passages that could be case economically, using sand cores.

Cores in the permanent mold process may be of gray iron, steel, sand or plaster. Metal cores may be either movable or stationary. Stationary cores must be perpendicular to the parting line to permit removal of the casting from the mold, and must be shaped so that the casting is readily

As originally designed, chaplets were required for core support. An improved design eliminated the need for chaplets by incorporating two holes which provided adequate support by the core itself and permitted better venting of the core gas. These two holes were later tapped and plugged.

Fig. 16

Chaplets may be used to position cores not readily supported by core-printed sections. As illustrated, chaplets prevent core sag before pouring and core float after pouring.

Fig. 15

freed. Metal cores not perpendicular to the parting line must be movable, so they can be withdrawn from the casting before it is removed from the mold. When baked sand or plaster cores are used in a permanent mold, the process is referred to as semipermanent molding. With sand or plaster cores, which are expendable, more complex shapes can be produced than with metal cores. The two cored passageways shown in Fig. 17 illustrate the limitations of cores in permanent mold castings. The cored passageway shown in Fig. 17(a) could be produced with metal cores. Sand or plaster cores would be needed to provide the inside radius of the passageway shown in Fig. 17(b), which would also be more expensive to produce. Cored holes in permanent mold castings can usually be held to closer tolerances, of both size and location, than in sand castings. Both movable cores and stationary cores can be machined to close dimensions and accurately located. Although sand cores in permanent molds are no more accurate dimensionally than in sand molds, in the metal molds they may be positioned and retained in place with more accuracy than in sand molds. In general, the problems of expendable cores in permanent mold castings and in sand castings are similar. The problems with metal cores are similar to those in die castings.

Investment Castings In sand casting, wood or metal patterns are used to make the necessary impression in the molding material. Although the pattern may be re-used, the mold is expendable. In investment casting, a metal die is used to produce wax or plastic patterns. These patterns, in turn,

Coring limitations of permanent mold castings are illustrated by this simple casting. (a) The corner radius of the cored passage must be sacrificed if metal cores are to be used to achieve production at the most economical level. (b) Plaster or sand cores can provide the inside radius but casting cost will be higher than for (a).

Fig. 17

Design for Economical Coring / 95 are used to produce ceramic molds, either solid or shell. Both the patterns and the ceramic molds are expendable. The crux of the molding problem involves the withdrawal of the wax or plastic pattern from the metal die and the withdrawal of cores, when used, from the pattern (Fig. 18). In other respects, the principles applicable to sand molding apply to investment molding. The wax pattern for investment molding shown in Fig. 19 incorporates an inside radius at the junction of the two cores. Metal cores A and B are required to be withdrawn from the wax pattern in the directions indicated in Fig. 19(a). Such withdrawal cannot be achieved without damaging the wax pattern. In the redesigned casting, shown in Fig. 19(b), the inside radius has been eliminated in order to circumvent this problem. To retain the original radius would have required a costly change from metal cores to “soluble cores” made of a material capable of being etched out or dissolved out of the wax pattern without damaging it. Figure 20 illustrates a curved core passage. Again, a metal core cannot be used in the design shown in Fig. 20(a). A two-section core,

Fig. 18

Metal tooling for making the wax or plastic pattern for tubular investment custing shown at left. Note draft to permit withdrawal of the core from the pattern.

(a) In making the wax pattern for investment molding, an inside radius is impossible using two cores; core A cannot be withdrawn without damaging the pattern. (b) Sharp corner permits easy withdrawal of cores.

Fig. 19

Fig. 22

eliminating the inside radius, is shown in the design of Fig. 20(b). Where grooves are required for tool clearance in subsequent machining, the correct procedure for incorporating these grooves into the design of a casting is that shown in Fig. 21(b). The method of Fig. 21(a) does not permit withdrawal of a metal core. In the original design of a four-blade propeller, Fig. 22(a), the hub was designed with a pronounced taper. This necessitated the use of four loose pieces, located at the hub and equally spaced between blades. Further details of this design, which more fully explain the location of the loose pieces, are given in Fig. 22(b). A simplification of the original parting line location, Fig. 22(c), is shown in Fig. 22(d). Note that the radii at the leading and trailing edges of the blade have been greatly modified in the revised design. When the hub taper was eliminated, use of a simple two-part mold was possible. The original design of a blade for a gas turbine called for notches in the cast airfoil section, as shown in Fig. 23(a). To provide such notches required the attachment of gates to a

(a) In making the wax pattern for investment molding, special coring procedure would be required, to produce radius Y. (b) By permitting the inside corner to be sharp, standard coring will suffice.

Fig. 20

contoured surface. The gate contacts were removed by an expensive grinding operation. A less expensive approach is shown in Fig. 23(b). The notch was eliminated from one end of the blade, permitting attachment of an adequate gate at this end. The large gate encouraged directional solidification and permitted feeding the casting as it froze. This gate was easily removed, and the required notch was subsequently machined in place. Figure 24 shows another example of simplified gate removal, in connection with the design of a two-blade propeller. Originally, Fig. 24(a), a hole was to be cored through the hub of the propeller. This required double gating to a contoured surface. By eliminating the cored hole and providing a simple pad at one end of the hub, the use of a single gate was made possible. Removal of this gate was inexpensive and there was no longer a risk of damaging the airfoil section. With some designs of this type, cast in some metals, gating on the hub as in Fig. 24(b) could be used with a core in position. Holes in Investment Castings. Holes of various shapes can usually be formed with little or no difficulty by the investment casting process. Limiting factors are the size of the opening, the ratio of the size of opening to the length of core, and the complexity of the shape as it affects metal fluidity and wall thicknesses. Generally, in investment casting, through holes of less than 3/32-in. diameter and blind holes of less than 3/16-in. diameter require special coring techniques that add significantly to the cost of the part. Blind holes that have a length-to-diameter ratio of more than 2:1 are more difficult to cast, because of breaking or shifting of the unsupported end of the core. General rules for designing cored holes in investment castings are embodied in the

(a) Tool relief undercut is incorrectly located here, because it prevents withdrawal of a metal core. (b) Relocation of the relief notch as shown permits easy withdrawal of core.

Fig. 21

Tapered hub on this investment cast propeller (a) would require special core pulls to free the wax pattern from undercuts (b). A simplified parting line would be possible if full radii were not required at blade edges (c) and (d).

96 / Casting Design and Performance dimensional recommendations of Table 1. Because other factors of the design can influence these values, and because some foundry methods have greater capabilities than others, these figures are useful primarily as a general guide. Seven typical examples of cored holes and cavities easily incorporated in investment castings are illustrated in Fig. 25. None of these coring situations presented any difficulty. Small Cavities. The investment casting process permits coring of cavities that would be too small to be cored in sand castings. Typical of such cavities is the internal oil slot shown in Fig. 26. When the part shown was produced as a sand casting, this slot had to be milled in, and the hole leading to it drilled. Producing the part as an investment casting permitted the slot to be cored, and no subsequent machining on it was required. The change in method resulted in total savings of 41¢ per casting. Abutting Cores. In investment mold castings, as in permanent mold castings (Fig. 17),

(a) In the original design of this investment cast turbine blade, careful removal of the gates located on the contoured surface was required, to maintain the airfoil shape. (b) Simplified gating and gate removal were made possible by eliminating one notch and gating into a flat surface. The notch was machined later.

Fig. 23

where holes connect internally to form a sharp corner, the use of special cores is not required. In an investment casting, a special core would be required for the curved hole in Fig. 17(b), whereas that in Fig. 17(a) could be produced with regular mold equipment. In die casting, the design shown in Fig. 17(a) would be possible to manufacture, but the design shown in Fig. 17(b) would be impossible. Where sharp corners will reduce product performance (as by turbulence in fuel valves, fraying of insulation in electrical wire housings, or signal error in radar components), fillet radii should be incorporated. This requires expendable cores, which are more expensive than steel core pins. Figure 27(a) shows an example of an investment casting that required expendable cores. As designed, this casting could not be produced with steel core pins in the die used to produce the wax or plastic patterns. To permit the use of steel core pins, the casting would have to be redesigned as shown in Fig. 27(b). Six core pins would then be necessary. They would be located and pulled from the pattern as shown. An added operation would be required to plug the hole at core 1 after casting. This redesign has five core-pin junctions. It would be virtually impossible to produce these castings without prominent flash at these junctions, because of the difficulty of maintaining firm butt-type joints. Removing the flash would be an added expense. Unequal Expansion of Mold and Core. In investment castings, unequal heating and expansion of the mold and core can cause nonuniformity of wall thickness. This is particularly likely in castings that have tortuous passages formed by cores completely surrounded by metal. Typical of such castings is the one shown in Fig. 28. When this casting was poured, the mold was able to dissipate heat at a greater rate than the core. Consequently, the core became hotter than the mold and expanded more. The rigid mold resisted the expansion of the core, causing the core to bend. This bending moved part of the core closer to one wall and farther away from the other, thus producing a thinner wall on one side of the casting than on the other. Recommended minimum tolerances for the thickness of walls enclosing tortuous cored passages in investment castings are as follows:

(a) The cored hole in this investment cast propeller hub necessitates a relatively expensive gating arrangement. Because the gates were attached to the blades, contour machining was necessary for complete removal of the gates. (b) By eliminating the cored hole, a simple and less costly gate is possible. (The center hole was subsequently drilled.)

Fig. 24

Length of hole, in.

Under 2 2 to 4 Over 4

Wall thickness tolerance, in.

þ0.005  þ0.010  þ0.012 

Designing to Eliminate Cores Useful as they are, cores add to the cost of producing castings; hence the designer should be alert to possibilities for avoiding them. A decision often depends on cost analysis. An example is shown in Fig. 29. In the original design of this casting, Fig. 29(a) a core is required to permit molding of the “hook” shape. The possible redesign shown in Fig. 29(b) would permit easy removal of the pattern from the sand, eliminate the need for a core, and effect a saving in molding cost. Another example is provided in Fig. 6 of the preceding article. Figure 30 shows a sand cast malleable iron wheel hub for which redesign eliminated a ring core and at the same time provided a stronger casting. As originally designed, Fig. 30(a), the eight ribs and eight small bosses prevented this casting from being molded with the parting line parallel to the axis of the hole. Furthermore, adjacent to the flange, the casting had a cross section smaller than either the flange or the extreme end of the casting. The undercut section that was thus formed prevented the pattern from being withdrawn from the mold in a direction perpendicular to the mounting flange. A ring core, as shown, was necessary to produce the shape. By revising the casting as shown in Fig. 30(b), the need for the ring core was eliminated and the shape could be withdrawn easily from the mold. By broadening the base of the tubular section the eight ribs were also eliminated. In the original design, the small diameter of the tubular section at the junction with the flange section was unable to withstand the forces of service. Eight strengthening ribs were required, to assure satisfactory performance of the casting in application. As redesigned, the broader base of the tubular section provided sufficient strength to permit elimination of the ribs. Figure 31 shows a sand cast launching frame, produced from 4330 steel, that was redesigned to eliminate cores. The original design of this casting called for the coring of two pockets approximately 100 mm (4 in.) deep, to reduce weight. In production, however, it was difficult to attain sound metal in this cored area. At the expense of adding weight, the cored pockets (and the cores) were eliminated in a redesign as shown. This not only corrected the defects but also facilitated production. The solid center was easily risered and encouraged a beneficial freezing pattern in the entire casting. Simplified Cores. Not all designs or redesigns can eliminate cores, but often cores can be simplified. This was accomplished for the fuel cell interconnect, an aluminum casting, shown in Fig. 32. A semi-permanent mold

Design for Economical Coring / 97 Table 1

General dimensional relations for cored cavities in investment castings

Length X, in.

Diameter Y, in.

Min wall, in.

For a typical ferrous casting 0.250 0.125 0.500 0.250 0.750 0.375 1.000 0.500 1.250 0.750 1.500 1.000 2.000 1.500 2.500 1.750

0.030 0.040 0.050 0.060 0.060 0.060 0.060 0.060

Inside diameter A, In.

As-east roundness (TIR), in.

0.092 to 0.250 0.251 to 0.500 0.501 to 1.000 over 1.000, add 0.020 in. per in. A diameter, in.

0.250 0.500 0.750 1.000

Coring Versus Drilling

0.012 0.016 0.020

B diameter, in.

As-cast concentricity (TIR), in.

0.750 1.000 1.500 2.000

0.008 0.010 0.016 0.020

Minimum diameter Y Depth W, in.

Ferrous, in.

Nonferrous, in.

0.188 0.250 0.500 0.625 0.750 0.750 1.000 1.000

0.188 0.250 0.312 0.375 0.438 0.500 0.562 0.625

0.250 0.500 0.750 1.000 1.250 1.500 2.000 2.500

Min draft

0 0 0 0 0 0 0 0

15´ 15´ 15´ 15´

H, in.

D, in.

Z, in.

Parallelism(a), in.

0.250 0.500 0.750 1.000 1.250 1.500 2.000 2.500

0.125 0.250 0.375 0.500 0.625 0.750 0.875 1.000

0.500 0.750 1.000 1.250 1.500 1.750 2.000 2.250

þ0.005  þ0.005  þ0.005  þ0.010  þ0.012  þ0.013  þ0.015  þ0.015 

(a) Parallelism may be regarded as the tolerance for dimension Z. T, in.

X, in.

a´, in.

Y, in.

b´, in.

0.125 0.250 0.125 0.250 0.500 1.000 2.000

0.250 0.500 0.750 1.000 0.500 0.500 2.500

þ0.004  þ0.005  þ0.005  þ0.005  þ0.008  þ0.010  þ0.015 

0.125 0.250 0.500 0.750 0.250 0.250 1.500

þ0.003  þ0.004  þ0.005  þ0.005  þ0.005  þ0.008  þ0.010 

Y, in.

X

a´ min

0.250 0.500 1.000

15 30 45

þ1  þ1   þ1 

X diameter, in.

Y, in.

Z

0.250 0.500 0.250 0.500 0.750 and up

0.500 0.750 0.500 0.750 1.500

15 30 60 90 120

R, in

/64 /32 1 /16 1/16 1 /16 1 /16 1 /8 1 1



þ0  þ0   þ0  þ0   þ0 

casting as originally designed, it required a contoured sand core to form the inside configuration shown. A revision to eliminate defects encouraged a further revision to simplify the core. By moving the wall to coincide with the inside contour of the flanges, a simple straight-through metal core sufficed. The redesigned castings (now permanent mold, because of the metal core) were more accurate and less costly.

30´ 30´ 30´ 45´ 60´

A, in.

B, in.

C, in.

D, in.

E, in.

n´, in.

0.500 0.750 1.000 1.250

1.000 1.500 2.000 2.500

0.250 0.500 0.750 1.000

0.750 1.000 1.500 2.000

0.375 0.500 0.750 1.000

þ0.004  þ0.006  þ0.008  þ0.010 

Although for many castings coring is essential to successful production, for others it is advisable to omit cores and to remove excess metal by other means. The choice may be based on considerations of soundness, dimensional accuracy, economy, or producibility. For example, if a casting is to have one or more round holes, these may be produced with greater accuracy or economy by subsequent boring or drilling, rather than by cores. As a general rule, it is often cheaper to core larger holes if an appreciable amount of metal can be saved, or if machining operations can be speeded up or eliminated. It is usually more expensive to core smaller holes unless a large amount of machining expense can be saved. Eliminating cored cavities in favor of subsequent drilling may also permit more efficient risering in some designs, thus preventing defects. In the casting illustrated in Fig. 33, The core adjacent to the gate impeded flow of metal into the mold. This caused cold shuts at the end of the casting opposite the gate. Dispensing with this core permitted metal to flow freely to all parts of the cavity, so that the cold shuts were eliminated and sound castings produced. The previously cored hole was later drilled. In the sand casting shown in Fig. 34, it was cheaper to core and finish-drill the 0.84-in.diameter hole than to cast solid and produce the hole entirely by drilling. The cost of metal saved was greater than the cost of coring. Actual savings in this instance amounted to 1 ½¢ for each casting. If the cored hole had been much smaller, however, this would not have been true, because the core cost would have remained relatively constant whereas the value of the metal saved would have been reduced in proportion to the volume of the hole.

98 / Casting Design and Performance

Fig. 25

Examples of cored holes and cavities easily produced in investment castings

Oil slot was easily cored when part was produced as an investment casting, but had to be milled in when part was sand cast. Total saving of 41¢ per casting was realized by changing from sand to investment casting.

Fig. 26

Because it was completely surrounded by molten metal, the core that formed the passage through this investment casting expanded more in length than the mold when the metal was poured. This caused bending of the core, which influenced the casting wall thickness.

Special coring practice involving additional cost would be required to produce this investment casting as shown in (a). Normal practice using metal core pins in the pattern die is possible for the casting as shown in (b). Design (b) is less expensive than (a).

Fig. 27

Fig. 28

Fig. 30

Fig. 29

A redesign that eliminated a core and reduced the cost of the casting

An improved design that eliminated one core and eight ribs from a sand casting. This resulted in a stronger, more economical part.

Design for Economical Coring / 99

A cored hole 0.84 in. in diameter was incorporated in this casting to reduce machining cost. It was less expensive to core and drill than to drill through solid metal.

Fig. 34

Fig. 31

By eliminating two cored pockets from this 4330 steel sand casting, a better solidification pattern was established and defects were eliminated.

A sand core was required to produce the center configuration of this semipermanent mold aluminum casting. By increasing the wall thickness and placing the inner face of the wall in line with the inner edge of the flanges, defects were eliminated and a simple metal core produced the center cavity.

Fig. 32

Fig. 33

Better flow of molten metal was obtained by eliminating the 0.406-in.-diameter core adjacent to the gate of this investment casting. This enabled the metal streams at the extreme end of the cavity to unite properly.

Casting Design and Performance Pages 101–119

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Casting Design and Geometry Michael Gwyn, Advanced Technology Institute and American Foundry Society Technical Department

CASTINGS possess many inherent advantages that have long been accepted by the design engineer and metal parts user. In terms of component design, casting offers a great amount of flexibility. The casting process permits the formation of streamlined, intricate, integral parts of strength and rigidity that are not obtainable by other methods of fabrication. The shape and size of the part are also primary considerations in the design and application of cast parts. In this category, the possibilities of metal castings are unsurpassed. The flexibility of cast metal design offers a wide scope in converting ideas into an engineered part. The freedom of design offered through the metalcasting process allows the designer to accomplish several tasks simultaneously. These include the following:  Freedom of design to optimize functionality

and manufacturability

 Net or near-net shape design  Intricate components can be produced as a

single cast part.

 Few restrictions on part weight or size  Almost all metals and alloys can be cast.  Optimal appearance

Structural design engineers who work successfully with castings commonly design in a narrow group of casting types poured from familiar alloys (such as the family of irons or the 300 series of aluminum) and molded from familiar metalcasting processes (such as green sand or nobake). Rules of thumb and lists of design considerations are available in many design handbooks and references as an aid for common design situations. However, close examination of these recommended rules often reveal conflicting guidance in many cases. For example, the use of gusseting instead of mass for stiffness may be labeled “recommended” in one set of design rules and “poor” in another. Further, when a design engineer goes from a familiar casting design realm to an unfamiliar one, unexpected results may occur. For example, when substituting aluminum bronze for ductile iron (for a given metalcasting mold process), there is likely to be trouble with the usual ductile ironstyle geometry. Good aluminum-bronze geometry is different from typical ductile iron geometry, and

the molding process may need to supplement the different geometry with heat-transfer techniques. Not suspecting this, the design may suffer from no-quotes, higher-than-expected prices, or metalcaster requests for design changes. How are design engineers to know the differences in a successful casting geometry between that of an aluminum bronze and a ductile iron? Moreover, if the design engineer did know that information, what would be the proper course of design action? The answer lies in a better understanding of the relationship among geometry, metallurgy, and physics.

Casting Design Parameters The key to consistent and good casting design is how to choose a suitable geometry. There are six parameters (based on physics) that underlie cost-effective casting design. Four are based on casting properties:    

Liquid metal fluid life Solidification shrinkage type Slag/dross formation tendency Pouring temperature

Two other design parameters are based on structural properties:  Section modulus  Modulus of elasticity

These six parameters, as a system, can drive the geometry of casting design from a process standpoint. Geometry is not only the result of product function design, but it also has a powerful influence on the castability of cost-effective castable designs that are economically produced, machined, and assembled into a final product.

Fluid Life Fluid life more accurately defines the alloy liquid characteristics than does the traditional term fluidity. Molten metal fluidity is a dynamic property that depends on several factors that include:  Mold materials and surface characteristics  Alloy composition (and alloying effects on

freezing range)

 Surface tension and surface films  Gas content and suspended inclusions     

(which may alter fluidity directly or indirectly by effects of surface tension) Degree of superheat Rate of pouring Rate of heat transfer to the surroundings Heat of fusion and solidification Viscosity

Viscosity is a factor, but fluidity should not be confused with viscosity. In fact, molten metals have a relatively low kinematic viscosity (viscosity divided by density). Fluid life is affected more by other factors that change as the alloy is delivered from a pouring ladle, diecasting chamber, and so on into a gating system and finally into the mold or die cavity. Heat transfer reduces the metal temperature, and oxide films form on the metal front as this occurs. Fluidity decreases most rapidly with temperature loss, and it can decrease significantly from the surface tension of oxide films. Temperature also is not the sole basis of fluidity at a given moment. For example, some aluminum alloys at 650 to 750  C (1200 to 1400  F) have excellent fluid life. However, some molten steels at 1650  C (3000  F) have much shorter fluid life. In other words, a molten alloy fluid life also depends on chemical, metallurgical, and surface tension factors. Fluid life can also be related to freezing range. In alloys with a wide freezing range, casting is more difficult, because there is a stronger tendency for nonuniform solidification. Extra effort may be needed with the use of chills to create a steeper thermal gradient in portions of the mold when casting alloys with a wide freezing range. Fluid life thus affects castability and the possible design configuration of a casting. The design limitation may be the minimum section thickness that can be cast reliably, the maximum length of a thin section, the fineness of cosmetic detail (such as lettering and logos), or the accuracy with which the alloy fills the mold extremities. However, it is important to understand that moderate or even poor fluid life does not limit the cost-effectiveness of design. Knowing that an alloy has limited fluid life, the designer will make sure the part features:

102 / Casting Design and Performance  Softer shapes and larger lettering  Finer detail in the bottom portion of the

mold, where metal arrives first, fastest, and generally hottest  Coarser detail in the upper portions of the mold, where the metal is slower to arrive and more affected by oxide films and solidification “skin” formation. Even an alloy with good fluidity, when overexposed to oxygen, may form a high-surface-tension oxide film that makes the fluidity die, rounding off the leading metal front as it flows.  More taper toward thin sections Some alloys, such as 356 aluminum, have been specifically designed metallurgically to enhance fluid life. In the case of 356, the high heat capacity of silicon atoms revives aluminum atoms as their fluid life begins to wane.

Solidification Shrinkage The three distinct stages of shrinkage as molten metals solidify are liquid shrinkage, liquidto-solid shrinkage, and patternmaker’s contraction. Liquid shrinkage is the contraction of the liquid before solidification begins. It is not an important design consideration, because the liquid metal will feed itself with additional metal Table 1

while in the mold. Feeding refers to liquid metal flowing into a mold to replace the volume of metal lost due to shrinkage. Feeding is more important for liquid-to-solid shrinkage. Liquid-to-solid shrinkage is the shrinkage of the metal mass as it transforms from the disconnected atoms and molecules of the liquid into the structured building blocks of solid metal. The amount of solidification shrinkage varies greatly from alloy to alloy. Table 1 provides a guide to the liquid-to-solid shrinkage of common alloys. As shown, shrinkage can vary from nearly no shrinkage to high shrinkage volumes. Alloys are further classified based on their solidification type: directional, eutectic-type, and equiaxed (Table 1). The type of solidification shrinkage in a casting is just as important as the amount of shrinkage. Specific types of geometry can be chosen to control internal integrity when solidification amount or types are a problem. Solidification on and perpendicular to the casting surfaces is known as progressive solidification. At the same time, solidification moves at a faster rate from the ends of the section(s) toward the source of feed metal (risers)—this is known as directional solidification. Directional solidification moves faster from the ends of the sections because of the greater amount of surface area through which the solidifying

metal can lose its heat. The objective is for directional solidification to beat out progressive solidification before it can block the source of the feed metal. Directionally solidifying alloys require extensive risering and tapering, but they also have the capability for excellent internal soundness when solidification patterns are designed properly. The eutectic-type alloy is the most forgiving of the three. Such alloys typically have less solidification shrinkage volume. Risers are much smaller and, in special cases, can be eliminated by strategically placed gates. The key feature with these alloys is the extended time that the metal feed path stays open. The plate solidifies more uniformly all over and all at once, similar to eutectic solidification. Eutectic-type alloys are less sensitive to shrinkage problems from abrupt geometry changes. Alloys that exhibit equiaxed solidification respond the most dramatically to differences in geometry. Shrinkage in these alloys tends to be widely distributed as micropores, typically along the center plane of a casting section. The reason is that solidification occurs not only progressively from casting surfaces inward and directionally from high-surface-area extremities toward lower-surface-area sections but also equiaxially via solidified islands in the middle of the liquid metal. These islands of solidification interrupt the liquid pathway of directional

Guide to the liquid-to-solid shrinkage of common alloys

Alloys are further classified based on their solidification type: directional, eutectic-type, and equiaxed. Solidification shrinkage Alloy group

Irons Gray iron Compacted graphite Ductile iron Austempered ductile iron I White iron Malleable iron High-alloy iron High-silicon iron Carbon and low-alloy steels Carbon Low alloy High-alloy steels Martensitic Partly austenitic Fully austenitic High manganese Superalloys Nickel base Cobalt base Nonferrous Aluminum 356, 357 Aluminum 201, 206

Fluid life

Type

Amount

Pour temperature 

Gas microporosity

Slag/dross

Excellent Excellent/good Good Good Excellent Excellent Fair/good Good

Eutectic-type Eutectic-type Eutectic-type/directional Eutectic-type/directional Directional Directional Equiaxed Equiaxed

Very small Small Small/moderate Small/moderate Moderate/large Moderate/large Moderate Moderate

2500–2600 2500–2600 2500–2600 2550–2650 2500–2600 2500–2600 2600–2700 2550–2650

Poor Poor

Directional Directional

Large Large

2850–3000  F 2850–3000  F

Poor/fair Fair Fair/good Fair

Directional Eutectic-type Equiaxed Equiaxed

Moderate/large Moderate Moderate Moderate/large

2850–2950 2800–2900 2775–2875 2725–2850

Poor Excellent

Directional Directional

Moderate Little

2600–2800  F 2600–2800  F

Moderate(+) tendency Moderate(+) tendency

Moderate Little/moderate

Excellent Good/fair

Eutectic-type Equiaxed

Little Moderate

1300–1400  F 1300–1400  F

High tendency High tendency

Moderate Moderate/large

F F  F  F

Higher Higher Higher Higher

tendency tendency tendency tendency

Little/moderate Little/moderate Little Little/moderate



Moderate (+) tendency Moderate () tendency Low tendency Moderate tendency High(++) tendency High(++) tendency

Large Large Little Moderate/large Very large Very large

Aluminum die-cast alloys are dominated by the diecasting process, and castability characteristics are less significant in this context. Magnesium ZE43 Excellent Directional Moderate 1300–1400 Magnesium ZK51A Excellent Directional Moderate 1300–1400 Magnesium AM60A Excellent Eutectic-type Little 1300–1400 Magnesium AZ91C Excellent Eutectic-type Little 1300–1400 Magnesium die-cast alloys are dominated by the diecasting process, and castablity characteristics are less significant in this context. Aluminum bronze Fair Equiaxed Moderate/large 2000–2150 Silicon bronze Fair Eutectic-type Little 1900–2050 Red brass 85-5-5-5 Fair (+) Equiaxed Moderate 1950–2100 Yellow brass Fair () Eutectic-type Moderate 1800–1950 Titanium Very good Eutectic-type Little 3300–3400 Zirconium Fair (+) Eutectic-type Little 3300–3400

F F  F  F  F  F  F  F 



F F  F  F 

 

F F F  F  F  F  

Low()tendency Low()tendency Low tendency Low(+)tendency Low()tendency Low()tendency Low(+)tendency Low(+)tendency

Little Little Little/moderate Little/moderate Little Little Little/moderate Moderate

Low tendency Low tendency

Moderate Moderate

Moderate Moderate Moderate Moderate

Moderate Moderate Little/moderate Little/moderate

tendency tendency tendency tendency

Casting Design and Geometry / 103 solidification. Gradually, the pathways freeze off, leaving micropores of shrinkage around and behind the islands that grew in the middle of the pathway. Larger risers, thicker sections, and tapering are counterproductive, causing micropores to coalesce into larger pores across more of the casting cross section. Microporosity is kept small and confined to a narrow midplane in the casting section by more thermally neutral geometry with smaller, further-spaced risers. A significant bilateral and reciprocal relationship exists between solidification shrinkage and geometry. Most simply, eutectic-type solidification is tolerant of a wide variety of geometries; the least reciprocity is required. Most complexly, equiaxed solidification requires the most engineering foresight in the choice of geometry and may require supplemental heat-transfer techniques in the mold process. In the middle lies directional solidification. While capable of the worst shrinkage cavities, it is the most capable of very high internal integrity when the geometry is properly designed. Well-planned geometry in a directionally solidifying alloy can eliminate not only shrinkage but the need for any supplemental heat-transfer techniques in the mold. In fact, the real mechanism behind the bilateral and reciprocal relationship between solidification shrinkage and geometry is heat transfer. All three modes of heat transfer—radiation, conduction, and convection—are involved in solidification of castings, and all three depend on geometry for transfer efficiency. Convection and conduction are very important in casting solidification, and transfer rates are highly affected by geometry. Thermal expansion/contraction, or patternmaker’s contraction, is the contraction that occurs after the metal has completely solidified and is cooling to ambient temperature. This contraction changes the dimensions of the casting from those of liquid in the mold to those dictated by the alloy rate of contraction. So, as the solid casting shrinks away from the mold walls, it assumes final dimensions that must be predicted by the pattern- or diemaker. This variability of contraction is another important casting design consideration, and it is critical to dimensional accuracy. Tooling design and construction must compensate for it. Achieving dimensions that are “just like the blueprint” requires the metalcaster’s patternand/or diemaker to be included. In some cases involving metal molds, such as diecasting and permanent mold casting, the die material is stronger than the casting while it cools. Geometric features may actually pin a casting in the die. If this occurs, the casting is deformed or stretched during cooling, making the prediction of the final geometric dimension impossible. The unpredictable nature of thermal expansion/contraction makes tooling adjustments inevitable. For example, a highly recommended

practice for critical dimensions and tolerances is to build the patterns/dies/coreboxes with extra material on critical surfaces, so that the dimensions can be fine-tuned by removing small amounts of tooling stock after capability castings have been made and measured.

Structural Properties As described, the properties of castability influence geometry, and a well-chosen geometry by designers can offset the metallurgical nature of the more difficult-to-cast alloys. Likewise, casting engineers can benefit from an understanding of the structural design requirements of a section. Geometry is the key to consistency, and both the designer and metalcaster must understand the interplay of geometry on structural design and castability. In terms of structural geometry, castings can easily apply shape to structural requirements. Most casting designs are used to statically or dynamically control forces. When designing a component structurally, a design engineer is generally interested in safely controlling forces through choice of allowable stress and deflection. Although choice of material affects allowable stress and deflection, a very significant factor is geometry. Geometry has a direct influence on the stiffness and stresses in a structure. Casting has wide capability in producing various shapes, and computer-generated solid models and rapid prototypes provide flexibility in the design stage. However, optimal geometry for allowable, uniform stress may not be acceptable geometry for castability. When a foundry engineer quotes a design that considered structural geometry only, requests for geometry changes are likely. At this point, the geometry adjustments for castability may be more substantial than the solid model software can “tweak.” The result can be no-quotes, higher-than-expected casting prices, or starting over with a new solid model. A practical solution to this problem is to concurrently engineer product geometry while considering structural, metalcasting, and downstream manufacturing needs. The result can be a casting geometry optimized for both functionality and manufacturability. The most efficient technique is to make engineering sketches or marked sections and/or views on blueprints. The idea is to explore overall geometry before locking into a solid model too quickly. Engineering sketches or markups are easy and quick to change—even dramatically—in the concurrent brainstorming process; solid models are not. Solid models can provide important analysis and refinement of design, but two key structural properties that should be emphasized for castable design are section modulus and modulus of elasticity. The modulus of elasticity defines the inherent stiffness or rigidity of a material,

while the section modulus (or section stiffness) depends on the geometric (section profile) configuration of the part. These two structural factors are critical in knowing how to choose a suitable geometry for a castable design. Modulus of elasticity is an important parameter in structural design, because it is directly related to geometric deflection under load. A larger modulus of elasticity means less deflection. Modulus of elasticity varies widely among materials, and it varies significantly among metals; that is, some metals are considerably stiffer than others. A given alloy family (say steel or aluminum) also tends to have the same modulus value; for example, the entire family of steels (carbon, low alloy, and high alloy) has the same elastic moduli value of approximately 30  106 psi. In contrast, all aluminum alloys have elastic moduli that are three times smaller ( 10  106 psi). Thus, a steel casting is three times stiffer than an aluminum casting of identical geometry, simply because steel is stiffer than aluminum. Section Modulus. Evaluation of section modulus depends on the geometric configuration of the part. As the design engineer well knows, the classic formulas for bending stress, torsional stress, and deflection are relatively simple. Each, however, contains the same parameter, section modulus, which is a function of shape and difficult to compute. However, a quick, simple way to compute or estimate section modulus can be helpful in evaluating options for a more castable geometry. The key is to estimate the area moment of inertia, which is the basic measure of section modulus (see the section “Quick Method for Estimating Area Moment of Inertia from Sketches” in this chapter).

Applying the System This six-faceted system (based on four casting properties and two structural properties) is capable of optimizing:  Geometry for castability, structure, and

downstream processing (machining and assembly)  Process geometry (risering, gating, venting, and heat-transfer patterns) in the mold The process geometry forms the casting geometry. Quickly sorting through possible casting and process geometries by marking up blueprints or by making engineering sketches is an effective way to rapidly improve system geometry. An elegant result of a good brainstorming markup session can be a solid model of the casting and its process geometry, the basis of rapid prototyping and/or computerized testing. The objective is to explore geometry possibilities, looking for an ideal shape that is both castable in the chosen alloy and allowable in stress and deflection for that alloy. As noted, there is

104 / Casting Design and Performance great variety in the four metallurgical characteristics that govern alloy castability. Similarly, great variety exists among metals in their allowable stress and deflection. Therefore, an ideal casting shape for all six of the casting design factors is not necessarily a trivial exercise. For alloys that have good castability, choosing geometry for allowable stress and deflection is the best place to start. For alloys with less-than-the-best castability, it is better to first find geometry that assists castability and then modify it for allowable stress and deflection. Not all alloys are like ductile iron, which is highly castable, relatively resistant to stress, and moderately resilient against deflection. For ductile iron, many geometries may be equally acceptable. Martensitic high-alloy steel has fair-to-poor castability but can have amazing resistance to stress and can tolerate very large deflections without structural harm. Therefore, structural geometry is easy to develop, but a coincidental castable shape is more difficult to design. Premium A356 aluminum has good castability but rather weak resistance to stress and low tolerance for deflection. Carefully chosen structural geometry, however, combined with solidification enhancements in the molding process, has resulted in extremely weight-effective A356 structural components for aircraft, cars, and trucks. Optimizing casting geometry using the sixparameter system is not difficult. The six casting and structural characteristics influence important variables in designing, producing, and using metal castings. These variables include:       

Casting method Design of casting sections Design of junctions between casting sections Surface integrity Internal integrity Dimensional capability Cosmetic appearance

Both the designer and metalcaster possess a vital ally to streamline any casting design: casting geometry. Casting geometry is the most powerful tool available to improve castability of the alloy and mechanical stiffness of the casting.

Design of Junctions A junction is a region in which different section shapes come together within an overall casting geometry. Simply stated, junctions are the intersection of two or more casting sections. The four junction types are L-, T-, X-, and Ydesigns. Designing junctions is the first step to finding castable geometry via the six-faceted system for casting design. There are major differences in allowable junction geometry, depending on alloy shrinkage amount and pouring temperature. One alloy may allow abrupt section changes and tight geometry, while another alloy requires considerable adjustment of junction geometry, such as radiusing, spacing, dimpling, and feeding.

Considerations of Secondary Operations in Design System-wide thinking also must include the secondary operations, such as machining, welding and joining, heat treating, painting, and plating. One aspect that affects geometry is the use of fixturing to hold the casting during machining. Frequently, the engineers who design machining fixtures for castings are not consulted by either the design engineer or the foundry engineer as a new casting geometry is being developed. Failure to do so can be a significant oversight that adds machining costs. If the casting geometry has been based on the four casting characteristics of the alloy, then the designer knows the likely surfaces for riser contacts and may have some idea of likely parting lines and core match lines. These surfaces and lines will be irregularities on the casting geometry and will cause problems if they contact fixturing targets. It is best to define the casting dimensional datums as the significant installation surfaces, in order of function priority, based on how the casting is actually used. Targets for machining fixtures should be consistent with these datum principles. Nothing is more significant in successful computer numerical control and transfer line machining of castings than the religious application of these datum fixture and targeting principles.

Drawings and Dimensions The tool that has had the most dramatic positive impact on the manufacture of parts that reliably fit together is geometric dimensioning and tolerancing (GD&T), as defined by ANSI Y14.5M—1994. When compared to traditional (coordinate) methods, GD&T:  Considers tolerances, feature-by-feature  Minimizes the use of the title block toler-

ances and maximizes the application of tolerances specific to the requirement of the feature and its function  Is a contract for inspection (tells the person reading the drawing how to measure the part), rather than a recipe for manufacture In other words, GD&T specifies the tolerances required feature-by-feature in a way that does not specify or suggest how the feature should be manufactured. This allows casting processes to be applied more creatively, often reducing costs compared to other modes of manufacture, as well as finish machining costs. GD&T encourages the manufacturer to be creative in complying with the dimensional specifications of the drawing, because the issue is compliance with tolerance, not necessarily compliance with a manufacturing method. By forcing the designer to consider tolerances

feature-by-feature, GD&T often results in broader tolerances in some features, which opens up consideration of lower-cost manufacturing methods, such as castings.

Factors That Control Casting Tolerances How a cast feature is formed in a mold has a significant effect on the tolerance capability of the feature. The following six parameters (in order of preference) control the tolerance capability of castings. Molding Process. The type of molding process has the greatest single influence on tolerance capability. How a given molding process is mechanized and the sophistication of its pattern or die equipment can refine or coarsen its base tolerance capability. Casting Weight and Longest Dimension. Logically, heavier castings with longer overall dimensions require more tolerance. These two parameters have been defined statistically in tolerance tables for some alloy families. Mold Degrees of Freedom. This parameter is least understood. Just as some molding processes have more mold components (mold halves, cores, loose pieces, chills, etc.) than others, some casting designs require more mold components. Each mold component has its own tolerances, and tolerances are stacked as the mold is assembled. More mold components mean more degrees of freedom, hence more tolerance. Good design for tolerance capability minimizes degrees of freedom in the mold for features with critical dimensions. Design engineers must be aware of this and consider what really needs to be controlled when tolerancing a design. For example, is a position tolerance really needed relative to a datum or another specific feature? Interdependent features with tight position tolerances often can be located in a single-piece portion of the molding tool, reducing the number of degrees of freedom and improving dimensional accuracy. Draft. It is common for casting designs to ignore the certainty of draft, including mold draft, draft on wax and/or Styrofoam patterns made from dies, and core draft. Since 1 of draft angle generates 0.017 in. of offset per inch of draw (approximately 0.5 mm/30 mm), draft can quickly use up all of a tolerance zone and more. Patternmaker’s Contraction. The uncertainty of patternmaker’s contraction is why metalcasters normally recommend producing first-article and production-process verification castings (sometimes called sample or capability castings) to establish what the dimensions really will be before going into production. A common consequence of patternmaker’s contraction uncertainty is a casting dimension that is out of tolerance, not because it varies too much but because its average value is too far from nominal. In other words, the dimension contracted more or less than what was expected.

Casting Design and Geometry / 105 Cleaning and Heat Treating. Many casting dimensions are affected by downstream processing. At the least, most castings are touched by abrasive cutting wheels and grinding—even precision castings. Many castings are heat treated, which can affect straightness and flatness. When considering the breadth and depth of the importance of geometry in casting design, from its influence on castability, the geometry of gating/risering, structural form, cosmetic appearances, and downstream fixturing, extensive brainstorming of geometry is highly recommended. The standard for optimal casting geometry is high, but the possibilities for geometry are limitless. Find ways of exploring geometry quickly, such as engineering sketching, before committing to a print or solid model.

Quick Method for Estimating Area Moment of Inertia from Sketches As noted, estimating the area moment of inertia is the basic measure of section modulus. The difficulty in computing area moment of inertia for casting shapes is one of the hidden reasons for the design and use of fabrications. Fabrications are made from building blocks of wrought shapes, such as I-beams, rectangular bars, angles, channels, and tubes. These shapes, which are simple and constant over their length, have area moments of inertia that are easy to calculate or are available in handbooks. Consequently, stress and deflection calculations are relatively easy. Fabricated designs, however, are heavy and nonuniform in stress compared to a well-designed casting for the same purpose. Section modulus (or, more specifically, the area moment of inertia) is a function of shape and difficult to compute. Therefore, a quick, simple way to compute or estimate the area moment of inertia can be useful in brainstorming for a casting geometry with both structural fitness and castability. Solid modeling is used for both structural analysis by design engineers and for mold filling and solidification analysis for the casting engineer. However, it can be helpful to markup and brainstorm engineering drawings before building a solid model. This can be an efficient structural evaluation of geometry in casting design. To take full advantage of engineering sketching/print marking as a way to brainstorm geometry, one must be able to quickly evaluate stress and deflection at important cross sections in the sketches Although there are five kinds of stress (tension, compression, shear, bending, and torsion), the interesting ones for complex structures are bending and torsion, and their equations are shown in Table 2. (If more than one type of stress is involved in the same section, the principle of superposition allows the individual stress types to be analyzed separately and then added together; once again, the larger of the stresses to be combined is usually from bending or torsion.)

Equations for deflection are more complexlooking and are different for each type of loading geometry. However, a simplified relationship for bending deflection can be used to estimate whether deflection will increase or decrease with a given geometry. The relationship for deflection (d) from bending around axis x-x is: dxx /

ðBending momentÞðSection lengthÞ2 E  Ixx

where E is the modulus of elasticity, and Ixx is the moment of inertia. This is not an exact analytical relation but rather a proportional (/) relation to estimate how changes in geometry affect deflection. If the area moment of inertia can be quickly estimated, engineering isometric sketches can be roughly evaluated in estimating the effects of stresses on deflection. Maximum tensile stress in bending is often most critical in structural design. Section modulus is defined as the area moment of inertia divided by the maximum distance from the center of bending (centroid) to the outermost edge of the casting cross section. Section modulus is similar to a stiffness index, because it considers not only magnitude of area moment of inertia but also maximum section depth. If maximum section depth increases faster than area moment of inertia, a geometry change can actually increase maximum tensile stress rather than reduce it. This index, termed section modulus, accounts for that potential problem. The estimation method recommended is based on three principles. One is intuitive, and the other two are from the mathematics of engineering mechanics. The Design Engineer’s Sense of Load Magnitudes and Component Size/Shape. Engineers routinely use this sense to sketch sized shapes that are in the ballpark of the final design. Casting engineers can learn this sense, and when they do, they become effective concurrent engineering partners in their customers’ casting designs.

The Equation for Area Moment of Inertia (Fig. 1). Although the calculus for an interesting casting cross section can be very difficult, the relationship expressed between depth of section (Y) and change in cross-sectional area (dA) is very simple. The position and shape of the two rectangles in Fig. 1 (top) demonstrate this simple yet powerful relationship. The change in shape of the inside of the tube (at bottom of Fig. 1) is an even more dramatic illustration. Calculations were not made in either case, but the qualitative impact of Y2dA on stiffness and stress is unmistakable. Area Moment of Inertia. Once the engineering sense of structural size and Y2dA has been applied qualitatively to a sketched cross section, the parallel axis theorem can be applied to simple building blocks in the cross section to estimate area moment of inertia quantitatively. A numerical value for area moment of inertia is required to calculate the stress level in the sketched cross section. The parallel axis theorem is illustrated in Fig. 2.

Specific Design Tips for Castings The optimal geometry for casting and its structural properties offers many opportunities for improving quality while decreasing costs

Table 2 General relationships between stress and geometry sxx

taverage

Bending around axis x-x tension/ compression, max Tension/ compression, max Average compression Shear

ttorsional

Shear, max

saxial sbearing

thorizontal Shear from bending at slice n-n

max Þ ¼ ðBending momentÞðY Ixx

force ¼ CrossAxial -sectional area

Bearing force ¼ Projected contact area force ¼ CrossShear -sectional area max Þ ¼ ðTorqueÞðRadius J

forceÞðStatic moment;partial areaÞ ¼ ðShear ðlxx ÞðWidth of partial area slice n-nÞ

Geometric aspects in calculating the area moment of inertia in rectangular coordinates (Ixx) or polar coordinates (J). The essence of area moment of inertia can be quickly gleaned from its base formula. Moment of inertia is an important geometric factor because stresses in bending and torsion are inversely proportional (Table 2).

Fig. 1

106 / Casting Design and Performance Area moment of inertia Parallel axis theorem

Area moment of inertia Axial - vs.- polar

+Y

Y

Rule of pythagoras 2

2

x

2

r =y + x 2

2

dA

2

r dA = y dA + x dA 2

2

2

òr dA = òy dA + òx dA –X

Ix-x

=

ò or

y

r

XMAX

YMAX Y2dA YMIN

Iy-y R

J

=

ò0

J

=

Ix-x

=

ò

+X

2

X dA YMIN

Find centroid of overall cross-section by weighted average of component shape

X

positions:

YG

R1 Y1

YG1

X

YG = Sum of AR1 Y1+ AR2 Y2

Axis 0

+ .... + ASHAPEn Yn

2

r dA + I y-y

Y YG12AR1

Divided by

IR1X-X = IR1 +

Sum of all areas

IR1X-X = b1h1 /12 + YG1 (b1h1)

3

2

Area moment of inertia Centroid position Rectangles

AXls 2

Area = bh

Y

Centroid h X

X b Y

X-X = h/2

Y-Y = b/2 Axls 1 Moment of inertia Ix-x = bh3/12 Iy-y = b3h/12

Parallel axis theorem and general equations in calculating area moment of inertia. Good casting designs typically have complex cross sections, and although the general equations for moment of inertia are not trivial, the parallel axis theorem provides an estimating technique.

Fig. 2

to add value. The objective of good casting design is to explore geometry possibilities, looking for an ideal shape that is both castable in the chosen metalcasting alloy and allowable in stress and deflection. For alloys that have good castability, choosing geometry for stress and deflection is the best place to start. For alloys with less-than-the-best castability, it is better to first find geometry that assists castability and then modify it for allowable stress and deflection.

Design for Functional Packaging Product engineers and designers are faced with the challenge of adding more functions at reduced costs, often with additional constraints on physical space. No matter how well an individual component or subsystem is designed, the product is useless if it does not fit into a limited envelope. This is especially true in transportation vehicles. Transportation system design is becoming more complex because the leading requirement has become functional packaging. Components must fit into a confined environment and survive interactions with mating and nearby components. Although functional packaging often is the foremost requirement in vehicle design, few product engineers and designers understand all aspects of functional packaging. Following are ten key points to consider when

developing a functional package in the case of components for transportation vehicles. Hard Points. Each vehicle has hard points used in its manufacture that typically remain fixed for many years across model changes and sometimes across completely different models. These points are used for locating and handling the vehicle as well as major assemblies during manufacture. The most obvious are the lifting points on vehicle frames used to carry them down the assembly line before suspensions and wheels have been mounted. Diesel engines for the truck industry are prime examples in which hard points for mounting the engine have remained constant across engine models for years. Begin every design with an understanding of the customer’s hard points. When developing a three-dimensional (3-D) computer-aided design (CAD) model, start by creating the customer’s hard points, then package the design around these constraints. If possible, model the hardware associated with the hard points as well. Mating Parts. The most obvious design consideration when developing a functional package is mating parts. Solid models of mating parts are routinely imported into CAD systems prior to developing a design. Although most designers and product engineers consider mating parts when developing a design, the allowable variation associated with them often is

overlooked. Dimensional variation due to tolerancing can cause numerous failure modes, including interference conditions and part mismatch. Using today’s design tools, 3-D models can be modified to examine these phenomena. Imported electronic models of mating parts can be modified to investigate the allowable variation of a given component. Nearby Parts. As with mating parts, other nearby components must be considered when developing a design package. The same issue of allowable variation often is neglected. As with mating parts, 3-D models can be modified to examine the maximum and minimum material conditions of nearby parts as well as positional variation allowed in tolerancing. Close Mechanical Members. When designing a product, nearby moving parts, such as linkages and control arms, must be understood. Often, random vibrations caused by road conditions may excite nearby components, causing movement. Many knocks and other noiserelated issues stem from such interactions. When developing a CAD design, model the 3-D envelope required by the movement of a nearby member as a solid space. The design must package around this space. Thermal Source. Transportation components and assemblies must be capable of surviving a wide range of temperatures. For example, the temperature in an engine compartment may be as low as 40  C (40  F) when starting a vehicle that has been parked outside in the cold winters of Manitoba, Canada. During operation a few minutes later, exhaust components, such as the catalytic converter and exhaust gas recirculation tube (EMF), may reach temperatures in excess of 600  C (1110  F). Consider the operating temperatures of mating and nearby components. When packaging a design, sufficient space must be allowed between a hightemperature member to avoid interactions. In some cases, heat shielding may be required, making packaging more difficult. EMF Sources. With the increased use of electronics in transportation products, the electromechanical output of components has become a consideration when packaging a design. If necessary, space must be maintained between nearby members to avoid EMF interactions. In some cases, EMF shielding may be required. Run Downs. Assembly of vehicle components and systems typically involves the use of threaded fasteners. When packaging a design, do not encroach on the space needed by tooling to “run down” a fastener. Consider the run downs needed on mating parts, nearby parts, and the component being designed. When developing a CAD design, model the 3-D cylinder required by a tool to run down a fastener. The design must package around this space. In some cases, the envelope needed for a rundown tool can be minimized by using specialty fasteners with internal drives, such as Allenheaded screws.

Casting Design and Geometry / 107 Future Parts. Often, components that are designed first occupy maximum space, making subsequently developed components more difficult to design and more costly to produce. When designing a component, consider what parts and assemblies will be added to complete the vehicle. The most neglected components have historically been wire harnesses and hoses of all types. Even fuel lines often are afterthoughts on many vehicles, designed after all other systems. Gaging Points. In many cases, transportation vehicle components and assemblies must be aligned or modified during manufacture. Although such cases are undesirable, they do exist. At least three points in space are required to locate a component. When developing a package, include and specify these points and allow clearance so that these points may be reached by the gage hardware. When developing a CAD design, the gage hardware should be modeled as a solid. This is yet another consideration that consumes functional design space. Service Window. Service requirements for transportation vehicles range from changing lubricants to replacing perishable components such as batteries and oil filters. Clearance for service tools must be computer-modeled as a solid space. A functional package must be developed around these pathways. Access to perishable components also must be maintained. Modeling extraction paths for such components is difficult. At the very least, a minimum clearance must be maintained to allow for the removal of such parts.

Add Strength, Not Mass Often, when designers and product engineers are faced with making a component stronger, the simplest solution at hand is to add thickness. Adding more material can be a quick fix. Using basic physical theory, doubling the material thickness in a select location should double the strength. However, such a basic, straightforward theory makes many assumptions that are rarely true. The theory that increasing thickness adds strength assumes that the added material is perfect. When using net shape manufacturing processes, this is rarely the case. Material integrity can change drastically as the wall thickness increases. This is particularly true with casting processes. Of the casting processes, diecasting is among the most sensitive to increases in material thickness. An atomized spray of metal is injected into the die, filling the cavity from the outside inward. The metal, which contacts the die, freezes quickly and creates a thin, high-integrity skin. The remaining molten metal freezes at a slower rate. As heat is removed at the die surface, the casting solidifies inward, leaving trace amounts of porosity in the center. This porosity is often

referred to as centerline porosity. As the wall thickness of a die-cast component increases, the amount of centerline porosity increases. Although actions can be taken in tool design and process settings to minimize the effects of centerline porosity, increases in wall thickness often create casting defects in the areas that need to be of the highest integrity. Slow-cooling casting processes, which use risers and other methods to minimize porosity, often exhibit extreme changes in material properties as the thickness is increased. This is primarily due to variations in the solidified microstructure, which differ as the distance from the mold surface increases. Near the surface is a thin chill zone in which the casting is of the highest integrity with a fine, equiaxed structure. Immediately following this zone is a columnar region that forms as heat flows from the liquid through the chill zone and into the mold. Note that columnar structures associated with different surfaces in the mold can run into each other. In the center of the casting is a large, grained, equiaxed region, which is the last portion to solidify. The material properties in each of these regions are very different. Adding quality material is the key to improving strength. Adding more of the high-integrity skin (chill zone) is the best way to add strength. To add more of the high-integrity skin, the surface area of the component must be increased. This can be done efficiently by using ribs of a uniform wall thickness. Ribs serve to improve component stiffness while reducing weight by removing material from the casting. Correct rib depth and spacing during component design depends on the overall component engineering. Ribs are recommended for reinforcing or stiffening most webbed sections and are required for thin webs. When designing castings with webs, ribs will provide the following benefits:  Reduce web thickness and save weight  Reduce web load deflection by strengthening

and stiffening

 Prevent warping of unsupported web areas  Prevent hot tearing or cracking of the casting

during solidification by acting as a chill An easy method for computing proper web section thickness does not exist. However, for both the casting design engineer and metalcaster, webs must be neither too thick nor too thin for the specific design. Web thickness depends on various factors, including web area, alloy used, structural requirements, and the number and geometry of joining webs and stiffeners. For good aluminum casting design, rib junctions should be staggered to avoid a large metal mass that will be present at the junction of multiple sections. This mass can cause uneven cooling of the casting (preventing directional solidification) and defects. Staggering the ribs reduces the strength of the section, but both the loss in strength and the amount of metal

can be held to the minimum at the intersections by staggering the rib junctions a distance less than or equal to the thickness of the rib. Another factor to consider in designing junctions in castings is the use of fillets at all sharp corners. By rounding junctions to the proper radii, fillets prevent cracks, tears, and shrinkage at re-entry angles. They also make corners more moldable for manufacturing and reduce stress concentrations in the casting when in service. Fillets must be large enough to meet engineering stress requirements and reduce stress concentrations but not so large that they cause shrinkage during casting solidification. When possible, ribs should be designed appreciably deeper than their widths. The recommended width for ribs in aluminum castings is 1.5 times the thickness of intersecting walls. The minimum rib width is equal to the wall thickness. Ribs in compression while in service offer a greater factor of safety than ribs in tension. Because thin ribs are subject to failure in service, they generally are regarded as poor design practice. Figure 3 provides four degrees of rib design for aluminum cast components. The rib atop the section in Fig. 3 is termed a bead and is rated highly as a design for load-carrying capability. However, it can be difficult to cast. To aid in manufacturing a component with a beaded rib, design the feature so it is located on the parting plane of the mold. Cross ribbing or ribbing on both sides of a casting (Fig. 4) is undesirable from a casting manufacturing perspective. Cross ribbing increases the rate of casting defects and the total cost of the component. This design also creates hot spots in the casting during cooling and solidification, creating feeding difficulties and potential shrinkage-related defects. If design requirements dictate the placement of cross ribbing on a casting, the designer should adhere to the recommended practice of staggering the location of the ribs, as shown in Fig. 4 (right). This will alleviate the problems encountered in casting cross ribs. Cored rectangular holes in ribs and webs (Fig. 5) promote difficulties during casting manufacturing. In ribs and/or webs, round or oval cored holes are preferred. However, if other shaped holes are required, all corners should be rounded to enhance casting solidification. In highly stressed webs, ribs, and walls, beading cored holes (Fig. 6) is critical for increasing the component service life. If a component has been designed with a heavy section that has been cored, ribs may be added for strength and to achieve a uniform wall thickness. During design stages, however, remember that if a component is designed with deep ribs, these ribs require adequate draft to ensure removal or ejection of the mold from the pattern during the casting manufacturing process. The goal in all designs should be to create a component without ribs, if possible. If the casting wall alone has adequate strength and

108 / Casting Design and Performance t

R= t 4

t

(a) Poor

(b) Acceptable R=t

t

T 4

T

3t

R= t 4

Cross ribbing (left) should be avoided in cast component design due to the hot spots and potential shrinkage defects it creates in casting solidification. Staggered ribs (right) provide for better manufacturing.

Fig. 4

t

R=t T+t 2

t R = T+t t 2

R=t

t

3t - 3T

(c) Good

(d) Excellent

Fig. 3

Four degrees of rib design for aluminum cast components. The beaded rib in is rated highly for load-carrying ability. This type of rib feature is cast most easily when located on the parting plane of the mold.

Fig. 5

When designing a cored hole between ribs or in a web design, design the hole to be round or oval with rounded corners (right). Rectangular holes (left) create manufacturing difficulties.

T

T

1.5T 1.5T

T

T

–2T

T

Fig. 6

Although the design on the left is an acceptable design for a cored hole in a highly stressed rib design, the design on the right is preferred to improve the service life of the component.

stiffness, the elimination of ribs results in lower stresses and improved stress distribution in the component.

Designing Castings with Uniform Wall Thickness In many castings, functional requirements dictate that walls be uniform or nearly uniform in thickness. However, if these castings are not properly designed, the molten metal will freeze prematurely during casting before all parts of the mold cavity are filled. Even if the mold cavity is properly filled, portions of the casting may not be fed before the onset of metal solidification, causing centerline shrinkage or porosity.

To alleviate the problem of incomplete mold fill with uniform section thickness castings, walls may be tapered with extra padding during the design, and/or ribs and webs may be added to the casting to provide feed paths for the molten metal. When pouring a flat plate with uniform wall thickness, the metal enters the mold cavity and attempts to flow in all directions. If the distance from the gate to the extremities of the mold cavity is too great, the metal freezes prematurely, resulting in misruns. When ribs are added to the plate, the metal flows readily to the extremities, and successful cavities are produced. The same success with the plate is attained when the casting gate is enlarged and uniformly tapered to the extremities. The metal flows into

the mold cavity at the heaviest section, thus preventing an initial chilling of the metal by the mold, as occurs with a small volume of metal. Since the molten metal in heavy sections retains more of its heat as it flows through the mold, it more readily can reach the extremities of the mold cavity. The goal of casting solidification design is that the volume of metal lost because of shrinkage can be replaced by the riser. Although measures are taken with uniform-thickness castings to ensure proper freezing, there is no assurance that freezing will start at points furthest from the gate and progress in an orderly manner to the gate. The result could have metal nearer to the gate freezing first, isolating pockets of molten metal, which results in voids or porosity at these pockets. One solution is the use of two risers to feed metal from opposite ends of the casting. However, risers add to the cost of producing a casting. Another option is designing feed paths to allow molten metal to flow easily to all sections of the mold cavity. If the enlarged sections are gradually tapered from the gate to the extremities, solidification follows an orderly progression through the casting, terminating at the riser. Compensation for normal shrinkage is provided throughout the casting, and sound metal is assured. The same beneficial freezing pattern is obtained by tapering the plate, provided the angle of taper is sufficient and the source of the feed metal is at the heaviest section. Under these conditions, a favorable freezing pattern is assured because the metal that has flowed the farthest (and is therefore the coldest) is deposited in the thin section. In contrast to a heavy section, the thin section of molten metal has less heat to transmit to the mold before freezing can start. Thus, the direction of solidification can be predicted in tapered sections, provided the metal enters at the heavy end. If the need for feed paths is anticipated in the early stages of casting design, padding can be made a functional part of the casting, or it can be located so that its removal adds little to no cost.

Designing Castings with Unequal Sections Casting processes provide design freedom by allowing engineers to place metal where needed to meet desired criteria. Typically, functional

Casting Design and Geometry / 109 demands of the part as well as induced stress distribution result in a nonuniform mass distribution in a designed structure. This means that a casting shape can assume nonuniform distribution of interconnected thick and thin sections. For example, if a heavy section is surrounded with thin sections, strategically located risers may be needed to feed isolated heavy sections. If each heavy section requires a separate riser, the result is lower yield, increased labor for casting cleaning, and increased final part cost. However, there are certain instances that can lend themselves to redesign to make the shape more readily castable without additional risers. In those circumstances, a designer’s knowledge of the metalcasting process or close collaboration between the designer and casting engineer can play a significant role in modifying the shape to suit the process. For example, one design option may be a feeding pad, which is essentially additional metal added to the casting that allows directional solidification to progress toward the side riser. The addition of feeding pads can result in a solid casting without a top riser, thus reducing yield. This type of modification can be made in many instances without affecting the part functionality and only minimally increasing its weight. In application, when the part is rotating around a central axis, the additional pad can be added on the opposite side to maintain dynamic balance. Another geometric modification that can be considered to improve component castability with isolated masses is to reallocate the existing geometric features. For example, ribs are used to provide additional structural support to isolated bosses. The size of the ribs is an important factor in casting, since it determines how far the molten metal can flow and how it will solidify. If the ribs are too thin, this will promote rapid cooling and premature solidification and will result in either a cold shut during pouring or inadequate feeding of the boss, causing shrinkage defects. Alternately, large ribs will add to the weight of the casting. Hence, ribs should be dimensioned properly to satisfy structural integrity while allowing castability.

Designing Thin Sections The lightweight structure of thin-wall castings allows for increased payload and reduced energy consumption in transportation applications. However, thin-wall castings also can pose manufacturability problems associated with mold filling. Rapid cooling of thin-wall sections of the casting reduces the fluidity of the molten metal, which could cause the molten metal to prematurely freeze before it can completely fill the mold cavity. The mold-filling capability in thin-wall casting depends on a number of different process- and design-related variables, including the type of metal, pouring temperature, casting shape, and metalcasting process.

The practical minimum wall thickness for a particular casting configuration can be determined based on experiments using inexpensive and accurately made loose patterns. Often, a metalcasting engineer will recommend a value to start at, based on prior experience. The wall thickness then should be adjusted, depending on if the defect rate is above or below the acceptable range. Typically, mold-filling problems are rarely an issue when the wall thickness exceeds 9.5 mm (0.375 in.). Alternately, when the wall thickness is reduced below 3.2 mm (0.125 in.), adequate mold filling is virtually impossible except for short distances. The problem is premature freezing of the molten metal. In such instances, a designer can allow for certain geometric features on the casting that can ensure fluidity of the molten metal for the entire length of the mold cavity. The chapter “Design Problems Involving Thin Sections” in this book describes various examples. In some cases, the selection of a different alloy can improve the castability of thinwalled castings. For example, aluminum can achieve thinner walls more successfully than steel. However, a drastic change in metal is not an option in many cases. If a casting is not structural or load bearing, a designer can take this fact into consideration to achieve thinner walls. Even in a similar group of materials, small deviations in alloy composition can have a significant impact on material fluidity. For example, 300-series stainless steels have better fluidity than 300-series stainless steels with the addition of molybdenum. Hence, whenever possible, the designer should allow a change in the material if that could lead to the desired casting configuration at a reduced cost.

Designing for Economical Coring One of the major advantages of the metalcasting process is its ability to produce components with complex internal cavities (such as undercuts and skewed internal walls). The shapes used to form internal cavities within the casting are called cores, which can be made from different materials (sand, ceramic, and metal) depending on the application. The same principles that govern moldmaking also are valid for making cores. The cores are made in coreboxes and must be removed from the corebox after curing, so they cannot have undercuts that would prevent them from being removed from the corebox. In addition, deep pockets in the corebox must have a sufficient draft angle to prevent high friction between the core and corebox during removal, which could cause a core to break. For deep pockets, additional loose pieces are used in the corebox to assist the core removal. Although cores increase the cost of castings, they provide a number of advantages. The most important is the shaping of internal passages,

which can be of almost any arbitrary shape and with nonuniform cross sections. Another advantage is shaping the geometry of the part to optimize wall thickness. Cores can provide casting geometry that can make stress distribution uniform throughout the part, thereby reducing its weight. This capability has not been extensively exploited in the past due to the complexity of stress analysis in highly complex shapes, but with the advancements in finite-element analysis tools, this design advantage of metalcastings becomes a significant advantage. Cores also can form zero draft angles on castings and can be used to allow for undercuts on the vertical walls of the casting to reduce or eliminate secondary machining. The process used for making the core also has significant influence on the core strength and its ability to hold under its own weight. For example, shell cores typically are stronger than oil-based cores and can allow more freedom with respect to core location within the mold. In addition, some geometries and coremaking processes allow the cores to be hollow, reducing the core weight and stresses. Cores can be placed within the mold cavity in a number of different ways, each offering certain advantages and also posing certain limitations. For example, a core for a bell-shaped casting could be in the drag section of the mold, with a portion protruding into the cope section. Alternatively, the core placed in the cope section may hang into the drag section. The first case allows for almost any core size, while the second limits the core size because of the stress in a core induced by its own weight. Sand cores are fairly weak and can break if not supported properly. Thus, it is important to support cores within the mold cavity in such a way as to prevent their fracture and/or excessive deflection. Due to limited flask size and gating system issues when the casting is very long, supporting the core may require horizontal positioning of the mold cavity and core within the mold. In this case, orientation of the core presents additional problems. The stresses caused by the weight of the core can be higher than the core tensile strength, thus causing the core to break. Even if the core does not break, the body force on the core will cause it to deflect downward before the metal is poured, while buoyancy forces will cause it to bend upward after the metal has been poured. To prevent or minimize these problems, designers redesign the casting to allow for additional support. The resulting stresses and deflection can be reduced by an order of magnitude with this simple modification. If necessary, holes on the casting can be plugged after casting and cleaning. Another method to support the core in the mold cavity to lower stress and minimize deflection is to use chaplets. Chaplets are metal shapes made of the same material or compatible material to the one used for casting. Since they are placed within the mold cavity, chaplets become an integral part of the casting. Hence, the casting engineer must ensure tight bonding is achieved between the casting and chaplet.

110 / Casting Design and Performance The stresses and deflection of cores caused by the body and buoyancy forces are of great importance in manufacturing good castings. These two phenomena depend on the core shape, core material, nature of core support within the mold, and metal being poured. To determine the stress distribution and deflection of the core before and after the metal has been poured, one must apply the basic principles of statics, strength of materials, and fluid dynamics to the particular core configuration. The most important factor for maximum core length is the type of core support within the mold cavity. Vertical placement of the core within the cavity allows for longer cores than horizontal orientation (due to the elimination of a buoyancy force acting on the core). However, this orientation of the core within the mold cavity can cause damage to the mold during mold assembly. In the case of horizontally positioned cores, the core supported on both ends can be longer than the one supported only on one end. In the case of the cantilevered core, the minimum length of the core print should be 33%. In addition, the core print section should be of sufficient mass to allow the core to remain in place for mold closure. The maximum allowable core length also increases as the core diameter increases (as strength of the core increases due to the increased cross section). The casting wall thickness also influences the core length (casting wall thickness decreases as the maximum recommended core length increases). This is caused by the decreased amount of metal in the mold cavity and thus reduced buoyancy force. The type of metal also has a certain level of influence on the recommended maximum length of the core. Steel has a higher density than most metals and results in a higher buoyancy force than aluminum and magnesium. As a result, the maximum recommended core length is smaller for steel than for aluminum and magnesium.

Core Design Principles Cored holes should be designed as simply as possible while satisfying functional requirements. Several basic principles exist that should be observed when designing cored holes:  Assure the strength of the core during core-

making, handling, and pouring.

 Provide sufficient mass to the core sections

to allow proper cooling during solidification.

 Avoid thin casting sections formed by cores.  Avoid or minimize the need for complex

core handling. This will result in a scrapped core or a defective casting if the core breaks during pouring, due to the high pressure exerted by the metal stream. In some cases, thin core sections can cause significant rework on the casting and/or a scrap rate of more than 50%. When designing narrow holes in a casting, designers must allow for the possibility that the particular feature should be machined instead of cast to allow for additional elements to strengthen the core. Thin core sections also pose a problem for gating and feeding the casting. The ingate should be placed in such a way that the incoming stream does not impinge directly onto the thin core section; otherwise, the core could fracture. Having too many thin core features will limit the possible location for ingates and thus proper gating of the casting. In addition, too many thin core sections protruding to the surface of the casting will affect the location of a riser (if required to feed the casting during solidification). Thin core sections surrounded by heavy metal sections should be avoided, due to the excessive amount of heat to which they are exposed without the ability to dissipate it properly. In this design, hot tears (cracks) can be expected at these locations. These cracks represent significant stress raisers in the casting, reducing its fatigue strength. Gas-related defects also are common in these configurations, due to the inability to properly vent thin cores. The gas formed by a sand binder will not be able to dissipate through the mold material and could end up in the mold cavity and the metal itself. As a result, gas porosity could be expected adjacent to thin and wide core sections. Another factor is metal penetration, which can be a common problem if there is premature decomposition of the core binder. Last, cleaning of narrow cored sections can be a very difficult, labor-intensive job. Designers must allow for an increase in the cored sections according to the specification provided by the foundry engineer and according to the casting process and type of alloy. There is a minimum casting wall thickness that can be physically or economically achieved. As noted, the minimum casting wall thickness is governed by a number of different variables, including casting process, type of alloy, and pouring temperature. Potential problems in the production of thin-wall castings are misruns and cold shuts. The cores that form thin-wall sections on the casting should be avoided or redesigned to provide uniform fill of the entire mold cavity.

and/or stacked cores.

 Minimize the number of required cores.

Cores are typically made from sand and can be fragile, depending on the cross-sectional area of the particular feature on the core and the core manufacturing method. Long, thin sections on the core can easily be broken during

Design Conversions to Casting Welded assemblies, fabrications, or forgings may be done more effectively by casting. How does one spot a conversion opportunity for casting? This chapter describes how one can

identify components worth considering as castings for possible savings or improved performance. On the floor of manufacturing and assembly operations, countless components are made up of several stamped, wrought, or machined metal parts. Engineers everywhere stipulate fabrications and forgings without weighing all of their options. Could those parts be redesigned to a single cast metal component for improved performance? Is there a potential performance gain or cost-savings that would make the redesign viable? The choice of whether a component is best manufactured as welded, assembled, fabricated, forged, machined, or cast is based on the component geometry, production costs, and requirements in application. This section looks at these issues and provides a framework for analyzing all manners of manufacturing as possible conversion candidates.

Weldments and Assemblies Time to market can be a major focus, and components often are designed to be manufactured by the process(es) that will deliver them with the shortest lead time. This leads to original equipment manufacturers (OEMs) designing and manufacturing weldments and assemblies of weldments in-house, using their available capacity and labor and eliminating the demands of sourcing a component (such as design changes and pattern costs). Design engineers generally feel more comfortable designing building blocks of simple wrought shapes. This is the method commonly taught in structural design, and manufacturing engineering and purchasing colleagues are comfortable with bringing weldments and assemblies of weldments into production. Another quick-to-market option is to hog-out (machine) a shape from stock to produce a required component, simplifying the sourcing process to a one-step machining operation. Although both approaches deliver the product to the customer in a time-efficient manner, they may not deliver it in the most cost-effective manner, particularly when volume increases. Weldments and assemblies of weldments create a significant volume of part numbers to produce (or buy), schedule, and track. They require high levels of labor to fixture (weld, bolt, align, etc.) and typically are not as dimensionally or structurally consistent due to the inherent variances in manufacturing. On the other hand, hog-outs have the advantage of known wrought stock mechanical properties (making them a popular choice for critical aerospace applications) but suffer from high costs, because a large percentage of the bar stock they are borne from ends up on the machining room floor. Because mechanical properties of casting alloys are not as well defined nor as widely available in properties literature (casting mechanical properties depend

Casting Design and Geometry / 111 on the process and involve liquid flow and solidification gradients), some structural designers feel more comfortable with wrought metal properties. As a result, however, the opportunity to learn more about the casting structural properties is often missed, along with the prospects of lighter, stronger, and more cost-effective parts. In prototype situations when time is of the essence, weldments, assemblies, and/or machining are sometimes selected over casting to bring the product to market quickly. The key is to convert these fabricated components to the most efficient manufacturing method when volume production is ready to begin. The first thing to look for in a possible conversion to casting is a component with a complex geometry. This could be a single component that was machined or forged, but more often than not, the most impressive cost/ weight-savings is with a series of stampings and/or other wrought shapes that are welded and/or bolted together. Components often are designed with the building-block mindset during the prototype stages due to design familiarity. However, the building blocks usually can be streamlined to one-piece cast components. The type of complex geometry components that make good conversion candidates to casting often have a high surface-area-to-volume ratio. In addition, they have a high number of inches of weld, which leads to high fabrication costs due to the weld time, material, and complicated fixtures. The complex geometry components often also have high aspect ratios, which mean that they are long with respect to their width (rangy components). These factors typically indicate where the most cost-savings potential is in a casting conversion. When a fabricated component has a high number of inches of weld in critical stress areas, a greater chance exists that problems in the weld itself or from microstructural effects in the neighboring heat-affected zones will cause failure in the field. High-stress welded joints are generally less capable in overload instances or cyclic fatigue than junctions formed in the metalcasting process. Another factor to consider with weldments or assemblies of weldments is the number of separate parts necessary to form one component. Beyond the inconsistency in mechanical and dimensional properties developed throughout the multipiece fabrication, the OEM must inventory all the part numbers and absorb the labor costs to manufacture the component. In addition, the in-house manufacturing facility is responsible for ensuring dimensional consistency from part to part as well as maintaining a low scrap rate during production. Many of these costs are found in burden rates or are activity-based and go unrecognized in a cursory value analysis of component cost. The last factor to consider is final component performance. The well-known mechanical properties of wrought metals are directional

(stronger in the wrought direction, weaker in the transverse direction) and may be compromised significantly by welding. A properly produced, high-integrity casting with isotropic mechanical properties (equal in all directions) enjoys uniformity of properties in continuous sections as well as junctions. In addition, visual appearance must be evaluated. While fabrications offer smooth surfaces, their welded junctions are not as pleasing to the eye as the continuity of a complex cast shape. If a component or series of components have been identified for possible conversion to casting, the next step is to bring in a metalcaster to determine the of the component castability and possible cost/weight reductions. Typically, redesigns to casting aim for a 40% cost reduction from the weldment, assembly, or hog-out to overcome the labor and lead-time benefits associated with the other manufacturing methods. The first step in identifying a casting partner is to determine the material in which the redesigned component could be made. Weldments and assemblies of weldments often are fabricated from wrought steel shapes because carbon and low-alloy steel is the most easily weldable of metals and provides high toughness, strength, ductility, yield stress, and stiffness. The problem is that steel often is chosen solely on its weldability. If the component is going to be a casting instead of a weldment, then weldability may not be as important. Therefore, other materials should be considered. Potential questions to ask to determine the correct material include: What material is required for the application from a physical and mechanical property standpoint? Will the component be used in a severe service application with high fatigue and impact resistance, or is it a cosmetic part? Does it need to adsorb vibration? Does it need to control sparks? Does it need to resist corrosion? Does it need to be lightweight? Steel is the material of choice when high impact resistance and fatigue life are required for severe service applications. In addition, if the cast component is going to be welded to another component, then steel is a logical choice. Generally, however, the opportunity exists to replace a steel weldment with a casting from another alloy family, such as iron or aluminum. If the component is going to be used in a fatigue application but is not a severe service application, then iron may be the logical choice because of its castability (ease of casting). Ductile iron, in particular, has made a name for itself in the conversion of steel fabrications to castings, because it combines excellent castability with reasonable toughness. When weight is an issue without severe service application, aluminum often is the casting material choice. With new technology in alloying and molding, aluminum is increasingly becoming a material of choice in important fatigue applications, such as replacing stamped steel fabrications with squeeze-cast aluminum in automotive suspension components.

Regardless of choice of alloy family, the most significant factor in the success of castings in structural applications is the use of geometry to control stress and stiffness. Only the casting process offers so much variety in shape at low cost. Even though some alloys are stiffer than others (for example, steel is three times stiffer than aluminum), it is stiffness from shape that makes castings a breakthrough metal product form. Once a material is selected, design engineers must then go to their purchasing colleagues to begin the search for a casting partner to assist in the redesign of the component to casting. Ask purchasing colleagues to find casting sources who appreciate the importance of structural geometry in casting design. Such a casting producer will offer the most overall capability in cost reduction and component improvement when converting stamped steel fabrications, weldments, and weldment assemblies to metal castings. When examining a weldment or assembly for a possible conversion to casting, the following are seven considerations to determine the feasibility of the redesign:  Geometry: 











Is the component complex enough to warrant a redesign to casting? Length of welding required: The more length of weld needed to make up a component from several smaller components, the more likely it is a candidate for conversion to casting. Number of parts: Is the component made up of several smaller parts that have been attached together? If it is, would a single casting reduce the cost associated with part number inventory as well as free up in-house operations for other, more important manufacturing tasks? Dimensional consistency: How tight are the tolerances for the component? Is warping ever an issue? The inherent dimensional inconsistency with fabrications is eliminated with a single cast component. Scrap rates: Are scrap rates high for the component during in-house fabrication? If so, the component may be better manufactured out-of-house via a different production method. Appearance: Is the component visible? Does it require a clean, streamlined appearance with a fine surface finish? Field problems: Does the component fail in use? Are the stresses exerted on it too great? Castings provide consistent properties throughout each component, eliminating much of the variability found in assemblies and weldments.

Fabrications Geometry can help identify successful casting design from former fabrications. Even for castings that do not serve a meaningful

112 / Casting Design and Performance structural purpose, geometry is critical. Geometry for casting design takes into account alloy castability, deflection, stress concentration, risering/gating/venting, and efficient machining and assembly. As a result, well-selected casting geometries are capable of surpassing other methods, particularly fabrications, in making metal shapes that are limited by their process. Flexibility in geometry choice is more restricted among fabricating, as well as stamping, forging, or machining from stock. Since stress is a function of loads applied and geometry of structure (geometry alone controls the amount of stress for a given system of forces on a structure; material choice controls how much stress is allowable), castings make efficient, cost-effective structures and allow the most freedom of geometry. In redesigning fabrications to castings, opportunities exist to use casting geometry to simplify design and reduce weight and cost by designing for more uniform stress throughout the part. Since fabrications are made from building blocks of uniform cross section, they tend to be overdesigned in sections away from the critical stress areas. They also tend to suffer from heat-affected properties and stress-concentration factors in the weld joint areas. By using engineering sketching with basic stress analysis techniques, design engineers can formulate an estimated casting design geometry ready for validation via finite-element stress analysis and casting process analysis using a solid model. Since casting design geometry selection requires brainstorming, sketching is the best way to find a shape that is structurally efficient, castable, and manufacturable downstream. A dimensioned, well-proportioned sketch also is an effective platform for building a solid model. Alternative design methods based on engineering sketching are viable for the following reasons. First, dimension-engineered sketches are probably the most efficient starting point for construction of solid models. Second, sketching allows the integration of the four disciplines of good casting geometry:    

Geometry of castability Geometry of structure Process geometry Geometry to minimize/reduce machining and assembly

cost

of

Third, a metalcasting geometry can be quickly sketched using the designer’s sense of scale, proportion, and load-carrying capability of shape and material. Last, reality checks of sections where stress and/or deflection are a concern can be quickly calculated using classic stress analysis; methods of estimating þ20%  make this step practical. Computer solid modeling can validate or refine the estimates. The method of sections, vector algebra, the principle of superposition, and methods of transforming stress and checking allowable transformed stress are all highly graphical

techniques. Nothing in stress analysis is more valuable than a simple sketch of the loads and reactions expressed in a free-body diagram. This methodology as a precursor, coupled with today’s model building and simulation technology, is the most effective pathway to efficient structural metalcasting designs. Remember, this method uses engineering sense of scale and load-carrying capability combined with computer modeling to verify intuition. The goal is to obtain reasonable estimates of geometry that will be relatively uniformly stressed and will respond well to the alloy to be cast. The goal is not to compete with the computer; it is to help the computer finish the job quicker. When performing a component redesign from fabrication to casting, do not start from the fabricated form. Start with a clean sheet of paper and think only about what the functional areas and surfaces of the component design must be. The free-body diagrams at important slices in the overall structure come together to form efficient casting design. A principle in stress analysis called superposition allows loads and reactions to be analyzed separately and then added up. Some of them are positive and negative, one partially canceling another out. It is very much like algebra—factoring the equation in clever ways makes the analysis easier. In a broad sense, the stress analysis of metal castings can be described as “the unique capability to easily and cost-effectively change area moment of inertia over section length.” This capability—plus the ready ability to soften surface shapes where stresses concentrate—underlies the power of castings to control stress and deflection in structural applications. Some important elements and steps of the principles of engineering mechanics for linearly elastic solids at rest are summarized as follows. Stick-Figure Sketches. Draw a picture of the loads on the structure and the reactions to the loads using stick figures. Recognize that these loads and reactions are vectors with magnitude and direction. They also work like algebra because  þ signs are important. Using Method of Sections. This is the basis of free-body diagrams. Any loaded static structure can be sliced at any location in any plane, and reactions can be defined at the slice to hold the remaining part of the structure in equilibrium. The reactions at the slice are the basis of classical stress analysis. Stress Correlates to Geometry. Stress is 100% dependent on geometry. Choice of material determines how much stress can be tolerated, but the amount of stress always is strictly dependent on shape. This simple fact is why metal castings are significant as structural components. The choice of low-stress geometry is helpful in crack-sensitive structures because geometry reduces the crack propagation driving force. A misconception about using geometry to lower and improve the orientation of stress is that “lowering stress in one region of a structure

simply raises it in another region.” The fact is that increasing area moment of inertia lowers the stress at that section without necessarily affecting any other section negatively. Stress Correlation to Area. Areas are the key parameter in geometry to control shape for stress and deflection. The essence of area moment of inertia can be gleaned quickly from its base formula. Considering the axial formula, as cross-sectional area becomes taller in the Y-direction, the area moment of inertia (Ixx) becomes larger quickly because the distance (Y) is squared. At the same time Y2 becomes significantly larger (if the cross section starts becoming wider), then Ixx increases even faster (dA is the change in cross-sectional area with each small increase in Y; when Y starts to become a large distance, a big change in area width becomes important). This formula explains why I-beams, with their skinny center web but wide flanges separated substantially from the center, are good at controlling stress. Considering the polar form for the area moment of inertia (J), as the radius of a cross section increases, J becomes larger quickly because the radius is squared. For resisting torque and reducing torsional stresses, intuitively the area in the center of the shape is relatively meaningless. This explains why cylindrical tubes such as drive shafts and torque tubes are effective in controlling torsional stress. Deflection. Controlling deflection involves both stiffness from area moment of inertia and choice of material. Stiffness of metalcasting alloys (or any material) is defined by Young’s modulus, also called modulus of elasticity. Formulae for deflections, even for relatively simple beams, are complex-looking equations. However, the relationships for limiting or allowing deflection are as simple as the formulas for bending stress or torsional stress. An interesting example of these relationships in deflection is the design of a cast aluminum front suspension component for an automobile. The modulus of elasticity for aluminum is relatively low, 10  106 lbf/in.2 (68,900 MPa). Therefore, the objective should be to keep deflection low. To do so, the formula calls for the casting geometry to be “stubby,” that is, short in section length relative to stiffness in area moment of inertia. On the other hand, steel castings often are used as stress nodes in complex welded steel fabrications. They are designed to absorb stress and allow deflection in a way that protects the welds, particularly from fatigue stress. In this case, the casting design can take advantage of the high modulus of elasticity of steel (30  106 lbf/in.,2 or 206,800 MPa), and the shape can be rangy, with longer sections relative to the section area moment of inertia, allowing flex. Simplifying Stresses with Mohr’s Circle. Using Mohr’s circle, complicated stress combinations can be transformed into the maximum

Casting Design and Geometry / 113 shear stress and principal stresses (maximum tensile or compressive stress) to ensure that stresses are kept below safe limits for the material involved. If a stress is found to be too high for static overload, fatigue life, or propagation of known or potential cracks, simply change the cross-sectional area of the casting (even a dramatic change in cross section) so the maximum shear stress and/or principal stresses are low enough to be safe. Mohr’s circle and Tresca’s hexagon (which checks maximum shear and principal stresses against allowable stress) are graphical methods. They fit the overall technique of using sketches to get in the “good casting design ballpark” and let the computer finish the job. Steel castings and fabrications can seem interchangeable because they share many qualities; cast steels and wrought steels have such similar mechanical properties that the American Society of Mechanical Engineers Code does not differentiate between steels on the basis of their manufacturing process but by their chemical composition. However, noticeable differences do exist between the two that can affect the design and cost-effectiveness of a component. Wrought products (rolled or forged) exhibit a characteristic known as directionality. This characteristic, also known as anisotropy, means that a component has strength and ductility in the working direction but has lower transverse properties. Cast steel products do not exhibit directionality; rather, they can be described as isotropic. Steel castings can be stressed in any direction without concerns over the lower strength, ductility, and toughness that are exhibited in the transverse direction of wrought products. Designers of fabrications must be aware of the directional properties and incorporate them into the component design, or it could become overstressed when a load is applied in the transverse direction. When compared to their wrought equivalents, the mechanical properties of cast steels are approximately the average value of the longitudinal and transverse directions in the wrought product. Transverse properties always are lower than the properties obtained in cast steels, which means that casting offers more design flexibility than fabrications. Properties for wrought products are determined by performing mechanical tests in the longitudinal direction. Tests in the transverse direction usually are performed only when specially requested. Steel castings are tough and ductile, contrary to the common belief that they are brittle and subject to abrupt failure. Brittleness is a function of metallurgy, not of process, and steels are not brittle alloys. Because steel castings are isotropic, uniformly heat treated, and more stress relieved than a fabrication, longer fatigue lives and more deformation without sudden failure are common in castings. One factor that must be considered in any analysis of the design of fabrications using

wrought products such as bar plate or tube is that the welds usually are placed in the highest stressed location in the component. Unless the fabricator follows a design code, most welds are placed at high-stressed section changes or design features such as corners, which limit the load-bearing strength of the component. To combat this, welded fabrications often can be redesigned to one-piece castings. Then, the casting, which welds as well as or better than the fabrication, can allow the designer to locate the welds away from the highly stressed areas. Often, the weldability of the materials used in fabrications is taken as a given. They must be weldable, otherwise they would not be used. In addition, there also is the misconception that wrought steels are easier to weld than cast steels. Cast steels have a lower susceptibility to underbead cracking—the cracking that occurs from the introduction of hydrogen into the liquid metal. The isotropic qualities of the casting provide a good (not too hard) welding surface. Welds made in the production of castings are almost always stress relieved. Castings tend to undergo heat treatment (a strengthening process) after welding occurs, keeping the component strong even in the weld areas. Fabrications usually are not stress relieved after welding, which can lead to a weaker area around the weld. Field welds can be difficult to stress relieve due to the location or size of the part, but if the component is a casting, it can be designed to place the unstress-relieved weld in a lower stress location, which improves the component life. It has been shown that when castings are used at junctions such as a node on an oil production platform, the weight of the connecting area can be reduced by as much as 50%, stemming from the elimination of joints and welds to attain the required strength. In addition, the position of the welds also is moved out of the high-stress area, and the welds become simpler (circumferential instead of complex and irregular in form). Placing a weld in a position that makes it easier and simpler to perform also can minimize the nondestructive testing required and the number of discontinuities or stress raisers associated with welding in difficult positions and inaccessible areas. Designers sometimes say that it is easier to design fabrications from plate and bar because it is easier to visualize a component made as a series of right-angled connections. Visualizing a component where there is total freedom of form can be more difficult—it requires the designer to think in three dimensions. Yet, this freedom permits designers to design only what the component needs—no extra material, edges, or welds. Conventional fabrication generally is a compromise between material availability, fabrication capability, design codes and engineering requirements. However, creative design of

castings often will enable steel sections to be tailored to meet specific engineering loads, thus improving engineering efficiency. This often leads to substantial engineering benefits and weight reduction elsewhere in the surrounding steelwork through the elimination of offset work points and their associated bending moments. Alignment or dimensional problems due to production are likely to be greater with fabrications than castings, because of the distortion that may occur during manufacture. Straightening operations carried out on components can have a more detrimental effect on fabrications because they will be plastically deformed in the high-stress areas. These high-stress areas often are associated with the welded joint. Although castings also will be plastically deformed during straightening, the location of the deformation is unlikely to be in a notched area such as a weld bead. Because steel castings typically are welded into a larger fabricated steel structure, it is important to consider deflection. Allowing steel castings to flex protects the weld joints in the base structure from fatigue failure. The combinations of section length, depth, and cross section in geometry permissible with the high stiffness and high yield stress capabilities of steel can resist fatigue and protect mating sections from early fatigue failure. Allowing a steel casting to flex can reduce the stress concentration in the casting weld connections to a base structure. Identifying an appropriate cast shape opens the door for designers to the opportunity for novel designs and shapes that challenge traditional fabricated concepts. Casting design should simplify the component shape as far as possible, satisfying the basic engineering requirement while reducing the overall size of the component where possible. This can be achieved by eliminating or adjusting offset work points, local stiffening, and deepened sections in plate, box girders, and tubes—all of which are necessary in fabrications to achieve acceptable designs. When designing, it is important to establish geometry that will be workable during secondary operations. This includes casting designs for welding into a larger fabricated assembly, considering the design of weld-joint geometries (compatible mating of casting and plate thicknesses and stress distribution of weld geometries) and thinking about assembly features. For example, with fabrications the capability to position and hold several pieces of plate in a weld fixture is more difficult than forming a mold cavity for casting. Castings allow designers to buy shape cheaply. If the desired component is mostly steel and the costs involved are mostly material, a fabrication should be used. However, the more components, the more linear inches of welding required and the more machining required per individual component, the more attractive casting becomes. Designs or

114 / Casting Design and Performance fabrications that comprise the most pieces and the most welds are ideal candidates for a casting conversion. In general, castings provide tighter tolerances and better mechanical performances. They allow the designer to shape the component exclusively for the project at hand, with no extra pieces or sections. Complex components and assemblies that can be consolidated into fewer parts become cost-effective as castings. Limiting assembly reduces cost. Castings often weigh less because the geometry can be tailored to the actual component requirements instead of being restricted by the capabilities of bars and sheets. The cost bases for fabrications and castings are different. An increase in steel thickness, shape complexity, and stiffening in a fabrication pushes up costs because of the amount of welding and nondestructive testing necessary. This also can be affected by the increased risk associated with the stress relief of highly restrained, heavy sections. Conversely for castings, castability is enhanced with increased section size, and with optimal design, the cost/ton will decrease with increased weight.

Dimensional Control Dimensional accuracy is a critical factor to be considered through the casting design, tool making, pouring, and postprocessing of the casting part. To control these sources of error, the mechanism of each error source must be understood.

Metal Shrinkage Shrinkage is a leading error source in casting. Different materials may have different contraction behaviors and thus need different shrinkage compensation factors. Shrinkage or contraction also occurs in the following stages:  Contraction in volume of the liquid metal as

it cools to solidification temperature

 Shrinkage or decrease in volume, as the liq-

uid metal changes state and becomes a solid

 Contraction of the solid metal as it cools fur-

ther, to room temperature

 Metallurgical phase changes in the solid

state that may be accomplished by volume changes, for example, in the case of ferrous castings, the austenite-to-pearlite transformation  Volume and dimensional changes that may occur if the casting is heat treated, to modify its properties To improve the accuracy of shrinkage compensation, it is necessary to determine the real sources of contraction and theoretically understand the shrinkage behaviors. Volume Contraction of Liquid Metal. This is of no practical significance, because castings are made with risers and gating systems that contain a reservoir of molten metal that can

enter the mold cavity to replace volume changes, due to liquid contraction. Solidification Shrinkage. The change (decrease) in volume in passing from the liquid condition to the solid condition cannot be avoided. It varies, according to the type of metal, and in some gray cast iron (hypereutectic) is almost completely offset by an accompanying expansion, resulting from graphite being deposited from solution. In the case of metals such as steel, white irons, or aluminum, if may be as high as 6% by volume, and this calls for an adequate supply of feed metal, if shrinkage cavities or porous areas are to be avoided. Crystallization proceeds from the cooler outer surface of a casting, and crystal dendrites grow in a pool of liquid, which supplies the intrametal to overcome the volume change. The problem occurs in the latter stages of solidification, when insufficient liquid metal is available, and the viscosity of the liquid is so high that it cannot adequately flow into the area between growing crystal dendrites. The result is a shrinkage cavity. The metalcaster may control the solidification rate, so that the last area to solidify is located in the riser or feedhead, which is subsequently cropped from the casting through several techniques: use of chillers, adjusting the speed of gating, and directional solidification to avoid trapping the shrinkage cavity in a critical area. Directional solidification occurs when the molten metal in a casting solidifies in such a manner that liquid feed metal is always available for that portion that is just solidifying. When using a riser or feedhead to provide feed metal, the operator always makes this riser considerably heavier than the section of the casting that has to be fed. Doing this ensures earlier solidification in the casting and later solidification of the riser, which can then provide the liquid feed metal. The engineer must realize that if the casting is too complex and that if the foundry operator has to resort to too many extra steps to provide a sound casting, then the casting may cost a little more than it would if it were less complex and if it were designed allowing for these metallurgical phenomena. Sound, accurate castings of high integrity begin with a proper design, which can be properly cast in the first place. Solid Metal Contraction. This contraction is what the foundry allows for when a shrinkage rule is used for the proper dimensions of pattern equipment. The shrinkage of different metals varies and may be as high as ¼ in./ft to as low as 1/10 in./ft. The appropriate shrinkage rule will be used in dimensioning the pattern; that is, in a 12 in. long casting, for example, to be cast in a metal with a 1/8 in./ft shrinkage, the pattern will actually measure 121/8 in. Unfortunately, shrinkage of any metal is uniform in all directions; thus, a long, thin casting may tend to shrink more in terms of length than it will in terms of width or thickness. Only experience and general foundry practices dictate exactly how and where the shrinkage will occur. The only practical way to take care of

this problem is to provide generous machine allowances or be a little more realistic in the tolerances specified in a casting. Where it is imperative that high accuracy is required in the final casting, patterns must be built and prototypes must be cast, so that the shrinkage can be accurately determined and allowed for in that specific casting. One of the biggest problems with solid contraction lies in the development of residual stresses and strains in the casting. Obviously, it is impossible for all areas of a casting, particularly a complex one, to solidify uniformly at the same time. The position is created where thin ribs, for example, may solidify and contract before adjacent, thicker sections have experienced much of their contraction. The net result is the introduction of differential shrinkage and residual stress into the castings. This is the chief reason why complex castings requiring a high degree of accuracy and precision should be stress relieved. Stress relieving is the procedure of subsequently heating the casting into the plastic range, whereby it can deform and accommodate the stresses that have been built up during initial solidification. Where the design is particularly bad and where contraction is hindered, the ever-present problem of cracking and actual breakage can manifest itself. Metallurgical Solid Changes. Many metals, particularly the ferrous ones, go through phase changes in the solid condition that involve dimensional changes; for example, cast iron and steel will transform from austenite (stable phase above 1330  F) to pearlite or ferrite during cooling, with an accompanying change in volume. From the foundry standpoint, these changes produce and severely aggravate casting stresses. Here again, all the foundry operator can do is persuade the casting designer to avoid abrupt section changes or to use stress-relieving heat treatments in order to remove residual stresses. In many cases, these contraction and solid-state changes are predictable and will result in a uniform degree of warping and distortion in a casting. If this is uniform, then it can be allowed for by suitable compensating variations in the dimensions of the original pattern equipment. Volume and Dimensional Changes in Heat Treatment. When a solid casting is heated to anneal or otherwise heat treat it, it is always, except perhaps in the case of stress relief, heated to a temperature above that where the metallurgical solid-state changes occur. This means it is going to go through a temperature range where volume changes will occur and where stresses will be reintroduced into the casting. Here again, the only resort is to heat and cool slowly and carefully, to avoid abrupt changes in temperature. The designer must understand and allow for the fact that the casting will be heat treated, and therefore, the design must avoid complicated and abrupt changes in casting sections. In some heat treatments (for example, where a white cast iron or a partially white cast iron, such as malleable iron, is annealed to remove carbides and replace them with graphite), the dimensional changes that occur will be permanent. Usually,

Casting Design and Geometry / 115 graphitization is accompanied by a growth in dimension, which, again, must be allowed for in the original pattern design. Metals that are white and as-cast, such as malleable iron, may often be made from a pattern made to a standard rule, because the contraction during cooling in the mold may be completely offset by the graphitization during subsequent heat treatment.

Dimensional Changes of Casting Molds During the casting process, it stands to reason that some dimensional changes of the mold can occur when a mold cavity is heated by molten metal. Fortunately, these are quite small and are insignificant in the scheme of things. Volume changes can result from the pressure of molten metal on the mold cavity coupled with a decrease in strength of the mold from heat. This strength decrease causes what is commonly called mold wall movement. Most sand molds are made from silica, which exhibits an expansion as it is heated. Unfortunately, at a temperature of approximately 1000  F, it undergoes a phase change that involves a slight contraction. Therefore, it is possible for certain areas of the mold to be expanded while other areas are contracted. This sets up severe strains in the molds, and the foundry operator must allow for the strain by the degree of mold ramming, the selection of the grain size of the silica sand, and by the judicious use of binders that may contract and offset the silica contraction. In sand casting, many individual processing steps are needed to produce a sand mold. Such steps include sand preparation, coremaking, moldmaking, mold and core assembly, and finishing. Each of these steps, as well as other factors such as pattern wear, will contribute to the overall dimensional variability of casting features. When discussing dimensional errors or variability, it is important to distinguish between dimensional accuracy and dimensional variability. Dimensional accuracy—an indication of how close the casting dimension is to the actual target value— is often referred as a system error. The main causes for poor dimensional accuracy are pattern equipment errors, pattern wear, and casting contraction uncertainty, much of which can be corrected before a production run. Dimensional consistency—the variation of individual casting dimensions about the mean casting dimension—is often referred to as random error. Many metalcasting process variables contribute to the dimensional inconsistency of castings. It is therefore important that significant sources of dimensional variability are identified and that appropriate process controls are established to minimize them.

Design Case Studies Following are three different metalcasting design projects and how they went from idea to design to production.

Example 1: Design Conversion to Ductile Iron Casting for a Pour Chute Pivot Frame Integrating a multipiece part into a casting is often done to reduce assembly, machining, and management costs. However, the decision to convert a part to a casting is only the first step. During the conversion process, design engineers must make several key decisions in order to achieve the most optimal design. The best material and casting process must be chosen, design iterations should be performed with the casting process capabilities in mind, and time should be spent on discovering whether more advantages can be obtained through metalcasting—either by enhancing the part performance or reducing part cost further. A manufacturer of concrete mixer trucks for the heavy construction industry faced these choices when it converted a portion of its pour chute to a one-piece casting after a proposed redesign showed it would reduce costs and improve the part. The casting design team consisted of representatives of the track manufacturer, metalcasting facility, pattern shop, and machine shop and focused on three imperatives: design for performance, design for production, and design for cost. Three casting design issues played a major role in meeting these design imperatives:

 Meet or exceed the strength and durability of

the original weldment

 Choose an alloy grade for sufficient tough

 





ness and ductility for fatigue and impact resistance Design the casting with the selected alloy grade to keep stresses below the maximum design stress Minimize machining steps and costs Include the critical dimensional features— machined center hole in the hub, machined mounting holes in the triangular frame, and as-cast locking holes on the locking plate Achieve general tolerances of 0.75 mm (0.03 in.), as-cast hole tolerances of 0.5 mm (0.02 in.) and machined tolerances of 0.13 mm (0.005 in.) Achieve a clean and smooth surface finish with no visible surface defects or grinding marks

The resulting one-piece ductile iron casting weighed 18.14 kg (40 lb) and consisted of a central hub, a flat locking plate, a square mounting frame, and a triangular support frame (Fig. 8). The envelope dimensions of the pivot frame casting were 48.3 by 35.6 by 15.2 cm

 Select a ductile iron grade that meets the

mechanical requirements

 Optimize the design for stress reduction and

weight savings

 Develop a casting tool design that produces

flaw-free frames at the best cost Knowing the Application. The cement truck components must be rugged and durable to withstand mixing and pouring stresses as well as road shock and vibration over the life of the truck. The metal pour chute on the rear of the truck carries the concrete mix from the drum down to the concrete forms. The chute is designed to swing across a 160 arc to facilitate easy pouring into the forms without repositioning the truck. An integral part of the pour chute is the pivot plate frame. Originally, the pivot plate frame was a steel assembly consisting of nine fabricated and machined parts that were welded together (Fig. 7). The welded assembly performed to mechanical specification but was less than ideal for cost, assembly, variation in form and fit, and inventory management. A near-net shape one-piece casting design was proposed that would cost less, improve form and fit, and reduce stresses for longer life and improved durability. Part Requirements. The pivot plate frame supports the pouring chute of a concrete mixer truck, serves as a pivot point, and locks in position during the pour. The critical and production requirements for the cast plate frame were:

The original welded pivot frame assembly, shown here with the pivot shaft attached, consisted of nine pieces welded together.

Fig. 7

Fig. 8

The one-piece ductile iron casting resulted in a 50% production cost-savings.

116 / Casting Design and Performance (19 by 14 by 6 in.), the center hub had a diameter of 75 mm (3 in.) and a height of 90 mm (3.5 in.), and the minimum wall thickness was 9.5 mm (0.375 in.). The center hub of the plate frame secured the pivot shaft on which the plate and chute rotate. The locking plate had seven lock holes for securing the chute at different swing angles, and the two bushings on the top of the triangular frame secured the pouring chute to the pivot frame. Choosing the Material. The original assembly was welded from steel plates and bar stock, but ductile iron was chosen for the casting because it met the performance requirements while featuring a lower casting production cost compared to a steel alloy. Ductile iron exhibits a linear stress-strain relation, a considerable range of yield strengths, and ductility. It was clear that ductile iron was the metal of choice, but the casting design team had to choose among three grades of the ductile iron specified in ASTM A536 (Table 3). The ASTM A536 ductile iron casting specification is based on demonstrated mechanical properties that depend on the proper cast iron microstructure. High-ductility grades have a ferritic microstructure, intermediate grades have a mixed ferrite and pearlite microstructure, and high-strength grades have a primarily pearlite microstructure. The performance and production requirements of the pivot frame called for an alloy grade with sufficient toughness and ductility for fatigue and impact resistance. The microstructure of A536 grade 2 is primarily ferritic, so it has more than enough ductility/elongation and just misses the hardness requirement. However, it does not meet the minimum requirements for ultimate tensile strength and yield strength. Both A536 grades 3 and 4 feature mixed ferrite/pearlite microstructures, and both meet and exceed the requirements for tensile strength, yield strength, and hardness. However, grade 4 falls short on ductility and elongation. Because A536 grade 3 fulfills all the requirements, including ductility, it was selected as the bestchoice alloy for this casting. Optimizing Design with Finite-Element Analysis. One of the design advantages of metalcasting is the freedom to optimize cross sections and shapes beyond the limits of welded

Table 3

plates and bar stock. The design team used this advantage to design the triangular support frame of the pivot plate frame for the pouring chute. The new design needed to meet two objectives:  Keep the tensile stresses in critical sections

below the calculated stresses in the original welded design  Minimize the weight of the overall casting to keep production costs down and save weight on the truck Three-dimensional computer-aided design is the key to rapidly optimizing the design for mechanical performance and weight reduction. This also reduces the first-part time. The design team developed three designs and used finiteelement analysis to evaluate and optimize the stresses in the triangular support frame (where the stresses were highest). The designs included a plate with perimeter ribs, a frame with cross ribs, and a plate with ribs and cutouts (Fig. 9). Design A, which used a flat plate in the support section with perimeter ribs for stiffening and stress control, met target stress requirements at the joint where the triangular frame met the straight bars, but the weight of the design was higher than desired. Design B, which used a heavy frame in the support section with cross ribs for stiffening and stress control, markedly reduced the weight, but the stresses were excessive at the joint where the triangular frame met the straight bars.

Example 2: Design Conversion to a Die-Cast Aluminum Fan Housing

Grades of ductile iron alloy A536 A536 grade 2

Properties

Ultimate tensile strength, ksi Yield strength, ksl Elongation % Hardness, BHN

A536 grade 3

A536 grade 4

Performance requirement

65-45-12

80-55-06

100-70-03

>80

65 min

80 min

100 min

>55

45 min

55 min

70 min

>6 >195

12 min 156–216

6 min 187–255

3 min 241–302

Design C used a flat plate in the support section, with perimeter ribs and two long ribs in the back for stiffening. In addition, three cutouts in the center panel reduced the overall weight without increasing stresses in the critical sections. This design was selected as the final design because it kept the stresses within limits and met the weight target. Rapid prototyping using a high-speed 3-D printing system was used to check form-fit function on the final design. Saving in the Final Steps. Green sand molding was chosen as the best-value mold method to cast the pivot frame chute. Metalcasting engineers used orientation and risers to produce the desired directional flow in the mold. After casting, the component was shotblasted to remove residual sand on the surface, flash lines and a riser stub were ground off, and the casting was painted prior to machining. Three features on the pivot frame casting required separate machining steps. The inner diameter (6.35 cm, or 2.5 in.) of the center bore was rough drilled and finish drilled to specification. The inner diameters (2.26 cm, or 0.89 in.) of the two bearing holes were finish drilled to specification. The new casting, which converted a ninepiece weldment into a one-piece component, produced a 50% cost-savings on each part, compared to the original assembly, welding, and machining costs. The tool payback was four months of production. The one-piece design reduced part inventory, technical data management costs, and production-to-delivery time. Quality-management principles were applied at each stage of the casting process. Emission spectrographic analysis of the furnace and ladle chemistry assured precise control of the alloy composition, dimensional checks were performed on all features, visual examinations tested the surface appearance, and soundness checks were performed with sectioning on prototype castings. The converted component featured better form and fit with less dimensional variation, lower stresses, and longer life, compared to the weldment.

Three designs were developed for the triangular support frame of the pivot frame (clockwise from top): Design A featured a plate with perimeter ribs; design B featured a frame with cross ribs; and design C featured a plate with ribs and cutouts.

Fig. 9

In cold climates, tents, sheds, and buildings without central heating use individual space heaters to provide warmth. However, the hot air from the heaters rapidly rises to the ceiling without actually warming the living space of these uninsulated structures. One solution is a thermoelectric fan (TEF), which directly circulates warm air throughout the shelters. This is an efficient way to heat the whole area without being a drain on the power source. The TEF design hinged on the production of a die-cast fan housing, which provided structural support and protection for the fan, while also acting as a heat sink and thermal heat

Casting Design and Geometry / 117 conduit. Product designers from the TEF manufacturer worked closely with the diecasting supplier in each step of the casting design to achieve a final concept that was functional and castable. Knowing the Requirements. One of the most important elements of the TEF was its ability to conduct heat efficiently, so thermal properties of the diecasting were of top concern. The TEF sits on a space heater and circulates the hot air from the heater downward toward the floor before it rises to the ceiling, reducing thermal variations in the shelter and providing warmer conditions for the occupants. With the TEF, the space heater is able to run at a lower fuel rate while achieving the proper temperature levels throughout the enclosure. The bottom surface of the TEF is heated by direct contact with the hot stove, absorbing a small amount of heat that is converted by a solid-state thermoelectric generator into electricity, which drives an axial motor. The fan is directly attached to the upper shaft of the motor. No other external power source is needed. The fan housing is a 35.6 cm (14 in.) diameter, open-top cylinder with a complex slotted and finned base (Fig. 10). The cylinder is 25.4 cm (10 in.) tall and weighs 3.4 kg (7.5 lb). The wall thicknesses of the housing range from 2.24 cm (0.88 in.) at the base to 0.31 cm (0.124 in.) on the sidewalls. Slots in the base provide paths for the forced air to exit close to the hot surface of the stove, picking up heat and moving down away from the fan. In addition to high thermal conductivity to maximize heat transfer, the casting needed to have controlled wall thickness and uniform metal fill for structural integrity and rigidity. The machined features of the housing had to be free of porosity for good thermal conductivity and to prevent off-tolerance drilling and tapping. Dimensional tolerances for precisionmachined features needed to be  þ0.127 mm (þ 0.005 in.) to meet form-and-fit requirements.  The diecasting process was chosen to produce the fan housing because of its ability to

produce complex shapes with close tolerances, limited need for machining, integrated fastening elements such as bosses and studs, and its ability to produce castings at high rates with low cost. Annual production for the housing was 1500 units. The diecasting engineers had three design focuses: performance, castability, and cost, which hinged on a few design issues, including:  Choosing a metal alloy that best met the per-

formance and castability requirements

 Optimizing the diecasting process to reduce

porosity

 Tailoring the thermal management process

for quality and reproducibility Choosing an Alloy. The efficiency of the thermoelectric generator depended on establishing the temperature difference between its top and bottom. So, the aluminum housing had to act as a heat sink, pulling the maximum amount of heat away from the top of the generator and transferring it to the fan air. This required an alloy with high thermal conductivity for the housing. A range of different metal alloys is available to the diecasting engineer, and the performance requirements and manufacturability issues must be weighed together to find the right fit. In this case, the designers needed an alloy with high thermal conductivity, but for mechanical performance, they also needed it to be both strong and stiff. Because the housing had a machined mating surface and numerous screwholes to be drilled and tapped, an alloy with easy machinability was sought. From the diecasting perspective, the thin walls of the fins and ribs in the base required an alloy that would flow and fill the die cavity rapidly and easily but not solidify too quickly. Early solidification in the die can produce casting defects, such as lack of fill and cold shuts. The design engineers determined that three types of metal alloys could be considered for this application: zinc ZA8, aluminum A380, and the proprietary aluminum Thermalcast 130

of the diecasting facility (Table 4). All the alloys met the mechanical property requirements of strength, ductility, and modulus. ZA8, however, had a thermal conductivity 4% short of the requirement and weighed 150% more than the aluminum alloys. A380, a common diecasting alloy, featured an ultimate tensile strength of 47 ksi, a yield strength of 23 ksi, and a density of 2.63 g/cm3, (0.095 lb/ in.3), but the thermal conductivity was 20% lower than the design target. The Thermalcast 130 aluminum alloy had a high thermal conductivity of 130 W/mK, with the mechanical properties of conventional aluminum diecasting alloys. As a result, it was chosen as the metal to use in casting the fan housing. Controlling Metal Flow. After choosing an alloy, the diecasting engineers determined the best flow control and gate design for the diecast component. The diecasting facility uses a vacuum-assisted process in which vacuum evacuation of the die cavity reduces gas entrapment during metal injection and decreases porosity in the casting. The result is a higherquality casting; however, the vacuum system is not a substitute for good diecasting design in the engineering of the die cavity, runners, gates, and overflows. Control of metal flow in the die is a key factor in producing sound castings. Metal must flow rapidly and uniformly into the die, minimizing turbulence and entrapped air. A key feature in die design is the positioning of the runners and gates—the passages that feed molten metal into the die cavity. Well-designed gates are positioned to permit rapid flow into the die cavity, minimizing turbulence and long flow paths for the molten metal. For the fan housing, the diecasting engineers opted to position the gate at the center of the cavity, directly into the center hub of the casting (Fig. 11). With the centershot, the metal flows uniformly and rapidly from the center hub into all the thin-wall ribs and outer rim, producing complete fill. Producing the Final Part. After a final gating design was achieved, the housing was cast in a three-plate die (Fig. 12) with a cold chamber

Table 4 Properties of zinc ZA8, aluminum A380, and proprietary aluminum Thermalcast 130 Properties

The fan housing diecasting features wall thicknesses that range from 0.88 to 0.124 in. The interior view of the fan housing shows the slots of the base that provide paths for the forced air to exit close to the hot surface of the stove, picking up heat and moving down away from the fan.

Fig. 10

Density, lb/in.3 Thermal conductivity, W/mK Ultimate tensile strength, ksl Tensile yield strength, ksl Ductility, % elongation Modulus, Msl

Zinc Target ZA8

Al Thermalcast 130

Al A380

0.095 130

0.096 96

120

0.24 115

>40

54

46

47

>20

42

24

23

>3

6–10

4

4

>10

10.2

10.3

10.3

118 / Casting Design and Performance gating system. Once the housing was cast and trimmed, it was checked for dimensional tolerances and surface condition. Then, it was bead blasted and machined. The base hub of the housing was finish machined to a 32 root mean square finish for good thermal conduction. Finally, the necessary mounting and assembly holes were drilled and tapped, and the housing was finished with a protective black powder coat for corrosion protection and appearance.

Example 3: Coil Rotor for Traction Motor of Railroad Wheel Assembly This example is a step-by-step look at the decision process for casting of a coil rotor for a diesel locomotive engine. Several decisions are needed on how the part will be cast that may affect the final product, even if the design may specify casting as the production method. The casting design engineers for this component focused on three imperatives:

The diecasting designers opted for a centershot gate for molten metal flow. This allowed metal to flow uniformly and rapidly into all the thin-wall ribs and outer rim.

Fig. 11

Fig. 12

 Design for performance  Design for production/castability  Design for cost

To meet the design imperatives, engineers had to select a steel composition that met the performance and casting requirements, choose a molding system that achieved tolerance and cost targets, plan an orientation in the mold that optimized the precision features in the casting, and design a gating and riser system that ensured soundness in the casting. The armature coil rotor is an integral part of each traction motor used in a diesel locomotive. A diesel locomotive is equipped with a turbocharged diesel engine that delivers the shaft horsepower to the main electrical generator to convert into electrical power to drive the locomotive. The electrical power from the generator then is distributed to the traction motors, each of which converts the electrical energy into torque on the drive wheels through a reduction gear case. The armature coil rotor has a complex shape disc on one end of the rotor shaft, which serves as an end plate containing the windings and acting as a structural feature. The coil rotor disc is the most complex casting in the rotor assembly. The 38.1 cm (15 in.) diameter disc weighs 54.4 kg (120 lbs) and has a slotted and ribbed main body and a 14.5 cm (5.7 in.) high, 17.8 cm (7 in.) diameter center hub with a 11.9 cm (4.7 in.) inside diameter center hole. The outer rim is 7.6 cm (3 in.) high and 1.27 cm (0.5 in.) thick. The slots in the disc aid in cooling the windings, and the disc has a range of section thicknesses from 4.83 cm (1.9 in.) to as thin as 0.48 cm (0.19 in.) in the ribs. The nominal performance requirements for the wheel are:  Ultimate tensile and yield strengths of 60

and 30 ksi

 Elongation of 22% and reduction of 30%

 Hardness of 120 to 163 BHN  Surface finish of 250 to 500 root mean

square on as-cast surfaces

 Dimensional tolerances of þ0.08 mm (+/

 0.003 in.) on machined surfaces high-magnetic-permeability steel for magnetic performance

 Low-carbon,

Alloy Selection. Due to the range of section thicknesses in this casting, an alloy with excellent fluidity was required. Also needed was an alloy with low manganese, phosphorus, and sulfur content to minimize the magnetic hysteresis losses. The original equipment manufactures specifications originally called for a low-carbon steel (1015) based on the mechanical requirements. However, that specific carbon level in the alloy has insufficient fluidity at the pour temperature for smooth and rapid metal flow into this complex mold. The metalcasting facility suggested an alternate alloy (1020) that contained a higher carbon level while possessing the same levels of manganese, phosphorus, and sulfur without the loss of mechanical properties. The higher carbon level reduces the melt temperature and gives improved fluidity in the mold at the pour temperature. Choosing a Molding Method. This casting could be produced by two types of molding methods: green sand or shell mold. Each of the two molding methods has relative capabilities, advantages, and costs, as shown in Table 5. After comparing the two molding methods according to the requirements for surface finish, section thickness, and heat-transfer efficiency for the support rotor, it was determined shell molding was the best choice. Green sand molding can meet the baseline requirements for the rotor as-cast tolerance, surface finish, detail level, and tool cost, but it does not meet the requirement for minimum wall thickness or heat-transfer efficiency. Shell molding, on the

The fan housing was cast in a three-plate die with a cold chamber gating system. Shown here is the cover half (left) and ejector half (right) of the die.

Casting Design and Geometry / 119 Table 5 Relative capabilities, advantages, and costs of green sand and shell molding Target

As-cast dimonsional tolerance across 1 in. Nominal surface finish Minimum section thickness Heat-transfer efficiency Intricacy of detail Tool/pattern cost

Green Sand

Shell-Mold

þ0.030 (0.08 cm) 

þ0.030 (0.08 cm) 

þ0.008 (0.02 cm) 

500 0.19 in. (0.48 cm) High Fair Medium

500–900 0.25 in. (0.64 cm) Low (sand) Fair Low(wood)

300–500 0.15 in. (0.38 cm) High (metal fill) Very good Medium (metal)

Note: The performance numbers for green sand and shell mold are typical. Actual values for a given metalcasting facility could be higher or lower depending on the equipment and expertise at the facility.

Fig. 13

Four gates on the rim

other hand, meets all the functional requirements and offers a thin-wall capability and heat-transfer efficiency that cannot be achieved with green sand. The higher cost of the metal tool for the shell mold was acceptable, given the production run and benefits. Planning the Component Orientation. The orientation of the part in the mold is an important factor in producing a sound casting. Casting defects, when they occur, tend to rise and segregate in the cope of gravity castings, so machined surfaces and critical features should be molded in the drag. In the coil rotor disc, the thin walls next to the hub and the ribs on the base of the rotor are critical features that

Two-gate approach. The two-gate approach was chosen for this component in order to facilitate proper filling.

Fig. 14

carry a significant load and must be flaw-free for strength and machinability. The rotor is oriented in the horizontal plane, so the parting line is perpendicular to the rotor centerline. Designing the ribs in the cope section would allow the possibility for inclusions or porosity to occur in the ribs of the casting, putting the ribs and machined surfaces at risk. By orienting the molding with the component ribs down in the drag section, inclusions or porosity that may form would segregate in the top of the casting in the center hub, where they will have minimal effect on mechanical

properties. The rotor ribs will have no defects with this orientation. Designing the Gating System. Proper design of the gating system is critical for uniform, controlled metal flow. Nonuniform, long path, and/or slow metal flow may produce unfilled sections or solidification shrinkage in the casting. Too many gates and runners also will require additional molten metal and decrease the casting yield. Casting engineers were faced with two choices for a gating system. One placed four gates on the outside rim of the casting (Fig. 13), and another placed two gates in the center hub of the casting (Fig. 14). In the first approach, the four gates on the rim gave a very long metal flow path into all the ribs. With the four-rim gates, a test casting showed poor fill into the ribs, with resulting rib misruns and defects. The test casting with the two-gate approach, however, revealed excellent metal fill. The two gates on the center hub gave fast, direct metal flow into all the ribs. This approach was chosen. Because the rotor rim has significant thermal mass and will solidify slowly, a continuous source of molten metal must feed into the rim during solidification to prevent shrinkage porosity. Perimeter risers feed molten metal into the rim. Smaller risers would normally solidify more quickly than the rim, which would shut off metal feed into the rim before it is fully solidified. To overcome this problem and to provide constant molten metal to the rim, the risers are surrounded with a 13 mm (0.5 in.) thick exothermic sleeve. The sleeve is a source of heat for the riser and keeps the riser molten and feeding metal into the rim as the rim solidifies. After solidification is complete, the rotor casting is finished and machined before it is inspected for quality, dimensions, and finish.

Casting Design and Performance Pages 121–132

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Thin Sections THIN SECTIONS save weight and may thus contribute to a more favorable strength-toweight ratio. By requiring a smaller volume of metal, thin walls (within practicable limits) may also lower casting costs, particularly when an expensive alloy is being poured. Realization of the advantages of thin walls in a casting depends largely on certain physical limitations of metal in the casting process. These limitations are concerned with the fluidity of molten metal as it affects mold filling, with the processing required for metal soundness, with differences in alloy and in casting process, with distortion and heat treating problems (in some designs), and with the cost of foundry engineering and development. Guidelines on the minimum thicknesses to which a cast section can be designed depends on various factors. For example, Table 1 lists guidelines intended to prevent an increase of casting cost because of thinness of section. Other guidelines may be based on the actual limits of the process, rather than cost, but often fail to define the applicable processing conditions. Because so many aspects of design (such as the area of the thin-wall section, its proximity to a gate and the availability of heavy sections to conduct molten metal to it) have an influence on the actual minimum wall obtainable for any particular casting, recommendations such as those in Table 1 have limited usefulness when applied to a specific part. They do, however, provide the casting designer with a starting point, by familiarizing him with minimum thicknesses that normally can be produced efficiently. The mutability of what was presumed to be a practical minimum wall thickness is demonstrated by the experience derived in producing the large cover casting shown in Fig. 1. Cast in 17-4 PH stainless steel by a ceramic mold process, this part was approximately 21 in. long by 6 in. wide and weighed about 15 lb. Minimum weight was desirable, and it was assumed initially that a wall thickness of 0.150 in. was the safe and practical minimum. Machined aluminum patterns were mounted on aluminum cope and drag plates, and castings produced were sound and free of cold shuts. Subsequently, one flat mounting face of the pattern was machined to reduce the casting wall thickness to 0.080 in., a reduction of about 47%. Even with this large reduction, only a small percentage of castings showed signs of

cold shuts. (However, it was apparent that a further reduction in wall thickness would probably lead to difficulty in filling the mold.) Another general example is the following minimum wall thicknesses proposed for areas approximately 1 in. square in investment castings, depending on the metal being cast: Aluminum alloy Beryllium copper Low-alloy steel

0.050 in. 0.040 in. 0.060 in.

At least one producer of investment castings considers wall thicknesses about 0.010 in. less than those indicated to be feasible where smaller areas are to be fed. Sections with an area larger than 1 sq in. may require heavier walls than those indicated. Three examples of investment castings that were produced successfully with wall Table 1

thicknesses less than are usually recommended are shown in Fig. 2. Figure 2(a) illustrates an aluminum investment casting (alloy 356) that contained, in several locations, walls 0.031 in. thick; Fig. 2(b), a beryllium copper casting incorporating in three locations tubular sections having walls 0.031 in. thick; and Fig. 2(c), a 4140 steel investment casting with three walls 0.050  þ 0.010 in. thick. All three castings were successfully produced in quantity, incorporating wall thicknesses less than are usually considered minimum for economical production with these alloys.

Determination of Thinnest Wall Prior to the use computer-assisted simulation, practical estimates of minimum wall thickness of a casting could sometimes be established

Typical minimum section relations for three metals in five casting processes (a) Minimum section thickness, in.

Casting method

Sand Permanent mold Investment Die casting Plaster mold

Aluminum

0.125  þ 0.031 0.093 þ 0.015  0.062  þ 0.010 0.062 þ 0.010  0.080  þ 0.015

Magnesiu

Steel

0.156  þ 0.031 0.125 þ 0.015  0.062  þ 0.010 0.093 þ 0.010  ...

0.187  þ 0.031 ... 0.093  þ 0.010 ... ...

(a) For compatibility with economical production. Not the thinnest producible by the processes.

Fig. 1

A 17-4 PH stainless steel ceramic mold casting, the wall thickness of which was reduced, from the presumed practicable minimum of 0.150 to 0.080 in., without appreciably affecting the soundness of castings produced

122 / Casting Design and Performance

Fig. 2

Three investment castings that incorporate walls of less than recommended minimum thickness

reliably by producing sample castings from inexpensive loose patterns accurately produced. On these experimental patterns, wall thicknesses would progressively increased or diminished, or revisions may be made to compensate for errors in calculating contraction, at a fraction of the cost entailed in revising production patterns. This method was used for the casting shown in Fig. 3. For this flanged elbow casting, the problem was to determine the minimum practical wall thickness, using aluminum alloy 356 and the plaster mold process. Wall thickness was reduced progressively by modifying the pattern equipment, and castings were produced with three different wall thicknesses. Significant variables were carefully controlled, to eliminate any extraneous influence on the tests. The results of the tests were evaluated in terms of the percentage of rejected castings, as follows: Wall, 0.040 in. 0.060 0.080

Fig. 3 0.080-in.

An aluminum elbow (alloy 356) cast by the plaster mold process to three different thicknesses to determine the effect of wall thickness. Rejections were 80% with 0.040-in. wall, 35% with 0.060-in., and 10% with

Rejections, 80% 35 10

Because the percentage of rejected castings is normally reflected in the selling price of acceptable castings, the price of the thinnerwall castings with higher rejection rates could be expected to increase proportionately. Also, the problem of meeting production schedules would become increasingly critical as wall thickness decreased. These are disadvantages that the designer must evaluate in relation to the known advantages of a wall thickness approaching the absolute minimum. Stainless Steel. Figure 4 shows a thin-wall austenitic stainless steel sand casting with a number of design features that helped to simplify casting problems and permit production at minimum cost. The casting is of relatively uniform section thickness, except for the heavy boss at the top. This boss was ideally placed to facilitate filling the sand mold and to avoid

A thin-wall sand casting produced from austenitic stainless steel. One section of the casting required two revisions in wall thickness to bring rejection rate to an acceptable level. Rejections were 50% with 0.060-in. wall, 25% with 0.075-in., and 5% with 0.090-in.

Fig. 4

distortion as the steel solidified and cooled. The casting has good stability and rigidity, and its slightly curved surfaces minimize distortion. As originally designed, the walls in the lower part of the casting were 0.060 in. thick. However, because of difficulty in eliminating porosity in the walls, caused by the back pressure of mold gas, rejection rate was approximately 50 per cent. Increasing the thickness of these

walls to 0.075 in. reduced the rejection rate to about 25 per cent. When the wall thickness was increased to 0.090 in., the rejection rate was further decreased to 5 per cent. Even the thickest wall (0.090 in.) is well below the minimum usually suggested in tables like Table 1, which are based on economy rather than absolute process capability. This casting illustrates what can be done when the design of a casting is favorable for producibility.

Design Problems Involving Thin Sections / 123

Thin-Wall Steel Sand Castings In all casting processes, castings that are designed so that rapid and orderly flow of metal to all parts of the mold is promoted are easiest to produce with a minimum of flaws or defects. However, the specific casting process can influence metal flow, and, because of inherent differences, each casting process exerts its particular influence on the production of thinwall castings. The relative unsuitability of green sand molds to the casting of thin-wall sections is considerably magnified in the casting of either alloy steels or superalloys. For this reason, most aircraft castings made of these materials are produced in shell, ceramic or baked sand molds. An analytical approach to the design of thinwall steel castings to be produced in green sand molds necessitates an understanding of four major production problems involved: mold filling, distortion, soundness, and the influence of design on heat treating problems. Mold-filling problems become increasingly acute as wall thicknesses are decreased to less than 8/16 in. Above this thickness, mold filling is seldom critical. When wall thickness is decreased to less than 1/16 in., adequate mold filling becomes almost impossible, except for extremely short distances, and misruns result because of the freezing of metal before the mold is filled. To achieve minimum weight, the designer will often incorporate minimum section thicknesses within a casting, in all locations where high strength is not a requirement. For instance, in the turbine bearing housing shown in Fig. 5 (a green sand casting made of alloy steel), eight posts, each incorporating a stiffening rib, connect the top and bottom portions. It was calculated that a rib thickness of 0.100 in. would be adequate to meet strength requirements. However, ribs of this thickness could not be filled completely; all castings produced were rejected for misruns. No practical improvement was possible in foundry processing. Consequently,

Eight stiffening ribs in this alloy steel green sand casting were originally 0.100 in. thick; misruns made all castings produced unacceptable. Increasing thickness of these ribs to 0.125 in. almost completely eliminated rejections for misruns.

the patterns were revised to increase rib thickness to 0.125 in.; rejections for misruns were almost entirely eliminated. Alloy Selection. Important in the problem of mold filling are the fluidity and ability to feed of the metal specified. These properties of castability are low in steels and superalloys, as compared with light-metal alloys. And among steels, the castability of one alloy differs from that of another. As a result, a casting that can be produced successfully in one alloy may not be castable in another. Among foundries, the comparative moldfilling abilities of metals are controversial, and relative ratings of metals in regard to this characteristic are to some extent empirical. Existing disagreement reflects the many factors that may affect the ability of metal to flow in a mold. From its experience with many alloys, one foundry lists alloys in order of decreasing ability to fill thin cavities as follows:  Cobalt-base alloys (except precipitation-

hardening alloys)

 Nickel-base alloys (except precipitation   

hardening alloys) 11 to 13% Cr steels Series 4300 alloy steels Series 300 stainless steels Series 300 stainless steels molybdenum

containing

Rigging and processing procedures vary from foundry to foundry, and foundries generally specialize in respect to the range of castings and alloys with which they prefer to work. This should be recognized by the engineer when he designs a casting with rigid requirements that may tax the capabilities of the process best suited to produce the casting. The aft-fuselage speed-brake hinge shown in Fig. 6 was designed to be sand cast from 4340 steel and heat treated to a tensile strength of 180,000 to 200,000 psi. It was to be produced to close tolerances and rigid standards of inspection. The foundry accepted the job with the understanding that the material could be changed from 4340 steel to type 431 stainless, and that certain of the more stringent tolerances could be modified. The change in material was requested because of the foundry’s extensive experience with stainless steel. Although type 431 stainless costs considerably more than 4340, this increase was more than offset by

the higher percentage of acceptable castings produced in stainless steel. Friction between molten metal and mold may seriously hinder mold filling. It becomes increasingly troublesome as the ratio of mold surface to metal volume increases, a condition that is characteristic of thin-wall castings. A high ratio promotes premature freezing in thin sections. Mold gases may create back pressures that retard the flow of metal. These gases usually originate from the burning of organic moldbonding material and from the heated air in the mold cavity. Gases are expelled through the mold wall in proportion to the static head of metal in the sprue and risers, depending also on the permeability of the mold material. Designs that must be oriented in the mold so as to have dome-like thin sections are especially likely to create problems in venting of mold gases, which can generate back pressure and cause the metal to freeze before the mold is filled. The most obvious solution in the foundry is to use a higher pressure head, in order to force metal into the cavity more rapidly, but this may cause undesirable turbulence of the metal at or near its point of entry into the mold. Mold-Filling Problems. An unusual moldfilling problem that, because of the design of the casting, had no satisfactory solution was encountered in the production of the alloy steel part shown in Fig. 7. This casting was molded in the position shown, because, had the mold been inverted, the heavy section at the junction of the shaft and plate sections could not have been fed through the long shaft. Because of its diameter, the entrance to the saucer shape is small in proportion to the volume of metal that must pass through it. Thus, the metal broke into separate streams and trickled down to various points on the periphery of the plate section. Such separate streams oxidized rapidly and resulted in misruns, cold shuts, and “BB shot”. (The BB-shot defect refers to drops of molten metal that separate from the main body of fluid metal, freeze prematurely, and fail to fuse with the main body of metal on recontacting it.) A possible solution might have been provided by the addition of a reservoir (indicated by the broken line, B, in Fig. 7) surrounding the periphery of the plate. Such a ring would have been removed subsequently by machining. Although the wall thickness of 0.060 in. would probably

Fig. 5

This close-tolerance sand casting, although designed to be cast from 4340 steel, was produced in type 431 stainless because of the foundry’s greater experience with stainless steel. Increase in material cost was more than offset by higher percentage of acceptable castings.

Fig. 6

124 / Casting Design and Performance adding to the thickness of the thin sections and by tapering the walls to the center of the volute (E). Illustrating another mold-filling problem, the steel casting shown in Fig. 9 indicates how obstructions to the flow of metal as it fills the mold can result in casting defects. In this casting, metal entering the mold cavity through the heavy section (B) flowed into the thin plate area and flanged sections. In pouring, the mold was tilted, with the heavy section down, in order to prevent a trickling of metal into the plate and flanged areas. To assure metal flow to all parts of the mold, this casting was poured rapidly. However, as the wall of metal advanced to each of the three sand cores C, an eddy was formed behind each core, as indicated by the shaded areas, resulting in defects. The defects became increasingly critical as the thickness of the plate was decreased. The possible revision of core shape shown in the lower portion of Fig. 9 was not permissible. Consequently, in order to produce a sound casting, it was necessary to eliminate the coring of holes and to drill these holes. Another problem in mold filling results from the cascading of molten metal down one side of the mold for a casting designed with a ladder shape. For the original design illustrated in Fig. 10, metal must enter the mold cavity at points A and rise evenly into the two outer walls. In

walls less than about 3/16 in. thick, this is difficult to assure. If the outer walls fill unevenly, metal will flow across a cross member B, and cascade down the opposite wall. As descending metal makes contact with metal rising in the wall, bubbles are likely to be trapped, and cold shuts and other defects may appear. The weight of metal in a thin-wall casting is usually insufficient to create enough pressure to force entrapped air out through the mold walls. The result is voids and porosity. Thus, this 8-in.-long casting as originally designed, with a 0.060-in. wall thickness and five cross members, is incapable of being properly filled. The problem can be solved in at least three ways. First, the outer wall thickness can be increased to 0.187 in. or more, as in Fig. 10 (a). Second, the inner cross members may be slanted, parallel to each other, as in Fig. 10(b), eliminating most of the danger of cascading. Third, the inner cross members may be slanted in opposite directions to form a truss, as in Fig. 10(c). Distortion is a major problem in the production of thin-wall steel castings. During solidification and in cooling to room temperature, distortion may be caused by variations in the cooling rate and in the contraction of sections of different thickness in the casting. In fragile castings, distortion may be caused by mold restraint. It may also occur during heat treating.

An alloy steel sand casting that was impractical to produce because its inverted saucer shape contributed to cold shuts and misruns.

Thin-wall volute castings such as this present problems of mold filling because molten metal must travel a long distance to fill the thin walls.

have caused the metal to freeze before it could flow into the reservoir, increasing the plate thickness to 0.120 in. should have permitted the flow of metal into the reservoir. Unfortunately, this approach would have been impractical; the additional metal would have had to be removed by machining to a curved surface. The cast volute shown in Fig. 8 is a common configuration that illustrates another mold-filling problem. There are three ways to gate such a volute. The first and most common method is to gate at the inside of flange A. The second is to gate at flange B. The third method is to gate at both flanges A and B. If the mold is filled at flange A, the metal will take a relatively long time to fill the thinwall section before filling flange B. Thus, if the wall of the volute is less than about 3/16 in. thick, the metal is likely to freeze before filling the outside flange of the casting. If the mold is filled at flange B, the same problem exists, with the added danger that metal will cascade at angle C, producing cold shuts and other defects. Finally, if the mold is filled at flanges A and B, entrapped air and mold gases in the thin-wall section may cause bubbles and porosity. Also, if the distance from C to D exceeds about 6 in., this thin section will be very difficult to feed well enough to make it sound. The broken lines in Fig. 8 indicate a method for improving the castability of such a part by

Fig. 7

This thin-wall steel casting required rapid pouring to fill the mold completely. Three cores obstructed the free flow of metal, creating eddies that resulted in defects. Redesign of cores as shown, had it been otherwise acceptable, would have solved the problem of metal flow.

Fig. 8

Fig. 9

Fig. 10

Uneven filling of the sides of the “ladder” configuration of this casting as originally designed caused metal to flow through the cross members and cascade down either side. Three revised designs are shown.

Design Problems Involving Thin Sections / 125 Soundness. Because of the small cross-sectional area of a thin wall, it is difficult for a riser to replace the metal lost in shrinkage. In very thin sections, feeding may be less of a problem, because the likelihood of centerline shrinkage is slight. Examples of Good Design. The chine fitting (a steel casting) shown in Fig. 11 is an excellent example of good casting design. The heaviest section is at the center, where it can be fed readily. The legs of the casting are tapered in steps. All wall sections of uniform thickness are less than 3 in. long and can be fed adequately from the center, with metal soundness assured. Changes in section thickness are not abrupt, and all inside corners are provided with generous fillets to resist tearing. Radii on outside corners serve to maintain constant wall thickness. A stout tie bar crosses the center, where the danger of distortion is greatest. The walls of the casting are thick enough to resist distortion. Finally, the radii at points C are large enough so there is no cracking or tearing

in this area. This casting was producible without difficulty. Another well-designed steel casting is shown in Fig. 12. This part can be fed with a riser at each end, and at each side at points A; thus, the casting can be filled from the center or either end. Wall thicknesses are approximately 0.12 in., and the T-sections are easy to feed, because they are not more than 4 in. long nor more than 2 in. from the nearest riser. As originally designed, the casting was weak at points B. However, this deficiency was readily corrected by the addition of outer ribs. The revised design has good rigidity, with all areas of stress well braced and with ample corner radii and fillets. The casting did not distort in production. The wing spar casting shown in Fig. 13, although large and rather complex, is another example of good design. Basically, the design is that of a box section with a flat side approximately ¼ in. thick. One end of this steel casting is closed and varies in thickness from 0.178 to 1.50 in. Except for the round bearing support at the other end, the remainder of the casting varies in thickness from 0.187 to 0.250 in.

An example of good design in terms of foundry producibility. The heaviest section of this cast steel chine fitting is at its center and is easily risered. From the center, each section tapers in steps to the extreme ends.

Fig. 11

Fig. 12 risers

A box-shaped wing spar casting in which heavy sections can be easily risered from the top, the two sides, and the two ends. This steel casting demonstrates the rule that areas of increased mass should be restricted to five sides of a basically cubic configuration.

A well-designed steel casting in which all sections are accessible to properly located

Most of the heavy boss sections are located on the inside of the flat plate area. In casting, the part was inverted so that the bosses could be fed through the plate area, thus making this the cope surface. All massive areas could be reached from the top, the two ends, or the sides. The design of this casting exemplifies the principle that areas of increased mass should be restricted to five sides of a basically cubic configuration. All corners of the casting have generous fillets and radii. The box-like shape resists distortion in cooling from the liquid and during heat treatment. The numerous blanks attached to the horizontal ribs are intended to provide specimens for test purposes, and can be removed easily without damaging the casting. For thin-wall castings, such specimens provide more reliable test results than do specimens that are separately cast.

Thin-Wall Aluminum and Magnesium Castings Light loads that can be adequately supported by aluminum or magnesium castings incorporating thin sections. An example of an aircraft part is shown in Fig. 14, sand cast in aluminum alloy 356, this part was designed as an access door for the fuselage of an airplane. On the basis of experience with similar shapes, the practical minimum wall thickness for this casting was established at 0.08  þ 0.01 in. To assist in conducting the molten metal to all parts of the casting, however, the ribs were designed to a thickness of 0.12  þ 0.01 in. A substantial number of these casting were produced, with rejections from misruns or cold shuts held to acceptable limits. This degree of success indicated that the wall thicknesses chosen for the configuration did not, in fact, exceed process capabilities. In thin-walled castings of certain designs, the light weight of aluminum fails to provide sufficient metallostatic pressure to force the metal into the thin sections. In such castings, it may be necessary to choose a process that provides for the application of additional pressure to the metal.

Fig. 13

Fig. 14

An aluminum sand casting (alloy 356) with a minimum wall thickness for its shape. Heavier ribs assisted in providing adequate metal flow and kept rejections at an acceptable low level.

126 / Casting Design and Performance The design for the fin-shoe slide casting shown in Fig. 15 called for maintenance of thin-wall sections (0.100 in., tapering to 0.040 in.), a surface roughness not exceeding 250 micro-in., and the use of an aluminum alloy capable of providing a minimum tensile strength of 32,000 psi, a minimum yield strength of 20,000 psi, and a minimum elongation of 5 per cent. These mechanical property requirements were to be checked on the basis of separately cast test bars taken from the same heat from which the casting was poured. To accommodate the minimum wall thickness desired (0.040 in.), the investment process with gravity pouring was selected. Although its mold-filling ability is inferior to that of aluminum alloy 356, alloy 195 was first chosen, primarily on the basis of mechanical properties. This combination of alloy and process proved unsuccessful; all castings were rejected because of cold shuts and misruns. Next, a close-tolerance dry sand mold containing three equally spaced cavities around a sprue was tried. By spinning the mold at approximately 800 rpm, enough pressure was exerted on the molten aluminum to produce sound castings. The resulting mechanical properties and surface conditions were satisfactory also. Analyzing the production record of this part, it was concluded that even a 0.095-in. section tapering to 0.035 in. might have been attempted without incurring an excessive number of rejections. Core Movement. In many castings whose designs incorporate cored holes, the shifting of cores as the castings are poured can reduce wall thickness to the point where complete filling of the mold becomes difficult. To avoid a high rate of rejections, such designs should be modified to improve core stability. Even those castings in which a wide variation in wall thickness is permissible may encounter a high rejection rate if core movement results in the production of walls that are too thin to be cast satisfactorily. For example, because of core movement in the aluminum sand casting shown in Fig. 16, the 5/16-in. wall

was reduced to 3/32-in. in some areas. To maintain the 5/16-in. wall thickness, additional support was provided for the cores. Effect of Area. In aluminum and magnesium castings, as in all other castings, the total casting area has an important influence on the successful production of thin walls. This is evidenced by a comparison of the castings shown in Fig. 17 and 18. The casting of Fig. 17 is 3 by 2 by 1 in., with uniform walls 0.080 in. thick. This AZ63A magnesium casting was easily produced in shell molds (the rejection rate was less than 5 per cent from all causes). Compare this with the AZ91 magnesium sand casting illustrated in Fig. 18. The wall thickness of the body of the latter casting was established at 0.080  þ 0.010 in.—the same as for that in Fig. 17. However, because of the greater size of the casting in Fig. 18, ribs 0.100  þ 0.010 in. thick were required, to assist in conducting metal to all areas of the mold. In addition, it was necessary to produce this casting under closely controlled conditions. It is probable that, even with the ribs, the 0.080-in. wall would have proved too thin for ordinary commercial production and a revision to thicken the wall would have been necessary. Strength-to-Weight Ratio. Because they are rigid and have a high strength-to-weight ratio, truss-type castings are desirable for many instrument applications. However, production problems arise when the metal does not flow readily through thin walls in such castings to all parts of the mold. The pilot’s scope and sight support shown in Fig. 19 was produced from magnesium alloy AZ91 in a close-tolerance dry sand mold. The first castings produced were rejected for misruns, cold shuts and hot tears. It was obvious that the heavy ends, acting as manifolds, should be made still heavier, so that the metal could retain most of its heat as it flowed from the heavy sections to the thin sections of the mold cavity. It was also evident that the frame sections of the casting would have to be made heavier to

Insufficient pressure from a static head of molten aluminum caused cold shuts and misruns that resulted in rejection of all investment castings of the above design. By centrifuge casting in a close-tolerance dry sand mold, rejections were reduced to an acceptable level.

Fig. 15

improve metal flow; consequently, ribs were added to these thin sections. Because the junctions, and thus the increased mass (and accompanying hot areas), extended the length of the thin truss sections, the metal was able to flow to all parts of the casting. No solidification problems resulted, because freezing progressed along the length of the ribs and feed metal was available through the junction area. Heat in the ribs kept the frame sections molten for a longer time. Thus, the junction areas of the frame sections were given a longer time to solidify and build up strength to resist the contractive forces that were causing hot tears. To prevent hot spots at those points where the newly added ribs joined with other components of the casting, the ribs were cut short and were not permitted to join with the abutting walls. However, they were brought close enough to these walls so the metal could utilize the ribs as flow channels, and only a normal percentage of rejections occurred from all causes. Feeding Through Thin Sections. When metal must travel through a thin section of a

Core movement in this aluminum sand casting caused reduction of a 5/16-in. wall to 3/32 in. Additional core support was required, to hold desired wall thickness.

Fig. 16

A thin-wall casting that was successfully produced in a shell mold from magnesium alloy AZ63A. Rejection rate was less than 5%.

Fig. 17

Design Problems Involving Thin Sections / 127

Sand casting of AZ91 magnesium alloy, in common with the one shown in Fig. 17, had an 0.080-in. wall. Because of the considerably greater size of this casting, however, thicker ribs were required, to assist in conducting the metal to all parts of the mold.

Fig. 18

In this sand casting, the 3/32-in. wall froze with microporosity and shrinkage. By increasing the thickness to 5/32 in., defects were eliminated. Aluminum alloy 355

Fig. 20

Complicated thin-wall truss-type casting was produced economically by adding heavy ends to act as manifolds, and small ribs to act as guides for the molten metal. By not junctioning the ribs with the abutting walls, hot spots were avoided. Alloy was AZ91 magnesium; process was close-tolerance dry sand molding.

Fig. 19

casting to fill an adjoining heavy section, temperature gradients and solidification rates may be difficult to control. In the aluminum sand casting of Fig. 20, produced for an aircraft structural application, the bottom part of the mold cavity was filled through the 3/32-in. wall. The chilled bottom flange section filled satisfactorily, but all castings were rejected because of microporosity and shrinkage in the thin section. Establishing a sharper temperature gradient would have accentuated the defects. Side risering was impractical, because of the trimming and finishing problems involved. A feasible solution was to increase the 3/32-in. wall to 5/32 in. This increase in thickness, aided by the chills at the bottom face of the flange, created a freezing pattern that permitted adequate

feeding of all sections as the casting froze progressively from the bottom to the top flange.

Thin-Wall Permanent Mold Castings In the permanent mold process, the degree of as-cast surface smoothness required has a direct effect on the section thickness that can be achieved. Surfaces of the metal mold must be covered with an insulating coating so that they are protected against the damaging effects of thermal shock that occur when the molten metal is poured. Thin sections are coated with materials of various compositions and thicknesses that have certain heat-transfer characteristics. Heavy sections are coated with other types of materials, with different characteristics. These mold coatings

influence the directional solidification of the casting. Expendable Cores. One advantage of the permanent mold casting process is that sand or plaster cores can be used instead of metal cores. This helps to overcome many design limitations. For example, in a wave-guide casting requiring a smooth inside surface, the use of permeable plaster cores achieves a surface finish better than that which could be obtained with coated metal cores. Plaster also has excellent insulating qualities and does not prematurely chill the metal in thin sections. Sand cores permit the casting of tortuous passages, or of chambers or passages that are larger in section than the opening through the wall of the casting. Because the sand cores are expendable, removing them presents no major problem. Nonuniform walls in a permanent mold casting are usually difficult to produce efficiently, especially when a heavy section must be fed through a thin section. In the aluminum casting illustrated in Fig. 21, gating and risering the outer heavy section was routine, but the inner heavy section of the casting had to be filled and fed through the five thin ribs. To fill this inner section properly, the metal had to be poured quickly. This caused sinks to appear on the outer surfaces of the casting. Rapid pouring also created excessive turbulence. Increasing the thickness of the rib opposite the gate would have permitted the molten metal to flow with normal streamline velocity through the rib to all parts of the mold, and would also have allowed the inner sections of the casting to be fed more readily during the cooling cycle. Nonuniform wall thickness necessitated by molding and machining requirements caused

128 / Casting Design and Performance shrinkage defects in the permanent mold casting shown in Fig. 22. This aluminum casting (alloy 355) was part of an aircraft fuel system and had to be leakproof at a pressure of 35 psi. Minimum casting weight was mandatory, because the assembly was at its maximum acceptable weight. The shrinkage defects occurred at the junction of the hose-attachment tube A with the body of the casting. Stock required for machining increased the thickness of the wall of this tube. Furthermore, the internal core of this tube had to be tapered, to facilitate its withdrawal. This coring requirement also necessitated that a sharp corner remain at the junction of the large core with the small core that was withdrawn from the tubular section; this, too, increased the mass of metal at the junction. Compare the junction of tube A, with all the attendant difficulties it presented, with the

junction of the larger tube section B. Here, the direction of withdrawal of the casting from the mold permitted a blend of the tube section with the casting wall, thus avoiding the defects encountered with tube A. Influence of Large Areas. Again with reference to the casting shown in Fig. 22, the interrelated requirements of pressure-tightness and smoothness would have made production in a permanent mold difficult had the casting been made larger but with the wall thickness left unchanged. Thin walls encompassing large areas cannot be easily fed; this may give rise to centerline shrinkage or microporosity. To prevent premature freezing and assist the flow and feeding of metal to these thin sections, the mold coating in the thin-wall area must be heavier. However, as has been noted previously, heavier coatings cause rougher surfaces;

thus is demonstrated the basic incompatibility between thin sections and surface smoothness. Minimum Section Thickness. The casting shown in Fig. 23 taxed the capabilities of the permanent mold process. This casting, 3.703 in. in diameter by 30 in. long, with a uniform wall thickness of 0.148 in., was designed for a bazooka assembly. For portability, the unit had to be light in weight; the aluminum alloy specified (7% Mg, 0.1% Ti, 0.1% Mn, 0.05% Be) was sluggish and hot short. The various bosses adjoining the large thin-wall section caused cracking unless all processing conditions were optimum. Castings of this configuration and size are difficult to produce, and careful attention must be given to mold coatings. Fortunately, smoothness was not a requirement for this casting, although general good appearance was necessary. Producibility could have been improved by an increase in the wall section—specifically, by adding metal to the inside of the wall, from which it could be bored out later with negligible contribution to cost, as a machining operation was already required on the inside surface. Good Design. The aluminum casting shown in Fig. 24 illustrates good design for the permanent mold process. The 5.6 mm (7/32-in.) center web is heavy enough to allow complete filling of the mold and feeding of the center section. It also provides sufficient wall around the inserts to prevent shrinkage cracks. Changes in section thickness are gradual, radii are large, and areas for risering are adequate. If weight were a problem, the heavy sections could be made thinner and certain cored areas

Fig. 21

In this aluminum permanent mold casting, thin connecting ribs made it difficult to feed the heavy center section. Increasing the size of the rib opposite the riser would have improved feeding.

Fig. 22

Molding and machining requirements necessitated nonuniform walls at the junction of small tube A with the body of this casting. Shrinkage defects resulted at this junction. Large tube B and body of casting were blended smoothly. No defects occurred at this spot. Permanent mold casting of aluminum alloy 355

Design Problems Involving Thin Sections / 129 could be increased in size. However, making the web section thinner would be undesirable. Another example of good design is the front canopy bow casting shown in Fig. 25. This structural casting, produced in aluminum, was designed so that optimum location of gates was possible. Metal could flow to all parts of the permanent mold cavity without the need for excessive mold coatings to prevent premature freezing. By taking advantage of the maximum chilling effect of the mold, the necessary strength in the metal was developed. In addition, all surfaces were produced to a finish of 125 micro-in. or better.

Thin-Wall Investment Castings The investment casting process is capable of producing intricately shaped castings of almost

This aluminum casting is representative of the maximum capabilities of the permanent mold process for the wall thickness indicated.

Fig. 23

Fig. 24

any metal, and can assist the designer in minimizing casting weight, obtaining maximum strength-to-weight ratio, and producing small parts. In designing for minimum wall thickness, the fluidity of the alloy and its ability to flow in the mold are important considerations. Of almost equal importance is the solidification range of the alloy in relation to proper feeding of a section. The surface area of exposed molten metal and the feeding distance are also important in determining the castability of thin sections. Table 2 shows typical minimum-wall-thickness relations for a tube 40 mm (1½) in. long produced from various metals by investment casting. These recommendations reflect the experience of the industry as a whole. Although these minimum thicknesses may be subject to modification, the relative mold-filling capabilities of the different metals will remain as indicated in the table. In any casting in which heavy sections are separated by thin sections, it is likely that there will be problems of filling the cavity and of feeding the metal during solidification. Figure 26 illustrates how a simple revision in the design of an investment casting overcame these problems. Produced from 8630 alloy steel, this casting had thin sections connecting the two heavy end sections. This made it extremely difficult to fill the mold cavity completely; misruns and shrinkage occurred frequently. Because this casting was gated on the three end bosses, widening of two of the three bosses

This aluminum casting illustrates good design for the permanent mold process. There is adequate metal around the cast-in-place inserts, and the web is heavy enough to permit feeding of the center section.

shown in Fig. 26 placed a larger area of the boss directly over posts A. This created a direct path for metal flow to the opposite heavy section and eliminated cold shuts. Increasing the size of these bosses added enough additional heat to the mold to keep the thin sections liquid for a longer period. In turn, the gated areas could feed the previously isolated heavy end for a longer period, thus eliminating shrinkage defects. The third boss was increased in area for uniformity of appearance only. Chilling Effect of the Mold. Even though the molten metal is poured into a heated mold, investment casting presents the same possibilities for premature freezing of thin walls as other casting processes. A high ratio of area to thickness in a section encourages quick cooling and freezing of the molten metal. However, the distance the metal must travel to fill the section must also be considered. A heavier section would feed into a thin section 13mm (½ in.) long by 50 mm (2 in.) wide with considerably less difficulty than into a thin section 2 in. long by ½ in. wide, even though the surface areas (1 sq in.) and the thicknesses were the same. The top chart in Fig. 27 shows recommended and minimum section thicknesses in relation to the longest dimension of a casting; the lower chart shows minimum diameter of cylindrical cores as a function of maximum length of core, for both blind holes and through holes. Since many factors influence the maximum and minimum limits of any particular design detail, this information should be considered approximate, for use only as a guide. Mold and Metal Temperature. The pouring temperature of an alloy is selected partly on the basis of the thickness of the section to be poured. Metal is usually poured at the lowest possible temperature, to reduce the likelihood of casting defects from gas, dross, metal-andmold reaction, and other deleterious effects that depend on temperature and time at temperature. However, in pouring thin sections, it is often necessary to increase pouring temperature, in order to prolong the time interval between pouring and solidification. Usually, a cooler mold results in a better surface finish, because of the faster freezing of the metal skin and the elimination—or at least

Table 2 Typical minimum section relations for a tube 40mm (1½ in.) long produced from various metals by investment casting

Good casting design, permitting ideal location of gates, made for efficient production of this aluminum permanent mold casting. Although this casting had a wall thickness of 0.093 in., the recommended minimum for the material and process (Table 1), all surfaces were produced to a finish of at least 125 micro-in.

Fig. 25

Metal

mm

Min wall, in.

Carbon steel Series 300 stainless steel Series 400 stainless steel Aluminum alloy Magnesium alloy Aluminum bronze (10% Al) Berylllum copper Cobalt-chromium alloy

1.52 1.27 1.65

0.060 0.050 0.065 0.050 0.050 0.060 0.040 0.050

1.27

From “How to Design and Buy Investment Castings,” Investment Casting Institute, 1960, p 132

130 / Casting Design and Performance

Fig. 26

Providing clear channels for metal flow in the mold simplified production of this casting and permitted gating at one end only. Investment mold cast; 8630 steel

reduction—of chemical reactions that may take place at the interface of the metal and the investment. However, this same cooling effect reduces the chance of filling extremely thin sections. Consequently, it may be necessary to relax the surface finish requirements in order to obtain the thin wall desired. Mold permeability partially provides for the release of entrapped gases and plays an important part in the determination of wall thickness. Two principal types of molding material are used by manufacturers of investment castings: a plaster-base investment, for the lower-melting nonferrous alloys; and a silica-base investment, for steel and other alloys poured at higher temperatures. The silica-base investment in solid mold work is more permeable than the plaster-base investment. Because of this, plaster-base molds for nonferrous castings are often spun centrifugally, to impart a higher force to the molten metal entering the mold. This is done particularly with the light alloys, where the molten metal in the gating system provides insufficient

pressure to force the casting metal through thin sections. Prevention of Defects. In the investment process, as in other processes, misruns and cold shuts may be encountered in the production of castings with thin walls. In the steel investment casting illustrated in Fig. 28, cold shuts occurred in the thin section, which was farthest from the gate. The metal was required to flow around the core into the cavity, and down into this 0.13-in. section. When the metal reached this point, it was too cold for the two streams to unite satisfactorily. As a result, all castings produced were rejected. Eliminating the large cored hole, which restricted metal flow, solved this problem. Hot tearing and cold cracking also may be encountered in thin sections of investment castings. Hot tearing may occur when a section of the casting develops insufficient strength to resist the contractive forces exerted as the metal cools in the mold. Cold cracking and distortion may occur if a highly stressed area is too thin to withstand the forces that develop during

cooling. Uneven section thickness gives rise to stress from unequal cooling rates, which can cause distortion or cracking of the casting. Although few castings can be designed to have completely uniform sections, uniformity should be an objective of most designs, and abrupt changes in section thickness should be avoided. Cooperative Designs. Difficulties in attaining soundness in thin walls may be minimized through design. One simple solution is to redesign other sections of a casting so that minimum wall thicknesses are not called for. For example, in a casting that has several heavier sections but employs thin sections to reduce mass, it is sometimes possible to reduce the heavier sections and add to the thinner sections. This may result in a total weight even less than that of the original part. Designing for Economy. Uniform thin sections present problems that often increase the cost of the casting. Tapering thin sections as little as 2 or 3 per side will increase producibility. (The increasing thickness should, of course, be in the direction of the heavier section from which metal is being fed.) This taper will encourage directional solidification of the metal from the thinnest to the heaviest section. An alternate method is to provide flow paths in the form of ribs. If these ribs can remain, they contribute little to cost; if they must be removed, additional expense is incurred. As an example, producing the 4140 steel investment casting shown in Fig. 29 presented a problem. Even though the most favorable gating practice was used, it was impossible to fill the 0.09-in. leg sections, and misruns were common. Strictly in terms of producing a sound casting, two solutions were apparent. In the first, and more desirable, solution, the legs would be tapered by increasing their thickness progressively as they approached the junction with the body. Normally, such additional metal could be added to the inside, the outside, or both sides of each leg. In this casting, however, adding metal was not permissible. The mating components of the assembly of which this casting was to be a part were in production, and it was mandatory that the specified dimensions be observed. The second solution would be to add ribs that would act as channels to permit the molten metal to flow quickly to the extremities of the legs. These actually were added, and are shown in phantom in Fig. 29. It was practical to add these ribs only to the outside faces, because their removal to a close tolerance was necessary before shipment. This increased the unit cost of the casting. However, adding the ribs solved the problem of filling the legs, and many castings were produced successfully. Later, a similar casting, shown in Fig. 30, was designed with tapered legs such as were suggested in connection with the casting shown in Fig. 29. Both castings were produced in solid investment molds from 4140 steel, and both performed the same function in application. However, through improved design, the casting

Design Problems Involving Thin Sections / 131

Fig. 27

(a) Recommended and minimum section thicknesses with respect to the longest dimension of a casting. (b) Minimum diameter of cylindrical cores with respect to maximum length of core.

Fig. 28

In this steel investment casting, elimination of the round cored hole did away with cold shut defects in the 0.13-in. section, which were caused by the meeting of two metal streams that were too cold to unite

properly.

in Fig. 30 cost approximately 27% less than the other, and delivery time was shorter. Feeding Ribs. In the investment process, heavy sections separated by thin sections can be filled by adding gates and runners to channel metal directly to these heavy sections. However, this increases the cost, because it involves adding more gates and runners to the wax patterns and entails removing all traces of the gates from the castings. It was shown in Fig. 29 that a thin section can be filled more efficiently by adding ribs in strategic locations. Ribs work equally well in conducting metal across a thin section to feed a heavy section that would otherwise be isolated. Although the ribs may add a small amount to the cost of the pattern die, they reduce over-all production cost by eliminating

gates and runners that would otherwise be necessary. The addition of ribs permits the adjacent thin sections to be made even thinner, because the thin sections are no longer needed as primary feed paths. The ribs also add rigidity to the thin sections and can be considered as part of the working structure of the casting. However, their main advantage is cost reduction. For example, the inner gimbal investment casting shown in Fig. 31 was produced from 8620 steel with relatively heavy sections at both ends and in the center areas. These heavy sections were connected by a thin (0.093-in.) wall that incorporated a small rib. The first production castings revealed that the cross-sectional area of the thin-wall body was not large enough, even with the small ribs, to permit the

entire casting to be fed from one end or from the center hubs. As shown in Fig. 31, six gates were necessary to produce sound castings. In subsequent production, it was established that sound castings could be produced by increasing the size of the ribs and using only two gates. These gates were located at one end, each opposite a rib, as shown. Castings thus produced cost 28% less than castings produced to the original design. Alloy selection is governed not only by the mechanical properties desired, but also by the ability of the metal to fill a thin section. The two alloys capable of producing the thinnest sections are probably beryllium copper and silicon brass. These alloys have been poured successfully in wall thicknesses of 0.010 in. for a distance of 0.040 to 0.050 in. A thin-wall beryllium copper casting is illustrated in Fig. 2(b). Aluminum alloy 356 also is excellent for casting thin sections, although less so than beryllium copper or silicon brass. A ferrous alloy excellent for thin sections is SAE D2, a high-carbon high-chromium airhardening tool steel. This alloy is extremely fluid. The series 300 stainless steels and many of the superalloys for high-temperature applications have good castability and are capable of filling sections of minimum thickness under specific conditions. Among the series 400 stainless steels, type 440 is best, followed by 420 and 430. Type 410, often specified as AMS 5350, is only fair from the standpoint of castability, and will not produce minimum wall thicknesses sometimes desired. Despite its limitations for thin walls, type 410 is often used for aircraft castings because of its corrosion resistance and strength. The low-alloy and carbon steels have only fair ability to produce thin sections, and their fluidity and castability decrease with decreasing carbon content. As carbon content is increased to about 0.50%, there is a corresponding increase in fluidity and ability to fill thin sections, but above 0.50% C there is a tendency for gassing, which increases with the carbon content. Thin-wall castings have been produced successfully using “RH” Monel, particularly when their design is compatible with optimum production conditions. Such a casting is shown in Fig. 32. Although this casting has a wall only 0.035 in. thick, its design permits ideal gating at its center, and several foundries have produced it successfully. Unlike thin sections in permanent mold castings, thin sections in investment castings will usually have the best surface attainable. Because the cooling effect of the mold at the thin sections is usually greatest, the metal in contact with the mold freezes almost instantly. This restricts the amount of time for a possible chemical reaction between the mold and the metal, thus limiting surface imperfections. Rapid freezing reduces the depth that the molten metal can penetrate into any cracks or other defects in the mold precoat. However,

132 / Casting Design and Performance

To encourage metal flow in the thin sections of this 4140 steel investment casting, the legs were tapered as shown. Compare with similar casting shown in Fig. 29

Fig. 30

By increasing the size of the casting ribs, four gates were eliminated from this investment casting (8620 steel), at a cost saving of 28%.

Fig. 32

The ideal gating area at the center of this casting permitted walls thinner than usually recommended in “RH” Monel investment castings.

Fig. 31

To encourage metal flow in the 0.09-in. legs of this 4140 steel investment casting, ribs were added as shown. These were removed before shipment.

Fig. 29

because a more fluid metal will penetrate mold defects more than a sluggish metal under the same conditions, the choice of a fluid alloy to fill thin sections may result in a rough finish on the casting, particularly in heavy sections that freeze slowly.

Casting Design and Performance Pages 133–138

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Uniform Sections IN MANY CASTINGS, functional requirements dictate that walls be uniform or nearly uniform in thickness. Many problems in producing castings having walls of uniform thickness are associated with the premature freezing of molten metal before all parts of the mold cavity have been filled. In other castings, it may be possible to fill the mold cavity completely, but centerline shrinkage or areas of porosity may be encountered because portions of the casting cannot be fed before the onset of freezing. In such castings, certain walls may be tapered by the addition of padding, which may be incorporated into the design, or ribs or webs may be added to provide feed paths. The function of ribs or of sections of walls that have been enlarged in providing flow and feed paths for the molten metal as it fills the mold and cools is shown schematically in Fig. 1. In pouring a flat plate, Fig. 1(a), the metal enters the cavity and attempts to flow in all directions. If the distance from the gate to the extremities of the mold cavity is too great, the metal freezes prematurely, and misruns result. When ribs are added to the plate, Fig. 1(b), the metal flows readily to the extremities of the mold cavity, and successful castings are produced. The same success is attained when the plate is enlarged at the gate and uniformly tapered to the extremities, as shown in Fig. 1(c). The metal flows into the mold cavity at the heaviest section, thus preventing an initial chilling of the metal by the mold, such as occurs with a small

Fig. 1

volume of metal. Because the molten metal in heavy sections retains more of its heat as it flows through the mold, it can more readily reach the extremities of the mold cavity. A related problem is involved in the mechanics of cooling and freezing of molten metal and the attendant volumetric shrinkage. Referring to Fig. 1(a), let us assume that the metal is poured and fills the mold. For some castings where chilling of metal by the mold is possible, there is no assurance that freezing will start at points farthest from the gate and progress in an orderly manner to the gate so that the volume of metal lost because of shrinkage can be replaced from the reservoir (riser) intended for that purpose. In fact, a casting such as this may be expected to contain sections that are isolated from the feed source by metal that is closer to the gate. The metal nearer the gate freezes first and thus surrounds and isolates these molten pockets. Shrinkage of the molten metal in these isolated pockets creates voids or porosity. With two risers feeding metal from opposite ends of the casting, more sound metal would be obtained in the plate, and with a riser feeding from each edge of the plate, a still greater percentage of sound metal could be obtained. However, risers add to the cost of producing a casting, and a method that minimizes the number of risers is more economical. This can sometimes be accomplished by providing feed paths within the casting configuration, as shown in Fig. 1(b). These paths permit the molten

metal to flow easily to all sections of the cavity. If the enlarged sections are gradually tapered from the gate to the extremities, solidification follows an orderly progression through the casting and termigates in the riser. Compensation is provided for normal shrinkage throughout the casting, and sound metal is assured. The same beneficial freezing pattern is obtained by tapering the plate, Fig. 1(c), provided the angle of taper is sufficient and the source of feed metal is at the heaviest section. Under these conditions, a favorable freezing pattern is assured because the metal that has flowed the farthest (and is therefore the coldest) is deposited in the thin section. In contrast to a heavy section, the thin section of molten metal has less heat to transmit to the mold before freezing starts. Thus, the direction of solidification can be predicted in tapered sections, provided the metal enters at the heavy end. If the need for feed paths is anticipated in the early stages of the design of a casting, padding may be made a functional part of the casting, or it may be located so that its removal adds little or no cost. Consider the two investment castings shown in Fig. 29 and 30 of Chapter 9 (“Design Problems Involving Thin Sections”). For successful production of the casting shown in Fig. 29, ribs were required on each of the legs. These provided feed paths that encouraged complete filling of the cavity and resulted in

Flat plates (a) are difficult to produce completely filled and with sound metal. By adding ribs (b), or tapering the plate (c), the problems are eliminated or substantially ameliorated.

134 / Casting Design and Performance sound castings. The casting shown in Fig. 30 of Chapter 9 was designed with each leg tapered. This permitted easy filling and feeding of the legs and induced a satisfactory freezing pattern. Although the addition of ribs to the first casting was satisfactory from a foundry point of view, removal of the ribs was necessary for the proper functioning of the part. In the second casting, the taper was satisfactory in terms of both production and function, because the casting was designed with this in mind. In the magnesium casting shown in Fig. 19 of Chapter 9, the problem was to distribute metal efficiently to all flow channels and all parts of the mold before freezing, and to accomplish this objective with a minimum amount of metal. (This was an aircraft application, and no excess weight was permitted.) The solution entailed greatly enlarging the two ends into which the metal was gated, thus allowing them to function as manifolds. Increasing their size kept the metal hotter and permitted unrestricted flow to each connecting member of the casting. As shown, ribs were added to each thin strut section, to provide a larger flow channel. Because of the larger mass, the rib and strut junctions developed hot spots in the mold and helped to prevent premature freezing. The heavy manifold ends were later machined off, but the small ribs remained intact.

Sand Castings Produced in a dry sand mold, the steel elbow casting shown in Fig. 2 was originally designed with a uniform wall 1½ in. thick. This casting, intended for use at high temperature and high pressure, was produced with a rejection rate of 16%, because of shrinkage porosity caused by insufficient feeding of molten metal to all parts of the casting during solidification. ASTM class 1 radiographic soundness was specified for all areas except the unflanged end, where no defects whatever were permitted because of a subsequent welding operation for connecting the elbow to an adjacent member of the assembly. To overcome this feeding problem, the casting was revised by adding ½ in. to the wall thickness at the flanged end and tapering the wall uniformly to its original thickness at the opposite end. This change lowered the rejection rate from 16% to 6%. The gate valve casting of Fig. 3 suggests another method for improving the flow of metal and feeding all sections adequately. These stainless steel castings were produced in sand molds and were intended for use in nuclear reactors, necessitating maximum attainable soundness. All castings were inspected radiographically, and any discernible defect was cause for rejection. Although this casting was redesigned primarily to reduce the cost of producing it, addition

As originally designed, with a constant wall thickness, this sand cast steel elbow fitting was producible only at a 16% rejection rate, because of shrinkage porosity due to insufficient feeding. Redesign, in which walls were increased at the flanged end and tapered, permitted more successful production; the rejection rate dropped to 6%. Dimensions in inches

Fig. 2

Fig. 4

of the 75 mm (3/4-in.) webs also reduced the distance the metal had to travel as it cooled, and minimized shrinkage porosity caused by impeded feeding. The webs functioned efficiently as manifolds to distribute molten metal to all parts of the mold cavity. Because they abutted the tubular sections, the webs reduced the distance the molten metal had to flow to reach the central portion of the casting. The webs also conducted metal to the flanges, which were used as flow paths to feed those parts of the casting nearest the flanges. These heavier sections then fed molten metal to the thinner sections. Although the webs were approximately the same thickness as the flanges, the webs were the last parts of the casting to freeze. The greater volume of metal that passed through the web-forming section of the mold heated it to a higher temperature than was reached in the flange-forming sections. A related casting is shown in Fig. 4. These sand cast stainless steel fittings for a nuclear reactor also were revised to increase efficiency of production. The addition of the web reduced by one the number of risers required to obtain

The application of this sand cast stainless steel valve body required completely sound metal. The uniform walls were fed readily from the risers through the flanges and the webs. Webs were added to the original design.

Fig. 3

Adding the webs and tapering the tapping beads in the elbow fitting (a) and the tee fitting (b) assisted in distributing the molten metal and created a favorable freezing pattern that permitted adequate feeding of all sections of these stainless steel sand castings during solidification.

Design Problems Involving Uniform Sections / 135 sound metal. However, in these castings there were no flanges to assist the webs in the distribution of metal. Instead, internal padding of the thread beads provided the necessary flow and feed paths. As shown, the padding was tapered, with the heaviest part adjacent to the web and the thinnest part diametrically opposite—that is, at the point farthest from the webs. The elbow castings, Fig. 2 and 4(a), and tee castings, Fig. 3 and 4(b), benefited from a more orderly solidification of molten metal as a result of redesign. Directional solidification was assured by the uniform temperature gradients in the molds, created by both the size of the casting sections and the flow path of the metal. Freezing of the metal started at the point farthest from the risers and progressed uniformly to the risers. All sections of the castings were adequately fed, and sound metal was obtained. The padding on the elbow and tee castings in Fig. 4(a) and (b) was machined away when the thread beads were bored to the desired diameter before being tapped. The AZ91 magnesium casting of Fig. 5 illustrates a straightforward approach to padding. This casting was produced in dry sand molds. The minimum castable section thickness, arbitrarily established at 0.130 in., was heavy enough to provide the necessary strength. However, all castings produced were defective because of cold shuts and shrinkage. The draft angle on the pattern was then increased from one degree to three degrees, and 0.125 in. was added to all wall thicknesses. With these revisions, all castings produced were sound. Although such a freehanded use of padding is not generally recommended, especially for aircraft castings, circumstances may occasionally warrant it. (The unwanted weight that was added to the casting shown in Fig. 5 was justifiable only because delivery of these castings was of utmost importance.) Problems presented by walls of uniform thickness are not always attributable to the distance the metal must flow or be fed, but sometimes result from the way in which the walls are shaped. Figure 6(a), a section of a large aluminum sand casting for aircraft use, illustrates a tortuous configuration that caused turbulence as the metal flowed through the mold, resulting in the entrapment of oxides. Micro-shrinkage also occurred, and the rejection rate of this expensive casting was excessively high. By relieving the sharpness of the angles of the casting walls, as shown in Fig. 6(b), free flow of metal was obtained. This eliminated both oxide inclusions and microporosity and resulted in the production of sound, acceptable castings.

Shell Mold Castings As illustrated by previous examples, casting problems presented by uniform wall thicknesses are not always attributable to the distance metal must flow. Shape, too, exerts a

strong influence. The casting shown in Fig. 7, a hollow spherical casting with a heavy flange, is an apparent exception to rules relating to feeding distance. Because the distance from the edge of the heavy flange to the bottom of the sphere is 4 in., it might be assumed that the casting could be fed satisfactorily from the

heavy flange. However, as originally designed (Fig. 7a), it could not be. The total area of the thin section greatly exceeded that of the flange section from which it was to be fed, and feeding was inadequate; the part was unsound below points A, or about two thirds of the way down the wall of the sphere.

In this sand casting as originally designed, excessively thin walls resulted in cold shuts and shrinkage. Overcorrecting by nearly doubling all wall thicknesses and by increasing the draft angle eliminated defects but added unnecessarily to the weight. Alloy was AZ91 magnesium

Fig. 5

A partial section of a large, intricately shaped aluminum casting in whose original design (a) sharp bends retarded metal flow and created turbulence. The resulting inclusions and porosity caused a high rejection rate. Revised design (b), in which bends were eliminated or were made less acute, was producible without defects.

Fig. 6

In this flanged spherical casting as originally designed (a), feeding was restricted a short distance from the riser, resulting in unsoundness below points A. Tapering the walls, as shown in (b), provided adequate feeding. This alloy steel casting was produced in a shell mold.

Fig. 7

136 / Casting Design and Performance

Walls of uniform thickness did not deter successful production of this casting, because it was designed so as not to exceed the capabilities of the metal (magnesium alloy AZ63A) or the process (shell molding) that was used in producing it. Rejection rate was less than 5%

Fig. 8

Fig. 9

To assure soundness, this wall had to be tapered from the top to a point not more than 1½ in. from the bottom of the sphere, as indicated in Fig. 7(b). Such a taper should be no less than 1½ ; a taper of 3 would virtually guarantee soundness. When a large sphere is cast it is usually necessary to provide risers at various points around the sphere in order to feed from lower wall sections as well as from the top. Not all castings require padding or the addition of ribs to obtain sound metal. When the normal capabilities of metal and process are not exceeded, uniform walls are practical and do not defy successful production. A good example of a successfully produced casting with walls of uniform thickness is shown in Fig. 8. This AZ63A magnesium part, 3 in. long,

In this permanent mold casting, functioning of heavier ribs with thin walls induced hot tears and shrinkage at the junctions. Uniformity of wall thickness would have eliminated these defects.

2 in. wide, and 1 in. high, with 0.080-in.-thick walls, was produced in shell molds with a rejection rate of less than 5% for all causes.

Permanent Mold Castings Castings produced in metal molds have an advantage, in one respect, over those produced in sand or ceramic molds. Water lines can be incorporated in the mold to create a planned chilling effect on specific sections of the casting. By adjusting the rate of water flow, the degree of chilling can be controlled accurately. This permits some leeway in the variation of wall thicknesses that can be incorporated in a single casting. This possibility of chilling the casting should not be used as a substitute for good design. Regardless of the casting process, a well designed casting can be produced more easily and more economically than one of unfavorable design, even though the unfavorable design does not tax the capabilities of the process. The permanent mold casting shown in Fig. 9 was structural in use and required radiographic quality. The inner wall was designed to have minimum thickness, to save weight. Because of the taper required for molding, ribs were heavier at the junction with the inner wall than was functionally necessary. Eliminating shrinkage areas and cracks at these junctions was a problem. Two of the ribs at the parting line were still heavier, because of their being used for the attachment of gates and risers. These ribs were insulated with extra mold coating, because smoothness of the as-cast surface was not mandatory. By water-cooling the inner steel core to obtain some chilling of the metal, the casting was made to good quality. The design of this casting could have been improved by increasing the thickness of the inner wall to equal that of the heavier adjoining ribs. This uniformity of wall thickness would have eliminated the shrinkage defects. Size can determine whether a casting design incorporating uniform walls is practical or impractical. The effect of large areas of uniform wall thickness on producibility can be illustrated with reference to the casting shown in Fig. 10. This casting was relatively easy to produce to the dimensions indicated. However, if it had been several times larger, with no change in wall thickness, progressive solidification would have had to be induced by varying the thickness of the mold coating, even if a wall thickness of 4.06 mm (0.160 in.) were otherwise adequate. Such practice is difficult, time-consuming and expensive. An alternate method would be to incorporate a taper in the walls to help produce the necessary solidification pattern.

Investment Castings This casting was easy to produce in a permanent mold to the dimensions indicated. However, if all dimensions other than wall thickness were increased proportionally, this casting would be difficult to produce. The enlarged casting would require tapered walls or controlled variation in the thickness of the mold coating, in order to permit efficient production.

Fig. 10

Uniform walls in an investment casting are readily produced if the casting is not too large, in relationship to wall thickness, and if gates

Design Problems Involving Uniform Sections / 137

This 356 aluminum alloy casting, produced in a centrifuged plaster-base investment mold, represents minimum wall thickness for the shape, alloy and process.

Fig. 12

Fig. 11

A well-designed aluminum investment casting, incorporating uniform walls and - efficient gating areas

A critical application necessitated class 1A quality for this type 347 stainless steel investment casting. Hot tearing at the junctions and porosity in the body, which resulted from too-rapid solidification within the uniform wall of the tubular section, were eliminated by tapering the body toward each flange and increasing the fillet radii at the junctions.

Fig. 13

Fig. 14

can be placed at desirable locations. The body of the aluminum investment casting illustrated in Fig. 11 incorporates a uniform wall thickness of 0.09 in. throughout, even at the point where the wall changes direction. The irregularly shaped flange also is maintained at this same thickness. The rectangular flange (3.8 mm, OR 0.15 in. thick) is just enough heavier to permit efficient gating and transfer of metal to all sections. This part was easy to produce, and virtually no defects were encountered. The 356 aluminum alloy part illustrated in Fig. 12 was produced in a centrifuged mold and heat treated to the T6 condition. This casting is part of the leading-edge assembly of an aileron; all walls are 0.090 in. thick. The foundry considered the 0.090-in. wall to be of minimum practical thickness for this configuration. Decreasing thickness below 0.090 in. would result in rejected castings, and the rejection rate would increase with decreasing wall thickness. By centrifuging the mold, pressure is exerted on the molten metal, forcing it rapidly into all parts of the mold cavity. Centrifugal action is often used to produce castings with uniform or thin walls that cannot be produced with static pressure alone. Lack of favorable directionality of freezing was a problem with the type 347 stainless steel investment casting illustrated in Fig. 13. This inlet by-pass valve adapter was intended to carry fuel close to a hot part of an airplane engine. Failure of the casting could cause a fire or explosion, with disastrous loss of the plane

The flat surfaces of this investment casting as originally designed (a) created problems of thickness tolerance and surface defects from cracking of the mold precoat. A redesign (b) eliminated the flat surfaces and the attendant defects.

138 / Casting Design and Performance and its occupants. Consequently, this part had to be produced to class 1A quality. As originally designed, the casting did not lend itself to efficient production. The problem was not in filling the mold but in controlling solidification. With a uniform thickness of 0.100 in., the body of the casting solidified too fast for the fillet sections at the junctions of the body and the flanges to develop enough strength to withstand the contractive forces of the solidified body. There were hot tears at the junctions and porosity in the 0.100-in. walls. To correct these defects, the casting wall was maintained at its original thickness (0.100 in.) at a point midway between the two flanges and was uniformly tapered to a thickness of 0.150 in. at each flange. Isolated sections

were thus eliminated. To overcome hot tears, the fillet radii at the junctions were increased from 0.10 to 0.25 in. With a slight modification in processing, the redesign permitted acceptable castings to be produced, and rejections were few. Although uniform walls are desirable in many castings, in one facet of production of investment castings, uniform walls create problems. The 4130 steel casting shown in Fig. 14(a) illustrates two such problems. The first involved cavitation, or contraction of heavy sections of the wax pattern. As a result of this pattern defect, the center of the casting was depressed below the permissible thickness variation. The second problem, although it appeared to be one of processing related to

precoating of the mold, was basically one of design. On smooth, flat surfaces, a mold precoat is likely to crack when the molten metal is poured. If it cracks, but does not peel or wash, the result is a blemish on the surface of the casting where the molten metal fills the crack. Peeling or washing of the precoat causes larger areas of surface imperfections. The solution to these problems lies in designing to avoid flat surfaces. Figure 14(b) shows how this was done successfully for this particular casting. The ¼-in.-diam riblike raised section along the center of the part acted as a flow channel to assure complete filling of the heavier section of the casting.

Casting Design and Performance Pages 139–146

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Unequal Sections DEPENDING on casting conditions, certain masses in casting sections may be too large and too remote from a riser to be fed through the intervening sections and made sound. This chapter deals with the design of uneven, enlarged and “isolated” sections for soundness, strength and economy. During the slow solidification of a casting, a thin skin of frozen metal forms like a shell around the outer part of the mold cavity soon after the liquid metal is poured. This rigid shell becomes, in effect, a mold for the remainder of the casting, and the volume lost by shrinkage of the metal as it solidifies progressively must be replaced from some feeding source, to avoid shrinkage porosity. The function of the feeding source is to supply liquid metal to the mass continuously until it is frozen solid without porosity. In this process, the freezing front inside the mass must not be cut off from its feeding source before the mass is solid. Sound metal can be produced only if proper directional solidification is attained by design and foundry engineering. When it is discovered in the foundry that a casting as designed is practicably or economically unproducible because of unequal or isolated sections, the designer has two recourses available: 1. Thicken (pad) the thin section that serves as a feed path to the heavier remote section 2. Reduce the mass of the remote thick section This is not to imply that risering of isolated sections is not the most practical solution for some castings. However, risering is a function of foundry engineering, rather than of design engineering. Where the casting shape is virtually dictated by specific requirements of weight, space or function, risering the isolated areas may be the only acceptable method of producing the required casting successfully. As an example, Fig. 1, an aluminum sand casting illustrates riser requirements where heavy sections are isolated from a common feed source and must be risered independently. This casting was designed with close tolerances, for aircraft use. Requirements for sound metal and minimum weight necessitated the use of individual risers for all heavy sections, rather than the addition of padding to the ribs or walls, to permit feeding two or more heavy areas from one

riser. Because risers add to the cost of a casting in several ways (time is consumed in providing them, they reduce the yield of metal poured, and time is required for their removal), every effort should be made to design so that risers are not required, or, if this is impossible, to restrict their use to a minimum. To produce the casting shown in Fig. 1, eighteen risers were needed, to assure the desired high level of quality. In this instance, they were necessary and justifiable. However, in many other castings referred to in this chapter the need for risers could be greatly reduced. The rest of this chapter will deal with problems of designing isolated heavy sections that are functionally essential and with the two most efficient solutions to these problems over which the designer has control: (a) providing flow and feed paths, and (b) reducing the mass of the isolated sections.

the section through which the metal is being fed), which is considerably less than the maximum distance that can be fed and still provide sound metal. Because of the difficulty of risering the isolated boss, the casting would not have been producible unless other components of the design permitted use of the required padding.

Designs Requiring Padding or Other Feed Paths Many castings benefit from an increase in section size that will create a flow path to fill completely a section that is difficult, or otherwise impossible, to fill, or that will provide an effective feed path from a riser to an isolated heavy section. The functional requirements of a casting may demand the use of padding because of the impracticability of risering an isolated section. The sand cast steel gear housing shown schematically in Fig. 2 contains an internal web that supports a bearing boss totally inaccessible to direct risering. A preliminary casting, made as a test for unsoundness, showed the anticipated defect: shrinkage porosity in the isolated boss. Castings were then produced with chills incorporated in the mold. Although the chills modified the shrinkage defects, they did not eliminate them. It was necessary, therefore, to add padding, as shown, to obtain adequate feeding of the boss. This change completely eliminated shrinkage defects, without the use of chills. In this casting, it was necessary to feed a distance of about 3½t (3½ times the thickness of

An aluminum sand casting (alloy 356) that required 18 risers, as shown, to achieve sound metal in the heavy sections. The necessity for minimum weight precluded the use of padding to reduce the number of risers.

Fig. 1

A boss isolated from feed metal in this steel sand casting could not be produced with sound metal, even though chills were used. By padding the web, as shown, chills were eliminated, and sound metal was obtained in the boss.

Fig. 2

140 / Casting Design and Performance A casting that presented a similar problem, and for which the same solution was used for reasons of economy, is illustrated in Fig. 3. This wheel casting, produced in a sand mold, incorporates a central hub and a side boss. Both required feeding to obtain soundness. It was possible to riser the central hub, as shown in Fig. 3(b). However, an analysis of cost indicated substantial savings could be realized by padding and feeding the central hub from the side riser, which would still produce the required soundness in the central hub, Fig. 3(c). Using two risers, the metal yield (the amount of metal in the “cleaned” casting, compared with the amount poured into the mold) was

A malleable iron sand cast wheel (a) that could be produced with two risers (b) at a yield of 30% of the metal poured, or with one riser and padding (c) at a yield of 45% All dimensions in inches.

Fig. 3

approximately 30 percent. Using only one riser and adding the required padding increased the yield to approximately 45 percent. The central riser required a core to form the ingate area and prevent mold wash. Producing this core and setting it in the mold entailed additional cost. The lower yield of the two-riser system, plus the cost of the core, made it distinctly advantageous to add the padding and feed the central hub from the riser that fed the boss at the outer flange. This part was cast from malleable iron, and probably less peripheral risering was required than would have been needed for an identical steel casting. This requirement, however, would depend on the level of quality desired. With either metal, padding would be more economical than risering. The feeding distance to the central hub of this casting was approximately 4½t. By rule of thumb, the total thickness of the pad and the casting (at the location of padding) should be not less than one-fifth of the metal feeding distance. Or, stated in another way, metal can, as a rule, be properly fed for a distance of five times its thickness. This is not a true maximum; castings of optimum design have been fed to distances of 10t or 15t without sacrificing soundness. The 5t rule, however, is a safe generalization and a good point at which to start. If the function of the part involves rotation, it is often possible to add material diametrically opposite the padded section, to counterbalance the pad. Because this increases the strength of the casting, it may permit a reduction in thickness of the web section to the point where the final weight of the casting would remain essentially unchanged. In such a design, the padding would become a functional part of the casting. An efficiently designed casting may utilize necessary component sections as flow and feed paths for molten metal. This is exemplified by the sand cast malleable iron gear housing shown in Fig. 4. In this casting as originally designed, Fig. 4(a), three ribs provided

necessary rigidity and support for the two bearing bosses (one on each side of the casting). With the part molded as indicated by the parting line, the boss in the cope side of the mold was easily risered and presented no problems. The boss in the drag half of the mold, however, could be fed only by appending risers to the flange opposite the ribs and feeding metal to the boss through the ribs. Because the ribs were too small in cross section, the boss could not be fed adequately, with the result that the metal in the boss was porous, and, in actual operation of the part, oil from inside the housing seeped through these pores. A revision, Fig. 4(b), removed the center rib and distributed its metal equally between the two side ribs. This increased the rib width from 0.50 to 0.75 in. Where previously the metal in the boss was fed a distance of 7t, it now was fed only 4½t. This improvement was made without adding any weight to the casting. The two heavier ribs provided adequate strength (equal to that of the three lighter ribs), and the defects were eliminated because these ribs were now heavy enough to permit adequate feeding during solidification of the bosses. When padding is required in order to produce sound, acceptable castings, but must subsequently be removed, every effort should be made to add the padding where its removal will not be difficult. Tubular-shaped castings, in particular, lend themselves to this treatment. Consider, for example, the ball-joint casting of Fig. 5. Produced from stainless steel in a sand mold, this casting was required to meet stringent quality standards. Each casting was inspected radiographically, and all discernible defects were cause for rejection. As originally designed, all castings were rejected because of shrinkage porosity caused by an unfavorable freezing pattern. An analysis of the casting revealed that simple tapering of the walls in two areas (see section A-A and B-B of Fig. 5) would permit placing risers so that all sections would taper

As illustrated, casting members may be used as feed paths for motten metal. The ribs in the original design (a) were too thin to permit adequate feeding of the bearing boss, causing it to be porous. By eliminating the center rib and distributing that metal to the two outer ribs (b), the boss was adequately fed, and the porosity encountered with the original design was eliminated.

Fig. 4

Design Problems Involving Unequal Sections / 141

Fig. 5

Fig. 6

Padding, properly placed, permitted production of this stainless steel sand casting to required soundness

Unusual padding of the center cavity of this stainless steel shell mold casting simplified production. Without the padding, complicated risering would have been necessary and would have caused high residual stresses.

toward a riser and all riser contacts could be removed easily. The metal added in the form of padding also was removed easily, during the machining operations already required in order to produce the desired surface on the ball section of the casting and on the mounting face of the flange. The addition of the tapered padding resulted in a reduction in rejection rate from 100% (porosity) to only 3% (all causes). A unique approach to padding a casting in order to feed isolated heavy sections more efficiently is shown in Fig. 6. Designed for production in a shell mold, this 17-4 PH stainless steel valve body has five heavy sections that appear to require risering for soundness. However, this amount of risering would add substantially to the cost of the casting. Also, the resulting

complex freezing pattern undoubtedly would create high internal stresses, which, in turn, would cause distortion before and during machining. Because the machined surfaces of the casting required high dimensional accuracy, distortion might have been a serious problem. To determine the amount of padding required, several castings were produced with a standard, straight-through core that allowed 1/8 in. of additional metal on the inside diameter of the casting for machining stock. The castings were rigged with a large riser at the top and a side riser (also serving as the sprue), as shown in the lower view of Fig. 6. From the experience gained, the foundry estimated that castings produced in this manner would average about 50 percent scrap because

of shrinkage defects. In contrast, it was estimated that producing the part with heavy padding, as shown, would reduce the rejection rate to less than five percent. Machining costs increased due to the extra metal to be removed, but this cost increase was less than the costs with a 50 percent rejection rate. The heavy padding was adopted for production. This casting and its gating and risering system demonstrate several desirable casting features. When metal is poured into the mold at the sprue, the pool at the bottom of the sprue serves to quiet the action of the metal, and turbulence is minimized as metal enters the mold cavity. Although there will be some cascading as the metal enters the body-forming section of the mold cavity and drops to the bottom flange section, entrapped gases, if present, will rise as the metal fills the mold and move to the top center riser. All dross and foreign matter can also be floated into this riser, leaving only sound metal in the casting section. One exception is the casting flange immediately adjacent to the ingate. At this location, the mold must be vented to permit entrapped gas to escape easily. All sections of the casting taper toward the top riser; this assures directional solidification and enables the sections to be fed from the top riser. An exception is again the heavy section immediately adjacent to the gate; however, because the gate is large, this sprue will function as a riser and will feed molten metal as it is required during freezing. As was expected, shrinkage porosity and impurities were found only in the risers and the metal to be removed by machining. After machining, castings were sound and completely satisfactory. Heavy sections that might have been isolated from a feed source with regular coring procedures, were fed easily because of the extreme wall thickness. Wall-thickening is practical from a cost stand-point when the added metal can be readily removed by boring, turning or other simple machining.

142 / Casting Design and Performance The lifting-eye casting of Fig. 7(a) appears to have a simple configuration; however, it presented several production problems. Because a cored cavity was located in the large flat face, it was necessary to select a parting plane that was incompatible with the most efficient gating and risering. With the riser placed as shown, and with padding added to encourage complete feeding of the casting during cooling and freezing, premature freezing was encountered in the area adjacent to the cored cavity. Despite the placement of exothermic sleeves around the riser, the use of metal inserts rammed-up in the core and heated just before they were inserted in the mold, and the starting of pouring almost immediately after the mold was closed, the rejection rate was 25 percent of the castings poured, and the remainder of the castings required extensive welding to make them usable. Rejections were caused mainly by internal and external shrinkage defects. Elapsed time between receipt of the order and delivery of acceptable castings was six months. A request by the foundry to eliminate the cored hole was granted. The patterns were revised to make the parting plane identical with the mounting face, as shown in Fig. 7(b). The riser was moved to the center of the mounting face. With these changes, there was a gradual transition in casting section area: the section farthest from the riser was the smallest, and the area immediately under the riser was the largest. By using an exothermic sleeve around the riser, the metal was kept molten and properly fed the casting during the entire cycle. With these revisions, the rejection rate of these castings dropped to only two percent, and no repair welding was required. This permitted a 32 percent reduction in the selling price of the castings. The redesign enabled the foundry to deliver acceptable castings 10 days after receipt of the order. Isolated heavy sections in aluminum castings may require feed pads in order to produce sound metal. Because of the relatively low melting point of aluminum, the quick freezing of heavy sections is often accomplished by means of metal chills. However, chills are sometimes impractical, and they always increase cost. Feed pads are often more practical. The end use of the aluminum sand casting illustrated in Fig. 8 required that four bosses be incorporated in its design. Castings were rejected because of porosity in these bosses, where sound metal was essential. Chills, conforming in shape to the bottom half of the bosses, were incorporated in the core that formed the inner configuration of the casting, but only part of the porosity was eliminated by this revision. Although increased weight was objectionable, four ribs were added, one over each boss, as shown in Fig. 8(b), to allow metal to be fed to the bosses from the risers on the top flange. Chills were used, to minimize the size of feed pads. Although risers could have been added to the outside of the vertical wall immediately

Fig. 7

Although apparently of simple shape, this lifting eye, a stainless steel sand casting, could not be produced efficiently to the original design (a). Redesign (b) eliminated a cored hole and facilitated foundry production.

Fig. 8

Four bosses isolated from all practical sources of feed metal in this aluminum alloy sand casting (a) were produced to the required soundness by adding minimum ribs (a weight consideration) and conforming

chills (b).

opposite each boss, this would have introduced problems of patternmaking and molding and could have required a special cleaning or trimming operation on the rough castings. Cost was the determining factor in the decision to

pad and chill, as shown, rather than to add four outside risers. A similar problem, involving a wall too thin to feed adjacent heavier sections, is illustrated by the aluminum sand casting in Fig. 9.

Design Problems Involving Unequal Sections / 143 Because the part was to serve as a structural member, each casting was inspected to stringent radiographic and fluorescent-penetrant standards. As originally designed, with a 3/32-in. wall thickness, all castings were rejected because of shrinkage defects in this wall. In an attempt at successful production of these castings, a circular runner was incorporated, and gates were added at several points along the periphery of the upper flange. Four risers were placed as indicated, on the top flange over the wall, to provide a feed source. The molten metal entered the mold cavity at the gates, traveled through the flange, through the vertical wall sections, and into the lower flange. The large amount of metal that passed through the mold section forming the 3/32in. wall contributed considerable heating effect to that mold section and allowed the 3/32-in. wall to remain open just long enough to feed the larger sections below. However, because of its thin cross section, the 3/32-in. wall froze almost

Shrinkage in this aluminum sand casting (alloy 355) where the wall was too thin was eliminated by increasing the wall to 5/32 in.

Fig. 9

Fig. 11

simultaneously at all locations, and the metal lost to the lower sections could not, in effect, be replaced; the wall exhibited severe microshrinkage. Adding side risers to the wall was impractical because they would complicate the coring required by undercuts formed by the flanges. Also, because of the thinness of wall, distortion would have been difficult to avoid when the side riser contacts were subsequently machined. Bottom gating was tried. The result was gross

Fig. 10

Padding this aluminum permanent mold casting facilitated efficient foundry production.

porosity in the 3/8-in. wall, as well as microshrinkage in the 3/32-in. wall. Thus, the only practical solution was to increase the thickness of the 3/32-in. wall to 5/32 in. This increase completely eliminated microporosity. The bottom flange was chilled, to induce solidification before the metal in the wall section froze. Although this was an aircraft casting, increasing the wall thickness and thereby adding unwanted weight was mandatory in order to produce acceptable castings. Permanent mold castings require flow and feed paths similar to those required by sand castings. The example shown in Fig. 10 was produced in a simple two-piece mold with the parting line as indicated. A three-piece mold would have been required if the hub had been risered directly. This might have required elimination of the 0.65-in.-diameter cored hole, which, in turn, would have had to be drilled in a subsequent operation, at extra cost. These castings were originally produced without the feed pad, and all were rejected for shrinkage porosity at the top of the hub. This condition was attributed to localized heating of the mold by the molten metal entering at the riser-sprue and flowing directly downward as the mold was filled. A feed pad 1/16 in thick by ½ in. wide was added. The heating of this section of the mold by molten metal, and the enlarged size of the feed pad, combined to keep the metal in the pad molten until the hub had solidified. Freezing then progressed through the pad to the riser. Sound castings were produced, capable of passing leakage and radiographic inspections. Another example of a permanent mold casting that benefited from added feed paths is the aluminum piston shown in Fig. 11. As originally designed, Fig. 11(a), the walls of the piston were of uniform thickness, which encouraged centerline porosity in the skirt section and some porosity in the wristpin bosses. Porosity was attributed to the simultaneous freezing of all sections of the skirt and the freezing of the wall section above and around the wrist-pin bosses before the freezing of the bosses. Consequently, feeding of metal to the skirt sections and bosses was inadequate.

Uniform walls of this permanent mold cast aluminum piston (a) resulted in simultaneous freezing, which caused porosity and centerline shrinkage. By tapering the walls and adding ribs for metal feed paths to the bosses (b), acceptable castings were easily produced.

144 / Casting Design and Performance To correct this condition, the skirt of the casting was tapered from the original thickness of 0.19 in., immediately adjacent to the dome, to 0.12 in. at the extreme end. Four ribs, two at each boss, were added, to serve as feed paths. The taper in the skirt assured satisfactory progressive solidification of metal in that section. The four ribs provided adequate feed paths for the bosses and provided additional strength in the wrist-pin bosses. The aluminum casting shown in Fig. 12 also was made in a permanent mold. As part of an aircraft fuel system, the casting was produced to exacting quality requirements, one of which was retention of aircraft fuel at a pressure of 35 psi. This pressure-tightness required perfectly sound metal; any porosity or defects would permit leakage. Molding was complicated by the location of the small tubular outlet. With this outlet located away from the optimum parting plane, a core (or a loose piece) was required in the mold to form the outside contours of the outlets. The draft requirements of this core and the smaller core necessary to form the inside of this tubular section combined to create a heavy section remote from any practical riser location. Porosity in this area resulted in a 40 percent rejection rate. To improve producibility, although at the penalty of adding undesirable weight, a small amount of padding was added to the wall connecting the tubular section with the mounting flange. A riser was appended to the flange, to feed the heavy section at the junction of the wall and the tubular section. With these revisions, sound castings were produced. This casting illustrates the desirability of considering molding requirements in the initial stages of product design. Figure 13 shows a permanent mold casting of aluminum alloy 355 which might have presented difficult production problems but did not because the problems were anticipated and

Fig. 12

eliminated in design. The key to success was the adequate size of the posts connecting top and bottom sections. The design permitted use of a comparatively simple three-piece mold and a central core. The parting lines of the mold ran radially from the center of the casting on planes that bisected the solid posts. Gates were located at one post only, as shown. Risers over each solid post could successfully feed all of the lower part of the casting, because of the adequate size of the posts. The heavy section bordering the small insert near the base of the casting was made sound by the chilling effect of the insert and because of its location under a solid post. Investment castings benefit from the incorporation of feed paths primarily in terms of cost reduction. In the investment process, additional runners can often be added to the rigging, to conduct metal to a part of the casting that cannot be filled or adequately fed through adjacent sections. An excellent example of cost reduction associated with the proper use of feed paths is demonstrated by the casting shown in Fig. 14 (also discussed in greater detail as Fig. 31 in Chapter 9). As originally designed, this casting closely approached but did not reach the conditions for production at maximum efficiency. It had relatively heavy sections at each end in the form of rings or flanges, and in the center area in the form of hubs. Experimental castings revealed that, even with the small ribs included, the cross-sectional area of the casting body was too small to permit the entire casting to be fed from one end or from the hubs. Consequently, six gates were required, to feed all sections adequately. Castings were later produced with the ribs increased in size, as shown in the redesign of section A. Only two gates, located at one end on the flange and opposite the ribs, were required. Production experience indicated that

castings made with the larger rib could be produced at a saving of 28 percent.

Designs That Reduce the Mass of a Remote Section Unlike the examples cited in the first part of this chapter, some castings can be improved in producibility by reducing the size of remote heavy sections, thus lessening the need for feed metal and making it possible for existing casting sections to provide the necessary feed paths. In reducing the size of an isolated section, the first benefit is a reduction in weight, and thus a reduction in metal cost. The size reduction can also decrease the number of risers required. Still another benefit is the elimination of certain potential casting defects. For example, the steel lever shown in Fig. 15(a), a sand casting for an aircraft application, was difficult to produce because hot tears developed in the thin wall connecting the hub and the fork. Usually, shrinkage defects also developed in the heavy base of the fork. Risering the heavy base eliminated the shrinkage defects but aggravated the hot tearing. The revision shown in Fig. 15(b) gave acceptable castings. In this redesign, the web separating the two fingers is tapered, with the heaviest section adjoining the hub. Only one riser, located at the hub, was required, to feed all sections of the casting adequately. The redesign used somewhat less metal, with no sacrifice in strength. Housings. Too often, cast housings are developed by specifying walls of uniform thickness spaced at a definite distance from the enclosed mechanism, to provide uniform clearance. To these walls, bosses, ribs and other necessary components are appended without careful planning. The resulting configuration is

Molding and production of this aluminum permanent mold casting (a), part of an aircraft fuel system, were complicated by the small tubular section that protruded at an angle. Padding (b) and core pulls (c) were necessary to produce the part.

Design Problems Involving Unequal Sections / 145

Fig. 13

Good design facilitated production of this permanent mold casting from aluminum alloy 355. (See text.)

Fig. 14

Six gates were required to produce this gimbal ring, an 8620 steel investment casting. By increasing the size of the ribs, four gates were eliminated and cost was decreased 28%.

then dimensioned and given to the foundry to produce as a casting. Seldom will this method of design yield a shape compatible with efficient production in the foundry. The malleable iron casting shown in Fig. 16 is a gear housing for a power-steering mechanism. Six risers were required, to obtain the desired soundness of metal. A large number of castings were produced in sand molds to the original design. To increase efficiency of production, the design was reviewed to determine whether the number of risers could be reduced. It was found that risers could be minimized by squaring the gear cavity and narrowing the mounting bosses, as shown in the redesign. These revisions permitted the elimination of three of the six risers. An aluminum hydraulic pump body (alloy 319), produced as a permanent mold casting, incorporated a heavy section that could not be fed unless padding were added to both the riser and the casting body (see section B-B, original design, Fig. 17). This heavy section surrounded the cored hole shown in sections C-C and D-D. Without padding, the heavy section was cut off from feed metal in the riser because of premature freezing of the metal at the junction of the heavy section and the outer casting wall. Although acceptable castings were produced with this padding, it was necessary to machine off the padding, which increased the cost. By reducing the volume of the heavy section, as shown in sections B-B and D-D, the casting was produced successfully without padding. To assist in filling the mold cavity, the bottom wall of the casting was increased in thickness from 1/8 to 5/32 in., as shown by section A-A.

Redesign eliminated a remote heavy section and encouraged a beneficial freezing pattern in this sand cast steel lever.

Fig. 15

146 / Casting Design and Performance

Fig. 16

A revision in design of this sand cast malleable iron gear housing reduced the mass of metal at each of the four mounting lugs. This decreased the riser requirement from six to three and substantially reduced cost

Decreasing the mass of metal around the cored hole (D-D) and increasing the thickness of the plate section of the casting (A-A) permitted efficient production of this aluminum pump body (alloy 319) in a permanent mold without padding (B-B).

Fig. 17

Casting Design and Performance Pages 147–153

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Junctions* IN MOST CASTINGS there are junctions between intersecting component members. It is customary to incorporate fillets at these junctions, to eliminate sharp inside corners and the attendant problems of stress concentration. The designer should be cognizant of the casting production problems introduced by junctions. Designs should strive for maximum soundness while retaining compatibility with other casting requirements. For practical purposes, and discounting the initial chilling effect of the mold, the rate at which molten metal solidifies is the same. for a thinner section ( 1 in. or 25 mm) as for a thicker section (4 in. or 100 mm). Because, in general, metal in a casting solidifies from the mold surfaces toward the center of a section, thin sections will freeze before heavy sections. Hence, where a thin section is immediately adjacent to a heavy section, and where the only source of feed metal for the heavy section is through the thin section, the thin section can freeze first and shut off the source of feed metal for the heavy section. With no source of feed metal to replace the volume lost because of metal shrinkage, such a heavy section will be porous. Also, with the onset of freezing, the thinner section starts to contract while the heavy section is still solidifying. This can cause stress to develop in the casting, resulting in cracks at the hottest (weakest) section. Different sections of a casting can be conveniently compared by inscribed circles, as shown in Fig. 1. The ratio of the areas of the inscribed circles is an indication of relative freezing times. The large sections finish freezing proportionately later than the small sections. The influence of fillets in increasing the mass of a junction can also be seen in Fig. 1. As shown, a 7/8-in. (22-mm) circle can be inscribed at an L-junction of two 3/4-in. (19-mm) members having no fillet Under the same conditions except for a ½-in. (13 mm) fillet, a 11/16-in. (27 mm) circle can be inscribed. This represents a 47 percent increase in area, volume and mass of metal. The ability of a sand or ceramic mold to absorb heat must be considered when junctions are being designed. In general, heat from the molten metal is transferred into the mold in a direction perpendicular to the interface, or radially if the interface is curved. Figure 2 shows schematically the directions of heat transfer in

two different types of L-junctions. At the outer surface of the junction shown in Fig. 2(a) there is more mold material available to absorb heat from the molten metal than is available adjacent to the straight portions of the casting that are well away from the corner. Consequently, the metal at this part of the junction will cool at a somewhat faster rate than will the metal in the remote straight sections of the casting. However, the metal at the inside corner of the junction will cool at a much slower rate than the metal at the outside corner or in the straight areas, because there is much less mold material to absorb the heat of the molten metal. As a result, the metal in the junction near the inside corner surface will be last to cool and will be cut off from feed metal if the source is through the straight section of the casting. A shrinkage defect will result, as shown. When the corner is provided with a fillet and uniform wall thickness, the shrinkage cavity is reduced in size. When the fillet radius equals the wall thickness of the junctioning members of the casting, as in Fig. 2(b), shrinkage voids usually are eliminated, even under the most adverse conditions, because the metal cools uniformly. Theoretically, for steel castings fillet radii should equal the thickness of the heaviest of the junctioning wall sections when the thickness of this wall is ½ in. (13 mm) or less. For wall thicknesses of ½ to 1 in. (13 to 25 mm), a radius equal to 3/4 of the heaviest wall thickness usually

Fig. 2

*Adapted from Casting Design, American Society for Metals, 1962

Fig. 1

Inscribed circles facilitate comparison of metal mass in casting junctions.

Schematic representation of the effect of junction shape on heat transfer from solidifying metal to the mold

148 / Casting Design and Performance

Fig. 3

Mold restraint to normal contraction creates areas of stress concentration in casting (a). Designs that permit unhindered contraction (b) are relatively free from residual stresses.

Tapering fillets from a riser to the extremities of a casting encourages soundness and often permits reduction in the weight of a casting

Fig. 4

In designing to prevent shrinkage cavities, one of the most difficult problems is to know when and where risers may be used by the foundry. The positioning of risers is a matter of judgment based on the experience of the foundry engineer. Also, other considerations, such as tooling-point locations, may prevent a riser from being placed in the most advantageous position. Thus, the casting may benefit if it is designed with the assumption that none of the junctions can be fed directly. As an example, consider the typical casting section shown in Fig. 2(a): If a riser could be located directly in contact with the junction so that it could feed the enlarged mass, no shrinkage cavities would occur. However, if the only source of feed metal is through one of the arms, shrinkage defects would occur, because the arms would freeze first and cut off the heavy junction from the feed metal. Hence, when feasible, a junction similar to Fig. 2(b) is preferred.

Junction Elements Fig. 5

Comparison of defects obtained in L-junctions of castings produced under conditions controlled in such a manner as to encourage the defects. Section size of tests: 3 by 3 in.

is sufficient. For wall thicknesses greater than 1 in. (25 mm), a fillet radius of ½ of the heaviest wall thickness normally will suffice. In practice, the stresses developed in the casting during solidification and cooling are more important in determining fillet size than is the wall thickness of the casting in the area of the fillet. Where adequate or even partial feeding of the junctions is achieved, smaller fillet radii than those mentioned above may be allowable if the design of the part does not introduce severe stresses from solidification or cooling. For example, the fillet radii required in a casting of the I-beam type, as in Fig. 3(a), are larger than those required in T-sections, Fig. 3(b). In the cooling of the I-beam type of casting, stresses are set up because the contraction of the web section is restrained by the mold material between the standing flanges of the I-beam section. Thus, a relatively large fillet at the junctions of the I-beam components is required, to prevent hot tearing at the fillet area. In the

T-section, because the casting is free to contract, no cooling stresses are set up by the mold material; therefore, smaller fillets are practical. When fillets are required in an area where the junctions can serve as flow and feed paths for the molten metal, it is advantageous to taper the fillets so that they are largest in the area where the riser will be placed and smallest at the point farthest from the riser. This tapering assists in creating a beneficial freezing pattern of the molten metal. Freezing will begin at the areas farthest from the riser and will proceed uniformly toward the riser. The junctions, being heavier than the adjacent sections, will freeze after these adjacent sections and thus will serve as sources for feed metal. Because of the tapering of the fillet in the junctions, freezing will progress toward the riser; this assures adequate feeding of the junctions by the riser. This principle, which is illustrated in Fig. 4, often permits a reduction in the weight of a casting.

A common understanding of the problems of junction and fillet design have helped designers and metalcasters evaluate the specific effects of design variables on the soundness of metal in different types of intersections. The influence of definite junction and corner configurations on the size and number of defects is briefly described here from one of the first systematic studies by Briggs et al (AFS Transactions, vol. 46, 1939, p 605). The early work is still valid and instructive. Steel casting junctions were considered for the five joint types, represented by the letters L, T, V, X and Y. All other configurations at corners could be considered modifications of one or more of these five. Designs in the forms of these letters, and their influence on shrinkage, were studied in detail by producing actual castings, summarized here in Fig. 5 to 9. Except where indicated, all casting sections were 3 by 3 in. (75 by 75 mm). The arms were about 24 in. (61 cm) long, with the risers placed at the extreme ends. The risers were utilized as sprues for pouring. Plain carbon basic electric

Design Problems Involving Junctions / 149

Fig. 6

Fig. 7

Comparison of defects in T-junctions produced under controlled conditions to encourage the defects. Section size of tests: 3 by 3 in. unless otherwise noted

Comparison of defects obtained in V-junctions of castings produced under conditions controlled in such a manner as to encourage the defects. Section size of tests: 3 by 3 in. Members intersect at 45 degrees.

steel was poured into dry sand molds from a teapot ladle. All castings were horizontal when poured. Heavy sections at the junctions could be fed only through the thinner arms. Three-inch (75 mm) sections were chosen because they were large enough to exhibit pronounced defects and were less sensitive to variations in the mold and metal temperature than thinner sections would have been. Junctions were modified by using radii of different sizes in the corners, or by varying the thickness of one of the junctioning legs as shown. The castings were examined radiographically. The L-junctions in Fig. 5 show that defects were eliminated by using a ½-in. (13 mm) fillet and reducing the wall thickness at the corner. No detectable defects were present in the L-junction where the radius of the fillet was equal to the wall thickness and a uniform wall was maintained at the corner. In such a design, however, a centerline weakness similar to that in a uniform wall of sufficient length may be

present at the interface where the two freezing fronts meet. The T-junctions shown in Fig. 6 indicate that defects can be eliminated in this type of junction only by coring a hole at the center of the junction of the two members. It is doubtful, of course, whether this is a practical solution in production castings. Any junction through which a hole could be cored could probably be risered to provide adequate feed metal. A depression in the crossarm of the Tjunction substantially reduces the defect but does not eliminate it. Because the size of the defect increases as the fillet increases (causing the mass to increase), the smallest practical fillet compatible with other casting requirements is recommended. V-junctions (Fig. 7) formed by uniform sections will not be free from shrinkage cavities, primarily because a hot spot is readily developed in the mold at the junction between the two sections. Mold sand, a poor conductor of heat, cannot conduct the heat away from

the sand enclosed between the sections of the casting as quickly as the heat is transferred from the molten metal into this mold section. Consequently, this portion of the mold becomes hot enough to cause the adjacent metal to become the last to freeze. This condition exists at the inside corner of all junctions; however, because of the acute angle formed by the V-junction, the condition is more pronounced. The defects decrease in area as the inside radius at the V-junction increases in size. As in the L-junction, a slightly reduced wall thickness at the V-junction encourages sound metal. X-junctions, as shown in Fig. 8, could not be designed in this study so that all defects were eliminated. By coring a hole through the junction, however, the defects were greatly reduced in size. As with the T-section, coring may not be a practical solution. It may be possible to attach a riser to the core print area as easily as it would be to set a core. Y-junctions as shown in Fig. 9, produced defects unless a triangular hole was cored through the heavy section of the junction. Although a round hole substantially reduced the defective area, it did not eliminate the defects completely. Note that the areas of the defects remained almost constant regardless of the design alteration of the uncored Y-junction. In this study, all junctions would have had substantially sound metal if risers had been located directly over the junctions. However, where feeding of the mass of molten metal at the junctions was inadequate during cooling and solidification, shrinkage defects occurred. Figure 10 shows defects found at junctions in actual production castings. These substantiate the data from the many trial castings in Fig. 5 to 9. Chills can decrease the area of corner defects. Although in many castings chills may be of value in obtaining the desired soundness of metal, the designer should not depend on them to produce acceptable castings. Chills add (often substantially) to the cost of

150 / Casting Design and Performance

Fig. 8

Comparison of defects obtained in X-junctions produced under controlled conditions to encourage the defects. Section size of tests: 3 by 3 in. unless otherwise noted. Center lines of offsets are 8 in. apart.

Fig. 9

Comparison of defects obtained in Y-junctions produced under controlled conditions to encourage the defects. Section size of tests: 3 by 3 in. Members intersect at 60 degrees.

casting production, and increase the dimensional variation of the areas which the chills contact. It is preferable to obtain the desired level of soundness by design.

Junctions in Aluminum Castings Aluminum produces defects in the same areas but of smaller magnitude than steel,

because less heat is transferred from solidifying aluminum than from steel. The mold material could dissipate the heat from aluminum more easily, and there would be less likelihood of hot spots forming. Although more liberties can be taken with the design of junctions for an aluminum casting than for a steel casting, junctions should still be fed adequately if shrinkage cavities and porosity are to be avoided.

As an example, the problems encountered in producing the aluminum alloy sand casting, a primary flight-control-mechanism support, shown in Fig. 11 were concerned with the junctions of the ribs with the walls of the rectangular tube sections (view C-C), and with the L-junctions in the rectangular tube section (view A-A and section B-B). In the original design, extensive shrinkage porosity was encountered at these junctions. Because class

Design Problems Involving Junctions / 151

Fig. 10

Defects in production castings verify results obtained in the test castings shown in Fig. 5 to 9.

Fig. 11

An aluminum alloy sand casting (a primary support for a flight-control mechanism) that required redesign of several junctions before sound, acceptable castings could be produced

1-A x-ray soundness was specified for this casting, the defects were cause for rejection. A review of the processing revealed no practical way to supply feed metal to these areas. Consequently, a design revision was necessary

to reduce the mass at the junctions of the ribs with the tubular section (see locations A, B and C in views C-C for both original and revised design), and to alter the L-junctions of the tubular sections (see view A-A and section B-B).

This reduction in mass at the junctions (locations A, B and C), the increase in the inside radius, and the establishment of a uniform wall at the L-junctions of the tubular sections resulted in the production of sound castings.

152 / Casting Design and Performance If this casting were to be produced from steel, further reduction of the mass of metal at some of the junctions would undoubtedly be required.

T-Junctions Versus Y-Junctions For a given material, section thickness and fillet radius, T-junctions present fewer solidification problems than Y-junctions. This becomes apparent by comparing the areas of the inscribed circles in the T-junction and Y-junction shown in Fig. 12. With wall thicknesses identical and the same size fillets at each junction, the area of the circle inscribed in the Y-junction is about 35% greater than the area of the circle in the T-junction. If a junction is not easily fed, it is obvious that the better design would incorporate T-junctions rather than Y-junctions wherever possible. In many castings, this is accomplished easily by the simple design expedients shown in Fig. 13. The 17-4 PH stainless steel casting shown in Fig. 14 provides an excellent example of the advantages of the T-junction over the Y-junction. Produced as an investment casting, this structural aircraft part required class 1-A x-ray soundness. Of the castings produced to the original design, 87 percent were rejected because of defects at the Y-junctions formed by the casting body and the arms that extend from the body at an angle of about 45 . It was not feasible to add runners to the rigging to supply feed metal to these problem junctions, because removal of the stubs where runners would contact the casting would have entailed expensive machining operations. Consequently, the casting was redesigned to reduce the relatively large mass of metal formed by the Y-junctions. This was accomplished by revising the arms so that T-junctions (similar to the "Recommended" junction in Fig. 13a) were formed with the tubular body. This revision can be seen by comparing the original and the redesign shown in Fig. 14. After the redesign of junctions, no castings were rejected.

Comparison of the mass of metal of Y- and T-junctions with all walls the same thickness and all radii equal. The inscribed circle in the Y-junction (a) is about 35 percent larger than inscribed circle in junction (b).

Fig. 12

Two practical methods of revising Y-junctions to T-junctions, to reduce the mass of metal at the junctions. With all walls the same thickness and all radii equal, in (a) the circle inscribed in the Y-junction is approximately 23 percent larger than the circle inscribed in the T-junction. In (b) the circle inscribed in the Y-junction is approximately 26 percent larger than the circle inscribed in the T-junction.

Fig. 13

Y-junctions caused porosity in this 17-4 PH stainless steel investment casting. (a) Revising to T-junctions. (b) eliminated the cause of the defects.

Fig. 14

Junctions of Unequal Sections Where walls of two different thicknesses form a junction, it is advisable to incorporate a gradual increase in the thickness of the thinner section so that it is about equal in thickness with the heavier section at the point of juncture. Recommended methods of achieving this blending in T-junctions and L-junctions whose intersecting walls are unequal in thickness are shown in Fig. 15. At an L-junction, if one or both of the outside surfaces are to be machined, the outside radius must be omitted and the corner left sharp. The sharp corner, however, is likely to contain larger shrink defects than would be present with the rounded corner. Figure 16 shows recommended methods for blending abrupt changes in the thickness of a

Fig. 15

Recommended designs for proper blending of unequal sections for a T-junction (a) and an L-junction (b)

Fig. 16

Recommended designs for blending abrupt changes in wall thickness

Design Problems Involving Junctions / 153

Fig. 17

High residual stress and defects encountered in original design (a) were eliminated by redesign (b) of this winch end-plate, sand cast of steel.

wall to avoid or minimize stress concentration. Figure 16(a) shows a method recommended where the thicker section is less than 50 percent larger than the thin section; Fig. 16(b) shows a method recommended where the thicker section is more than 50 percent larger than the thin section. An example of blended junctions is provided by the gate valve casting shown in Fig. 3 of the article “Design Problems Involving Uniform Sections.” This casting incorporated every advantageous design detail so as to obtain the highest degree of soundness in the metal. Even though the junctions of the flanges and the valve body were adequately fed, they were

properly blended to avoid the possibility of stress concentration. In the steel sand casting shown in Fig. 17 astute design eliminated the junctions and thus the defects they encouraged. The original design, Fig. 17(a); encouraged shrinkage porosity at the junctions of the ribs and the flat plate section. Those castings not rejected because of porosity were highly stressed internally because of the uneven cooling rate resulting from the heavy section of metal at each junction. A redesign that replaced the opposing ribs with radial corrugations is shown in Fig. 17(b). These corrugations increased the strength of the casting enough to permit a reduction in the wall

thickness of the hub. This brought the hub and plate sections closer to a uniform cooling rate and substantially reduced the internal stresses previously encountered. Other advantages were a 24 percent reduction in weight and a 22 percent reduction in the cost of producing the casting. In aircraft-wheel design, similar simplifications have been achieved. The intricate, highly ribbed wheel structures of Fig. 9 were replaced by wheels of simpler design; the simplified wheels had increased service life under critical fatigue loading, and were lower in cost.

Casting Design and Performance Pages 155–161

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Distortion DISTORTION and hot tearing are closely related. The forces that cause distortion can also cause hot tearing, if they are great enough or are applied at a critical point. These forces result from differences in the times at which various parts of a casting freeze and start to contract, and from the resistance offered by unyielding mold sections. The solidification time differences and the mold resistance both act to restrain normal contraction of the metal. This restraint can result in residual stress, distortion or hot tearing, depending on the magnitudes of force and restraint.

Distortion Due to Differences in Solidification Times Figure 1 shows three castings that illustrate ways in which distortion can occur when a thin member solidifies and contracts considerably before a thick member. The casting in Fig. 1(a) contained a heavy center section, making it easy to feed and fill. However, the thick and thin sections froze in such a way as to produce severe distortion. The side plates froze and started to contract considerably in advance of the heavier center section. Then, as the center section solidified and cooled, contraction was restrained by the thin sections. This created stress in the side plates, causing them to buckle upward, as shown by phantom lines in section A-A.

Fig. 1

Although this distortion could be partially controlled by increasing the thickness of the 0.06-in. (1.5 mm) section to 0.187 in. (4.75 mm) (with subsequent machining off the excess material), the 12-in. (30 cm) radius was difficult to machine to close tolerances in production. This part could have been cast without excessive distortion if its cross section had been redesigned in any of the five ways suggested in Fig. 1(a). These redesigns are numbered in the order of preference, on the basis of potential foundry difficulties. The practicability of the slender core in the fifth choice is questionable. A cast fitting made of a precipitationhardening stainless steel is shown in Fig. 1(b). Although the molten steel had good fluidity, and no difficulty was encountered in producing a sound casting, distortion was a serious problem. The junction of the two walls formed an area of increased cross section, and it was necessary to place a riser adjacent to this junction to assure sound metal. Because of the riser, metal temperature in this corner was higher than at any other location in the mold. The walls away from the corner solidified first, while the increased mass at the junction and the retarded cooling rate of the metal at the surface of the fillet caused the fillet surface to freeze later than the sharp outside corner. Distortion resulted, as indicated by the phantom lines in Fig. 1(b). Although the casting could be straightened in a press after annealing, the

same pattern of distortion returned after subsequent hardening. Distortion could have been reduced by providing a radius at the outside corner, as indicated by the dotted line A. This would have created a uniform section thickness. If greater stiffness were required, gussets could have been added, as indicated at B. The steel casting shown in Fig. 1(c) also distorted because of the different thicknesses of intersecting members. The two side frames formed by T-sections solidified rapidly and contracted in length while most of the heavy center section was still fluid. When the heavy center section finally did solidify, it shortened dimension C, resulting in the distortion indicated by the phantom line in the illustration. At least two solutions are possible for the distortion problem presented by this casting. First, compensation for the distortion can be made in the pattern, in the direction opposite to that of the observed distortion. This pattern correction would, of necessity, be based on a guess and a hope that the part would come out straight, as cast. Such a solution may be costly to develop, and accuracy cannot be guaranteed. The second solution consists of increasing the section thickness of the end member and subsequently machining this section to the desired size. This would necessitate increasing the tolerance on the unmachined surfaces of the T-section in that area by a minimum of

Three castings in which distortion occurred as a result of differences in the solidfication times of sections of unequal thicknesses.

156 / Casting Design and Performance 0.06 in. (1.5 mm), to allow for the observed distortion. As an alternative to such a tolerance, excess metal could be machined from both the inside and outside surfaces of the T-section. Distortion of a casting can be caused or eliminated by the gating and risering system used in producing the casting. This is demonstrated by the casting shown in Fig. 2. This thin-wall casting, which was 58 in. long, required several gates along its edge, as shown, to fill the mold completely. Had a continuous runner been used, the casting would have frozen before the runner froze, because: 1. the runner would have been heavier than the casting. 2. the mold surrounding the runner would have been hotter than the mold surrounding the casting, because more molten metal would have passed through the runner; and 3. at the time the mold was completely filled, the metal in the casting would have been the coldest, because it would have traveled the farthest and consequently have passed through more unheated mold areas. When the casting froze and became rigid, the contractive forces of the runner would have bowed the casting toward the runner. This possibility of distortion was anticipated, and two short runners were used rather than one long one. No discernible distortion occurred. Gating and risering can influence distortion in still another way. In general, the flow path

of the molten metal, as determined by the gate location, controls the direction of solidification. This has a direct bearing on the contraction of the metal, as is shown in the charts of Fig. 3; which record the distortion that was encountered in the cored holes of two investment castings. As cast, the holes were somewhat elliptical, the measurements of dimension B being greater than those of dimension A. Most castings are designed to be strong enough so that dimensions will not be thrown out of tolerance by the contractive forces developed during solidification and subsequent cooling. Nevertheless, without distorting appreciably, a casting may develop residual stresses from solidification and cooling that will cause failure in use if the sum of the service stresses and the residual stresses exceeds the yield strength or the tensile strength of the cast metal. An example is the sand cast hub shown in Fig. 4. Although this casting was produced from malleable iron, the problem and solution are applicable also to other metals. As it was originally designed, Fig. 4(a), the six radial ribs froze before the adjoining casting sections, even though, at 3/8-in. (9.5 mm) thickness, they were heavier than the 11/32-in. (8.7 mm) tubular section. Approximately 25 per cent of the castings were rejected because of hot tears at junctions of these ribs with other sections of the casting. Under static load, many of the approved castings failed: the ribs, under tension, separated from the casting, and the hub was completely torn off the flange or disk section. The residual

Had a continuous runner been used in producing this 58-in.-long casting, contractive forces would have caused distortion. This result was anticipated, and two short runners, as shown, were used; no discernible distortion occurred.

Fig. 2

Fig. 3

stresses introduced by the differential in freezing time had weakened these junctions enough to cause malfunction of the part. As redesigned, Fig. 4(b), and with the rib thickness reduced to 1/4 in. (6 mm) the casting was easily produced to the required strength level, for two main reasons: freezing time was more nearly uniform throughout the casting; and the various members contracted more evenly, without pulling on each other in the mold. The original design was fed through a bore riser, located inside the barrel section, whereas the redesign was fed with two external risers. The revisions in design and the more effective distribution of metal to the outside of the flange made for a more producible casting that was also more serviceable. In the redesign, Fig. 4(b), uniformity in freezing and cooling of the different sections of the casting was encouraged by the incorporation of ribs reduced to 1/4-in. (6 mm) thickness. Although in the redesign these ribs were in a location highly influenced by surrounding metal, because of their diminished thickness they apparently reached solidification at about the same time as the abutting sections. The importance of designing to obtain nearly uniform cooling in casting members is further demonstrated by the casting shown in Fig. 5. This aluminum sand casting, a turbine air intake, is approximately 26 in. long, 28 in. in over-all depth and 12 in. wide (66  71  30 cm) at the widest section. The sections shown in Fig. 5 are taken through each of the two struts which extend from a central hub to the end configuration. Only the sections through the struts are shown, as the rest of the casting has no direct relation to the subject at hand. In approximately 30 per cent of these castings produced to the original design, (a) and (c) of Fig. 5, hot tearing occurred at the junctions of the cross ribs and the outside walls, because of hot spots. The uneven thickness of the outer wall contributed to the occurrence of defects. The pattern was revised to alter thicknesses of the sections forming these junctions to those shown in the redesign, (b) and (d) of Fig. 5. Increasing the thickness of the cross ribs from 0.25 to 0.26 in. (6.35 to 6.6 mm), and reducing the 0.30-in. (7.6 mm) thickness of outer walls at the junction to equal that of the now 0.26-in.-thick ribs, resulted in a cooperative freezing pattern. After this revision, no castings were rejected for hot tears.

Distortion from roundness of the cored holes in these two investment castings was related to the location of gates

Design Problems Involving Distortion / 157

Fig. 4

A sand cast malleable iron hub that required redesign and also rerigging to minimize residual stresses and permit successful application of the part. The problem was at the rib junctions.

Fig. 5

Cross sections of the struts of a turbine air intake, sand cast in aluminum alloy 356. When produced to the original design, 30% of the castings were rejected because of hot tears. Redesign shown eliminated these defects

Hot tearing in ribs or webs surrounded by heavier sections may occur also when these ribs or webs freeze later than the heavy sections. As an example, the investment cast 4130 steel gimbal ring of Fig. 6 had an excessively high rejection rate because of hot tears in a web connecting two heavy sections. Because the section of the mold which formed this “pocket” was almost completely surrounded by heavier sections of metal, a hot spot developed in the mold and prevented solidification of the enclosed web section until after the adjacent heavier sections had begun to contract. The simplest solution was to extend the web out from the enclosed area, thereby encouraging faster solidification of the thin section. This change allowed the web to develop enough strength to resist the contractive forces applied to it. The added material was removed by subsequent machining, to provide the clearance necessary for another member of the assembly. Another type of distortion that is attributable to nonuniform freezing occurs in plate sections of uniform thickness. This is the concave distortion in a flat or nearly flat surface that is commonly known as “oil canning”, and is the result of stresses set up as the plate freezes. There is no simple way to eliminate such distortion in a uniformly flat plate casting; consequently, the designer should compensate for it by providing sufficient tolerance or by employing a design of greater inherent stiffness. Figure 7 shows the amount of oil canning that may normally be anticipated in cast steel plates 10 in. square as a function of plate thickness. The magnitude of oil canning encountered in a production casting is illustrated in Fig. 8. This disk-shaped casting was produced from stainless steel in a shell mold. The distortion occurred because the molten metal at the thinner inside diameter froze first and contracted ahead of that at the heavier outside ring. When this heavier ring froze and contracted, the force applied to the inner section caused the thin section to move out of position to the extent shown. Wheels, such as the sand cast magnesium airplane wheel shown in Fig. 9(a), present a challenge to a casting designer. A delicate balance is required among all sections, to encourage cooperative freezing and contraction. If the central hub and the spokes freeze first and contract too far in advance of the rim section, hot tears may appear at the junctions of rim and spokes. Or, if enough metal has frozen at the junctions to resist the contractive forces of the spokes, residual stresses may develop, of sufficient magnitude to cause failure of the casting in service. Also, in this type of casting, freezing and contraction of the spokes before the rim causes a constrictive force to be applied to the spokes by the rim when it starts to contract. At a certain point, the forces may become great enough to displace the hub. One solution is to make the spokes thick enough so that all sections of the casting will freeze and contract nearly simultaneously. Another solution is to make the cross section of the spokes U-shaped, thus providing

158 / Casting Design and Performance enough strength to resist the contractive force of the rim.

Distortion Due to Mold Restraint

In this investment cast gimbal ring, of 4130 steel, a finger of mold material surrounded by molten metal became excessively hot and retarded the freezing of the web section where shown. The resulting hot tearing condition was corrected by redesigning the web so that it extended into a cooler section of the mold.

Fig. 6

Fig. 7

Oil-canning distortion in a flat cast plate as related to plate thickness

Fig. 8

The magnitude of oil canning encountered at location A in production of a stainless steel shell mold casting

Fig. 9

(a) Distortion here resulled from mold restraint, (b) and (c) two possible preventive redesigns.

A second major cause of casting distortion, particularly in fragile castings, is the restraint imposed by the mold as the casting cools and contracts. In its most severe form, this type of distortion is difficult to eliminate except by redesigning or by adding tie bars. The use of tie bars is discussed in a later section of this chapter. The sand casting shown in Fig. 9(a) presented no problems as regards mold filling, feeding or soundness. The two standing flanges A in the original design solidified and contracted without difficulty. However, when the bottom plate cooled and contracted, the flanges, separated by mold sand, were bent outward, as indicated by the phantom lines. Although some distortion was observed in the 0.50-in (13 mm)-thick bottom plate, it was of sufficient thickness to remain within assigned tolerance. If the bottom plate were reduced in thickness, the distortion would increase as the thickness decreased and would magnify the displacement of the standing flanges. The distortion experienced with this type of casting may be avoided by connecting the tops of the two standing flanges with a later-removable cross-member, as in Fig. 9(b), which is preferable, or by adding gussets to the sides of the flanges, as in Fig. 9(c), which would be less effective. Mold restraint to normal contraction of the cooling metal was responsible for hot tears in the aluminum semipermanent mold casting shown in Fig. 10. In this aircraft casting as originally designed, the heavy end flanges were connected by a thinner wall, which joined the flanges at the mid-point of their width. Also connecting the flanges were a series of tapping bosses. Because of the several variations in wall thickness, hot tears appeared in about 20% of the castings at the junctions of the thin wall with the heavier flanges; the thin wall froze and contracted in advance of the heavy tapping bosses and the flanges. With the outer sections of the mold made of metal and the core made of baked sand, the mold was impervious to any crushing force that might be developed by the contraction of these thin sections. In addition, with the flanges hotter than the thin sections, not enough strength developed in the junction fillets to resist the contractive forces. Hot tears resulted. Here, the presence of heavy sections, the heat-retaining ability of the mold, and the rigidity of the mold all combined to create difficulties. As a consequence, the casting was revised. By moving the connecting wall inward until its inner surface was congruent with the inner configuration of the mounting flanges, a metal core, which had the ability to chill the casting, could be used. To encourage uniform freezing, the

Design Problems Involving Distortion / 159

Mold restraint coupled with nonuniform freezing of the various sections of this aluminum (alloy 356) semipermanent mold casting resulted in hot tears that were responsible for a 20% rejection rate. Moving the wall and increasing its thickness as shown corrected the trouble.

Fig. 10

Fig. 12

The addition of ribs to this brass investment casting corrected dimensional inaccuracies between upright sections, caused by mold restraint to normal metal contraction. Without ribs, the upright sections became tilted outward and were out of tolerance.

Fig. 11

thickness of the connecting wall was increased and the flanges were reduced in thickness. These revisions reduced rejections—for all causes—to five per cent. Investment Castings. The restraint that the mold exerts on the forces of contraction can be a problem in investment castings also. In the brass investment casting of Fig. 11, mold restraint prevented the two upright sections from moving toward each other as the base cooled and contracted. As a result, the upright sections became tilted outward and were out of tolerance. By incorporating ribs as shown, the mold restraint was modified, and dimensionally acceptable castings were produced. When such ribs are used, their proportions must be designed with care. Improperly proportioned ribs may freeze before the adjoining members and thus promote, rather than prevent, distortion or hot tears.

Tie Bars Tie bars are expendable additions to a casting that are employed to equalize contraction in the mold and to provide rigidity during subsequent machining or heat treating. Many castings are producible within tolerance and without

distortion only by the strategic placement of one or more tie bars. Four tie bars were used to good advantage in the casting shown in Fig. 12. Produced from type 410 stainless steel in a ceramic mold, this casting exhibited good casting qualities, except that mold restraint affected dimensions A. By incorporating tie bars as shown in phantom in Fig. 12, dimensional stability was maintained in these areas of the castings. In addition, by leaving these tie bars in place until after the casting was heat treated, possible distortion from this processing operation was avoided. Another casting that exemplifies the salutary effects of tie bars is the cockpit-fairing-glass frame shown in Fig. 13. This aluminum semipermanent mold casting made use of tie bars in a major way. As originally designed, without tie bars, the casting warped excessively. Because of this lack of rigidity, problems were also encountered in machining. To remedy these problems, one center tie bar was incorporated, but with little success. Next, two tie bars were tried, one at each side of the casting. These corrected some of the distortion, but the percentage of acceptable castings was still too low to be economical. In addition, it was necessary to cast test coupons with each casting so that mechanical properties could be measured.

Tie bars in a stainless steel casting

These problems were solved by incorporating three tie bars as shown in Fig. 13. The center bar was removed after casting but before machining, and test coupons were produced from this bar. To maintain rigidity and eliminate machining problems, the two outside tie bars were permitted to remain on the casting until the part had been completely machined. However, because of residual stresses, the casting would spring out of shape when the tie bars were removed. This problem was solved by allowing the tie bars to remain in place until the casting was assembled into position on the aircraft, after which they were removed. Without tie bars, this casting probably could not have been produced in usable condition. Investment castings also may be held within specified tolerances by the use of tie bars. Two such castings are shown in Fig. 14. The 4130 steel casting, Fig. 14(a), required a small “standard” tie bar, positioned as shown, in order that the 3/8 -in. dimension between the side sections could be held. The 1020 steel casting, Fig. 14 (b), required a tie bar to assist in maintaining the 10/32 -in. dimension between the two side sections. A gate was devised that functioned also as a tie bar, thus presenting a manufacturing economy because it eliminated the cost of removing a special tie bar. (The gates required removal whether or not they served as tie bars.) The multipurpose gate was possible because of the flat ends of the side sections. Because they are as much a part of the wax pattern as is any other section of an investment casting, tie bars help to maintain dimensional stability in the wax pattern in the same manner as they do in the casting. Because distortion of a wax pattern is reproduced in the casting, the foundry-man must provide some other means of maintaining accuracy in wax patterns if tie bars are impractical. Sometimes, metal spacers are used between two parts of the wax pattern to assure maintenance of a dimension. If the hardening of the wax pattern (after it has been removed from the pattern die) is such that it

160 / Casting Design and Performance

This aluminum (alloy 356) semipermanent mold casting would probably not have been producible or machinable to a usable condition without the use of the three tie bars shown. Two of these tie bars were left in place until after the part (a cockpit-fairing-glass frame) was assembled in the aircraft, because of distortion from residual stresses.

Fig. 13

Produced as one part, machined, then given a T6 heat treatment, this sand cast aluminum (alloy 356) clamp distorted beyond usable condition. Changing to a T7 heat treatment minimized the distortion.

Fig. 15

Increasing section thickness in the problem areas of this permanent mold casting would result in better flow of metal in the mold, better feeding during solidification, and the elimination of defects due to shrinkage.

Fig. 16 would open a dimension rather than close it, a spacer may be set in place and a weight placed on the wax, so that the members of the wax pattern are forced into the desired dimension.

Distortion in Heat Treating Distortion that occurs in heat treating may necessitate some form of compromise in design or processing, so that a usable part is obtained. With the mounting clamp shown in Fig. 15, an aluminum (alloy 356) sand casting, a revision of the choice of heat treatment was required, in order to curtail the distortion that had occurred as a result of the first-selected process and made most of the heat treated castings unusable.

Produced as one part, this casting was machined, given a T6 heat treatment, and separated into two clamps. On separation, the castings would open up 1/8 to 1/4 in. (3 to 6 mm). The distortion occurred because of stresses introduced during quenching in the T6 heat treatment; the artificial aging step in this heat treatment did not effect enough stress relief to prevent subsequent distortion. Consequently, the heat treatment was revised to T7, and more stable castings were obtained. Although a T7 heat treatment takes more time and is more

Two investment castings that required tie bars to assure dimensional accuracy. In (a), a “standard” tie bar, which had to be removed after casting, added to the cost of production. In (b), the gate functioned also as a tie bar, and, since gate removal is a normal requirement, no additional cost was incurred for this operation.

Fig. 14

expensive than T6, the results may justify the added expenditure, as they did with this part. The permanent mold casting shown in Fig. 16 was used in a structural application and had to be sound and flat. It was solution heat treated, quenched, and aged. Difficulties were encountered in keeping the casting within flatness tolerance, and with shrinkage in the 3/8-in. (9 mm) web between the heavy bosses. The problem here was one of molten metal distribution and of supplying feed metal to replace the volume lost by shrinkage. A simple solution would be to increase the thickness of the webs between the bosses. This would permit better feeding of the casting, eliminating the shrinkage, and would reduce warpage in heat treatment by modifying the stress pattern.

Effect of Alloy The alloy to be cast can have a significant influence on distortion. Figure 17 shows a plaster mold casting for which a 4.8% Zn, 2.3% Mg aluminum alloy was specified. This material, which hardens at room temperature in about 20 days, had successfully eliminated distortion

Design Problems Involving Distortion / 161

Using a 4.8% Zn, 2.3% Mg aluminum alloy, this plaster mold casting distorted out of tolerance and was not producible to the original design. Using alloy 356, the casting was produced as designed, and distortion was within acceptable limits.

Fig. 17

that had occurred in heat treatment when aluminum alloy 356 had been used. However, the casting shown in Fig. 17 was not producible to the original design when the Al-Zn-Mg alloy was used; it was necessary to cast the center bar solid (so as to obtain sound metal in this section), and to fill the mold cavity and obtain the required pressure-tightness. This revision introduced considerable distortion. The thin sections froze and progressed through part of their contraction cycle before the heavier bar section froze. Thus, the contraction of the heavier section was resisted by the thinner sections, and this interplay of forces was strong enough to distort the casting 0.020 in. (0.51 mm) out of position at each end. Many processing variations were investigated, in an effort to correct this problem. It was found that chills, located as shown, reduced distortion so that the castings were within the specified tolerance. For comparison, castings were poured to the original design from aluminum alloy 356. All these castings were sound and passed the pressure-tightness test, and distortion was sufficiently low that the parts stayed within tolerances, as cast. To minimize distortion, the

castings were quenched in an air blast, rather than in a liquid. The Al-Zn-Mg alloy has a higher pouring temperature and does not feed isolated areas of the casting from risers as well as alloy 356. However, when a casting is so designed as to encourage distortion in heat treatment, the AlZn-Mg alloy, because of its ability to age harden at room temperature, may offer an advantage over alloy 356.

Casting Design and Performance Pages 163–170

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Corrosion of Cast Irons* CAST IRON is a generic term that identifies a large family of ferrous alloys. Cast irons are primarily alloys of iron that contain more than 2% carbon and 1% or more silicon. Low raw material costs and relative ease of manufacture make cast irons the least expensive of the engineering metals. Cast irons can be cast into intricate shapes because of their excellent fluidity and relatively low melting points and can be alloyed for improvement of corrosion resistance and strength. With proper alloying, the corrosion resistance of cast irons can equal that of stainless steels and nickel-base alloys in many services. Because of the excellent properties obtainable with these low-cost engineering materials, cast irons find wide application in environments that demand good corrosion resistance, such as in water, soils, acids, alkalis, saline solutions, organic compounds, sulfur compounds, and liquid metals.

Basic Metallurgy of Cast Irons The metallurgy of cast irons is similar to that of steels except that sufficient silicon is present to necessitate use of the iron-silicon-carbon ternary phase diagram rather than the simple iron-carbon binary diagram. Figure 1 shows a section of the iron-iron carbide-silicon ternary diagram at 2% Si. The eutectic and eutectoid points in the iron-silicon-carbon diagram are both affected by the introduction of silicon into the system. In the 1 to 3% Si levels normally found in cast irons, eutectic carbon levels are related to silicon levels as follows: %C þ 1/3ð%SiÞ ¼ 4:3

on the manner in which the carbon segregates in the microstructure. Higher silicon levels favor the formation of graphite, but lower silicon levels favor the formation of iron carbides. The form and shape in which the carbon occurs determine the type of cast iron (Table 1). The structure of the metal matrix around the carbon-rich constituent establishes the class of iron within each type of iron. As in steel, the five basic matrix structures occur in cast iron: ferrite, pearlite, bainite, martensite, and austenite. Ferrite is generally a soft constituent, but it can be solid solution hardened by silicon. When silicon levels are below 3%, the ferrite matrix is readily machined but exhibits poor wear resistance. Above 14% Si, the ferritic matrix becomes very hard and wear resistant but is essentially nonmachinable. The low carbon content of the ferrite phase makes hardening difficult. Ferrite can be observed in cast irons on solidification but is generally present as the result of special annealing heat treatments. High silicon levels promote the formation of ferritic matrices in the as-cast condition. Pearlite consists of alternate layers of ferrite and iron carbide (Fe3C, or cementite). It is very strong and tough. The hardness, strength, machinability, and wear resistance of pearlitic

Influence of Alloying Alloying elements can play a dominant role in the susceptibility of cast irons to corrosion attack. The alloying elements generally used to enhance the corrosion resistance of cast irons include silicon, nickel, chromium, copper, and molybdenum. Other alloying elements, such as vanadium and titanium, are sometimes used,

(Eq 1)

where %C is the eutectic carbon level, and %Si is the silicon level in the cast iron. The metallurgy of cast iron can occur in the metastable iron-iron carbide system, the stable iron-graphite system, or both. This causes structures of cast irons to be more complex than those of steel and more susceptible to processing conditions. An appreciable portion of carbon in cast irons separates during solidification and appears as a separate carbon-rich constituent (e.g., graphite, iron carbides) in the microstructure. The level of silicon in the cast iron has a strong effect

matrices vary with the fineness of its laminations. The carbon content of pearlite is variable and depends on the composition of the iron and its cooling rate. Bainite is an acicular structure in cast irons that can be obtained by heat treating, alloying, or combinations of these. Bainitic structures provide very high strength at a machinable hardness. Martensitic structures are produced by alloying, heat treating, or a combination of these practices. Martensitic microstructures are the hardest, most wear-resistant structures obtainable in cast irons. Molybdenum, nickel, manganese, and chromium can be used to produce martensitic or bainitic structures. Silicon has a negative effect on martensite formation, because it promotes the formation of pearlite or ferrite. Austenitic structures are typically found in the Ni-Resist cast irons and the austempered ductile irons. Austenite is a face-centered cubic atomic structure created primarily by alloying with austenite-forming elements such as nickel. Austenite is generally the softest and more corrosion-resistant matrix structure. However, the carbon-enriched austenite of austempered ductile iron has higher hardness and other unique characteristics over conventional ductile irons (Ref 1).

Table 1 Summary of cast iron classification based on carbon form and shape Type of cast iron

Fig. 1

Section of the iron-iron carbide-silicon temary phase diagram at 2% Si

*Reprinted from ASM Handbook, Volume 13B: Corrosion: Materials, p43–50

White cast iron Malleable cast iron Gray cast iron Ductile cast iron Compacted graphite cast iron

Carbon form and shape

Iron carbide compound Irregularly shaped nodules of graphite Graphite flakes Spherical graphite nodules Short, fat, interconnected flakes (intermediate between ductile and gray cast iron)

164 / Casting Design and Performance but not to the extent of the first five elements mentioned. Silicon is the most important alloying element used to improve the corrosion resistance of cast irons. Silicon is generally not considered an alloying element in cast irons until levels exceed 3%. Silicon levels between 3 and 14% offer some increase in corrosion resistance to the alloy, but above approximately 14% Si, the corrosion resistance of the cast iron increases dramatically. Silicon levels up to 17% have been used to enhance the corrosion resistance of the alloy further, but silicon levels over 16% make the alloy extremely brittle and difficult to manufacture. Even at 14% Si, the strength and ductility of the material is low, and special design and manufacturing parameters are required to produce and use these alloys. Alloying with silicon promotes the formation of strongly adherent surface films in cast irons. Considerable time may be required to establish these films fully on the castings. Consequently, in some services, corrosion rates may be relatively high for the first few hours or even days of exposure, then may decline to extremely low steady-state rates for the rest of the time the parts are exposed to the corrosive environment (Fig. 2). Nickel is used to enhance the corrosion resistance of cast irons in a number of applications. Nickel increases corrosion resistance by the formation of protective oxide films on the surface of the castings. Up to 4% Ni is added in combination with chromium to improve both strength and corrosion resistance in cast iron alloys. The enhanced hardness and corrosion resistance obtained is particularly important for improving the erosion-corrosion resistance of the material. Nickel additions enhance the resistance of cast irons to corrosion by reducing acids and alkalis. Nickel additions of 12% or greater are necessary to optimize the corrosion resistance of cast irons. The Ni-Resist group are high-nickel alloys (13.5 to 36% Ni) having high resistance to wear, heat, and corrosion. Nickel is not as common an alloying addition as either silicon or chromium for enhancing the corrosion resistance in cast irons. It is much more important as a strengthening and hardening addition. Chromium is frequently added alone and in combination with nickel and/or silicon to increase the corrosion resistance of cast irons. As with nickel, small additions of chromium are used to refine graphite and matrix microstructures. These refinements enhance the corrosion resistance of cast irons in seawater and weak acids. Chromium additions of 15 to 35% improve the corrosion resistance of cast irons to oxidizing acids, such as nitric acid (HNO3). Chromium increases the corrosion resistance of cast iron by the formation of protective oxides on the surface of castings. The oxides formed will resist oxidizing acids but will be of little benefit under reducing conditions. High-chromium additions, similar to higher-silicon additions, reduce the ductility of cast irons.

Copper is added to cast irons in special cases. Copper additions of 0.25 to 1% increase the resistance of cast iron to dilute acetic (CH3COOH), sulfuric (H2SO4), and hydrochloric (HCl) acids as well as acid mine water. Small additions of copper are also made to cast irons to enhance atmospheric-corrosion resistance. Additions of up to 10% are made to some high-nickel-chromium cast irons to increase corrosion resistance. The exact mechanism by which copper improves the corrosion resistance of cast irons is not known. Molybdenum. Although an important use of molybdenum in cast irons is to increase strength and structural uniformity, it is also used to enhance corrosion resistance, particularly in high-silicon cast irons. Molybdenum is particularly useful in hydrochloric acid (HCl). As little as 1% Mo is helpful in some high-silicon irons, but for optimal corrosion resistance, 3 to 4% Mo is added. Other Alloying Additions. In general, other alloying additions to cast irons have a minimal effect on corrosion resistance. Vanadium and titanium enhance the graphite morphology and matrix structure and impart slightly increased corrosion resistance to cast irons. Few other additions are made to cast irons that have any significant effect on corrosion resistance.

Influence of Microstructure Although the graphite shape and the amount of massive carbides present are critical to mechanical properties, these structural variables do not have a strong effect on corrosion resistance. Flake graphite structures may trap corrosion products and retard corrosion slightly in some applications. Under unusual circumstances, graphite may act cathodically with regard to the metal matrix and accelerate attack. While the structure of the matrix has a slight influence on corrosion resistance, the effect is small compared to that of matrix composition. In gray irons, ferrite structures are generally the least corrosion-resistant, and graphite flakes

exhibit the greatest corrosion resistance. Pearlite and cementite show intermediate corrosion resistance, while an austenitic structure imparts higher corrosion resistance. Shrinkage or porosity can degrade the corrosion resistance of cast iron parts by acting as natural crevices. The presence of porosity permits the corrosive medium to enter the body of the casting and can provide continuous leakage paths for corrosives in pressure-containing components.

Commercially Available Cast Irons Based on corrosion resistance, cast irons can be grouped into the following five classes. Unalloyed gray, ductile, malleable, and white cast irons represent the first and largest class. All of these materials contain carbon and silicon of 3% or less and no deliberate additions of nickel, chromium, copper, or molybdenum. As a group, these materials exhibit a corrosion resistance that equals or slightly exceeds that of unalloyed steels, but they show the highest rate of attack among the classes of cast irons. These materials are available in a wide variety of configurations and alloys. Major ASTM standards that cover these materials are listed in Table 2. Low- and moderately alloyed cast irons constitute the second major class. These irons contain the iron and silicon of unalloyed cast irons plus up to several percent of nickel, copper, chromium, or molybdenum. As a group, these materials exhibit two to three times the service life of unalloyed cast irons. Austempered ductile iron (ADI) is the newest group of alloys in this category, and they have some unique properties. The ADI delivers twice the strength of conventional ductile irons for a given level of elongation. In addition, ADI offers exceptional wear and fatigue resistance (Ref 1). Table 2 ASTM standards that include unalloyed cast irons Standard

A A A A

47 48 74 126

A A A A

159 197 220 278

A 319 A 395

Fig. 2

Corrosion rates of high-silicon cast irons as a function of time and corrosive media

A A A A A A A A

476 536 602 716 746 823 842 874

Materials/products covered

Ferritic malleable iron castings Gray iron castings Cast iron soil pipe and fittings Gray iron castings for valves, flanges, and pipe fittings Automotive gray iron castings Cupola malleable iron Pearlitic malleable iron castings Gray iron castings for pressure-containing parts for temperatures up to 345  C (650  F) Gray iron castings for elevated temperatures for nonpressure-containing parts Ferritic ductile iron pressure-retaining castings for use at elevated temperatures Ductile iron-castings for paper mill dryer rolls Ductile iron castings Automotive malleable iron castings Ductile iron culvert pipe Ductile iron gravity sewer pipe Statically cast permanent mold gray iron castings Compacted graphite iron castings Ferritic ductile iron castings suitable for low-temperature service

Corrosion of Cast Irons / 165 Major ASTM standards that cover these materials are listed in Table 3. High-nickel austenitic cast irons represent a third major class of cast irons for corrosion service. These materials contain large percentages of nickel and copper and are fairly resistant to such acids as concentrated sulfuric (H2SO4) and phosphoric (H3PO4) acids at slightly elevated temperatures, hydrochloric acid at room temperature, and organic acids such as acetic (CH3COOH), oleic, and stearic. When nickel levels exceed 18%, austenitic cast irons are nearly immune to alkali or caustics, although stress corrosion can occur. Highnickel cast irons can be nodularized to yield ductile irons. Major ASTM standards that cover these materials are listed in Table 4. High-chromium cast irons are the fourth class of corrosion-resistant cast irons. These materials are basically white cast irons alloyed with 12 to 35% Cr. Other alloying elements may also be added to improve resistance to specific environments. When chromium levels exceed 20%, high-chromium cast irons exhibit good resistance to oxidizing acids, particularly nitric acid (HNO3). High-chromium irons are not resistant to reducing acids. They are used in saline solutions, organic acids, phosphate mining, marine, and industrial atmospheres. These materials display excellent resistance to abrasion, and, with proper alloying additions, they can also resist combinations of abrasives and liquids, including some dilute acid solutions. High-chromium cast irons are covered in ASTM A 532. In addition, many proprietary alloys not covered by national standards are produced for special applications, such as wear components in mining operations or slurry pumps. High-silicon cast irons are the fifth class of corrosion-resistant cast irons. The principal alloying element is 12 to 18% Si, with more

Table 3 ASTM standards that include lowalloyed cast iron materials Standard

A 159 A 319 A 532 A 897

Materials/products covered

Automotive gray iron castings Gray iron castings for elevated temperature for nonpressure-containing parts Abrasion-resistant cast irons Austempered ductile iron castings

than 14.2% Si needed to develop excellent corrosion resistance. Chromium and molybdenum are also used in combination with silicon to develop corrosion resistance to specific environments. High-silicon cast irons represent the most universally corrosion-resistant alloys available at moderate cost. When silicon levels exceed 14.2%, high-silicon cast irons exhibit excellent resistance to H2SO4, HNO3, HCl, CH3COOH, and most other mineral and organic acids and corrosives. These materials display good resistance in oxidizing and reducing environments and are not appreciably affected by concentration or temperature. Exceptions to universal resistance are hydrofluoric acid (HF), fluoride salts, sulfurous acid (H2SO3), sulfite compounds, strong alkalis, and alternating acid-alkali conditions. High-silicon cast irons are defined in ASTM A 518 and A 861.

Forms of Corrosion Cast irons exhibit the same general forms of corrosion as other metals and alloys:           

Uniform or general attack Galvanic or two-metal corrosion Crevice corrosion Pitting Intergranular corrosion Selective leaching (graphitic corrosion) Erosion-corrosion Stress corrosion Corrosion fatigue Fretting corrosion Microbiological

Graphitic Corrosion. A form of corrosion unique to cast irons is a selective leaching attack commonly referred to as graphitic corrosion or graphitization. Graphitic corrosion is observed in gray cast irons in relatively mild environments in which selective leaching of iron leaves a brittle graphite network. Selective leaching of the iron takes place because the graphite is cathodic to the iron, and the gray cast iron structure establishes an excellent galvanic cell. While graphitic corrosion of gray

Table 5

Relative fretting resistance of cast iron

Poor Note: Because most cast iron standards make chemical composition subordinate to mechanical properties, many of the standards listed in Table 2 may also be used to purchase low-alloyed cast iron materials.

Table 4 ASTM standards that include highnickel austenitic cast iron materials Standard

A 436 A 439 A 571

Materials/products covered

Austenitic gray iron castings Austenitic ductile iron castings Austenitic ductile iron castings for pressurecontaining parts suitable for low-temperature service

cast iron is considered a form of selecting leaching, its mechanism on a microstructural level is similar to galvanic corrosion. This form of corrosion generally occurs only when corrosion rates are low. If the metal corrodes more rapidly, the entire surface, including the graphite, is removed, and more or less uniform corrosion occurs. Graphitic corrosion can cause significant problems because, although no dimensional changes occur, the cast iron loses its strength and metallic properties. Thus, without detection, potentially dangerous situations may develop in pressure-containing applications. Graphitic corrosion is observed only in gray cast irons. In both nodular and malleable cast iron, the lack of graphite flakes provides a more favorable anode/cathode ratio and no network to hold the corrosion products together. By maximizing the area of the anodic component while decreasing the area of the cathodic constituent, the potential for galvanic (graphitic) corrosion has been reduced. Because graphitization is so common with cast iron and it compromises the structural integrity of the metal, instrumentation using eddy-current measurements has recently been developed to detect and measure it (Ref 2). Fretting corrosion is commonly observed when vibration or slight relative motion occurs between parts under load. The relative resistance of cast iron to this form of attack is influenced by such variables as lubrication, hardness variations between materials, the presence of gaskets, and coatings. Table 5 compares the relative fretting resistance of cast iron under different combinations of these variables. Pitting and Crevice Corrosion. The presence of chlorides and crevices or other shielded areas presents conditions that are favorable to the pitting and crevice corrosion of cast iron. Pitting has been reported in such environments as dilute alkylaryl sulfonates, antimony trichloride (SbCl3), and calm seawater. Alloying can influence the resistance of cast irons to pitting and crevice corrosion. For example, in calm sea-water, nickel additions reduce the susceptibility of cast irons to pitting attack. High-silicon cast irons with chromium and/or molybdenum offer enhanced resistance to pitting and crevice

Average

Aluminum on cast iron Magnesium on cast iron Cast iron on chrome plate

Cast iron on cast iron Copper on cast iron Brass on cast iron

Laminated plastic on cast iron Bakelite on cast iron

Zinc on cast iron Cast iron on silver plate

Cast iron on tin plate Cast iron on cast iron with coating of shellac

Cast iron on copper plate Cast iron on amalgamated copper plate Cast iron on cast iron with rough surface

Source: Ref 3

Good

Cast iron on cast iron with phosphate coating Cast iron on cast iron with coating of rubber cement Cast iron on cast iron with coating of tungsten sulfide Cast iron on cast iron with rubber gasket Cast iron on cast iron with Molykote lubricant Cast iron on stainless with Molykoke lubricant

166 / Casting Design and Performance corrosion. Although microstructural variations probably exert some influence on susceptibility to crevice corrosion and pitting, there are few reports of this relationship. Intergranular attack is relatively rare in cast irons. In stainless steels, in which this type of attack is most commonly observed, intergranular attack is related to chromium depletion adjacent to grain boundaries. Because only the high-chromium cast irons depend on chromium to form passive films for resistance to corrosion attack, few instances of intergranular attack related to chromium depletion have been reported. The only reference to intergranular attack in cast irons involves ammonium nitrate (NH4NO3), in which unalloyed cast irons are reported to be intergranularly attacked. Because this form of selective attack is relatively rare in cast irons, no significant references to the influence of either structure or chemistry on intergranular attack have been reported. Erosion-Corrosion. Fluid flow by itself or in combination with solid particles can cause erosion-corrosion attack in cast irons. Two methods are known to enhance the erosioncorrosion resistance of cast irons. First, the hardness of the cast irons can be increased through solid-solution hardening or phasetransformation-induced hardness increases. For example, 14.5% Si additions to cast irons cause substantial solid-solution hardening of the ferritic matrix. In such environments as the sulfate liquors encountered in the pulp and paper industry, this hardness increase enables highsilicon iron equipment to be successfully used, while lower-hardness unalloyed cast irons fail rapidly by severe erosion-corrosion. Use of martensitic or white cast irons can also improve the erosion-corrosion resistance of cast irons as a result of hardness increases. Second, better inherent corrosion resistance can also be used to increase the erosion-corrosion resistance of cast irons. Austenitic nickel cast irons can have hardnesses similar to unalloyed cast irons but may exhibit better erosion resistance because of the improved inherent corrosion resistance of nickel-alloyed irons compared to unalloyed irons. Microstructure can also affect erosion-corrosion resistance slightly. Gray cast irons generally show better resistance than steels under erosion-corrosion conditions. This improvement is related to the presence of the graphite network in the gray cast iron. Iron is corroded from the gray iron matrix as in steel, but the graphite network that is not corroded traps corrosion products; this layer of corrosion products and graphite offers additional protection against erosion-corrosion attack. Flow-induced corrosion stemming from fluid velocity alone is another type of erosion-corrosion for steels and cast irons. In certain services where unalloyed or low-alloyed cast irons are used, their corrosion resistance is due to the formation of a thick, poorly adherent corrosion product rather than the usual passive oxide layer associated with the more common

corrosion-resistant alloys. Examples of such situations are concentrated sulfuric or hydrofluoric acids. In these services, the cast irons develop, respectively, a thick iron sulfate film or iron-fluoride film, and at low velocities these films remain intact and provide protection. However, at velocities greater than a couple feet per second, these films are washed away, allowing further corrosion of the cast irons. Microbiologically induced corrosion (MIC) is the corrosion of metals resulting from the activity of a variety of living microorganisms, which, as a result of their growth or metabolism, either produce corrosive wastes or participate directly in electrochemical reactions on the metal surfaces. This phenomenon is often associated with biofouling and corrosion of buried structures. Soils containing sulfate concentrations support conditions where MIC of cast iron pipe can occur. An Australian study estimates that 50% of all failures of buried metal were due to microbiological causes (Ref 4). Prevention is difficult, but cathodic protection and the use of protective coatings can be beneficial (Ref 5). Stress-corrosion cracking (SCC) is observed in cast irons under certain combinations of environment and stress. Because stress is necessary to initiate SCC and because design factors often limit stresses in castings to relatively low levels, SCC is not observed as often in cast irons as in other more highly stressed components. However, under certain conditions, SCC can be a serious problem. Because unalloyed cast irons are generally similar to ordinary steels in resistance to corrosion, the same environments that cause SCC in steels will likely cause problems in cast irons. Environments that may cause SCC in unalloyed cast irons include these solutions (Ref 6):  Sodium hydroxide (NaOH)  Sodium hydroxide-sodium silicate (NaOH          

Na2SiO2) Calcium nitrate (Ca(NO3)2) Ammonium nitrate (NH4NO3) Sodium nitrate (NaNO3) Mercuric nitrate (Hg(NO3)2) Mixed acids (H2SO4-HNO3) Hydrogen cyanide (HCN) Seawater Acidic hydrogen sulfide (H2S) Molten sodium-lead alloys Acid chloride Oleum (fuming H2SO4)

Graphite morphology can play an important role in SCC resistance in certain environments. In oleum, flake graphite structures present special problems. Acid tends to penetrate along graphite flakes and corrodes the iron matrix. The corrosion products formed build up internal pressure and eventually crack the iron. This problem is found in both gray cast irons and high-silicon cast irons, which have flake graphite morphologies. It is not seen in ductile cast irons that have nodular graphite shapes.

Resistance to Corrosive Environments No single grade of cast iron will resist all corrosive environments. However, a cast iron can be identified that will resist most of the corrosives commonly used in industrial environments. Cast irons suitable for the more common corrosive environments are discussed as follows. Sulfuric Acid. Unalloyed, low-alloyed, and high-nickel austenitic as well as high-silicon cast irons are used in H2SO4 applications. Use of unalloyed and low-alloyed cast irons is limited to low-velocity, low-temperature concentrated (> 70%) H2SO4 service. Unalloyed cast iron is rarely used in dilute or intermediate concentrations, because corrosion rates are substantial. In concentrated H2SO4, as well as other acids, ductile iron is generally considered superior to gray iron, and ferritic matrix irons are superior to pearlitic matrix irons. In hot, concentrated acids, graphitization of the gray iron can occur. In oleum, unalloyed gray iron will corrode at very low rates. However, acid will penetrate along the graphite flakes, and the corrosion product that forms can build up sufficient pressure to split the iron. Interconnecting graphite is believed to be necessary to cause this form of cracking; therefore, ductile and malleable irons are generally acceptable for oleum service. Some potential for galvanic corrosion between cast iron and steel has been reported in 100% H2SO4. High-nickel austenitic cast irons exhibit acceptable corrosion resistance in room-temperature and slightly elevated-temperature H2SO4 service. As shown in Fig. 3, their performance is adequate over the entire range of H2SO4 concentrations, but they are a second choice compared to high-silicon cast irons. High-silicon cast irons are the best choice among the cast irons and perhaps among the

Corrosion of high-nickel austenitic cast iron in H2SO4 as a function of acid concentration and temperature. Source: Ref 6

Fig. 3

Corrosion of Cast Irons / 167 commonly available engineering material for resistance to H2SO4. This material has good corrosion resistance to the entire H2SO4 concentration range at temperatures to boiling (Fig. 4). Rapid attack occurs at concentrations over 100% and in service containing free sulfur trioxide (SO3). High-silicon cast irons are relatively slow to passivate in H2SO4 service. Corrosion rates are relatively high for the first 24 to 48 h of exposure and then decrease to very low steady-state rates (Fig. 2). Nitric Acid. All types of cast iron, except high-nickel austenitic iron, find some applications in HNO3. The use of unalloyed cast iron in HNO3 is limited to low-temperature, lowvelocity concentrated acid service. Even in this service, caution must be exercised to avoid dilution of acid because the unalloyed and low-alloyed cast irons both corrode very rapidly in dilute or intermediate concentrations at any temperature. High-nickel austenitic cast irons exhibit essentially the same resistance as unalloyed cast iron to HNO3 but cannot be economically justified for this service. High-chromium cast irons with chromium contents over 20% give excellent resistance to HNO3, particularly in dilute concentrations (Fig. 5). High-temperature boiling solutions attack these grades of cast iron. High-silicon cast irons also offer excellent resistance to HNO3. Resistance is exhibited over essentially all concentration and temperature ranges, with the exception of dilute, hot acids (Fig. 6). High-silicon cast iron equipment has been used for many years in the manufacture and handling of HNO3 mixed with other chemicals, such as H2SO4, sulfates, and nitrates. Contamination of HNO3 with HF, such as might be experienced in pickling solutions, may accelerate attack of the high-silicon iron to unacceptable levels. Hydrochloric Acid. Use of cast irons is relatively limited in HCl. Unalloyed cast iron is unsuitable for any HCl service. Rapid corrosion occurs at a pH of 5 or lower, particularly if appreciable velocity is involved. Aeration or oxidizing conditions, such as the presence of metallic salts, result in rapid destructive attack of unalloyed cast irons, even in very dilute HCl solutions.

High-nickel austenitic cast irons offer some resistance to all HCl concentrations at room temperature or below. High-chromium cast irons are not suitable for HCl services. High-silicon cast irons offer the best resistance to HCl of any cast iron. When alloyed with 4 to 5% Cr, high-silicon cast iron is suitable for all concentrations of HCl at temperatures up to 28  C (80  F). When high-silicon cast iron is alloyed with chromium, molybdenum, and higher silicon levels, the temperature for use can be increased (Fig. 7). In concentrations up to 20%, ferric ions (Fe3+) or other oxidizing agents inhibit corrosion attack on highsilicon cast iron alloyed with chromium. At over 20% acid concentration, oxidizers accelerate attack on the alloy. As in H2SO4, corrosion rates of high-silicon cast iron are initially high in the first 24 to 48 h of exposure then decrease to very low steady-state rates (Fig. 2). Phosphoric Acid. All cast irons find some application in H3PO4 service, but the presence of contaminants must be carefully evaluated before selecting a material. Unalloyed cast iron finds little use in H3PO4, with the exception of concentrated acids. Even in concentrated acids, use may be severely limited by the presence of fluorides, chlorides, or H2SO4. High-nickel cast irons find some application in H3PO4 at and slightly above room temperature. These cast irons can be used over the entire H3PO4 concentration range. Impurities in the acid may greatly restrict the applicability of this grade of cast iron. High-chromium cast irons exhibit generally low rates of attack in H3PO4 up to 60% concentration and are commonly used in the phosphate mining industry where abrasion resistance is needed. High-silicon cast irons show good-toexcellent resistance at all concentrations and temperatures for pure acid. The presence of fluoride ions (F) in H3PO4 makes the high-silicon irons unacceptable for use.

Corrosion of high-silicon cast iron in H2SO4 as a function of acid concentration and temperature

Fig. 5

Fig. 4

Organic acids and compounds are generally not as corrosive as mineral acids. Consequently, cast irons find many applications in handling these materials. Unalloyed cast iron can be used to handle concentrated acetic acid, CH3COOH, and fatty acids but will be attacked by more dilute solutions. Unalloyed cast irons are used to handle methyl, ethyl, butyl, and amyl alcohols. If the alcohols are contaminated with water and air, discoloration of the alcohols may occur. Unalloyed cast irons can also be used to handle glycerine, although slight discoloration of the glycerine may result. Austenitic nickel cast irons exhibit adequate resistance to CH3COOH, oleic acid, and stearic acid. High-chromium cast irons are adequate for CH3COOH but will be more severely corroded by formic acid (HCOOH). High-chromium cast irons are excellent for lactic and citric acid solutions. High-silicon cast irons show excellent resistance to most organic acids, including HCOOH and oxalic acid, in all temperature and concentration ranges. High-silicon cast irons also exhibit excellent resistance to alcohols and glycerine. Alkali solutions require material selections that are distinctly different from those of acid solutions. Alkalis include sodium hydroxide (NaOH), potassium hydroxide (KOH), sodium silicate (Na2SiO3), and similar chemicals that contain sodium, potassium, or lithium. Unalloyed cast irons exhibit generally good resistance to alkalis—approximately equivalent

Fig. 6

Corrosion of high-chromium cast iron in HNO3 as a function of acid concentration and temperature. Source: Ref 6

Corrosion of high-silicon cast iron in HNO3 as a function of acid concentration and temperature

Isocorrosion diagram for two high-silicon cast irons in HCl. A, Fe-14.3Si-4Cr-0.5Mo; B, Fe16Si-4Cr-3Mo

Fig. 7

168 / Casting Design and Performance to that of steel. These unalloyed cast irons are not attacked by dilute alkalis at any temperature. Hot alkalis at concentrations exceeding 30% attack unalloyed iron. Temperatures should not exceed 80  C (175  F) for concentrations up to 70% if corrosion rates of less than 0.25 mm/yr (10 mils/yr) are desired. Ductile and gray iron exhibit approximately equal resistance to alkalis. However, ductile cast iron is susceptible to cracking in highly alkaline solutions, but gray cast iron is not. Alloying with 3 to 5% Ni substantially improves the resistance of cast irons to alkalis. High-nickel austenitic cast irons offer even better resistance to alkalis than unalloyed or low-nickel cast irons. High-silicon cast irons show good resistance to relatively dilute solutions of NaOH at moderate temperatures but should not be applied for more concentrated conditions at elevated temperatures. High-silicon cast irons are usually economical over unalloyed and nickel cast irons in alkali solutions only when other corrosives are involved for which the lesser alloys are unsuitable. High-chromium cast irons have inferior resistance to alkali solutions and are generally not recommended for alkali services. Atmospheric corrosion is basically of interest only for unalloyed and low-alloy cast irons. Atmospheric corrosion rates are determined by the relative humidity and the presence of various gases and solid particles in the air. In high humidity, sulfur dioxide (SO2) or similar compounds found in many industrialized areas and chlorides found in marine atmospheres increase the rate of atmospheric attack on cast irons. Cast irons typically exhibit very low corrosion rates in industrial atmospheres—generally under 0.13 mm/yr (5 mils/yr)—and the cast irons are usually found to corrode at lower rates than steel structures in the same environment. White cast irons show the lowest rate of atmospheric corrosion of the unalloyed cast irons. Pearlitic cast irons are generally more resistant that ferritic cast irons to atmospheric corrosion. In marine atmospheres, unalloyed cast irons also exhibit relatively low rates of corrosion. Low alloy additions are sometimes made to improve corrosion resistance further. Higher alloy additions are even more beneficial but are rarely warranted. Gray cast iron offers some added resistance over ductile cast iron in marine atmospheres. Corrosion in Soils. Cast iron use in soils, as in atmospheric corrosion, is basically limited to unalloyed and low-alloyed cast irons. Corrosion in soils is a function of soil porosity, drainage, and dissolved constituents in the soil. Irregular soil contact can cause pitting, and poor drainage increases corrosion rates substantially above the rates in well-drained soils. Neither metal-matrix nor graphite morphology has an important influence on the corrosion of cast irons in soils. Some alloying additions are made to improve the resistance of cast irons to attack in soils. For example, 3% Ni additions to cast iron are made to reduce initial attack in cast irons in poorly drained soils. Alloyed cast

irons would exhibit better resistance than unalloyed or low-alloyed cast irons but are rarely needed for soil applications, because unalloyed cast irons generally have long service lives, particularly if coatings and cathodic protection are used. Anodes placed in soils for impressed current cathodic protection are frequently constructed from high-silicon cast iron. The highsilicon cast iron is not needed to resist the basic soil environment but rather to extend service life when subjected to the high electrical current discharge rates commonly used in cathodic protective anodes. Several thousands of miles of cast iron pipe have been buried underground for decades, handling water distribution and collection for hundreds of municipalities. Much of this pipe is reaching the end of its useful life. Fortunately, technologies have been developed to line cast iron pipe in situ with polymer linings such as polyurethane or cement mortar (Ref 7, 8). These cure-in-place systems provide an economical alternative to open trench replacement, and the old cast iron pipe can still provide many years of structural integrity for the polymer or cement liners. Corrosion in Water. Unalloyed and lowalloyed cast irons are the primary cast irons used in water service. The corrosion resistance of unalloyed cast iron in water is determined by its ability to form protective scales. In hard water, corrosion rates are generally low because of the formation of calcium carbonate (CaCO3) scales on the surface of the iron. In softened or deionized water, the protective scales cannot be fully developed, and some corrosion will occur. In industrial waste waters, corrosion rates are primarily a function of the contaminants present. Acid pH waters increase corrosion, but alkaline pH waters lower rates. Chlorides increase the corrosion rates of unalloyed cast irons, although the influence of chlorides is small at a neutral pH. Seawater presents some special problems for cast irons. Gray cast iron may experience graphitic corrosion in calm seawater. It will also be galvanically active, that is, anodic, in contact with most stainless steels, copper-nickel alloys, titanium, and chrome-molybdenum nickel-base alloys. Because these materials are frequently used in seawater structures, this potential for galvanic corrosion must be considered. In calm seawater, the corrosion resistance of cast iron is not greatly affected by the presence of crevices. However, intermittent exposure to seawater is very corrosive to unalloyed cast irons. Use of high-alloy cast irons in water is relatively limited. High-nickel austenitic cast irons are used to increase the resistance of cast iron components to pitting in calm seawater. Chromium containing high-silicon cast iron is used to produce anodes for the anodic protection systems used in seawater and brackish water. Corrosion in Saline Solutions. The presence of salts in water can have dramatic effects on the selection of suitable grades of cast iron. Unalloyed cast irons exhibit very low corrosion

rates in such salts as cyanides, silicates, carbonates, and sulfides, which hydrolyze to form alkaline solutions. However, in salts such as ferric chloride (FeCl3), cupric chloride (CuCl2), stannic salts, and mercuric salts, which hydrolyze to form acid solutions, unalloyed cast irons experience much higher rates. In salts that form dilute acid solutions, high-nickel cast irons are acceptable. More acidic and oxidizing salts, such as FeCl3, usually necessitate the use of high-silicon cast irons. Chlorides and sulfates of alkali metals yield neutral solutions, and unalloyed cast iron experiences very low corrosion rates in these solutions. More highly alloyed cast irons also exhibit low rates but cannot be economically justified for this application. Unalloyed cast irons are suitable for oxidizing salts, such as chromates, nitrates, nitrites, and permanganates, when the pH is neutral or alkaline. However, if the pH is less than 7, corrosion rates can increase substantially. At a lower pH with oxidizing salts, high-silicon cast iron is an excellent material selection. Ammonium salts are generally corrosive to unalloyed iron. High-nickel, high-chromium, and high-silicon cast irons provide good resistance to these salts. Other Environments. Unalloyed cast iron is used as a melting crucible for such low-melting metals as lead, zinc, cadmium, magnesium, and aluminum. Resistance to molten metals is summarized in Table 6. Ceramic coatings and washes are sometimes used to inhibit molten metal attack on cast irons. Cast iron can also be used in hydrogen chloride and chloride gases. In dry hydrogen chloride, unalloyed cast iron is suitable to 205  C (400  F), while in dry chlorine, unalloyed cast iron is suitable to 175  C (350  F). If moisture is present, unalloyed cast iron is unacceptable in HCl and Cl2 at any temperature.

Coatings Four general categories of coatings are used on cast irons to enhance corrosion resistance: metallic, organic, conversion, and enamel coatings. Coatings on cast irons are generally used to enhance the corrosion resistance of unalloyed and low-alloy cast irons and to lessen the requirements for cathodic protection. Highalloy cast irons such as Ni-Resist or white irons are rarely coated. Metallic coatings are used to enhance the corrosion resistance of cast irons. These coatings may either be sacrificial metal coatings, such as zinc, or barrier metal coatings, such as nickel-phosphorus. From a corrosion standpoint, these two classes of coatings have important differences. Sacrificial coatings are anodic when compared to iron, and the coatings corrode preferentially to protect the cast iron substrate. Small cracks and porosity in the coatings have a minimal overall effect on the performance of the coatings. Barrier coatings

Corrosion of Cast Irons / 169 are cathodic compared to iron, and the coatings can protect the cast iron substrate only when porosity or cracks are not present. If there are defects in the coatings, the service environment will attack the cast iron substrate at these imperfections, and the galvanic couple set up between the relatively inert coating and the casting may accelerate attack on the cast iron. Metallic coatings may be applied to cast irons by electroplating, hot dipping, flame or thermal spraying, diffusion coating, or hard facing. Table 7 lists the metals that can be applied by these techniques. Zinc is one of the most widely used coatings on cast irons. Although zinc is anodic to iron, its corrosion rate is very low, and it provides relatively long-term protection for the cast iron substrate. A small amount of zinc will protect a large area of cast iron. Zinc coatings provide optimal protection in rural and arid areas. Other metal coatings are also commonly used on cast irons. Cadmium provides atmospheric protection similar to that of zinc. Tin coatings are frequently used to improve the corrosion resistance of equipment intended for food handling, and aluminum coatings protect against corrosive environments containing sulfur fumes, organic acids, salts, and compounds of nitrate-phosphate chemicals. Lead and leadtin coating are primarily applied to enhance the corrosion resistance of iron castings to H2SO3 and H2SO4. Nickel-phosphorus diffusion coatings offer corrosion resistance approaching that obtainable with stainless steel. Organic coatings can be applied to cast irons to provide short-term or long-term corrosion resistance. Short-term rust preventatives include oil, solvent-petroleum-based inhibitors and film formers dissolved in petroleum solvents, emulsified-petroleum-based coatings modified to form a stable emulsion in water, and wax. For longer-term protection and resistance to more corrosive environments, rubber-based coatings, bituminous paints, asphaltic compounds, or thermoset and thermoplastic coatings can be applied. Rubber-based coatings include Table 6

chlorinated rubber neoprene, and Hypalon (DuPont Dow Elastomers). These coatings are noted for their mechanical properties and corrosion resistance but not for their decorative appearance. Bituminous paints have very low water permeability and provide high resistance to cast iron castings exposed to water. Use of bituminous paints is limited to applications that require good resistance to water, weak acids, alkalis, and salts. Asphaltic compounds are used to increase the resistance of cast irons to alkalis, sewage, acids, and continued exposure to tap water. Their application range is similar to that of bituminous paints. Cast irons are also lined with thermoset and thermoplastics, such as epoxy and polyethylene, to resist attack by fluids. Fluorocarbon coatings offer superior corrosion resistance except in abrasive services. Fluorocarbon coatings applied to cast irons include such materials as polytetrafluoroethylene (PTFE), perfluoroalkoxy resins (PFA), polyvinyldene fluoride (PVDF), ethylene chlorotrifluoroethylene (ECTFE), ethylene tetrafluoroethylene (ETFE), and fluorinated ethylene polypropylene (FEP). Fully fluorinated fluorocarbon coatings resist deterioration in most common industrial services and can be used to 205  C (400  F), whereas partially fluorinated coatings are limited to approximately 150  C (300  F). Cast iron lined with fluorocarbon polymers can be very competitive with stainless, nickel-base, and even titanium and zirconium materials in terms of range of services covered and product cost. Conversion coatings are produced when the metal on the surface of the cast iron reacts with another element or compound to produce an iron-containing compound. Common conversion coatings include phosphate coatings, oxide coatings, and chromate coatings. Phosphate coatings enhance the resistance of cast iron to corrosion in sheltered atmospheric exposure. If the surface of the casting is oxidized and black iron oxide or magnetite is formed, the corrosion resistance of the iron can be enhanced, particularly if the oxide layer is impregnated with oil

or wax. Chromate coatings are formed by immersing the iron castings in an aqueous solution of chromic acid (H2CrO4) or chromium salts. Chromate coatings are sometimes used as a supplement to cadmium plating in order to prevent the formation of powdery corrosion products. The overall benefits of conversion coatings are small with regard to atmospheric corrosion. Enamel Coatings. In the enamel coating of cast irons, glass frits are melted on the surface and form a hard, tenacious bond to the cast iron substrate. Good resistance to all acids except HF can be obtained with the proper selection and application of the enamel coating. Alkaline-resistant coatings can also be applied, but they offer only marginal improvement in the resistance to alkalis. Proper design and application are essential for developing enhanced corrosion resistance on cast irons with enamel coatings. Any cracks, spalling, or other coating imperfections may permit rapid attack of the underlying cast iron.

Selection of Cast Irons Cast irons provide excellent resistance to a wide range of corrosion environments when properly matched with that service environment. The basic parameters to consider before selecting cast irons for corrosion services include:  Concentration of solution components in

weight percent

 Dissolved contaminants, even at parts per

million levels

 pH of solution  Solution temperature, potential temperature

extremes, and rate of change of temperature

 Degree of solution aeration  Percent and type of solids suspended in the

solution

 Duty cycle, continuous or intermittent oper-

ation or exposure

Resistance of gray cast iron to liquid metals at 300 and 600  C (570 and 1110  F)

Liquid metal

Liquid metal melting point,  C

Mercury Sodium, potassium, and mixtures Gallium Bismuth-lead-tin Bismuth-lead Tin Bismuth Lead Indium Lithium Thallium Cadmium Zinc Antimony Magnesium Aluminum

38.8 12.3 to 97.9 29.8 97 125 321.9 271.3 327 156.4 186 303 321 419.5 630.5 651 660

Resistance of gray cast iron(a) 300  C (570  F)

Unknown Limited Unknown Good Unknown Limited Unknown Good at 327  C (621  F) Unknown Unknown Unknown Good at 321 C (610  F)

600  C (1110  F)

Unknown Poor Unknown Unknown Unknown Poor Unknown Unknown Unknown Unknown Unknown Good Poor Poor at 630.5  C (1167  F) Good at 651  C (1204  F) Poor at 660  C (1220  F)

(a) Good, considered for long-time use. 10 mils/yr); Unknown, no data for these temperatures. Source: Ref 9

Table 7 Summary of metallic coating techniques to enhance corrosion resistance of cast irons Coating technique

Electroplating

Hot dipped Hard facing

Flame spraying Diffusion coating

Metals/alloys applied

Cadmium, chromium, copper, lead, nickel, zinc, tin, tin-nickel, brass, bronze Zinc, tin, lead, lead-tin, aluminum Cobalt-base alloys, nickel-base alloys, metal carbides, high-chromium ferrous alloys, high-manganese ferrous alloys, high-chromium and nickel ferrous alloys Zinc, aluminum, lead, iron, bronze, copper, nickel, ceramics, cermets Aluminum, chromium, nickelphosphorus, zinc, nitrogen, carbon

170 / Casting Design and Performance  Potential for upset conditions, for example,

temperature and concentration excursions  Unusual conditions, such as high solution velocity or vacuum  Materials present in the system and the potential for galvanic corrosion Although it is advisable to consider each of the parameters before ultimate selection of a cast iron, the information needed to properly assess all variables of importance is often lacking. In such cases, introduction of test coupons of the candidate materials into the process stream should be considered before extensive purchases of equipment are made. If neither test coupons nor complete service data are viable alternatives, consultation with a reputable manufacturer of the equipment or the cast iron, with a history of applications in the area of interest, should be considered. ACKNOWLEDGMENT This article is adapted from “Corrosion of Cast Irons”, by Donald R. Stickle, Flowserve Corporation, Corrosion, Volume 13, ASM Handbook, ASM International, 1987, p 566–572.

REFERENCES

SELECTED REFERENCES

1. J.R. Keough, Austempered Ductile Iron, Section IV, Ductile Iron Data, The Ductile Iron Society, 1998 (available on website) 2. Development of a Cast Iron Graphitization Measurement Device, NYGAS Technol. Briefs, Issue 99–690-1, Jan 1999 3. J.R. McDowell, in Symposium on Fretting Corrosion, STP 144, American Society for Testing and Materials, 1952, p 24 4. P. Ferguson and D. Nicholas, Corros. Australas., April 1984, p 12 5. S.L. Chawla and R.K. Gupta, Materials Selection for Corrosion Control, ASM International 1993, p 56–59 6. E.C. Miller, Liquid Metals Handbook, 2nd ed., Government Printing Office, 1952, p 144 7. New Lining System Upgrades Boston Gas Distribution System, Pipeline Gas J., April 1995, p 24–29 8. B.B. Hall, Rehabilitation of 1940’s Water Mains, Am. Water Works Assoc., Vol 91 (No. 12), 1999, p 91–94 9. R.I. Higgins, Corrosion of Cast Iron, J. Res., Feb 1956, p 165–177

 S.A. Bradford, CASTI Practical Handbook of     



  

Corrosion Control in Soils, co-published by CASTI and ASTM, 2000 “Corrosion Control of Ductile and Cast Iron Pipe,” 37254, NACE, 2001 Corrosion Data Survey, 6th ed., National Association of Corrosion Engineers, 1985 J.R. Davis, Ed., ASM Specialty Handbook: Cast Irons, ASM International, 1996 M.G. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986 “High Silicon Iron Alloys for Corrosion Services,” Bulletin A/2, Flowserve Corporation, Aug 1998 Properties and Selection: Irons, Steels, and High-Performance Alloys, Vol 1, ASM Handbook, ASM International, 1990 M. Szeliga, Corrosion of Ductile Iron Piping, NACE 1995 C.F. Walton, Ed., The Gray Iron Castings Handbook, A.L. Garber, 1957 C.F. Walton, Ed., Gray and Ductile Iron Castings Handbook, R.R. Donnelley & Sons, 1971

Casting Design and Performance Pages 171–173

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Corrosion of Cast Carbon and Low-Alloy Steels* STEEL CASTING COMPOSITIONS are generally divided into the categories of carbon and low-alloy, corrosion-resistant, or heat-resistant, depending on alloy content and intended service. Castings are classified as corrosion resistant if they are capable of sustained operation when exposed to attack by corrosive agents at service temperatures normally below 315  C (600  F). Carbon and low-alloy steels, the subject of this article, are considered resistant only to very mild corrosives, while the various high-alloy grades are applicable for varying situations from mild to severe services, depending on the particular conditions involved. For design and materials selection, the specific rate of corrosion may not be as important as the predictability and confidence in predicting a rate of corrosion. It can be misleading to list the comparative corrosion rates of different alloys exposed to the same corroding medium. In this article, no attempt is made to recommend alloys for specific applications, and the data supplied should be used only as a general guideline. Alloy casting users will find it helpful to consult materials and corrosion specialists when selecting alloys for a particular application. The factors that must be considered in materials selection include:  The principal corrosive agents and their

concentrations

 Known or suspected impurities, including

abrasive materials and their concentration

However, discretion and caution are suggested in evaluating the relative corrosion rates of various steels because of uncertainties in the results from controlled laboratory tests and simulated service condition tests, as well as anomalies in the intended environment. The best information is obtained from equipment used under actual operating conditions. Cast carbon and low-alloy steel and wrought steel of similar composition and heat treatment exhibit approximately the same corrosion resistance in the same environments. More detailed information applicable to cost alloys may be found in Ref 1 and 2. Plain carbon steel and some of the low-alloy steels do not ordinarily resist drastic corrosive conditions, although there are some exceptions, such as concentrated sulfuric acid (H2SO4).

Atmospheric Corrosion Unless shielded by a protective coating, iron and steel corrode in the presence of water and oxygen; therefore, steel will corrode when it is exposed to moist air. The rate at which corrosion proceeds in the atmosphere depends on the corroding medium, the conditions of the particular location in which the material is in use, and the steps that have been taken to prevent corrosion. The rate of corrosion also depends on the charac-

ter of the steel as determined by its chemical composition and heat treatment. To increase the corrosion resistance of steel significantly, amounts of alloying elements are increased. Small amounts of copper and nickel slightly improve the resistance of steel to atmospheric attack, but appreciably larger amounts of other elements, such as chromium and nickel, improve corrosion resistance significantly. The rate of corrosion of a material in an environment can generally be estimated with confidence only from long-term tests. A 15-year research program compared the corrosion resistance of nine cast steels in marine and industrial atmospheres. Table 1 shows the compositions of the cast steels tested. The cast steel specimens exposed were 13 mm (½ in.) thick, 100 by 150 mm (4 by 6 in.) panels with beveled edges. The surfaces of half the specimens were machined. Specimens of each composition and surface condition were divided into three groups. One group was exposed to an industrial atmosphere at East Chicago, IN, and the other two groups were exposed to marine atmospheres 24 and 240 m (80 and 800 ft) from the ocean at Kure Beach, NC. The weight losses of the specimens during exposure were converted to corrosion rates in terms of millimeters (mils) per year. The results of this research are shown in Fig. 1 to 4. These are uniform corrosion rates that do not apply to localized corrosion modes, such as

 Average operating temperature, including

variations even if experienced only for short periods  Presence (or absence) of dissolved oxygen or other gases in solution  Continuous or intermittent operation  Fluid velocity Each of these can have a significant effect on the service life of cast equipment, and such detailed information must be provided to make the appropriate materials selection. Many rapid failures are traceable to these details being overlooked—often when the information was available. Selection of the most economical alloy can be made by the judicious use of corrosion data.

Table 1

Compositions of cast steels tested in atmospheric corrosion Composition(a),%

Cast steel

Carbon, grade A Nickel-chromium-molybdenum 1Ni-1.7Mn 2% Ni Carbon, grade B 1% Cu 1.36Mn-0.09V 1.42% Mn 1.5Mn-0.05Ti

Nl.

Cu

Mn

Cr

V

C

Mo

P

S

Si

Other

0.10 0.56 1.08 2.26 0.03 0.04 0.01 0.01 0.01

0.13 0.13 0.08 0.12 0.03 0.94 0.15 0.13 0.11

0.61 0.80 1.70 0.77 0.65 0.87 1.36 1.42 1.48

0.21 0.60 0.08 0.19 0.10 0.11 0.08 0.16 0.04

0.03 0.04 0.04 0.03 0.04 0.07 0.09 0.04 0.03

0.14 0.26 0.27 0.17 0.25 0.28 0.37 0.37 0.33

trace 0.15

0.016

0.026

0.02 0.017 0.011

0.023 0.021 0.021

0.031 0.027 0.016

0.038 0.022 0.025

0.41 0.44 0.42 0.65 0.51 0.42 0.34 0.38 0.40

0.05Ti

(a) All compositions contain balance of iron. Source: Ref 3

*Reprinted from ASM Handbook, Volume 13B, Corrosion: Materials, S.D. Cramer, B.S. Covino, Jr., editors, p 51–53

trace

172 / Casting Design and Performance

Corrosion rates of various cast steels in a marine atmosphere. Nonmachined specimens were exposed 24 m (80 ft) from the ocean at Kure Beach, NC. Source: Ref 3

Fig. 1

Corrosion rates of various cast steels exposed at the 240 m (800 ft) site at Kure Beach, NC. Specimens were not machined. Source: Ref 3

Fig. 2

Corrosion rates for cast steels in an industrial atmosphere. Nonmachined specimens were exposed at East Chicago, IN. Source: Ref 3

Fig. 3

crevice corrosion, pitting, or local galvanic coupling. Figure 5 shows the results of another portion of this project. Corrosion rates for a 3-year exposure of various cast steels, wrought steels, and malleable iron in both atmospheres are compared. The following conclusions can be drawn from these tests:  The condition of the specimen surface has









no significant effect on the corrosion resistance of cast steels. Unmachined surfaces with the casting skin intact have corrosion rates similar to those of machined surfaces, regardless of the atmospheric environment. The highest corrosion rate occurs in the marine atmosphere 24 m (80 ft) from the ocean, with lower but similar corrosion rates occurring in the industrial atmosphere and the marine atmosphere 240 m (800 ft) from the ocean. The corrosion rate of cast steel decreases as a function of time, because corrosion products (scale and rust coating) build up and act as a protective coating on the cast steel surface. However, the corrosion rate of the most resistant cast steel (2% Ni) is always less than that of lesser corrosion-resistant cast steels. Cast steels with small amounts of copper or chromium, or slightly larger amounts of nickel, have corrosion resistance superior to that of cast carbon steel with manganese as an alloying element, when exposed to the atmospheres (Ref 3). Increasing the nickel and the chromium contents of cast steel increases the corrosion resistance in all three of the atmospheric environments.

All cast steels have greater corrosion resistance than malleable iron in industrial atmospheres and are superior or equivalent to the

Corrosion rates of machined and nonmachined specimens of cast steels after 7 years in three environments. The effect of surface finish on corrosion rates is negligible. Source: Ref 3

Fig. 4

Comparison of corrosion rates of cast steels, malleable cast iron, and wrought steel after 3 years of exposure in two atmospheres. Source: Ref 3

Fig. 5

Table 2 Corrosion of cast carbon and alloy steels in steam at 650  C (1200  F) for 570 h Composition, % Type of steel

Carbon Carbon-molybdenum Nickel-chromium-molybdenum 5Cr-molybdenum 7Cr-molybdenum(a) 9Cr-1.5Mo (a) Not a cast steel. Source: Ref 3

C

0.24 0.25 0.21 0.20 0.35 0.28 0.22 0.27 0.11 0.23

Cr

0.64 0.73 5.07 5.49 7.33 9.09

Ni

2.13 2.25

Average penetration rate Mo

mm/yr

mlls/yr

0.49 0.49 0.26 0.26 0.47 0.43 0.59 1.56

0.3 0.28 0.3 0.25 0.25 0.25 0.1 0.1 0.05 0.025

12 11 12 10 10 10 4 4 2 1

Corrosion of Cast Carbon and Low-Alloy Steels / 173 Table 3 Petroleum corrosion resistance of cast steels 1000 h test in petroleum vapor under 780 N (175 lb) of pressure at 345  C (650  F) Weight loss Type of material

mg/cm2

Cast carbon steel Cast steel, 2Ni-0.75Cr Seamless tubing, 5% Cr Cast steel, 5Cr-1W Cast steel, 5Cr-0.5Mo Cast steel, 12% Cr Stainless steel. 18Cr-8Ni

3040 2370 1540 950 730 6.4 2.1

mg/ln.2

196 153 99.2 61.5 47 100 30

Table 4 Corrosion of cast steels in waters Corrosive medium

Tap water Seawater Alternate immersion and drying Hot water 0.05% H2SO4 0.50% H2SO4

Corrosion factor(a)

Exposure time, months

Fe-0.29C-0.69Mn-0.44SI

Fe-0.32C-0.66Mn-1.12Cr

Fe-0.11C-0.41Mn-3.58Cr

2 6 2 6 2 6 1 2 6 2

100 100 100 100 100 100 100 100 100 100

85 73 60 80 93 109 100 71 89 223

58 61 26 40 30 25 64 68 102 61

Source: Ref 3

(a) Corrosion factor is the ratio of average penetration rate of the alloy in question to Fe-0.29C-0.69Mn-0.44Si steel. Source: Ref 3

wrought steels in this environment. The corrosion rate in the marine atmosphere depends primarily on the alloy content. The cast carbon steel is much superior to the AISI 1020 wrought steel but is slightly inferior to malleable iron (Ref 4).

Table 5

Corrosion of cast chromium and carbon steels in mineral acids Weight loss in 5 h 5% H2SO4

Steel

Carbon steel, 0.31% C Chromium steel, 0.30C-2.42Cr

5% HCl

5% HNO3

mg/cm2

mg/ln.2

mg/cm2

mg/ln.2

mg/cm2

mg/ln.2

2.7 4.9

17.42 31.6

2.1 5.41

13.55 34.9

80.79 47.36

521.1 305.5

Source: Ref 3

Other Environments Several low- and high-alloy cast steels have been studied regarding their corrosion resistance to high-temperature steam. Test specimens 150 mm (6 in.) in length and 13 mm (½ in.) in diameter were machined from test coupons and then exposed to steam at 650  C (1200  F) for 570 h. The steel compositions and test results are given in Table 2. Table 3 shows the resistance of cast steels to petroleum corrosion, and Tables 4 and 5 supply similar data relating to water and acid attack. These data show the value of higher chromium content for improved corrosion resistance.

ACKNOWLEDGMENT This article has been adapted from Raymond Monroe and Steven Pawel, Corrosion of Cast Steels, Corrosion, Volume 13, ASM Handbook, ASM International, 1987.

REFERENCES 1. T. Kodomd, Corrosion of Wrought Carbon Steels, Corrosion: Materials, Vol 13B,

ASM Handbook, ASM International, 2005, p 5–10 2. T. Oakwood, Corrosion of Wrought LowAlloy Steels Corrosion: Materials, Vol 13B, ASM Handbook, ASM International, 2005, p 11–27 3. C.W. Briggs, Ed., Steel Casting Handbook, 4th ed., Steel Founders’ Society of America, 1970, 662–667 4. C.W. Briggs, “Atmospheric Corrosion of Carbon and Low Alloy Cast Steels,” ASTM Corrosion Symposium, 1967

Casting Design and Performance Pages 175–183

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Corrosion of Cast Stainless Steels* Revised by Malcolm Blair, Steel Founders’ Society of America

CAST STAINLESS STEELS are usually specified on the basis of composition by using the alloy designation system established by the Alloy Casting Institute (ACI). The ACI designations, such as CF-8M, have been adopted by ASTM International and are preferred for cast alloys over the designations originated by the American Iron and Steel Institute (AISI) for similar wrought steels. The first letter of the ACI designation indicates whether the alloy is intended primarily for liquid corrosion service (C) or heat-resistant service (H). The second letter denotes the nominal chromium-nickel type, as shown in Fig. 1. As the nickel content increases, the second letter in the ACI designation increases from A to Z. The numerals following the two letters refer to the maximum carbon content (percent  100) of the alloy. If additional alloying elements are included, they can be denoted by the addition of one or more letters after the maximum carbon content. Thus, the designation CF-8M refers to an alloy for corrosion-resistant service (C) of the 19Cr-9Ni (F) type, with a maximum carbon content of 0.08% and containing molybdenum (M). Corrosion-resistant cast stainless steels are also often classified on the basis of microstructure. The classifications are not completely independent, and a classification by composition often involves microstructural distinctions. Cast corrosion- and heat-resistant alloy compositions are listed in Table 1.

Composition and Microstructure The principal alloying element in the highalloy family is usually chromium, which, through the formation of protective oxide films, results in the corrosion protection or stainless behavior. For most purposes, stainless behavior requires at least 12% Cr. Corrosion resistance further improves with additions of chromium to at least the 30% level. As indicated in Table 1, significant amounts of nickel and lesser amounts of molybdenum and other elements are often added to the iron-chromium matrix.

Although chromium is a ferrite and martensite promoter, nickel is an austenite promoter. By varying the amounts and ratios of these two elements (or their equivalents), almost any desired combination of microstructure, strength, or other property can be achieved. Varying the temperature, time at temperature, and cooling rate of the heat treatment also controls the desired results. It is useful to think of the compositions of high-alloy steels in terms of the balance between austenite promoters and ferrite promoters. This balance is shown in the widely used Schaeffler diagrams (Fig. 2). It should be noted that the Schaeffler diagram is used for welding and that the phases shown are those that persist after cooling to room temperature at rates consistent with fabrication (Ref 1, 2). The Schoefer diagram (Fig. 3) gives an indication of the amount of ferrite that may be expected based on the composition of the alloy in question. An ASTM standard provides note on the Schoefer diagram and methods for estimating ferrite content (Ref 3). The empirical correlations shown in Fig. 2 can be understood from the following. The field designated as martensite encompasses such alloys as CA-15, CA-6NM, and even CB-7Cu. These alloys contain 12 to 17% Cr, with adequate nickel, molybdenum, and carbon to promote high hardenability, that is, the ability to transform completely to martensite when cooled at even the moderate rates associated with the air cooling of heavy sections. High alloys have low thermal conductivities and cool slowly. To obtain the desired properties, a full heat treatment is required after casting; that is, the casting is austenitized by heating to 870 to 980  C (1600 to 1800  F), cooled to room temperature to produce the hard martensite, and then tempered at 595 to 760  C (1100 to 1400  F) until the desired combination of strength, toughness, ductility, and resistance to corrosion or stress corrosion is obtained (Ref 1, 2). Increasing the nickel equivalent (moving vertically in Fig. 2) eventually results in an alloy that is fully austenitic, such as CC-20, CH-20, CK-20, or CN-7M. These alloys are extremely ductile, tough, and corrosion resistant. On the

* Reprinted from ASM Handbook, Volume 13B; Corrosion: Materials S.D. Cramer, B.S. Covino, Jr., editors, p78–87

other hand, the yield and tensile strength may be relatively low for the fully austenitic alloys. Being fully austenitic, they are nonmagnetic. Heat treatment consists of a single step: water quenching from a relatively high temperature at which carbides have been taken into solution. Solution treatment may also homogenize the structure, but because no transformation occurs, there can be no grain refinement. The solutionizing step and rapid cooling ensure maximum resistance to corrosion. Temperatures between 1040 and 1205  C (1900 and 2200  F) are usually required (Ref 1, 2). Adding chromium to the lean alloys (proceeding horizontally in Fig. 2) stabilizes the dferrite that forms when the casting solidifies. Examples are CB-30 and CC-50. With high chromium content, these alloys have relatively good resistance to corrosion, particularly in sulfur-bearing atmospheres. However, being single-phase, consisting only of ferrite, they are nonhardenable, have moderate-to-low strength, and are often used as-cast or after only a simple solution heat treatment. Ferritic alloys also have relatively poor impact resistance (toughness) (Ref 1, 2). Between the fields designated M, A, and F in Fig. 2 are regions indicating the possibility of two or more phases in the alloys. Commercially, the most important of these alloys are the ones in which austenite and ferrite coexist, such as CF-3, CF-8, CF-3M, CF-8M, CG-8M, and CE-30. These alloys usually contain 3 to

Chromium and nickel contents in ACI standard grades of heat- and corrosion-resistant castings. See text for details. Source: Ref 1

Fig. 1

176 / Casting Design and Performance Table 1 Compositions of Alloy Casting Institute (ACI) heat- and corrosion-resistant casting alloys Composition (balance Iron)(b), % ACI designation

UNS No.

CA-15 CA-15M CA-40 CA-6NM CA-6N CB-30 CB-7Cu-1 CB-7Cu-2 CC-50 CD-4MCu CE-30 CF-3 CF-8 CF-20 CF-3M CF-8M CF-8C CF-16F CG-12 CG-8M CH-20 CK-20 CN-7M CN-7MS CW-12M CY-40 CZ-100 N-12M M-35 HA HC HD HE HF HH HI HK HL HN HP HP-50WZ HT HU HW HX

J91150 J91151 J91153 J91540 J91650 J91803 ... ... J92615 ... J93423 J92500 J92600 J92602 J92800 J92900 J92710 J92701 J93001 ... J93402 J94202 N08007 ... N30002 N06040 N02100 ... ... ... J92605 J93005 J93403 J92603 J93503 J94003 J94224 J94604 J94213 ... ... J94605 N08004 N08001 N06006

Wrought alloy type(a)

410 ... 420 ... ... 431 ... ... 446 ... ... 304L 304 302 316L 316 347 303 ... 317 309 310 ... ... ... ... ... ... ... ... 446 327 ... 302B 309 ... 310 ... ... ... ... 330 ... ... ...

C

0.15 0.15 0.20–0.40 0.06 0.06 0.30 0.07 0.07 0.50 0.04 0.30 0.03 0.08 0.20 0.03 0.08 0.08 0.16 0.12 0.08 0.20 0.20 0.07 0.07 0.12 0.40 1.00 0.12 0.35 0.20 0.50 0.50 0.20–0.50 0.20–0.40 0.20–0.50 0.20–0.50 0.20–0.60 0.20–0.60 0.20–0.50 0.35–0.75 0.45–0.55 0.35–0.75 0.35–0.75 0.35–0.75 0.35–0.75

Mn

1.00 1.00 1.00 1.00 0.50 1.00 0.70 0.70 1.00 1.00 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 2.00 1.50 1.00 1.00 1.50 1.50 1.00 1.50 0.35–0.65 1.00 1.50 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00

Si

1.50 0.65 1.50 1.00 1.00 1.50 1.00 1.00 1.50 1.00 2.00 2.00 2.00 2.00 1.50 2.00 2.00 2.00 2.00 1.50 2.00 2.00 1.50 2.50–3.50 1.50 3.00 2.00 1.00 2.00 1.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.50 2.00 2.50 2.50 2.50 2.50

P

0.04 0.04 0.04 0.04 0.02 0.04 0.035 0.035 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.17 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.04 0.03 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04

S

0.04 0.04 0.04 0.03 0.02 0.04 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04

Cr

Ni

11.5–14.0 11.50–14.0 11.5–14.0 11.5–14.0 10.5–12.0 18.0–21.0 14.0–15.5 14.0–15.5 26.0–30.0 24.5–26.5 26.0–30.0 17.0–21.0 18.0–21.0 18.0–21.0 17.0–21.0 18.0–21.0 18.0–21.0 18.0–21.0 20.0–23.0 18.0–21.0 22.0–26.0 23.0–27.0 19.0–22.0 18.0–20.0 15.5–20.0 14.0–17.0 ... 1.0 ... 8.0–10.0 26.0–30.0 26.0–30.0 26.0–30.0 18.0–23.0 24.0–28.0 26.0–30.0 24.0–28.0 28.0–32.0 19.0–23.0 24.0–28.0 24.0–28.0 15.0–19.0 17.0–21.0 10.0–14.0 15.0–19.0

1.00 1.00 1.00 3.5–4.5 6.0–8.0 2.00 4.5–5.5 4.5–5.5 4.00 4.75–6.00 8.0–11.0 8.0–21.0 8.0–11.0 8.0–11.0 9.0–13.0 9.0–12.0 9.0–12.0 9.0–12.0 10.0–13.0 9.0–13.0 12.0–15.0 19.0–22.0 27.5–30.5 22.0–25.0 bal bal bal bal bal ... 4.00 4.0–7.0 8.0–11.0 8.0–12.0 11.0–14.0 14.0–18.0 18.0–22.0 18.0–22.0 23.0–27.0 33.0–37.0 33.0–37.0 33.0–37.0 37.0–41.0 58.0–62.0 64.0–68.0

Other elements

0.5Mo(c) 0.15–1.00Mo 0.5Mo(c) 0.4–1.0Mo

... ... 0.15–0.35Nb, 0.05N, 2.5–3.2Cu 0.15–0.35Nb, 0.05N, 2.5–3.2Cu ... 1.75–2.25Mo, 2.75–3.25Cu ... ... ... ... 2.0–3.0Mo 2.0–3.0Mo 3  C min, 1.0 max Nb 1.5Mo, 0.2–0.35Se ... 3.0–4.0Mo ... ... 2.0–3.0Mo, 3.0–4.0Cu 2.0–3.0Mo, 1.5–2.0Cu 7.5Fe 11.0Fe 3.0Fe, 1.25Cu 0.26–0.33Mo, 0.60V, 2.50Co, 6.0Fe 28–33Cu, 3.5Fe 0.90–1.20Mo 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c), 0.2N 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 4.0–6.0W, 0.2–1.0Zr 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c)

(a) Cast alloy chemical composition ranges are not the same as the wrought composition ranges; buyers should use cast alloy designations for proper identification of castings. (b) Maximum, unless range is given. (c) Molybdenum not intentionally added

Fig. 2

Schaeffler diagram showing the amount of ferrite and austenite present in weldments as a function of chromium and nickel equivalents. Source: Ref 1

30% ferrite in a matrix of austenite. Predicting and controlling ferrite content is vital to the successful application of these materials. Alloys that contain both ferrite and austenite offer superior strength, weldability, and corrosion resistance compared to alloys that contain only austenite. Strength, for example, increases directly with ferrite content. Achieving specified minimums may necessitate controlling the ferrite within narrow bands. Figure 3 and Schoefer’s equations are used for this purpose. These duplex alloys should be solution heat treated and rapidly cooled before use to ensure maximum resistance to corrosion (Ref 1, 2). The presence of ferrite is not beneficial for every application. Ferrite tends to reduce toughness, although this is not of great concern, given the extremely high toughness of the austenite matrix. However, in applications that require exposure to elevated temperatures, usually 315  C (600  F) and higher, the metallurgical changes associated with the ferrite can be

Corrosion of Cast Stainless Steels / 177

Schoefer diagram for estimating the average ferrite content in austenitic iron-chromiumnickel alloy castings. Source: Ref 1

Fig. 3

Effect of chromium on oxidation resistance of cast steels. Speciments (13 mm, or 0.5 in., cubes) were exposed for 48 h at 1000  C (1830  F). Source: Ref 2

Fig. 4

Toughness of solution-annealed duplex stainless steel castings (closed symbols) and companion wrought alloys (open symbols) as a function of test temperature. Source: Ref 4

Fig. 6

Corrosion behavior of ACl H-type (heat-resistant) alloy castings in (a) air and in (b) oxidizing flue gases containing 5 grains of sulfur per 2.8 m3 (100 ft3) of gas. Source: Ref 2

Fig. 5

severe and detrimental. In the low end of this temperature range, the observed reductions in toughness have been attributed to carbide precipitation or reactions associated with 475  C (885  F) embrittlement. The 475  C (885  F) embrittlement is caused by precipitation of an intermetallic phase with a composition of approximately 80Cr-20Fe. The name derives from the fact that this embrittlement is most severe and rapid when it occurs at approximately 475  C (885  F). At 540  C (1000  F) and above, the ferrite phase may transform to a complex Fe-Cr-Ni-Mo intermetallic compound known as sigma (s) phase, which reduces toughness, corrosion resistance, and creep ductility. The extent of the reduction increases with time and temperature to approximately 815  C (1500  F) and may persist to 925  C (1700  F). In extreme cases, Charpy V-notch energy at room temperature may be reduced 95% from its initial value (Ref 1, 2). It has been demonstrated that

the impact properties of duplex stainless steels in the solution heat treated condition, in the cast and wrought form, are comparable (Fig. 4). More information on the metallography and microstructures of these alloys is available in Ref 5.

Corrosion Behavior of H-Type Alloys The ACI heat-resistant (H-type) alloys must be able to withstand temperatures exceeding 1095  C (2000  F) in the most severe high-temperature service. Chromium content is important to the corrosion behavior of these alloys. Chromium imparts resistance to oxidation and sulfidation at high temperatures by forming a passive oxide film. Heat-resistant casting alloys must also have good resistance to carburization.

More information on the corrosion of metals and alloys in high-temperature gases is available in Ref 6. Oxidation. Resistance to oxidation increases directly with chromium content (Fig. 5). For the most severe service at temperatures above 1095  C (2000  F), 25% or more chromium is required. Additions of nickel, silicon, manganese, and aluminum promote the formation of relatively impermeable oxide films that retard further scaling. Thermal cycling is extremely damaging to oxidation resistance, because it leads to breaking, cracking, or spalling of the protective oxide film. The best performance is obtained with austenitic alloys containing 40 to 50% combined nickel and chromium Figure 6 shows the behavior of the H-type grades. Sulfidation environments are becoming increasingly important. Petroleum processing,

178 / Casting Design and Performance 15M, CA-6NM, CA-6NM-B, CA-40, CB-7Cu-1, and CB-7Cu-2. These alloys are generally used in applications requiring high strength and modest corrosion resistance. Alloy CA-15 typically exhibits a microstructure of martensite and ferrite. This alloy contains the minimum amount of chromium to be considered a stainless steel (11 to 14% Cr) and as such may not be used in aggressive environments. It does, however, exhibit good atmospheric corrosion resistance, and it resists staining by many organic environments. Alloy CA-15M may contain slightly more molybdenum than CA-15 (up to 1% Mo) and therefore may have improved general corrosion resistance in relatively mild environments. Alloy CA-6NM is similar to CA-15M except that it contains more nickel and molybdenum, which improves its general corrosion resistance. Alloy CA-6NM-B is a lower-carbon version of this alloy. The lower strength level promotes resistance to sulfide stress cracking. Alloy CA-40 is a higher-strength version of CA-15, and it also exhibits excellent atmospheric-corrosion resistance after a normalize and temper heat treatment. Micro-structurally, the CB-7Cu alloys usually consist of mixed martensite and ferrite, and because of the increased chromium and nickel levels compared to the other martensitic alloys, they offer improved corrosion resistance to seawater and some mild acids. These alloys also have good atmospheric-corrosion resistance. The CB-7Cu alloys are hardenable and offer the possibility of increased strength and improved corrosion resistance among the martensitic alloys. General Corrosion of Ferritic Alloys. Alloys CB-30 and CC-50 are higher-carbon

and higher-chromium alloys than the CA alloys previously mentioned. Each alloy is predominantly ferritic, although a small amount of martensite may be found in CB-30. Alloy CB-30 contains 18 to 21% Cr and is used in chemical-processing and oil-refining applications. The chromium content is sufficient to have good corrosion resistance to many acids, including nitric acid (HNO3) (Fig. 8). Alloy CC-50 contains substantially more chromium (26 to 30%) and offers relatively high resistance to localized corrosion and high resistance to many acids, including dilute H2SO4 and such oxidizing acids as HNO3. General Corrosion of Austenitic and Duplex Alloys. Alloy CF-8 typically contains approximately 19% Cr and 9% Ni and is essentially equivalent to type 304 wrought alloys. Alloy CF-8 may be fully austenitic, but it more commonly contains some residual ferrite (3 to 30%) in an austenite matrix. In the solutiontreated condition, this alloy has excellent resistance to a wide variety of acids. It is particularly resistant to highly oxidizing acids, such as boiling HNO3. Figure 9 shows isocorrosion diagrams for CF-8 in HNO3, phosphoric acid (H3PO4), and sodium hydroxide (NaOH). The duplex nature of the microstructure of this alloy imparts additional resistance to stress-corrosion cracking (SCC) compared to its wholly austenitic counterparts. Alloy CF-3 is a reduced-carbon version of CF-8 with essentially identical corrosion resistance, except that CF-3 is much less susceptible to sensitization (Fig. 10). For applications in which the corrosion resistance of the weld heat-affected zone (HAZ) may be critical, CF-3 is chosen. Alloys CF-8A and CF-3A contain more ferrite than their CF-8 and CF-3 counterparts. Because the higher ferrite content is achieved by increasing the chromium/nickel equivalent ratio, the CF-8A and CF-3A alloys may have slightly higher chromium or slightly lower nickel contents than the low-ferrite equivalents. In general, the corrosion resistance is very similar, but the strength increases with ferrite content. Because of the high ferrite content, service should be restricted to temperatures below

Corrosion behavior of ACl H-type alloys in 100 h tests at 980  C (1800  F) in reducing sulfur-bearing gases. (a) Gas contained 5 grains of sulfur per 2.8 m3 (100 ft3) of gas. (b) Gas contained 300 grains of sulfur per 2.8 m3 (100 ft3) of gas. (c) Gas contained 100 grains of sulfur per 2.8 m3 (100 ft3) of gas; test at constant temperature. (d) Same sulfur content as gas in (c), but cooled to 150  C (300  F) each 12 h

Isocorrosion diagram for ACl CB-30 in HNO3. Castings were annealed at 790  C (1450  F), furnace cooled to 540  C (1000  F), and then air cooled to room temperature.

coal conversion, utility and chemical applications, and waste incineration have heightened the need for alloys resistant to sulfidation attack in relatively weak oxidizing or reducing environments. Fortunately, high chromium and silicon contents increase resistance to sulfur-bearing environments. On the other hand, nickel has been found to be detrimental to the most aggressive gases. The problem is attributable to the formation of low-melting nickel-sulfur eutectics. These produce highly destructive liquid phases at temperatures even below 815  C (1500  F). Once formed, the liquid may run onto adjacent surfaces and rapidly corrode other metals. The behavior of H-type grades in sulfidizing environments is represented in Fig. 7. Carburization. High alloys are often used in nonoxidizing atmospheres in which carbon diffusion into metal surfaces is possible. Depending on chromium content, temperature, and carburizing potential, the surface may become extremely rich in chromium carbides, rendering it hard and possibly susceptible to cracking. Silicon and nickel are thought to be beneficial and enhance resistance to carburization (Ref 7).

Corrosion Behavior of C-Type Alloys The ACI C-type stainless steels must resist corrosion in the various environments in which they regularly serve. The influence of the metallurgy of these materials on general corrosion, intergranular corrosion, localized corrosion, corrosion fatigue, and stress corrosion are discussed. General Corrosion of Martensitic Alloys. The martensitic grades include CA-15, CA-

Fig. 7

Fig. 8

Corrosion of Cast Stainless Steels / 179 400  C (750  F) due to the possibility of severe embrittlement. Alloy CF-8C is the niobium-stabilized grade of the CF-8 alloy class. This alloy contains small amounts of niobium, which tend to form carbides preferentially over chromium carbides and improve intergranular corrosion resistance in applications involving relatively high service temperatures. The development of niobium carbides is achieved through a heat treatment at 870 to 900  C (1600 to 1650  F); this is often referred to as a stabilizing heat treatment.

Fig. 9

Alloys CF-8M, CF-3M, CF-8MA, and CF3MA are 2 to 3% Mo-bearing versions of the CF-8 and CF-3 alloys. The addition of molybdenum increases resistance to corrosion by seawater and improves resistance to many chloride-bearing environments. The presence of molybdenum also improves crevice corrosion and pitting resistance, compared to the CF8 and CF-3 alloys. Molybdenum-bearing alloys are generally not as resistant to highly oxidizing environments when phases rich in molybdenum are formed (this is particularly true for boiling

Isocorrosion diagrams for ACl CF-8 in (a) HNO3, (b and c) H3PO4, and (d and e) NaOH solutions. (b) and (d) Tests performed in a closed container at equilibrium pressure. (c) and (e) Tested at atmospheric pressure

HNO3), but for weakly oxidizing environments and reducing environments, molybdenum-bearing alloys are generally superior. Alloy CF-16F is a selenium-bearing freemachining grade of cast stainless steel. Because CF-16F nominally contains 19% Cr and 10% Ni, it has adequate corrosion resistance to a wide range of corrodents, but the large number of selenide inclusions makes surface deterioration and pitting definite possibilities. Alloy CF-20 is a fully austenitic, relatively high-strength corrosion-resistant alloy. The 19% Cr content provides resistance to many types of oxidizing acids, but the high carbon content makes it imperative that this alloy be used in the solution-treated condition for environments known to cause intergranular corrosion. Alloy CE-30 is a nominally 27Cr-9Ni alloy that typically contains 10 to 20% ferrite in an austenite matrix. The high carbon and ferrite contents provide relatively high strength. The high chromium content and duplex structure act to minimize corrosion resulting from the formation of chromium carbides in the micro-structure. This particular alloy is known for good resistance to sulfurous acid and sulfuric acid, and it is extensively used in the pulp and paper industry. Alloy CG-8M is slightly more highly alloyed than the CF-8M alloys, with the primary addition being increased molybdenum (3 to 4%). The increased amount of molybdenum provides superior corrosion resistance to halide-bearing media and reducing acids, particularly H2SO3 and H2SO4 solutions. The high molybdenum content, however, renders CG-8M generally unsuitable in highly oxidizing environments. Alloy CD-4MCu is the most highly alloyed material in this group, with a composition of Fe-26Cr-5Ni-2Mo-3Cu. The chromium/nickel equivalent ratio for this alloy is quite high, and a microstructure containing approximately equal amounts of ferrite and austenite is common. The low carbon content and high chromium content render the alloy relatively immune to intergranular corrosion. High chromium and molybdenum provide a high degree of localized corrosion resistance (crevices and pitting), and the duplex microstructure provides

Fig. 10

Isocorrosion diagram for solution-treated quenched and sensitized ACl CF-3 in HNO3

180 / Casting Design and Performance Table 2 Duplex stainless steel corrosion test results

Grade(a)

CD-4MCuN

CD-3MN

Isocorrosion diagram for ACl CD-4MCu in HNO3. The material was solution treated at  1120 C (2050  F) and water quenched.

Fig. 11

CE-3MN

CD-3MWCuN

Heat

1 2 3 4 1 2 3 4 4(e) 1 2 3 3(e) 1 2 3

A 923C(b)

G 48(c)

Corrosion rate, mdd(d)

CPT,  C

CPT,  C

0.00 0.00 3.45 2.87 0.73 2.19 0.00 0.00 2.12 2.64 0.00 0.00 0.00 0.00 0.00 0.67

35 40 30 35 40 25 50 45 50 65 50 65 50 65 70 55

35 40 30 35 40 35 50 45 50 65 50 65 50 65 70 55

(a) ASTM A 890, solution annealed. (b) ASTM A 923 method C ferric chloride corrosion test. (c) ASTM G 48 method C critical pitting temperature (CPT) test, 6% FeCl3, 24 h. (d) Corrosion rate calculated from weight loss; mdd is mg/dm2/day. Maximum acceptable corrosion rate is 10 mdd; all specimens passed. (e) Centrifugally cast specimens

Isocorrosion diagram for ACl CD-4MCu in H2SO4. The material was solution annealed at 1120 C (2050  F) and water quenched.

Fig. 12 

SCC resistance in many environments. This alloy can be precipitation hardened to provide strength and is also relatively resistant to abrasion and erosion-corrosion. Figures 11 and 12 show isocorrosion diagrams for CD-4MCu in HNO3 and H2SO4, respectively. CD-4MCu does not require control of the nitrogen content, which can lead to excessive levels of ferrite that reduce the toughness of the material. The control of nitrogen within the range specified for CD-4MCuN eliminates this problem. The ASME Pressure Vessel Code recognized this fact and has replaced CD-4MCu with CD-4MCuN. The results of a series of corrosion tests on CD-4MCuN, CD-3MN, CE-3MN, and CD3MWCuN are shown in Table 2. The ASTM A 923 test detects the presence of detrimental intermetallic phases. The weight loss is associated with local depletion. The critical pitting temperature indicates the minimum temperature that pitting occurs. Grades CE-3MN and CD3MWCuN are known as superduplex stainless steels. A method of ranking the pitting resistance of duplex stainless steels has been developed. It is an empirical measure and is known as the pitting resistance number (PREN). The PREN is based on the composition of the alloy, and for super duplex stainless steels, the PREN should not be less than 40: PREN ¼ %  Cr þ 3:3  %Mo þ 16  %N

Table 2 shows the improved pitting resistance of these alloys. Fully Austenitic Alloys. Alloys CH-10 and CH-20 are fully austenitic and contain 22 to 26% Cr and 12 to 15% Ni. The high chromium content minimizes the tendency toward the formation of chromium-depleted zones during heat treatment in the sensitizing temperature range. These alloys are often used for handling paper pulp solutions and are known for good resistance to dilute H2SO4 and HNO3. Alloy CK-20 contains 23 to 27% Cr and 19 to 22% Ni and is less susceptible than CH-20 to intergranular corrosion attack in many acids after brief exposures to the chromium carbide formation temperature range. Alloy CK-20 possesses good corrosion resistance to many acids and, because of its fully austenitic structure, can be used at relatively high temperature. Alloy CN-7M, with a nominal composition of Fe-29Ni-20Cr-2.5Mo-3.5Cu, exhibits excellent corrosion resistance in a wide variety of environments and is often used for H2SO4 service. Figure 13 shows isocorrosion diagrams for CN-7M in H2SO4, HNO3, H3PO4, and NaOH. Relatively high resistance to intergranular corrosion and SCC make this alloy attractive for very many applications. Although relatively highly alloyed, the fully austenitic structure of CN-7M may lead to SCC susceptibility for some environments and stress states. Intergranular Corrosion of Austenitic and Duplex Alloys. The optimal corrosion resistance for these alloys is developed by solution treatment. Depending on the specific alloy in question, temperatures between 1040 and 1205  C (1900 and 2200  F) are required to ensure complete solution of all carbides and other high-alloy phases, such as s and w, that sometimes form in highly alloyed stainless steels. Alloys containing relatively high total alloy content, particularly high molybdenum content,

often require the higher solution treatment temperature. Water quenching from the temperature range of 1040 to 1205  C (1900 to 2200  F) normally completes the solution treatment. Failure to solution treat a particular alloy or an improper solution treatment may seriously compromise the observed corrosion resistance in service. Inadvertent or unavoidable heat treatment in the temperature range of 480 to 820  C (900 to 1500  F), such as caused by welding, may destroy the intergranular corrosion resistance of the alloy. When austenitic or duplex (ferrite in austenite matrix) stainless steels are heated in or cooled slowly through this temperature range, chromium-rich carbides form at grain boundaries in austenitic alloys and at ferrite/austenite interfaces in duplex alloys. These carbides deplete the surrounding matrix of chromium, thus diminishing the local corrosion resistance of the alloy. An alloy in this condition of reduced corrosion resistance due to the formation of chromium carbides is said to be sensitized. In small amounts, these carbides may lead to localized pitting in the alloy, but if the chromium-depleted zones are interconnected throughout the alloy or HAZ of a weld, the alloy may disintegrate intergranularly in some environments. If solution treatment of the alloy after casting and/or welding is impractical or impossible, the metallurgist has techniques to minimize potential intergranular corrosion problems. These include stabilizing of carbides by the addition of niobium, as described earlier, by cathodic protection, or by reducing the carbon content. The low-carbon grades CF-3 and CF-3M are commonly used as a solution to the sensitization incurred during welding. The low carbon content (0.03% C maximum) of these alloys precludes the formation of an extensive number of chromium carbides. In addition, these alloys normally contain 3 to 30% ferrite in an austenitic matrix. By virtue of rapid carbide precipitation kinetics at ferrite/austenite interfaces compared to austenite/austenite interfaces, carbide precipitation is confined to ferrite/ austenite boundaries in alloys containing a minimum of approximately 3 to 5% ferrite (Ref 8, 9). If the ferrite network is discontinuous in the austenite matrix (depending on the amount, size, and distribution of ferrite pools), then extensive intergranular corrosion will not be a problem in most of the environments to which these alloys would be subjected. An example of attack at the ferrite/austenite boundaries is shown in Fig. 14. These low-carbon alloys need not sacrifice significant strength compared to their high-carbon counterparts, because nitrogen may be added to increase strength. However, a large amount of nitrogen will begin to reduce the ferrite content, which will cancel some of the strength gained by interstitial hardening. Appropriate adjustment of the chromium/nickel equivalent ratio is beneficial in such cases. Fortunately, nitrogen is also beneficial to the corrosion resistance of austenitic and duplex stainless steels (Ref 10). Nitrogen seems to retard sensitization and

Corrosion of Cast Stainless Steels / 181

Fig. 13

Isocorrosion diagrams for solution-annealed and quenched ACl CN-7M in H2SO4, HNO3, NaOH, and H3PO4. (a), (b), (d), and (f) Tested at atmospheric pressure. (c) and (e) Tested at equilibrium pressure in a closed container. See Fig. 9 for legend.

Ferrite/austenite grain-boundary ditching in as-cast ACl CF-8. The specimen, which contained 3% ferrite, was electrochemical potentiokinetic reactivation tested, SEM micrograph. Original magnification 4550 . Source: Ref 9

Fig. 14

improve the resistance to pitting and crevice corrosion of many stainless steels (Ref 11). The standard practices of ASTM A 262 (Ref 12) are commonly implemented to predict and measure the susceptibility of austenitic and duplex stainless steels to intergranular corrosion. Table 3 indicates some representative results for CF-type alloys as tested according to practices A, B, and C of Ref 10 as well as two electrochemical tests described in Ref 13 and 14. Table 4 lists the compositions of the alloys investigated. The data indicate the superior resistance of the low-carbon alloys to intergranular corrosion. It also indicates that for highly oxidizing environments (represented here by A 262C-boiling HNO3), the CF-3 and CF-3M alloys are equivalent in the solutiontreated condition, but that subsequent heat treatment causes the corrosion resistance of the CF3M alloys to deteriorate rapidly for service in oxidizing environments (Ref 16). In addition, the degree of chromium depletion necessary to cause susceptibility to intergranular corrosion appears to increase in the presence of molybdenum (Ref 17). The passive film stability imparted by molybdenum may offset the loss

of solid-solution chromium for mild degrees of sensitization. Intergranular Corrosion of Ferritic and Martensitic Alloys. Ferritic alloys may also be sensitized by the formation of extensive chromium carbide networks, but because of the high bulk chromium content and rapid diffusion rates of chromium in ferrite, the formation of carbides can be tolerated if the alloy has been slowly cooled from a solutionizing temperature of 780 to 900  C (1435 to 1650  F). The slow cooling allows replenishment of the chromium adjacent to carbides. Martensitic alloys normally do not contain sufficient bulk chromium to be used in applications in which intergranular corrosion is likely to be of concern. Typical chromium contents for martensitic alloys may be as low as 11 to 12%. Pitting and Crevice Corrosion. Austenitic and martensitic alloys display a tendency toward localized corrosion in some environments. The conditions conducive to this behavior may be any situation in areas where flow is restricted and an oxygen concentration cell may be established. Duplex alloys have been found to be less susceptible. Localized corrosion is

182 / Casting Design and Performance Table 3 Intergranular corrosion test results for Alloy Casting Institute casting alloys Alloy(b)/Test results(c) Metallurgical condition

Solution treated

Simulated weld repair

Solution treated, held 1 h at 650  C (1200  F) As-cast

Test(a)

A 262A A 262B A 262C EPR JEPR A 262A A 262B A 262C EPR JEPR A 262A A 262B A 262C EPR JEPR A 262A A 262B A 262C EPR JEPR

CF8 (4)

CF8 (11)

CF8 (20)

CF8M (5)

CF-8M (11)

CF-8M (20)

P P P P P X X X X X X X X X X X X X X X

P P P P P X X X X X X X X X X X X X X X

P P P P P X X X X X X X X X X X X X X X

P P P P P X X X P P X X X X P X X X X X

P P P P P X X X P P X X X X X X X X X X

P P P P P X X X P P X X X X X X X X X X

CF-3 (2)

CF-3 (5)

CF-3 (8)

P P P P P P P P P P* P* P* P P P P P P P P P P P P P* P* P* P P P X X X P P P P P P X/P* X/P* X/P* P P P X X X P P P P** P** P** X/P* X/P* X/P* X/P P P

CF3M (5)

CF3M (9)

CF-3M (16)

P P P P P P P P P P X P X X/P P X P X X/P P

P P P P P P P P P P X P X P P X X X X/P P

P P P P P P P P P P X P X P P X P X P P

(a) See Ref 12 for details of ASTM A 262 practices. EPR, electrochemical potentiokinetic reactivation test: see Ref 12 for details. JEPR, Japanese electrochemical potentiokinetic reactivation test: see Ref 13 for details. (b) Parenthetical value is the percentage of ferrite. See Table 4 for alloy compositions. (c) P, pass; X, fail based on the following criteria: A 262A ditching