Advances in Building Energy Research

Citation preview

ABER_prelims

6/12/08

1:52 PM

Page i

Advances in Building Energy Research

ABER_prelims

6/12/08

1:52 PM

Page ii

Editor-in-Chief Mat Santamouris University of Athens, Greece

Editorial Board Professor O. Seppanen Technical University of Helsinki, Finland

Dr M. Sherman Lawrence Berkeley Laboratory, USA

N. Fintikakis Associate Editor for UIA, Greece

Dr P. Wouters Belgian Building Research Institute, Belgium

Professor F. Allard University of La Rochelle, France

Dr H. Akbari Lawrence Berkeley Laboratory, USA

Professor E. Maldonado University of Porto, Portugal

Professor Lee S. E. University of Singapore, Singapore

Professor A. Papadopoulos Aristotle University of Thessaloniki, Greece

Professor F. Nicols University of Strathclyde, UK

Dr E. Erell Ben Gurion University, Israel

Professor H. Yoshino Tohuku University, Japan

Professor F. Haghighat Concordia University, Canada

Professor R. Lamberts University of Santa Catharina, Brazil

Professor J. Clarke University of Strathclyde, UK

Professor A. Athienitis Concordia University, Canada

Professor J. Khedari University of Bangkok, Thailand

Professor F. Butera Technical University of Milan, Italy

Professor M. Wilson Metropolitan University, London

Professor K. Voss University of Wuppertal, Germany

ABER_prelims

6/12/08

1:52 PM

Page iii

Advances in Building Energy Research Volume 2

Editor-in-Chief Mat Santamouris

publishing for a sustainable future

London • Sterling, VA

ABER_prelims

6/12/08

1:52 PM

Page iv

First published by Earthscan in the UK and USA in 2008 Copyright © Mat Santamouris, 2008 All rights reserved ISSN ISSN ISBN

1751-2549 (Print) 1756-2201 (Online) 978-1-84407-517-1

Typeset by Domex e-Data, India Printed and bound in the UK by Cromwell Press, Trowbridge Cover design by Giles Smith For a full list of publications please contact: Earthscan Dunstan House 14a St Cross Street London EC1N 8XA, UK Tel: +44 (0)20 7841 1930 Fax: +44 (0)20 7242 1474 Email: [email protected] Web: www.earthscan.co.uk 22883 Quicksilver Drive, Sterling, VA 20166-2012, USA Earthscan publishes in association with the International Institute for Environment and Development A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Advances in building energy research / editor-in-chief, Mat Santamouris. v. cm. Includes bibliographical references. 1. Buildings–Energy conservation. I. Santamouris, M. (Matheos), 1956-TJ163.5.B84A285 2007 696–dc22 2007004087

The paper used for this book is FSC-certified. FSC (the Forest Stewardship Council) is an international network to promote responsible management of the world’s forests. Advances in Building Energy Research volume 2, 2008. Advances in Building Energy Research is published annually. Periodicals Postage Paid at Rahway, NJ. US agent: Mercury International, 365 Blair Road, Avenel, NJ 07001. POSTMASTER: Address changes to ADVANCES IN BUILDING ENERGY RESEARCH, 365 Blair Road, Avenel, NJ 07001.

ABER_prelims

6/12/08

1:52 PM

Page v

Contents List of Figures and Tables List of Acronyms and Abbreviations 1

Evolution of Cool-Roof Standards in the US

vii xiii 1

Hashem Akbari and Ronnen Levinson

2

A Review of Innovative Daylighting Systems

33

A. Tsangrassoulis

3

Physically Based Modelling of the Material and Gaseous Contaminant Interactions in Buildings: Models, Experimental Data and Future Developments

57

P. Blondeau, A. L. Tiffonnet, F. Allard and F. Haghighat

4

The Application of Urban Climate Research in the Design of Cities

95

Evyatar Erell

5

Solar Air Conditioning: A Review of Technological and Market Perspectives

123

S. Oxizidis and A. M. Papadopoulos

6

Experimental Methods in Ventilation

159

M. Sandberg, H. Lundström, H. O. Nilsson and H. Stymne

7

A Review of Optical Properties of Shading Devices

211

Athanassios Tzempelikos

Index

241

ABER_prelims

6/12/08

1:52 PM

Page vi

ABER_prelims

6/12/08

1:52 PM

Page vii

List of Figures and Tables FIGURES 1.1 1.2 1.3

1.4 1.5 2.1

2.2 2.3 2.4 2.5

2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21 3.1

Locations of the eight ASHRAE-defined climate zones in the US Locations of the 16 California climate zones Total savings per unit roof area (cooling equipment savings plus 15-year NPV of energy savings with TDV; US$/m2) by California climate zone, simulated for a prototypical non-residential Title 24 building with a cool low-sloped roof Adoption of commercial building energy codes by US states as of May 2007 Adoption of residential building energy codes by US states as of May 2007 Simulation of luminance distribution in a room with no shading system (upper) shaded by fabric roller (middle) and shaded by Venetian blinds (lower) Schematic representation of two geometrical characteristics: cut-off angle for solar rays and angle of acceptance for diffuse radiation Retrolux blind Okasolar-W system Creation and operational principle of a compound parabolic concentrator (CPC) using two parabolas and definition of maximum acceptance angle (α) Construction scheme of the Microlouver (by Siteco) Prismatic glazing shading operation Geometric configuration of a typical prismatic glazing section Laser-cut light-deflecting panel (LCP) operational principle Heliostat designed and installed at BartenBach LichtLabor, Innsbruck, Austria Inflatable heliostat prototype Mechanism for alignment of a solar tracker Definition of acceptance angle for fibre optics Himawari system HSL 3100 hybrid lighting system Solux daylighting system installed in the University of Athens during the Universal Fibre Optics research programme Heliobus system installed at Graubunden, Switzerland Horizontal light pipe Horizontal light pipe proposed by Canziani et al (2004) Prototype daylighting system developed by Whitehead (2006) Anidolic ceiling Elemental interaction phenomena between gaseous contaminants and water vapour in poly-dispersed porous materials

11 17

18 23 24

36 36 37 38

38 39 40 40 41 42 43 44 45 45 46 46 48 49 49 50 50 85

ABER_prelims

6/12/08

viii

5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 6.1 6.2 6.3 6.4 6.5 6.6 6.7

6.8

6.9 6.10 6.11 6.12 6.13 6.14 6.15

6.16

1:52 PM

Page viii

ADVANCES IN BUILDING ENERGY RESEARCH

Operational principles of heat pumps: (a) mechanically driven; (b) thermally driven Schematic representation of sorption refrigerator Basic cycle for adsorption and absorption in a (lnP, -1/T) diagram: (a) closed systems; (b) chemical reaction systems Temperature profile of a TCHP cycle Working media paths in a sorption system S-T diagram for a discontinuous sorption system process Entropy–temperature (S–T) diagram for a sorption heat pump Equivalent Carnot cycles of a sorption heat pump Cost of solar cooling systems v. vapour compression cooling Primary energy ratio of solar-assisted absorption and electrically driven vapour compression chillers Percentage of people who find air quality just acceptable at a given ventilation rate Sketch of the dependence of concentration and velocity upon ventilation flow rate The sequence outdoor ambient to the human Weekly average values of ventilation flow rate as a function of a season in a house ventilated by mechanical extract ventilation Distribution of air change rate in similar naturally ventilated dwellings (unbroken line) and similar mechanically extract ventilated (dashed line) (Left) A person simulator produces 100W of heat; (right) the photographer Magnus Mattsson A thermal manikin is exposed to two different environments, one actual with non-uniform and one ‘imaginary’ with uniform climatic conditions Supply air terminals: (top) High-velocity supply, medium-velocity supply; (bottom) low-velocity supply intended for displacement ventilation, local ventilation Interpretation of thermal length A flow dominated by buoyancy (Ars = 0.97): Overhead view of position of the front on the floor Velocity at a point in the occupied zone where there is a supply of isothermal heated air and cold air Assessment of velocity in a room ventilated by displacement ventilation Time constant Concentration depends upon both how the contaminant and the air are distributed Kitchen exhaust hood used for laboratory test of capture efficiency: Visualization with smoke and a laser light sheet; oil is sprayed into a hot pan by means of a spray-nozzle Concentration recorded in a room ventilated by displacement ventilation

127 128 129 132 133 135 136 137 142 145 160 161 161 162 162 163

168

170 171 173 174 174 175 176

177 178

ABER_prelims

6/12/08

1:52 PM

Page ix

LIST OF FIGURES AND TABLES

6.17

6.18 6.19 6.20 6.21 6.22 6.23 6.24 6.25 6.26 6.27 6.28 6.29 6.30 6.31 6.32 6.33 7.1 7.2 7.3 7.4 7.5 7.6

7.7 7.8 7.9 7.10 7.11 7.12 7.13 7.14

Bidirectional bulk density flow through a doorway studied in a two-dimensional model: The cooler air moves towards the left as a gravity current along the floor Unidirectional boundary layer flow through an opening Flow in a stairwell studied with the salt bath method: The density difference is made visible with the shadowgraph method A ventilation system, a ventilation system with infiltration, and purging flow rate Definition of displacement height (D) and roughness height (z0 ) Wind tunnel test of a hybrid-ventilated school at a scale of 1:200: In the background is shown the spires and the roughness elements Visualization by the sand erosion method of wind-driven flow through openings The heated full-scale thermal manikin with 33 individually controlled zones: This manikin was especially constructed for climate evaluation in 1991 Dummies simulating the heat generated by a person Manikin with 20 individually controlled segments Manikin with 120 independently controlled zones Illumination system, flow markers and a digital camera Particle image velocimetry (PIV) recording and interrogation process Orientation of the measuring plane Supply of warm air from a ceiling diffuser: Temperature distribution recorded with an infrared camera Components in a passive system Comfort zone diagrams adapted for Comfortina-type manikins Glazing system with N layers Between-glazing and exterior venetian blinds Discretization of venetian blinds Direct-to-direct transmittance Directly irradiated parts of venetian blind View of a cell formed by adjacent slats showing how the cell is divided into segments, si, for the calculation of direct solar transmittance; and side view of a cell showing case where some of the direct solar passes between adjacent slats without touching either of them Side view of horizontal blinds for calculating optical properties for sky and ground diffuse radiation Description of cut-off angles Examples of interior roller shades Spectral and directional transmission properties (different incidence angles) of a common roller shade Screen model rendering of intersecting orthogonal crossed cylinders Categorization of fabrics for draperies/shades Bidirectional reflection and transmission distribution functions Detection of the light transmitted through a sample

ix

179 180 180 181 182 183 184 184 185 186 187 195 196 197 197 199 203 214 217 218 220 220

222 223 226 229 230 231 233 234 235

ABER_prelims

x

6/12/08

1:52 PM

Page x

ADVANCES IN BUILDING ENERGY RESEARCH

TABLES 1.1 1.2

1.3 1.4 1.5

1.6

1.7

1.8

1.9

1.10

1.11

3.1 3.2 3.3 5.1 5.2

Cool-roof energy savings measured in six California non-residential buildings Thermal resistance of insulation below a cool roof (solar reflectance 0.55) that yields the same annual energy expenditure (cost at US$0.66/therm and US$0.08/kWh) as a low, medium or high level of insulation (3,11 or 38ft2 h °F BTU–1) below a conventional roof (solar reflectance 0.20) Roof thermal transmittance (U-factor) multipliers for cool roofs on buildings other than low-rise residential buildings Roof thermal transmittance (U-factor) multipliers for cool roofs on buildings other than low-rise residential buildings Multiplier by which the thermal transmittance (U-factor) of a residential roof assembly can be increased without raising annual energy use when the solar reflectance of the roof’s surface is increased to 0.55 (cool) from 0.10 (conventional) Ceiling thermal transmittance (U-factor) multipliers for residential cool roofs: It is possible that these multipliers should be replaced by their reciprocal to yield values less than or equal to unity Ceiling thermal resistances (ft2 h °F BTU–1) prescribed by ASHRAE Standard 90.2-2007 for ceilings under conventional (non-cool) and cool residential roofs, derived from ASHRAE Standard 90.2-2007: Reduced requirements for cool-roofed buildings are shaded Life-cycle cool-roof savings per unit roof area (cooling equipment savings plus 30-year NPV of energy savings with TDV; US$/m2) by California climate zone for a non-residential building with a steep-sloped roof Life-cycle cool-roof savings per unit roof area (cooling equipment savings plus 30-year NPV of energy savings with TDV; US$/m2) by California climate zone for a residential building with a low-sloped roof Life-cycle cool-roof savings per unit roof footprint area (cooling equipment savings plus 30-year NPV of energy savings with TDV; US$/m2) by California climate zone for a residential building with a steep-sloped roof Solar reflectance and thermal emittance requirements of a 2007 residential cool-roof rebate programme administered by two California utilities (Pacific Gas & Electric and Southern California Edison) Representative adsorption isotherm models for building applications Compilation of sorption and diffusion coefficients: Data fit with one-phase models (K
KH) Compilation of sorption and diffusion coefficients: Data fit with two-phase models (K
Kp) Main features of possible solar sorption cooling technologies Physical characteristics of several chillers of 400kW cooling capacity

3

9 10 10

13

14

15

19

20

21

28 66 69 72 138 143

ABER_prelims

6/12/08

1:52 PM

Page xi

LIST OF FIGURES AND TABLES

5.3 5.4 5.5 6.1

Strategic factors of the external environment Strategic factors of the internal environment Strategies for development (TOWS matrix) Description of the connection between the measured quantities and equivalent temperature

xi

148 149 151 167

ABER_prelims

6/12/08

1:52 PM

Page xii

ABER_prelims

6/12/08

1:52 PM

Page xiii

List of Acronyms and Abbreviations ABL ACT A/D AS ASHRAE BTDF BT(R)DF °C CAD CASE C6F6 CCD CEC CET CFC CFD CIE clo cm/s CO2 COP CPC CRP CSP CSUMM

2D 3D dB DDC DEM DSM EPA EPF ESCO ET °F ft2 g GAX

atmospheric boundary layer acoustic ceiling tile analogue/digital aspect ratio American Society of Heating, Refrigerating and Air-Conditioning Engineers bidirectional transmittance distribution function bidirectional transmission (or reflection) distribution function degrees Celsius computer-aided design Codes and Standards Enhancement hexafluorobenzene charge-coupled device California Energy Commission corrected effective temperature chlorofluorocarbon computational fluid dynamic(s) Commission Internationale de l’Eclairage average clothing centimetres per second carbon dioxide coefficient of performance compound parabolic concentrator carpet computer-simulated person Colorado State University Mesoscale Model (a hydrostatic, primitive-equation, three-dimensional Eulerian model of the Earth’s boundary layer climate) two dimensional three dimensional decibel direct digital control digital elevation model demand-side management US Environmental Protection Agency energy performance factor energy service company effective temperature; equivalent temperature degrees Fahrenheit square feet gram generator/absorber heat exchanger

ABER_prelims

xiv

6/12/08

1:52 PM

Page xiv

ADVANCES IN BUILDING ENERGY RESEARCH

GB GHG GW GWI GWP HCFC HFC HOE HVAC H/W Hz IAQ IAS IDC IECC IR ISO J K kg kHz kJ kJ/K kW kWh kWh/m2 lb ft–2 LCP LDA LEED LES LiBr m2 MHF MJ MJ/m2 m/s MTV MW N2 O NG NH3 NPV O&M ODP OSB

gypsum board greenhouse gas gigawatt global warming impact global warming potential hydro-chlorofluorocarbon hydrofluorocarbon holographic optical element heating, ventilation and air conditioning height-to-width ratio hertz indoor air quality ideal adsorbed solution integration of directional coefficients International Energy Conservation Code infrared International Organization for Standardization joule kelvin kilogram kilohertz kilojoule kilojoule per kelvin kilowatt kilowatt hours kilowatt hours per square metre pounds per square foot laser-cut light-deflecting panel laser Doppler anemometer Leadership in Energy and Environmental Design large eddy simulation lithium bromide square metre mean heat flux megajoule megajoules per square metre metres per second mean thermal vote megawatt nitrous oxide natural gas ammonia net present value operation and maintenance ozone-depletion potential oriented strand board

ABER_prelims

6/12/08

1:52 PM

Page xv

LIST OF ACRONYMS AND ABBREVIATIONS

Pa PB PC PCI PER PFC PG&E PHF PIER PIV PLY PMMA PMV ppb PPD PSV PV R&D RES RH RT s s–1 SAC SCE SET SF6 SMUD SRI SWOT TCHP TDV TES TOWS TVOC µ µm UHI US UV VCC VOC W W/m2 Wh/m2 WIS

pascal particle board personal computer park cool island primary energy ratio polyfluorocarbon Pacific Gas & Electric perceived heat flux Public Interest Energy Research Program particle image velocimetry plywood polymethyl methacrylate predicted mean vote parts per billion predicted percentage dissatisfied particle streak velocimetry photovoltaics research and development renewable energy source(s) relative humidity resultant temperature second per second solar-assisted cooling Southern California Edison standard effective temperature sulphur hexafluoride Sacramento Municipal Utility District solar reflectance index strengths, weaknesses, opportunities, threats thermochemical heat pump time-dependent valuation thermal energy storage threats, opportunities, weaknesses and strengths total amounts of VOC micro micrometre urban heat island United States ultraviolet vapour compression cooling volatile organic compound watt watts per square metre watt hours per square metre Advanced Window Information System

xv

ABER_prelims

6/12/08

1:52 PM

Page xvi

ABER_CH_01

7/11/08

10:19 AM

Page 1

publishing for a sustainable future

1

Evolution of Cool-Roof Standards in the US Hashem Akbari and Ronnen Levinson

Abstract Roofs that have high solar reflectance and high thermal emittance stay cool in the sun. A roof with lower thermal emittance but exceptionally high solar reflectance can also stay cool in the sun. Substituting a cool roof for a non-cool roof decreases cooling electricity use, cooling power demand and cooling equipment capacity requirements, while slightly increasing heating energy consumption. Cool roofs can also lower the citywide ambient air temperature in summer, slowing ozone formation and increasing human comfort. Provisions for cool roofs in energy efficiency standards can promote the buildingand climate-appropriate use of cool roofing technologies. Cool-roof requirements are designed to reduce building energy use, while energy-neutral cool-roof credits permit the use of less energy-efficient components (e.g. larger windows) in a building that has energy-saving cool roofs. Both types of measures can reduce the life-cycle cost of a building (initial cost plus lifetime energy cost). Since 1999, several widely used building energy efficiency standards, including ASHRAE 90.1, ASHRAE 90.2, the International Energy Conservation Code and California’s Title 24 have adopted cool-roof credits or requirements. This chapter reviews the technical development of cool-roof provisions in the ASHRAE 90.1, ASHRAE 90.2 and California Title 24 Standards, and discusses the treatment of cool roofs in other standards and energy efficiency programmes. The techniques used to develop the ASHRAE and Title 24 cool-roof provisions can be used as models to address cool roofs in building energy efficiency standards worldwide.

■ Keywords – cool roofs; solar reflectance; thermal emittance; solar reflectance index; building energy efficiency standards; ASHRAE 90.1; ASHRAE 90.2; California Title 24; International Energy Conservation Code (IECC); Leadership in Energy and Environmental Design (LEED); Energy Star; Florida Building Code (FBC); Revised Ordinances of Honololu (ROH); City of Chicago Energy Conservation Code

INTRODUCTION Roofs that have high solar reflectance (high ability to reflect sunlight: spectrum 0.3–2.5µm) and high thermal emittance (high ability to emit thermal radiation: spectrum 4–80µm) stay

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32 doi:10.3763/aber.2008.0201 ■ © 2008 Earthscan ■ ISSN 1751-2549 (Print), 1756-2201 (Online) ■ www.earthscanjournals.com

ABER_CH_01

7/11/08

2

10:19 AM

Page 2

HASHEM AKBARI AND RONNEN LEVINSON

cool in the sun. The same is true of roofs with lower thermal emittance but exceptionally high solar reflectance. Roofs that stay cool in the sun by minimizing solar absorption and maximizing thermal emission are hereafter denoted ‘cool roofs’.

BENEFITS OF COOL ROOFS Low roof temperatures lessen the flow of heat from the roof into the building, reducing the need for space cooling electricity in conditioned buildings. Since building heat gain through the roof peaks in mid to late afternoon, when summer electricity use is highest, cool roofs can also reduce peak electricity demand. Prior research has indicated that savings are greatest for buildings located in climates with long cooling seasons and short heating seasons, particularly those buildings that have distribution ducts in the plenum (Akbari, 1998; Konopacki and Akbari, 1998; Akbari et al, 1999). Cool roofs transfer less heat to the outdoor environment than do warm roofs (Taha, 2001). The resulting decrease in outside air temperature can slow urban smog formation and improve human health and outdoor comfort. Reduced thermal stress may also increase the lifetime of cool roofs, lessening maintenance and waste (Akbari et al, 2001). Earlier studies have measured daily air-conditioning energy savings and peak-power demand reduction from the use of cool roofs on buildings in several warm weather climates, including California, Florida and Texas. Cool roofs on non-residential buildings typically yielded measured summertime daily cooling energy savings and peak-power demand reductions of 10 to 30 per cent, though values have been as low as 2 per cent and as high as 40 per cent (see Table 1.1) (Konopacki et al, 1998). For example: ● Konopacki et al (1998) measured summer daily cooling energy savings per unit roof

area of 67, 39 and 4Wh/m2 (18, 13 and 2 per cent, respectively) for three California non-residential buildings – two medical offices in Davis and Gilroy and a retail store in San Jose. Assuming an aged solar reflectance of 0.55, estimated annualized cooling energy savings (daily savings × number of cooling days per year) were 6.4, 3.7 and 0.6kWh/m2 (16, 11 and 2 per cent, respectively), while peak-power demand reductions per unit roof area were 3.3, 2.4 and 1.6W/m2 (12, 8 and 9 per cent, respectively). ● Hildebrandt et al (1998) measured summer daily cooling energy savings of 23, 44 and 25Wh/m2 (17, 26 and 39 per cent, respectively) in an office, a museum and a hospice in Sacramento, California. Estimated annualized cooling energy savings were 1.3, 2.6 and 2.2kWh/m2, assuming an aged solar reflectance of 0.55. ● Konopacki and Akbari (2001) estimated summer daily cooling energy savings of 39Wh/m2 (11 per cent) and a peak-power demand reduction of 3.8W/m2 (14 per cent) in a large retail store in Austin, Texas. Estimated annualized cooling energy savings were 6.8kWh/m2, assuming an aged solar reflectance of 0.55. ● Parker et al (1998a) measured summer daily cooling energy savings of 44Wh/m2 (25 per cent) and a peak-power demand reduction of 6W/m2 (30 per cent) for a school building in Florida. Estimated annualized cooling energy savings were 4.7kWh/m2, assuming an aged solar reflectance of 0.55.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

Structure Plenum type Ceiling type Pre-coating condition Pre-coating solar reflectance Post-coating solar reflectance after one year Degraded (weathered) solar reflectance

Wood deck Ventilated plenum Tiles 25% granule loss and cracking 0.16

0.60

0.55

Metal deck Return plenum Tiles 25% granule loss and bubbling 0.24

0.60

0.55

0.55

0.60

Radiant barrier

fibreglass) Wood deck Ventilated plenum Tiles 25% granule loss and cracking 0.25

Asphalt capsheet with tan granules

3060 Built-up

0.55

0.60

Metal deck Return plenum Tiles 25% granule loss and bubbling 0.24

Asphalt capsheet with light grey granules 3.4 (R-19)

2290 Four-ply with capsheet

(D) SACRAMENTO OFFICE

0.55

0.60

Wood deck Ventilated plenum Tiles 25% granule loss and cracking 0.25

Asphalt capsheet with light grey granules None

455 Built-up gravel

(E) SACRAMENTO MUSEUM

0.55

0.60

Wood deck Ventilated plenum Tiles 25% granule loss and cracking 0.16

1.9 (R-11)

557 Composite shingle/ flat built up Asphalt capsheet with tan granules

(F) SACRAMENTO HOSPICE

10:19 AM

Insulation thermal resistance (m2 K/W)

2210 Built-up

(C) SAN JOSE RETAIL STORE

Asphalt capsheet with light grey granules 3.4 (R-19

Asphalt capsheet with light grey granules 1.4 (R-8 rigid)

2950 Built-up

(B) GILROY MEDICAL OFFICE

7/11/08

Material

Roof Area (m2) Type

(A) DAVIS MEDICAL OFFICE

TABLE 1.1 Cool-roof energy savings measured in six California non-residential buildings

ABER_CH_01 Page 3

Evolution of Cool-Roof Standards in the US 3

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

Plenum 39 (13%)

110 3.7

2.4

Conditioned space

67 (18%)

110 6.4

3.3

1.6

165 0.6

4 (2%)

Plenum

0.35 (R-2)

n/a

165 1.3

23 (17%)

Conditioned space

Uninsulated

(D) SACRAMENTO OFFICE

n/a

165 2.6

44 (26%)

Plenum

0.81 (R-4.6)

(E) SACRAMENTO MUSEUM

n/a

165 2.2

25 (39%)

Plenum

0.35 (R-2)

(F) SACRAMENTO HOSPICE

10:19 AM

Source: (a)–(c) Konopacki et al (1998); (d)–(f) Hildebrandt et al (1998)

0.81 (R-4.6)

Uninsulated

(C) SAN JOSE RETAIL STORE

7/11/08

Note: n/a = not available

Supply duct Insulation thermal resistance (m2 K/W) Location Savings Measured daily cooling energy savings (Wh/m2) Cooling days/year Degraded annual cooling energy savings (kWh/m2) Degraded peakpower demand reduction (W/m2)

(B) GILROY MEDICAL OFFICE

4

(A) DAVIS MEDICAL OFFICE

TABLE 1.1 Cool-roof energy savings measured in six California non-residential buildings (Cont’d)

ABER_CH_01 Page 4

HASHEM AKBARI AND RONNEN LEVINSON

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 5

Evolution of Cool-Roof Standards in the US

5

Cool roofs on residential buildings yielded measured summertime cooling energy savings and peak-power demand reductions that ranged from negligible to 80 per cent. For example: ● In a study of 11 Florida homes, Parker et al (1998b) measured average summer daily

cooling energy savings of 7.7kWh (19 per cent) per house and an average peakpower reduction of 0.55kW (22 per cent) per house. The daily electricity savings in individual houses ranged from 0.9kWh (0.2 per cent) to 15.4kWh (45 per cent) and the peak-power reduction ranged from 0.2kW (12 per cent) to 0.99kW (23 per cent). These initial savings resulted from increasing the solar reflectance of the shingle roofs to 0.70 from 0.08. ● Akbari et al (1997) measured summer daily energy savings of 14Wh/m2 (80 per cent) and peak demand savings of 3.8W/m2 (30 per cent) in a single-story, flat-roofed house in Sacramento. The savings resulted from increasing the solar reflectance of the roof to 0.70 from 0.18.

NEED FOR COOL-ROOF STANDARDS It is difficult for a building owner to assess the influence of roof properties on the lifetime cost of heating and cooling energy, which depends upon: ● ● ● ●

climate- and building-specific hourly uses of heating and cooling energy; hourly valuations of energy; the time value (discounting) of money; and the service life of the roof.

Building owners may also be unaware of the societal benefits of cool roofs, such as lower peak-power demand (reducing the likelihood of power failures on hot days) and lower outdoor air temperatures (improving comfort and slowing the formation of smog). Hence, without cool-roof standards, owners will tend to choose roofs that minimize initial construction cost, rather than the aggregate cost of construction and lifetime energy consumption. Provisions for cool roofs in energy efficiency standards promote their building- and climate-appropriate use, and also stimulate the development of energy-saving cool-roof technologies. For example, several manufacturers have introduced novel cool non-white roofing materials, including fibreglass asphalt shingles, clay and concrete tiles, and metal products (Akbari and Desjarlais, 2005). The development and long-term performance of cool-roof technologies are described by Akbari et al (2005a, 2005b), Levinson et al (2005b, 2005c, 2005d, 2007) and Berdahl et al (2008).

TYPES OF REQUIREMENTS IN STANDARDS Building energy efficiency standards typically specify both mandatory and prescriptive requirements. Mandatory requirements, such as practices for the proper installation of insulation, must be implemented in all buildings subject to the standard. A prescriptive requirement typically specifies the characteristics or performance of a single component

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

6

10:19 AM

Page 6

HASHEM AKBARI AND RONNEN LEVINSON

of the building (e.g. the thermal resistance of duct insulation) or of a group of components (e.g. the thermal transmittance of a roof assembly). All buildings regulated by a particular standard must achieve either prescriptive or performance compliance. A proposed building that meets all applicable mandatory and prescriptive requirements will be in prescriptive compliance with the standard. Alternatively, a proposed building can achieve performance compliance with the standard if: ● it satisfies all applicable mandatory requirements; and ● its annual energy use does not exceed that of a comparable ‘design’ (also known as a

‘standard’ or ‘reference’) building that achieves prescriptive compliance. Prescribing the use of cool roofs in building energy efficiency standards promotes the cost-effective use of cool roofs to save energy, reduce peak-power demand and improve air quality. Another option is to credit, rather than prescribe, the use of cool roofs. This can allow more flexibility in building design, permitting the use of less energy-efficient components (e.g. larger windows) in a building that has energy-saving cool roofs. Such credits are energy neutral, but may still decrease peak-power demand and improve air quality. They may also reduce the initial cost of the building. This chapter reviews the technical development of cool-roof provisions in the ASHRAE 90.1, ASHRAE 90.2 and California Title 24 building energy efficiency standards, and discusses the treatment of cool roofs in several other standards and energy efficiency programmes. The techniques used to develop the ASHRAE and Title 24 cool-roof provisions can be used as models to address cool roofs in building energy efficiency standards worldwide.

DEVELOPMENT OF STANDARDS In 1999, the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) first credited cool roofs on non-residential and high-rise residential buildings in ASHRAE Standard 90.1-1999: Energy Standards for Buildings Except Low-Rise Residential Buildings (ASHRAE, 1999). In 2001, ASHRAE amended its standards for low-rise residential buildings to credit cool roofs, implementing the revisions three years later in ASHRAE Standard 90.2-2004: Energy-Efficient Design of Low-Rise Residential Buildings (ASHRAE, 2004b). In January 2001, the state of California followed the ASHRAE approach by crediting in its ‘Title 24’ Energy Efficiency Standards for Residential and Non-Residential Buildings the use of cool roofing products on non-residential buildings with low-sloped roofs (CEC, 2001). In 2005, California upgraded Title 24 to prescribe minimum values of solar reflectance and thermal emittance for low-sloped roofs (i.e. roofs with a ratio of rise to run not exceeding 2:12) on non-residential buildings (CEC, 2006). As of June 2007, California is evaluating proposals to prescribe in the 2008 Title 24 standards minimum values of solar reflectance and thermal emittance for low-sloped roofs on non-residential buildings, and for both low-sloped and steep-sloped roofs on residential buildings. Other localities, such as Florida and Chicago, have adopted custom cool-roof requirements in their energy codes.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 7

Evolution of Cool-Roof Standards in the US

7

Note that the building envelope requirements of the ASHRAE and California Title 24 standards apply only to envelope components (e.g. roofs) that enclose conditioned spaces.

ASHRAE STANDARD 90.1 Recognizing the potential for solar-reflective roofs to reduce the conditioning energy use of commercial buildings, the ASHRAE Standard 90.1 committee organized a task force in 1997 to analyse the energy-saving benefits of cool roofs in different climates, and to propose modifications to the standard to account for the effect of roof solar reflectance. This section summarizes the cool-roof analysis performed for ASHRAE Standard 90.1 (Akbari et al, 1998).

Cool roofs versus roof insulation Solar-reflective roofs with high thermal emittance stay cool in the sun, reducing the flow of heat from the roof to the building’s conditioned space. This can decrease the need for cooling energy in summer and increase heating energy use in winter. The winter heating energy penalty is usually smaller than the summer cooling energy savings because in winter the sun is low, the days are short, the skies are often cloudy, and heating occurs mainly in early morning and early evening. Roof insulation impedes the flow of heat between the roof and the conditioned space, slowing both heating of the building when the roof is warmer than the inside air and cooling of the building when the roof is cooler than the inside air. One can develop an energy-neutral trade-off between the solar reflectance of the roof’s surface and the thermal transmittance of the roof assembly.

Survey of the radiative properties of roofing products The task force surveyed the solar reflectance and thermal emittance of various roofing products, including fibreglass asphalt shingles, elastomeric coatings, membranes, metal panels, clay tiles and concrete tiles. The solar reflectance of shingles ranged from 0.03 to 0.26, with most between 0.10 and 0.15. Roofing membranes, such as black single-ply roofing, smooth bitumen, grey single-ply roofing, and nominally white (actually grey) granule-surfaced bitumen exhibited solar reflectances of 0.06, 0.06, 0.23 and 0.26, respectively. Gravel roofs had solar reflectances of about 0.12 to 0.34, depending upon gravel colour. The thermal emittances of these non-metallic surfaces were about 0.8 to 0.9. Bare, shiny metal roofs have higher solar reflectance (about 0.60), but their low thermal emittances (about 0.10) make them as hot as a dark roof when wind speed is low. These data suggest that a conventional dark low-sloped roof could be conservatively assumed to have a solar reflectance of about 0.20. An asphalt-aluminium coating has a solar reflectance in the range of 0.30 to 0.61. A freshly applied white elastomeric coating typically has a solar reflectance of 0.60 to 0.85, while that of a new white single-ply roofing membrane usually exceeds 0.70. Soiling and weathering typically reduce the solar reflectances of elastomeric and membrane white roofs by about 0.10 to 0.15 within the first year, with little change in solar reflectance thereafter. It was therefore assumed that a ‘cool’ low-sloped roof should have an initial

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

8

10:19 AM

Page 8

HASHEM AKBARI AND RONNEN LEVINSON

solar reflectance not less than 0.70, an aged solar reflectance not less than 0.55, and a thermal emittance not less than 0.80.

Building energy simulations The DOE-2.1E building energy simulation programme (DOE-2, 2007) was used to estimate the influences of the solar reflectance of the roof’s surface and the thermal resistance of the roof’s insulation on the conditioning energy uses of residential and non-residential buildings with low-sloped roofs. The residential model applies to guest rooms in hotels, patient rooms in hospitals and high-rise residential apartments. The buildings were simulated with electric cooling; gas heating; low, medium and high levels of roof insulation (insulation thermal resistances of R = 3, 11 or 38ft2 h ºF BTU–1); a roof thermal emittance of 0.80; and roof solar reflectances of ρ = 0.05, 0.15, 0.45 and 0.75. The 19 simulation climates ranged from very hot to very cold. The thermal transmittance, or ‘U-factor’, of a roof assembly is the reciprocal of the sum of the thermal resistances of the roof assembly (including insulation) and its surrounding air films. In each climate, simulated values of annual cooling energy use (kWh), annual heating energy use (therms) and annual conditioning energy expenditure (US dollars at US$0.08/kWh and US$0.66/therm) were each regressed to the solar absorptance, α = 1 – ρ, of the roof’s surface, and to the thermal transmittance, U, of the roof assembly. Each climate-specific energy use or energy expenditure E was well fit by an expression of the form: E = C 0 + C1 α + C 2 U + C 3 U α .

[1]

This result was used to find combinations of roof solar absorptance α and roof assembly thermal transmittance U that yield equal annual energy expenditure E. It was also used to determine the extent to which increasing the solar reflectance of a roof from 0.20 (conventional roof) to 0.55 (aged cool roof) can decrease the need for roof insulation without increasing annual energy use. Table 1.2 shows for various cities the thermal resistance of insulation required under a cool roof to achieve the same annual energy use as low, medium and high levels of insulation below a conventional roof. The use of a cool roof reduced the need for insulation in all cases, though more so in hot climates than in cold climates.

Cool-roof credits ASHRAE Standard 90.1 permits both prescriptive and performance (‘energy cost budget’) compliance. ASHRAE Standard 90.1-1999 includes two forms of credits for a cool roof, defined as one with a minimum initial solar reflectance of 0.70 and a minimum thermal emittance of 0.75. For performance compliance, a cool roof on a proposed building is assigned a solar absorptance of 0.55 (solar reflectance of 0.45), while a non-cool roof on a proposed building and the roof on the design building are each assigned a solar absorptance of 0.70 (solar reflectance of 0.30). We note that the solar reflectance of 0.45 assigned to a cool proposed roof is less than that assumed in the preceding analysis; this

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 9

Evolution of Cool-Roof Standards in the US

9

TABLE 1.2 Thermal resistance of insulation below a cool roof (solar reflectance 0.55) that yields the same annual energy expenditure (cost at US$0.08kWh and US$0.66/therm) as a low, medium or high level of insulation (3, 11 or 38ft2 h °F BTU–1) below a conventional roof (solar reflectance 0.20) CDD50a

LOCATION Honolulu, HI Miami, FL Tampa, FL Phoenix, AZ Lake Charles, LA San Diego, CA Fort Worth, TX San Bernardino, CA Atlanta, GA San Francisco, CA Amarillo, TX Portland, OR Seattle, WA Boise, ID Vancouver, Canada Minneapolis, MN Halifax, Canada Bismarck, ND Anchorage, AK Edmonton, Canada

RESIDENTIAL BUILDING R=3 R = 11 0.0 3.5 0.1 4.3 0.4 4.9 0.9 6.2 0.7 5.6 0.1 4.2 1.1 6.7 0.9 6.0 1.0 6.4 1.7 8.1 1.4 7.4 2.0 8.6 2.2 9.2 1.9 8.4 2.2 9.1 2.4 9.7 2.4 9.7 2.3 9.4 3.0 10.9 2.8 10.4

9804 9261 8022 7858 6860 5170 6200 4854 4922 2486 4262 2320 1716 2748 1468 2701 1447 2222 684 880

R = 38 19.4 21.0 21.4 26.2 24.2 19.6 27.5 23.9 25.9 31.2 29.5 31.4 33.9 31.7 32.0 34.5 35.1 33.5 36.8 36.0

NON-RESIDENTIAL BUILDING R=3 R = 11 R = 38 0.0 3.9 16.5 0.3 4.5 18.2 0.5 5.0 19.2 0.9 5.9 22.0 0.7 5.5 21.3 0.2 4.2 16.5 1.1 6.4 23.5 0.9 5.7 21.1 1.0 6.1 22.4 1.3 6.9 24.8 1.4 7.1 26.1 1.8 8.0 27.4 1.9 8.3 28.6 1.8 7.9 27.7 1.9 8.3 28.5 2.1 8.9 31.2 2.2 9.2 32.2 2.2 9.0 31.5 2.6 10.0 34.4 2.5 9.7 33.3

Note: a = cooling degree days based on 50ºF Source: Akbari et al (1998)

may be a typographical error. The standard should be corrected to assign a solar reflectance of 0.55 (solar absorptance of 0.45) to a cool proposed roof. For prescriptive compliance, ASHRAE Standard 90.1-1999 (section 5.3.1.1) approximates the benefits of a cool-roof surface by adjusting the thermal transmittance of the proposed roof assembly. The standard includes the following adjustment to the thermal transmittance of the roof assembly with a cool surface: Uroof adj = Uroof proposed × F

[2]

where Uroof adj is the adjusted roof thermal transmittance for use in demonstrating compliance; Uroof proposed is the thermal transmittance of the proposed roof, as designed; and F is the roof thermal transmittance multiplier from Table 1.3. Since F ≤ 1, this has the effect of decreasing the assumed thermal transmittance (increasing the assumed thermal resistance) of a proposed roofing assembly with a cool surface.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10

10:19 AM

Page 10

HASHEM AKBARI AND RONNEN LEVINSON

TABLE 1.3 Roof thermal transmittance (U-factor) multipliers for cool roofs on buildings other than low-rise residential buildings HDD65a 0–900 901–1800 1801–2700 2799–3600 > 3600

(HDD18)b (0–500) (501–1000) (1001–1500) (1501–2000) (> 2000)

ROOF U-FACTOR MULTIPLIER 0.77 0.83 0.85 0.86 1.00

Notes: a = heating degree days based on 65ºF b = heating degree days based on 18ºC Source: ASHRAE 90.1-1999 (ASHRAE, 1999, Table 5.3.3.1B)

TABLE 1.4 Roof thermal transmittance (U-factor) multipliers for cool roofs on buildings other than low-rise residential buildings CLIMATE ZONE 1 2 3 4–8

ROOF U-FACTOR MULTIPLIER 0.77 0.83 0.85 1

Source: ASHRAE 90.1-2004 (ASHRAE, 2004a, Table 5.5.3.1)

Revisions ASHRAE Standard 90.1-2001 (ASHRAE, 2001) retains the same provisions for cool-roof credits. The current version of this standard, ASHRAE Standard 90.1-2004 (ASHRAE, 2004a), tabulates thermal transmittance multipliers by US climate zone (see Figure 1.1), rather than by heating degree days (see Table 1.4).

ASHRAE STANDARD 90.2 The procedure for incorporating the effect of roof solar reflectance in the ASHRAE Standard 90.2 residential standards was similar to that followed for ASHRAE Standard 90.1. This section summarizes the cool-roof analysis performed for ASHRAE Standard 90.2 (Akbari et al, 2000).

Building-energy simulations The Standard 90.2 task group used DOE-2.1E to simulate, in 29 climates, the influence of a solar-reflective roof on the heating and cooling energy uses of a residential building prototype used in previous 90.2 analyses and described by Akbari et al (2000). Parameters varied in the prototype buildings included presence or absence of an attic; duct location (attic or conditioned space);1 thermal resistance of duct insulation (2, 4 or 6ft2 h ºF BTU–1); roof solar reflectance (0.10, 0.25, 0.50 or 0.75); and thermal resistance of ceiling insulation

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

Evolution of Cool-Roof Standards in the US

FIGURE 1.1 Locations of the eight ASHRAE-defined climate zones in the US

10:19 AM

Source: ASHRAE (2007)

Zone 1 includes Hawaii, Guam, Puerto Rico, and the Virgin Islands

7/11/08

All of Alaska in Zone 7 except for the following Boroughs in Zone 8: Bethel Northwest Arctic Dellingham Southeast Fairbanks Fairbanks N. Star Wade Hampton Nome Yukon-Koyukuk North Slope

ABER_CH_01 Page 11

11

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

12

10:19 AM

Page 12

HASHEM AKBARI AND RONNEN LEVINSON

(1, 11, 19, 30 or 49ft2 h ºF BTU–1). Buildings were cooled electrically, and heated with an electric heat pump, electric resistance or natural gas. All roofs were assigned a thermal emittance of 0.80. In each climate, simulated values of annual cooling energy use (kWh), annual heating energy use (therms) and annual conditioning energy expenditure (US dollars at US$0.08/kWh and US$0.69/therm) were each regressed to the solar absorptance α of the roof’s surface, and to the thermal transmittance U of the roof assembly. Each climate-specific energy use or energy cost E was well fit by an expression of the form: E = C 0 + C1U + C 2 U 2 + C 3 U α.

[3]

This result was used to find combinations of roof solar absorptance and roof-assembly thermal transmittance that yield equal annual energy cost. It was also used to determine the multiplier by which the thermal transmittance of a roof assembly can be increased without raising annual energy use when the solar reflectance of the roof’s surface is increased to 0.55 (cool white steep-sloped roof) from 0.10 (conventional dark steepsloped roof). Table 1.5 shows this multiplier for various prototype configurations in the 29 US cities simulated. Multipliers exceeded unity in all but four cities, and were at least 0.94 in all cities – that is, all but four cities exhibited positive savings, and the penalties in cold climates were small.

Cool-roof credits ASHRAE Standard 90.2-2004 permits both prescriptive and performance (‘energy cost budget’) compliance. The standard includes two forms of credit for a cool roof, defined as a roof with a minimum initial solar reflectance of 0.65 and a minimum thermal emittance of 0.75; and/or a solar reflectance index (SRI) of at least 75 calculated under the medium wind-speed conditions specified by ASTM Standard E1980: Standard Practice for Calculating Solar Reflectance Index of Horizontal and Low-Sloped Opaque Surfaces (ASTM, 1998). SRI is a relative index of the steady-state temperature of a roof’s surface on a typical summer afternoon. SRI is defined to be zero for a clean black roof (solar reflectance 0.05, thermal emittance 0.90) and 100 for a clean white roof (solar reflectance 0.80, thermal emittance 0.90). Thus, warm surfaces have low SRI, and cool surfaces have high SRI. For performance compliance (section 8.8.3.1), a cool roof on a proposed building is assigned its actual solar absorptance, or possibly a solar absorptance of 0.35; the standard’s language is ambiguous. A non-cool roof on a proposed building and the roof on the design building are each assigned a solar absorptance of 0.20 (solar reflectance of 0.80). However, the authors believe the latter to be a typographical error, since the logical value would be a solar absorptance of 0.80 (solar reflectance of 0.20). For prescriptive compliance (section 5.5), ASHRAE Standard 90.2-2004 approximates the benefits of a cool-roof surface by adjusting the thermal transmittance of the proposed ceiling (the authors believe that ceiling may actually mean roof assembly). The standard

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 13

Evolution of Cool-Roof Standards in the US

13

TABLE 1.5 Multiplier by which the thermal transmittance (U-factor) of a residential roof assembly can be increased without raising annual energy use when the solar reflectance of the roof’s surface is increased to 0.55 (cool) from 0.10 (conventional) LOCATION

Honolulu, HI Miami, FL Brownsville, TX Phoenix, AZ Jacksonville, FL Tucson, AZ Lake Charles, LA El Paso, TX Los Angeles, CA San Diego, CA Las Vegas, NV Fresno, CA Charleston, SC Fort Worth, TX Fort Smith, AZ Sacramento, CA Albuquerque, NM Los Angeles, CA St Louis, MO Washington, DC Dodge, KS North Omaha, NE Denver, CO Winnemucca, NV New York, NY Bismarck, ND Redmond, OR Madison, WI Seattle, WA Fairbanks, AK San Francisco, CA

CDD65a

HDD65b

4329 4127 3563 3815 2657 2763 2624 2046 943 766 3067 1884 2010 2415 1895 1144 1211 470 1437 1044 1371 1051 623 604 1002 408 194 521 127 29 69

0 141 659 1154 1437 1554 1683 2597 1309 1076 2293 2602 2209 2304 3351 2794 4361 1291 5021 5233 5353 6047 6007 6444 5090 8666 6732 7495 4867 14095 3239

MULTIPLIER DUCTS DUCTS IN IN ATTICS CONDITIONED SPACE 2.62 2.62 2.19 2.19 1.60 1.60 1.53 1.53 1.52 1.52 1.49 1.49 1.46 1.46 1.42 1.42 1.38 1.38 1.37 1.37 1.37 1.37 1.34 1.34 1.33 1.33 1.31 1.31 1.24 1.24 1.22 1.22 1.19 1.19 1.16 1.16 1.11 1.11 1.09 1.09 1.09 1.09 1.08 1.08 1.06 1.06 1.06 1.06 1.05 1.05 1.02 1.02 1.01 1.01 1.01 1.01 0.97 0.97 0.97 0.97 0.94 0.94

ROOFS WITH NO ATTICS 1.69 1.67 1.61 2.39 1.59 2.76 2.02 1.68 1.64 1.69 1.65 1.56 1.58 1.64 1.63 1.61 1.40 1.55 1.50 1.23 1.34 1.27 1.33 1.28 1.28 1.25 1.26 1.18 1.12 1.09 0.98

Notes: Ducts in the attics have R-4 insulation (4 ft2 h °F BTU–1); ducts in the conditioned space are uninsulated a = cooling degree days based on 65°F b = heating degree days based on 65°F Source: Akbari et al (2000)

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

14

10:19 AM

Page 14

HASHEM AKBARI AND RONNEN LEVINSON

includes the following adjustment to the thermal transmittance of the ceiling under a cool roof: U ceiling adj = U ceiling proposed × Multiplier

[4]

where Uceiling adj is the adjusted ceiling thermal transmittance for use in demonstrating compliance; Uceiling proposed is the thermal transmittance of the proposed ceiling, as designed; and Multiplier is the ceiling thermal transmittance multiplier from Table 1.6. Since Multiplier ≥ 1, this has the effect of increasing the assumed thermal transmittance (decreasing the assumed thermal resistance) of a proposed roofing assembly with a cool surface. Hence, we believe the multiplier values to be in error. It is possible that each value in Table 1.6 should be replaced by its reciprocal to yield multipliers that do not exceed unity.

Revisions The current version of this standard, ASHRAE Standard 90.2-2007 (ASHRAE, 2007), retains the same cool-roof credits for performance compliance. However, the cool-roof credits for prescriptive compliance have been modified. Rather than specify ceiling thermal transmittance multipliers, the new standard prescribes reduced thermal resistances for ceilings under cool roofs in climate zones 1 to 3 (see Table 1.7).

CALIFORNIA TITLE 24 STANDARDS In 2001, cool-roof credits were added to California’s Title 24 Standards. The standards were upgraded in 2005 to prescriptively require cool roofs on non-residential buildings with low-sloped roofs. The California Energy Commission is currently (2007) considering adding prescriptive cool-roof requirements for all other buildings to the 2008 standards.

Cool-roof credits (2001) A Codes and Standards Enhancement (CASE) study prepared in 2000 by the Pacific Gas & Electric company indicated that cool roofs would cost-effectively save energy and

TABLE 1.6 Ceiling thermal transmittance (U-factor) multipliers for residential cool roofs: It is possible that these multipliers should be replaced by their reciprocal to yield values less than or equal to unity ZONE 1 2 3 4 5 6,7,8

CEILINGS WITH ATTICS 1.50 1.25 1.20 1.15 1.10 1.00

CEILINGS WITHOUT ATTICS 1.30 1.30 1.20 1.20 1.10 1.00

Source: ASHRAE 90.2-2004 (ASHRAE, 2004b, Table 5.5)

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 15

Evolution of Cool-Roof Standards in the US

15

TABLE 1.7 Ceiling thermal resistances (ft2 h ºF BTU–1) prescribed by ASHRAE Standard 90.2-2007 for ceilings under conventional (non-cool) and cool residential roofs, derived from ASHRAE Standard 90.2-2007: Reduced requirements for cool-roofed buildings are shaded CLIMATE ZONE

1 2 3 4 5 6 7 8

CEILINGS WITH ATTICS WOOD FRAME STEEL FRAME CONVENCOOL CONVENCOOL TIONAL ROOF TIONAL ROOF ROOF ROOF 30 20 30 20 30 24 30 24 30 27 30 27 38 38 38 38 43 43 43 43 49 49 49 49 49 49 49 49 52 52 52 52

CEILINGS WITHOUT ATTICS WOOD FRAME STEEL FRAME CONVENCOOL CONVENCOOL TIONAL ROOF TIONAL ROOF ROOF ROOF 13 10 19 10 22 17 19 17 22 18 22 18 22 22 22 22 26 26 26 26 38 38 38 38 38 38 38 38 38 38 38 38

Source: ASHRAE Standard 90.2-2007 (ASRAE, 2007, Tables 5.2 and 5.6.1)

reduce peak-power demand in California (Eilert, 2000). In January 2001, the state of California followed the approach of ASHRAE Standards 90.1 and 90.2 by adding a cool-roof credits to Title 24 (CEC, 2001). Roofs are considered cool if they have an initial solar reflectance not less than 0.70 and a thermal emittance not less than 0.75. An exception lowers this minimum initial solar reflectance requirement to 0.40 for tile roofs. Cool roofs were not made a prescriptive requirement. For performance compliance, a cool roof on a proposed building was assigned a solar absorptance of 0.45 (solar reflectance of 0.55). The roof of a standard (design) building was assigned a solar absorptance of 0.70 (solar reflectance of 0.30), as was a non-cool roof on a proposed building.

Prescriptive requirements for low-sloped roofs on non-residential buildings (2005) In 2002, the Berkeley Lab Heat Island Group began to investigate the prescriptive requirement in Title 24 of cool roofs for non-residential buildings with low-sloped roofs. The analysis approach was similar to that used to develop ASHRAE Standards 90.1 and 90.2. Steps included reviewing the physics of cool roofs; reviewing measurements of cool-roof energy savings reported in the literature; investigating the market availability of cool roofs; surveying cost premiums (if any) for cool roofs; reviewing roofing material durability; investigating the environmental consequences of cool roofs; and performing hourly simulations of building energy use to estimate the energy and peak-power demand savings potentials of cool roofs (Levinson et al, 2005a). A review of low-sloped roofing technologies indicated that cool options (solarreflective products or coatings) were available for nearly all types of low-sloped roofs,

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

16

10:19 AM

Page 16

HASHEM AKBARI AND RONNEN LEVINSON

including the three dominant products: built-up roofing, modified bitumen and single-ply membrane. A cool roof was defined as a roof with: ● an initial thermal emittance not less than 0.75 and an initial solar reflectance not less

than 0.70; and/or ● an initial thermal emittance (εinitial) less than 0.75 and an initial solar reflectance not

less than 0.70 + 0.34 × (0.75 – (εinitial).

The second term in this expression is the solar-reflectance premium required to ensure that under ASTM E1980 medium wind-speed conditions, the aged (weathered) temperature of a roof with low thermal emittance will not exceed that of an aged (weathered) cool roof with high thermal emittance. DOE-2.1E building energy simulations performed in California’s 16 climate zones (see Figure 1.2) indicate that the use of a cool roof on a prototypical California Title 24 nonresidential building with a low-sloped roof yields average annual cooling energy savings of 3.2kWh/m2, average annual natural gas deficits of 5.6MJ/m2, average source energy savings of 30MJ/m2, and average peak-power demand savings of 2.1W/m2. Total savings – initial cost savings from downsizing cooling equipment plus the 15-year net present value (NPV) of energy savings with time-dependent valuation (TDV) – ranged from US$1.90/m2 to US$8.30/m2 (see Figure 1.3). The typical cost premium for a cool low-sloped roof is US$0.0/m2 to US$2.2/m2. Cool roofs with premiums up to US$2.2/m2 are expected to be cost-effective in climate zones 2 to 16; those with premiums not exceeding US$1.9/m2 are expected also to be cost-effective in climate zone 1. Therefore, the 2005 California Title 24 Standards adopted a cool-roof prescriptive requirement in all California climate zones for nonresidential buildings with low-sloped roofs. Non-residential buildings with low-sloped roofs that do not meet this new prescriptive requirement may still achieve performance compliance.

Proposed prescriptive requirements for steep-sloped non-residential roofs, steep-sloped residential roofs and low-sloped residential roofs (2008) In 2005, the Berkeley Lab Heat Island Group began to investigate the merits of adding to the 2008 Title 24 Standards prescriptive requirements for the use of cool roofs on all other types of buildings, including non-residential buildings with steep-sloped roofs, residential buildings with steep-sloped roofs and residential buildings with low-sloped roofs. The methodology was similar to that used to consider the prescriptive requirement in the 2005 Title 24 Standards of cool low-sloped roofs for non-residential buildings. In these 2008 cycle analyses, the MICROPAS building energy simulation tool (MICROPAS, 2007) was used to simulate the hourly energy use of prototypical residential and small non-residential buildings (Akbari et al, 2006; Wray et al, 2006).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 17

Evolution of Cool-Roof Standards in the US

17

Source: Eley Associates

FIGURE 1.2 Locations of the 16 California climate zones

Steep-sloped roofs on non-residential buildings Berkeley Lab developed a non-residential prototype building that prescriptively complies with the 2005 Title 24 Standards. The energy use of this building was simulated with conventional and cool versions of three different steep-sloped (5:12) roofing products: fibreglass asphalt shingle, concrete tile and polymer-coated metal. Each conventional product had a solar reflectance of 0.10. The cool shingle had a solar reflectance of 0.25,

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

7/11/08

18

10:19 AM

Page 18

HASHEM AKBARI AND RONNEN LEVINSON

10 9 8.3

8 7 6 5.1

5.7

5.8

12

2

6.6

6.6

10

7

6.9

6.9

7.6

7.7

6

9

7.2

5.2

5 4

4.0

4.1

3

5

4.3

3 2

1.9

Cost premium (0 - 2.2 US$/m2)

Total savings [equipment + 15-year NPV energy] (US$/m2)

ABER_CH_01

1 0 1

16

4

11

141

31

5

8

California climate zone Source: Levinson et al (2005a)

FIGURE 1.3 Total savings per unit roof area (cooling equipment savings plus 15-year NPV of energy savings with TDV; US$/m2) by California climate zone, simulated for a prototypical non-residential Title 24 building with a cool low-sloped roof

while the cool tile and cool metal products each had a solar reflectance of 0.40. All products were assigned a thermal emittance of 0.90. Total savings per unit roof area – defined as initial cost savings from downsizing cooling equipment, plus the 30-year NPV of TDV energy savings – ranged from US$2.8/m2 to US$24.4/m2 across California’s 16 climate zones (see Table 1.8). The typical cost premium for a cool steep-sloped roofing product is US$0.0/m2 to US$2.2/m2. Cool roofs with premiums of up to US$2.2/m2 are expected to be cost-effective in all 16 climate zones. At the time of writing, California is considering including in its 2008 Title 24 Standards a prescriptive cool-roof requirement in all climate zones for non-residential buildings with steep-sloped roofs.

Low-sloped roofs on residential buildings Berkeley Lab developed a residential prototype building that prescriptively complies with the 2005 Title 24 Standards. The energy use of this building was simulated with conventional (ρ = 0.20) and cool (ρ = 0.55) versions of a low-sloped (horizontal) built-up roof. While the 2005 Title 24 Standards prescriptively require a sub-roof radiant barrier for residential buildings in some climate zones (2, 4 and 8 to 15), radiant barriers are not

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 19

Evolution of Cool-Roof Standards in the US

19

TABLE 1.8 Total savings per unit roof area (cooling equipment

savings plus 30-year NPV of energy savings with TDV; US$/m2) by California climate zone for a non-residential building with a steepsloped roof CLIMATE ZONE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SHINGLE 2.8 5.9 4.7 6.3 5.1 10.1 9.1 11.6 11.9 8.6 7.4 6.9 8.3 9.1 11.4 5.2

TILE 4.8 10.4 8.2 11.3 8.9 17.8 15.8 20.4 20.8 15.6 13.5 12.5 15.3 16.6 21.2 9.2

METAL 5.8 12.0 9.6 12.8 10.4 20.7 18.7 23.9 24.4 17.5 15.1 14.0 17.0 18.5 23.3 10.6

Note: Cool fibreglass asphalt shingles were assigned a solar reflectance of 0.25; cool concrete tiles and metal products, 0.40; and all conventional products, 0.10 Source: original material for this chapter

usually installed in houses with low-sloped roofs (in climates zones where radiant barriers are prescriptively required, a house without a radiant barrier would have to demonstrate performance compliance). Without a radiant barrier (typical of low-sloped roofs, in general, and pre-2000 construction, in particular), total savings – defined as initial cost savings from downsizing cooling equipment, plus the 30-year NPV of TDV energy savings – ranged from –US$2.4/m2 to US$8.2/m2 across California’s 16 climate zones (see Table 1.9). With a radiant barrier, the savings ranged from –US$2.5/m2 to US$4.7/m2. The negative savings occurred in coastal California climate zones with minimal summertime cooling requirements. The presence of a sub-roof radiant barrier reduces cool-roof energy savings, just as the presence of a cool roof reduces radiantbarrier energy savings. The typical cost premium for a cool roof is US$0.0/m2 to US$2.2/m2. Cool roofs with premiums of up to US$2.2/m2 are expected to be cost-effective in some climates zones. At the time of writing, California is considering including in its 2008 Title 24 Standards a prescriptive cool-roof requirement in hot Central Valley climates for residential buildings with low-sloped roofs.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

20

10:19 AM

Page 20

HASHEM AKBARI AND RONNEN LEVINSON

TABLE 1.9 Total savings per unit roof area (cooling equipment savings plus 30-year NPV

of energy savings with TDV; US$/m2) by California climate zone for a residential building with a low-sloped roof CALIFORNIA CLIMATE ZONE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

WITHOUT RADIANT BARRIER –2.4 1.9 –0.4 0.9 –0.6 0.7 1.2 2.9 4.2 5.9 5.3 4.2 5.9 4.9 8.2 2.3

WITH RADIANT BARRIER –2.5 –0.1 –1.1 –0.4 –1.4 –0.2 –0.1 0.9 1.8 2.5 2.5 1.6 3.0 2.1 4.7 0.1

Note: The cool roof was assigned a solar reflectance of 0.55; the conventional roof, 0.10. We note that while California’s Title 24 Standards prescribe the installation of sub-roof radiant barriers for residential buildings in California climate zones 2, 4 and 8 to 15, it is not a common building practice for homes with low-sloped roofs. The shaded values are appropriate to each climate zone’s radiant-barrier requirement Source: original material for this chapter

Steep-sloped roofs on residential buildings Berkeley Lab developed a residential prototype building that prescriptively complies with the 2005 Title 24 Standards. The energy use of this building was simulated with conventional and cool versions of three different steep-sloped (5:12) roofing products: fibreglass asphalt shingle, concrete tile and polymer-coated metal. Each conventional product had a solar reflectance of 0.10. The cool shingle had a solar reflectance of 0.25, while the cool tile and cool metal products each had a solar reflectance of 0.40. All products were assigned a thermal emittance of 0.90. The 2005 Title 24 Standards prescriptively require a sub-roof radiant barrier for residential buildings in some climate zones (2, 4 and 8 to 15); but they are not present on most existing houses. Without a radiant barrier (typical of pre-2000 construction), total savings – initial cost savings from downsizing cooling equipment, plus the 30-year NPV of TDV energy savings per unit roof area – ranged from –US$1.7/m2 to US$18.6/m2 across California’s 16 climate zones (see Table 1.10). With a radiant barrier, the savings ranged from –US$1.3/m2 to US$12.1/m2. Cool shingles incurred smaller saving (and penalties) than did cool tiles and cool metal products because the solar reflectance of the cool

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 21

Evolution of Cool-Roof Standards in the US

21

TABLE 1.10 Total savings per unit roof area (cooling equipment savings plus 30-year NPV of energy

savings with TDV; US$/m2) by California climate zone for a residential building with a steep-sloped roof CALIFORNIA CLIMATE ZONE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

WITHOUT RADIANT BARRIER SHINGLE TILE METAL –0.9 –1.7 –1.7 2.2 3.0 5.0 0.2 0.2 0.5 0.9 1.1 2.2 0.4 0.2 0.9 0.7 1.0 1.7 1.2 1.7 2.7 3.1 4.5 6.8 3.5 5.4 7.7 5.7 8.9 12.4 5.5 8.8 12.2 3.7 5.8 8.2 6.8 11.1 14.8 5.3 8.3 11.5 8.5 13.8 18.6 2.8 4.5 5.9

WITH RADIANT BARRIER SHINGLE TILE METAL –0.6 –1.2 –1.3 1.3 2.2 2.9 0.1 0.1 0.3 0.5 0.9 1.2 0.2 0.1 0.4 0.5 0.6 1.0 0.7 1.0 1.4 1.8 2.9 3.9 2.2 3.7 4.8 3.5 5.9 7.3 3.5 6.1 7.6 2.4 4.1 5.2 4.1 7.4 8.8 3.3 5.7 7.0 5.6 9.8 12.1 1.6 2.6 3.3

Note: Cool fibreglass asphalt shingles were assigned a solar reflectance of 0.25; cool concrete tiles and metal products, 0.40; and all conventional products, 0.10. While California’s Title 24 Standards prescribe the installation of sub-roof radiant barriers for residential buildings in climate zones 2, 4 and 8 to 15, it was not common practice in pre-2000 construction. The shaded values are appropriate to each climate zone’s radiant-barrier requirement. Source: original material for this chapter

shingle exceeded that of the conventional shingle by 0.15, rather than by 0.30. The negative savings occurred in coastal California climate zones with minimal summertime cooling requirements. The presence of a sub-roof radiant barrier reduces cool-roof energy savings, just as the presence of a cool roof reduces radiant-barrier energy savings. The typical cost premium for a cool roof is US$0.0/m2 to US$2.2/m2. Cool roofs with premiums of up to US$2.2/m2 are expected to be cost-effective in some climates zones. At the time of writing, California is considering including in its 2008 Title 24 Standards a prescriptive cool-roof requirement in hot Central Valley climates for residential buildings with steep-sloped roofs.

COOL-ROOF PROVISIONS IN OTHER STANDARDS AND PROGRAMMES Many US states have adopted building energy efficiency codes from ASHRAE Standard 90.1 or the International Energy Conservation Code (IECC). Aside from California, these include the cities of Atlanta (Georgia) and Chicago (Illinois); the states of Florida, Georgia, and Hawaii; and the territory of Guam. Cool-roof requirements have also been developed

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

22

10:19 AM

Page 22

HASHEM AKBARI AND RONNEN LEVINSON

by several voluntary energy-efficiency programmes, including the US Environmental Protection Agency (EPA) Energy Star label, the Leadership in Energy and Environmental Design (LEED) Green Building Rating System, and the cool-roof rebate programmes offered by the state of California and its utilities. An earlier report by Eley Associates (2003a) summarizes the history of cool-roof credits and requirements in these policies through 2003, with particular attention paid to the political process of their development. Here we review the treatment of cool roofs in several of these standards and programmes through 2007, focusing on technical details.

STANDARDS OTHER THAN ASHRAE 90.1, ASHRAE 90.2 AND TITLE 24 International Energy Conservation Code The 2003 International Energy Conservation Code (IECC) does not explicitly address the use of cool roofs. However, section 801.2 allows commercial buildings to comply with the 2003 IECC by satisfying the requirements of ASHRAE Standard 90.1, which, in turn, offers cool-roof credits. The 2003 IECC provides neither direct nor indirect cool-roof credits for residential buildings (ICC, 2003). The 2006 IECC retains the link to ASHRAE Standard 90.1 for commercial buildings, and explicitly offers cool-roof credits for residential buildings through performance compliance. Table 404.5.2(1) assigns to the roof on the reference residential building a solar absorptance of 0.75 (solar reflectance of 0.25) and a thermal emittance of 0.90, while the roof on the proposed building is assigned its proposed values of solar absorptance and thermal emittance (ICC, 2006). The adoption as of May 2007 of IECC and/or ASHRAE standards by individual US states is detailed in Figures 1.4 (commercial building codes) and 1.5 (residential building codes).

Chicago, Illinois To mitigate urban heat islands, the city of Chicago, Illinois, added a provision to section 18-13-303 of its 2001 Energy Conservation Code requiring that low-sloped roofs (those with a ratio of rise to run not greater than 2:12) exhibit an initial solar reflectance not less than 0.65, and a solar reflectance of at least 0.50 three years after installation. Mediumsloped roofs (those with a ratio of rise to run greater than 2:12 and less than or equal to 5:12) were required to have initial and three-year solar reflectances of at least 0.15. Both low- and medium-sloped roofs were required to have a minimum thermal emittance of 0.90. Roofs or portions of roofs that use photovoltaic, solar-thermal or roof garden systems were exempt from these requirements (Chicago, 2001). The cool-roof provisions of this code have been amended several times since 2001. The current (2007) code requires that low-sloped roofs installed by 31 December 2008 have an initial solar reflectance not less 0.25, while those installed after that day must use products that qualify for the US EPA Energy Star label (initial and aged solar reflectances not less than 0.65 and 0.50, respectively). Medium-sloped roofs must have an initial solar reflectance of at least 0.15. Chicago’s cool-roof standard has been weakened by the elimination of its thermal emittance requirement and the establishment of a very low minimum solar reflectance

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 23

Evolution of Cool-Roof Standards in the US

23

Source: Building Codes Assistance Project (BCAP, 2007)

FIGURE 1.4 Adoption of commercial building energy codes by US states as of May 2007

requirement for medium-sloped roofs. The effects of weakened standards on roof surface temperature are quantified below in the discussion of the Energy Star programme.

Florida The state of Florida first offered cool-roof credits for residential buildings in the 2001 edition of the Florida Building Code. The code’s whole-building performance method for compliance (Form 600A) multiplies the area of each envelope component by a ‘summer point multiplier’ and a ‘winter point multiplier’ to estimate its contributions to the summer cooling load and winter heating load. A cool-roof credit introduced in 2001 allows a proposed home with a white roof (solar reflectance ≥ 0.65, thermal emittance ≥ 0.80) to multiply its summer point multiplier by a credit factor of 0.55 and its ceiling winter point multiplier by a credit factor of 1.044 (FBC, 2001, sections 607.1.A.5 and 607.2.A.3.6). This reduces the estimated summer ceiling heat load of a white-roofed proposed home by 45 per cent and increases its estimated winter ceiling heat load by 4.4 per cent. The current (2004) version of the code retains this ‘white roof’ credit (FBC, 2004).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

24

10:19 AM

Page 24

HASHEM AKBARI AND RONNEN LEVINSON

Source: Building Codes Assistance Project (BCAP, 2007)

FIGURE 1.5 Adoption of residential building energy codes by US states as of May 2007

The 2007 code will use EnergyGauge USA FlaRes2007 (EnergyGauge, 2007), rather than point multipliers, to estimate the annual energy use of residential buildings (FBC, 2007). EnergyGauge USA FlaRes2007 is a building energy model based on DOE-2.1E that incorporates an improved attic model (Parker, 2005). The 2007 code will require that the annual energy budget of a proposed (‘as-built’) home not exceed that of a reference (‘baseline’) home whose roof has a solar reflectance of 0.25 and a thermal emittance of 0.90. Radiative properties that make the proposed roof cooler than the reference roof (e.g. a solar reflectance above 0.25) will permit the consumption of more energy elsewhere in the building. Conversely, radiative properties that make the proposed roof warmer than the reference roof (e.g. a solar reflectance below 0.25, or a thermal emittance less than 0.90) will require increased energy efficiency in other parts of the home. The solar reflectance of 0.25 to be assigned to the roof of the reference home in the 2007 code is greater than the reference roof solar reflectance of 0.15 used to generate the cool-roof credit factors present in the 2004 code (Parker, 2007). If the initial solar reflectance of the proposed home’s roofing product has not been measured by the manufacturer, it will be set to 0.04. If the initial thermal emittance is

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 25

Evolution of Cool-Roof Standards in the US

25

unmeasured, it will be assigned a value of 0.90 (FBC, 2007). We note that the latter provision can significantly overestimate the true thermal emittance of a bare-metal roofing product, which is typically less than 0.20. We address the influence of low thermal emittance on roof surface temperature in our discussion of the Energy Star programme. In 2004, the Florida Building Code adopted prescriptive and performance cool-roof credits for commercial buildings that are essentially the same as those in ASHRAE Standard 90.1-2004. The only difference is that the thermal transmittance multipliers used for prescriptive compliance in section 13.404.1.C.1 are mapped to Florida climate zones, rather than to US climate zones (FBC, 2004; Swami, 2007). The proposed 2007 Florida Building Code retains the same cool-roof provisions.

Hawaii Building energy codes in Hawaii are set by county ordinance, rather than by state law. Over 80 per cent of the state’s population live in the counties of Honolulu and Maui (Census, 2007). In 2001, the county of Honolulu amended its ordinances to prescriptively require that roof assemblies on low-rise residential buildings include at least one of the following: ● ● ● ●

insulation with a thermal resistance of 19ft2 h ºF BTU–1; 2 inches of continuous foam-board insulation; a radiant barrier and attic ventilation; or a cool roof with a solar reflectance not less than 0.70 and a thermal emittance not less than 0.75 (ROH, 2001, section 32-14.2).

The county of Maui followed suit in 2005 (MCC, 2004). The current (2007) ordinances of Honolulu and Maui retain this cool-roof requirement. The counties of Kauai and Hawaii neither credit nor prescribe the use of cool roofs on residential buildings. In 2001, 2002 and 2005, respectively, the counties of Honolulu, Kauai and Maui adopted cool-roof credits for commercial and high-rise residential buildings based on ASHRAE Standard 90.1-1999 (Wiig, 2007). The building envelope prescriptions for these buildings (e.g. section 32-8 of ROH, 2004) use a modified cool-roof credit adopted from the Hawaii Model Energy Code (Eley Associates, 2003b). By requiring that the product of the roof thermal transmittance (BTU ft–2 h–1 ºF–1), roof solar absorptance and a radiant barrier credit (0.33 if present, 1 if absent) be less than 0.05BTU ft–2 h–1 ºF–1, this provision allows for the use of reduced insulation under a roof with high solar reflectance. The envelope prescription does not set a minimum requirement for the thermal emittance of the roof. Hence, it may credit the use of a solar-reflective but hot bare-metal roof, as addressed in the discussion of the Energy Star programme. The county of Hawaii follows ASHRAE Standard 90.1-1989 for commercial buildings, which neither prescribes nor credits cool roofs.

Guam Guam’s code for non-residential and high-rise residential buildings (adopted in 1995) and its code for low-rise residential buildings (adopted in 2000) establish identical prescriptive

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

26

10:19 AM

Page 26

HASHEM AKBARI AND RONNEN LEVINSON

requirements for roofs on air-conditioned buildings (Eley Associates, 2007). For these buildings, ‘mass’ roof assemblies – that is, roofs made of concrete 4 inches or greater in thickness; having heat capacity per unit area greater than 7.0BTU ft–2 ºF–1; and/or weighing more than 35lb ft–2 – must have: ● a cool (‘high albedo’) surface of solar reflectance not less than 0.70 and thermal

emittance not less than 0.75; ● R-11 insulation (11ft2 h °F BTU–1) in the interior furring space; ● 2 inches of continuous insulation; or ● thermal transmittance not exceeding 0.12BTU h–1 ft–2 ºF–1.

Air-conditioned buildings with other types of roofs are required to have more insulation and/or a lower thermal transmittance than mass roofs, but cannot apply a cool-roof surface towards prescriptive compliance. Air-conditioned buildings that meet the code’s mandatory requirements but not its prescriptive requirements can achieve performance compliance via either ASHRAE Standard 90.1-1989 or via a building envelope trade-off option. The latter requires that the energy performance factor (EPF) of a proposed building not exceed that of a reference (‘budget’) building. The EPF of a building includes the EPF of its roof, which for mass roofs and roofs on metal buildings (buildings with metal sheathing and metal framing) is defined as the product of the roof’s area, thermal transmittance and solar absorptance. In EPF calculations, the solar absorptance of a proposed or reference roof is set to 0.30 if the roof is cool (solar reflectance ≥ 0.70, thermal emittance ≥ 0.75), or 0.70 otherwise. The roof of the reference building may or may not be cool, depending upon whether the roof assembly chosen for the reference building uses a cool surface to comply with the code (Eley, 2007).

VOLUNTARY ENERGY EFFICIENCY PROGRAMMES US EPA Energy Star label To qualify for its Energy Star label, the US EPA currently requires that low-sloped roofing products (those installed on roofs with a ratio of rise to run not exceeding 2:12) have initial and three-year aged solar reflectances not less than 0.65 and 0.50, respectively. Steepsloped roofing products (those installed on roofs with a ratio of rise to run greater than 2:12) must have initial and three-year aged solar reflectances not less than 0.25 and 0.15, respectively (EPA, 2007). We note that the Energy Star cool-roof requirements have two weaknesses. First, by specifying neither a minimum thermal emittance nor a minimum solar reflectance index, they permit the use of bare-metal roofs with high solar reflectance but low thermal emittance. Under ASTM E1980 medium wind-speed conditions, the surface of an aged bare-metal roof with a solar reflectance of 0.50 and a thermal emittance of 0.15 would be about 12K (22°F) hotter than that of an aged white roof with a solar reflectance of 0.50 and thermal emittance of 0.80. Second, the minimum three-year aged solar reflectance required for a steep-sloped roof (0.15) excludes only the hottest of roofing materials, such as granule-surfaced fibreglass asphalt shingles coloured with conventional dark pigments.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 27

Evolution of Cool-Roof Standards in the US

27

Many cool roofing products for steep-sloped roofs attain an aged solar reflectance of at least 0.30. Under these conditions, the surface of a roof with a solar reflectance of 0.15 and a thermal emittance of 0.80 will be 10K (18°F) hotter than that of a roof with a solar reflectance of 0.30 and a thermal emittance of 0.80.

LEED Green Building Rating System The Leadership in Energy and Environmental Design (LEED) Green Building Rating System assigns one rating point for the use of a cool roof in its Sustainable Sites Credit 7.2 (Heat Island Effect, Roof). LEED Version 2.0 (2001) requires a cool roof to either cover at least 75 per cent of its surface with materials that have initial and three-year aged solar reflectances of at least 0.65 and 0.50, respectively, and a thermal emittance of at least 0.90; or cover no less than 50 per cent of its surface with vegetation (GBC, 2001). LEED Version 2.1 (2002) requires a cool roof to either: ● cover at least 75 per cent of its surface with Energy-Star compliant products that also

have a thermal emittance of at least 0.90; ● cover no less than 50 per cent of its surface covered by vegetation; or ● cover at least 75 per cent of its surface with a combination of these two materials

(GBC, 2002). Compared to version 2.0, version 2.1 reduces the minimum initial solar reflectance required for steep-sloped roofs (ratio of rise to run greater than 2:12) to 0.25 from 0.65, and the minimum aged solar reflectance to 0.15 from 0.50. We note that the minimum thermal emittance requirement of 0.90 in versions 2.0 and 2.1 is unnecessarily high, as most high-emittance materials have thermal emittances in the range of 0.80 to 0.95. The LEED requirement of 0.90 tends to exclude many cool materials, such as white roofs, whose thermal emittances may lie slightly below 0.90. This issue is compounded by the high uncertainty (up to ±0.05) in measuring the thermal emittance of thermally massive materials. The less stringent minimum thermal-emittance requirement of 0.75 used in the ASHRAE and Title 24 Standards definitions of a cool roof is designed to include all high-emittance materials, most of which are expected (though not required) to exhibit thermal emittances of at least 0.80. LEED Version 2.2 (2005) uses SRI, rather than solar reflectance, thermal emittance or Energy-Star compliance, to qualify a non-vegetated cool roof (GBC, 2005). LEED Version 2.2 requires a cool roof to either: ● cover at least 75 per cent of its surface with products that have a minimum SRI of 78

(low-sloped roofs) or 29 (steep-sloped roofs); ● have at least 50 per cent of its surface covered by vegetation; or ● use a combination of vegetation and high-SRI materials that satisfy a particular

formula (GBC, 2005). We note that the SRI requirements in the current version of LEED (version 2.2) are about those achieved by a low-sloped roof with a solar reflectance of 0.65 and a thermal

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

28

10:19 AM

Page 28

HASHEM AKBARI AND RONNEN LEVINSON

emittance of 0.90, and by a steep-sloped roof with a solar reflectance of 0.28 and a thermal emittance of 0.90 (since the SRI of this cool low-sloped surface is actually 78.9, we recommend that its required SRI be increased to 79 from 78). We welcome both the simplicity of the SRI requirement and the ability to use truly cool materials whose thermal emittances are less than 0.90.

California cool-roof rebate programmes From 2001 to 2005, the state of California and several of its utilities offered rebates of US$0.10/ft2 to US$0.20/ft2 for the installation of cool roofs with initial solar reflectance not less than 0.70 and initial thermal emittance not less than 0.75. Since the current (2005) Title 24 Standards now prescriptively require cool roofs on non-residential buildings with low-sloped roofs, recent rebate programmes have focused on residential buildings. In January 2006, the Sacramento Municipal Utility District (SMUD) began offering a rebate of US$0.20/ft2 for the installation of a residential flat roof with a solar reflectance greater than 0.75 and a thermal emittance greater than 0.75. In May 2007, the SMUD programme was expanded to offer a rebate of US$0.10/ft2 for the installation of a steepsloped residential roof with a solar reflectance greater than 0.40 and a thermal emittance greater than 0.75 (SMUD, 2007). In January 2007, two California utilities – Pacific Gas & Electric (PG&E) and Southern California Edison (SCE) – began a new programme offering rebates of US$0.10/ft2 to US$0.20/ft2 for retrofitting existing homes in certain California climates with cool roofs (PG&E, 2007; SCE, 2007). The solar reflectance and thermal emittance requirements of this programme are detailed in Table 1.11. Note that all qualifying products must have a thermal emittance of at least 0.75. The low-sloped roof requirements of the PG&E/SCE programme are designed to promote the use of white roofs. The two levels of rebates for steep-sloped roofs (Tier 1: US$0.10/ft2 for solar reflectance between 0.25 and 0.39; Tier 2: US$0.20/ft2 for solar reflectance not less than 0.40) are designed to encourage the use of existing coolcoloured products (most of which lie in Tier 1) and to spur the development and sale of improved cool-coloured products (Tier 2).

TABLE 1.11 Solar reflectance and thermal emittance requirements of a 2007 residential cool-roof rebate

programme administered by two California utilities (Pacific Gas & Electric and Southern California Edison) ROOF SLOPE

REBATE TIER

Lowa Steepb Steepb

n/a Tier 1 Tier 2

INITIAL SOLAR REFLECTANCE ≥ 0.70 0.25–0.39 ≥ 0.40

INITIAL THERMAL EMITTANCE ≥ 0.75 ≥ 0.75 ≥ 0.75

Notes: a = ratio of rise to run less than or equal to 2:12 b = ratio of rise to run greater than 2:12 Source: Pacific Gas & Electric (PG&E, 2007); Southern California Edison (SCE, 2007)

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

REBATE (US$/FT2) 0.20 0.10 0.20

ABER_CH_01

7/11/08

10:19 AM

Page 29

Evolution of Cool-Roof Standards in the US

29

CONCLUSIONS Since the late 1990s, the quantification of energy savings offered by the use of cool roofs has led both ASHRAE and the state of California to add cool-roof credits and/or requirements to their energy efficiency standards for both residential and non-residential buildings. Many US states have adopted cool-roof credits from ASHRAE Standard 90.1 (1999 or later), IECC 2003 or IECC 2006. Several US cities and states other than California have developed custom cool-roof provisions to their energy standards. Voluntary energy efficiency programmes, such as the US EPA Energy Star label, the LEED Green Building Rating System of the US Green Building Council and rebate programmes offered by California and its utilities, have established qualifications for cool roofs. While cool-roof requirements have occasionally been too strict – for example, excluding many cool materials by setting a minimum thermal emittance of 0.90, rather than one of 0.75 – they are more often too lax. Particularly problematic are those definitions that (a) allow the use of hot bare-metal products on low-sloped roofs by specifying neither a minimum thermal emittance nor a minimum SRI; and/or (b) allow the use of all but the hottest materials on steep-sloped roofs by qualifying products with an aged solar reflectance as low as 0.15. We have also found ambiguities and outright errors in several cool-roof standards. These issues suggest that more care needs to be taken to ensure that cool-roof standards are both accurate and effective. The standards described in this chapter were developed primarily by workers at several US research laboratories. We expect that cool-roof standards will be further refined to incorporate improvements in building energy analysis and cool-roof technology. However, the methods used to develop the cool-roof provisions in the ASHRAE and California Title 24 standards can be used as models to address cool roofs in building energy efficiency standards worldwide.

AUTHOR CONTACT DETAILS Hashem Akbari: (corresponding author) Heat Island Group, Lawrence Berkeley National Laboratory, Berkeley, CA, 94720 US; [email protected] Ronnen Levinson: Heat Island Group, Lawrence Berkeley National Laboratory, Berkeley, CA, 94720 US

ACKNOWLEDGMENTS This chapter was supported by the California Energy Commission (CEC) through its Public Interest Energy Research Program (PIER) and by the Assistant Secretary for Renewable Energy under Contract No DE-AC02-05CH11231. We wish to thank Javier Ceballos, City of Chicago; Danny Parker, Jeff Sonne and Muthusasmy Swami, Florida Solar Energy Center; Howard Wiig, State of Hawaii; and Charles Eley, Architectural Energy Corporation for clarifying building codes.

NOTE 1 The location of distribution ducts can strongly influence the energy performance of cooling systems. Leaky and/or poorly insulated ducts in attics have exhibited delivery efficacies as low as

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

30

10:19 AM

Page 30

HASHEM AKBARI AND RONNEN LEVINSON

50 per cent (Jump and Modera, 1994). Delivery efficacy falls as the attic temperature rises. Parker et al (1998a) have developed a model to account for the influence of attic temperature upon the delivery efficacy of the distribution system.

REFERENCES Akbari, H. (1998) ‘Cool roofs save energy’, ASHRAE Transactions, vol 104, no 1B, pp783–788 Akbari, H. and A. Desjarlais (2005) ‘Cooling down the house’, Professional Roofing, March, www.professionalroofing.net/article.aspx?A_ID=609 Akbari, H., S. Bretz, H. Taha, D. Kurn and J. Hanford (1997) ‘Peak power and cooling energy savings of high-albedo roofs’, Energy and Buildings, vol 25, no 2, pp117–126 Akbari, H., S. Konopacki, D. Parker, B. Wilcox, C. Eley and M. Van Geem (1998) ‘Calculations in support of SSP90.1 for reflective roofs’, ASHRAE Transactions, vol 104, no 1B, pp984–995 Akbari, H., S. Konopacki and M. Pomerantz (1999) ‘Cooling energy savings potential of reflective roofs for residential and commercial buildings in the United States’, Energy, vol 24, pp391–407 Akbari, H., S. Konopacki and D. Parker (2000) ‘Updates on revision to ASHRAE Standard 90.2: Including roof reflectivity for residential buildings’, in ACEEE 2000 Summer Study on Energy Efficiency in Buildings, vol 1, Pacific Grove, CA, August, American Council for an Energy Efficient Economy, Washington, DC, pp1–11 Akbari, H., M. Pomerantz and H. Taha (2001) ‘Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas’, Solar Energy, vol 70, no 3, pp295–310 Akbari, H., R. Levinson and P. Berdahl (2005a) ‘Review of residential roofing materials, Part I: A review of methods for the manufacture of residential roofing materials’, Western Roofing Insulation and Siding, January/February, pp54–57 Akbari, H., R. Levinson and P. Berdahl (2005b) ‘Review of residential roofing materials, Part II: A review of methods for the manufacture of residential roofing materials’, Western Roofing Insulation and Siding, March/April, pp52–58 Akbari, H., C. Wray, T. T. Xu and R. Levinson (2006) ‘Inclusion of solar reflectance and thermal emittance prescriptive requirements for steep-sloped nonresidential roofs in Title 24’, http://energy.ca.gov/title24/2008standards/prerulemaking documents/2006-05-18_workshop/2006-05-19_NONRESDNTL_STEEP-SLOPED_COOL_ROOFS.PDF ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) (1999) ASHRAE Standard 90.1-1999: Energy Standard for Buildings Except Low-Rise Residential Buildings, SI Edition, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA ASHRAE (2001) ASHRAE Standard 90.1-2001: Energy Standard for Buildings Except Low-Rise Residential Buildings, SI Edition, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA ASHRAE (2004a) ASHRAE Standard 90.1-2004: Energy Standard for Buildings Except Low-Rise Residential Buildings, SI Edition, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA ASHRAE (2004b) ASHRAE Standard 90.2-2004: Energy-Efficient Design of Low-Rise Residential Buildings, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA ASHRAE (2007) ASHRAE Standard 90.2-2007: Energy-Efficient Design of Low-Rise Residential Buildings, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA ASTM (American Society for Testing and Materials) (1998) ‘ASTM E 1980-98: Standard Practice for Calculating Solar Reflectance Index of Horizontal and Low-Sloped Opaque Surfaces’, in Annual Book of ASTM Standards, vol 04.12, American Society for Testing and Materials, Philadelphia, PA BCAP (2007) ‘Status of residential and commercial building state energy codes’, www.bcap-energy.org Berdahl, P., H. Akbari, R. Levinson and W. A. Miller (2008) ‘Weathering of roofing materials – an overview’, Construction and Building Materials, vol 22, no 4, April, pp423–433

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

10:19 AM

Page 31

Evolution of Cool-Roof Standards in the US

31

CEC (California Energy Commission) (2001) 2001 Energy Efficiency Standards for Residential and Nonresidential Buildings, P400-01-024, California Energy Commission, Sacramento, CA CEC (2006) 2005 Building Energy Efficiency Standards for Residential and Nonresidential Buildings, CEC-400-2006-015, California Energy Commission, Sacramento, CA Census (2007) US Census Bureau State and County Quickfacts, http://quickfacts.census.gov Chicago (2001) ‘Amendment of Title 18 of Municipal Code of Chicago Concerning Energy Efficiency Requirements’, Journal of the City Council of Chicago, 6 June, p60939 Chicago (2007) City of Chicago Energy Conservation Code, Index Publishing Corporation, Chicago, IL DOE-2 (2007) Lawrence Berkeley National Lab DOE-2 website, http://simulationresearch.lbl.gov/dirsoft/d2whatis.html Eley (2007) Pers comm with Charles Eley, Architectural Energy Corporation, 17 August Eley Associates (2003a) Assessment of Public Policies Affecting Cool Metal Roofs, Final report prepared for the Cool Metal Roofing Coalition, www.coolmetalroofing.org/elements/uploads/casestudies/TMI_CaseStudy_28.pdf Eley Associates (2003b) Hawaii Commercial Building Guidelines for Energy Efficiency, www.archenergy.com/library/general/hawaiigl Eley Associates (2007) Guam Building Energy Code, http://eley.com/guam Eilert, P. (2000) High Albedo (Cool) Roofs: Codes and Standards Enhancement (CASE) Study, Pacific Gas & Electric report, www.energy.ca.gov/title24/2001standards/associated_documents/2000-11-17_PGE_CASE.PDF EnergyGauge (2007) EnergyGauge USA FlaRes2007 Energy and Economic Analysis Software, www.energygauge.com EPA (US Environmental Protection Agency) (2007) Roof Products Criteria for US EPA Energy Star Program, www.energystar.gov/index.cfm?c=roof_prods.pr_crit_roof_products FBC (Florida Building Commission) (2001) 2001 Florida Building Code, Florida Building Commission, Tallahassee, FL, www.floridabuilding.org FBC (2004) 2004 Florida Building Code, Florida Building Commission, Tallahassee, FL, www.floridabuilding.org/ FBC (2007) Proposed Modification to the Florida Building Code: Chapter 11, Energy Efficiency, www.dca.state.fl.us/FBC/thecode/Res_Chapter_11.rtf GBC (US Green Building Council) (2001) Leadership in Energy and Environmental Design Green Building Rating System for New Construction and Major Renovations (LEED-NC), Version 2.0, US Green Building Council, www.usgbc.org GBC (2002) Leadership in Energy and Environmental Design Green Building Rating System for New Construction and Major Renovations (LEED-NC), Version 2.1, US Green Building Council, www.usgbc.org GBC (2005) Leadership in Energy and Environmental Design Green Building Rating System for New Construction and Major Renovations (LEED-NC), Version 2.2, US Green Building Council, www.usgbc.org Hildebrandt, E., W. Bos and R. Moore (1998) ‘Assessing the impacts of white roofs on building energy loads’, ASHRAE Technical Data Bulletin, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA, vol 14, no 2, pp28–36 ICC (International Code Council) (2003) 2003 International Energy Conservation Code, www.iccsafe.org ICC (2006) 2006 International Energy Conservation Code, www.iccsafe.org Jump, D. and M. Modera (1994) Energy Impacts of Attic Duct Retrofits in Sacramento Houses, LBL-35375, Lawrence Berkeley National Laboratory, Berkeley, CA Konopacki, S. and H. Akbari (1998) Simulated Impact of Roof Surface Solar Absorptance, Attic, and Duct Insulation on Cooling and Heating Energy Use in Single-Family New Residential Buildings, LBNL-41834, Lawrence Berkeley National Laboratory, Berkeley, CA Konopacki, S. and H. Akbari (2001) Measured Energy Savings and Demand Reduction from a Reflective Roof Membrane on a Large Retail Store in Austin, LBNL-47149, Lawrence Berkeley National Laboratory, Berkeley, CA

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_01

7/11/08

32

10:19 AM

Page 32

HASHEM AKBARI AND RONNEN LEVINSON

Konopacki, S., L. Gartland, H. Akbari and L. Rainer (1998) Demonstration of Energy Savings of Cool Roofs, LBNL-40673, Lawrence Berkeley National Laboratory, Berkeley, CA Levinson, R., H. Akbari, S. Konopacki and S. Bretz (2005a) ‘Inclusion of cool roofs in nonresidential Title 24 prescriptive requirements’, Energy Policy, vol 33, no 2, pp151–170 Levinson, R., P. Berdahl and H. Akbari (2005b) ‘Solar spectral optical properties of pigments, part I: Model for deriving scattering and absorption coefficients from transmittance and reflectance measurements’, Solar Energy Materials & Solar Cells, vol 89, pp319–349 Levinson, R., P. Berdahl and H. Akbari (2005c) ‘Solar spectral optical properties of pigments, part II: Survey of common colorants’, Solar Energy Materials & Solar Cells, vol 89, pp351–389 Levinson, R., P. Berdahl, A. A. Berhe and H. Akbari (2005d) ‘Effects of soiling and cleaning on the reflectance and solar heat gain of a light-colored roofing membrane’, Atmospheric Environment, vol 39, pp7807–7824 Levinson, R., P. Berdahl, H. Abkari, W. Miller, I. Joedicke, J. Reilly, Y. Suzuki and M. Vondran (2007) ‘Methods of creating solarreflective nonwhite surfaces and their application to residential roofing materials’, Solar Energy Materials & Solar Cells, vol 91, pp304–314 MCC (Maui County Code) (2004) A Bill for an Ordinance Amending Title 16, Maui County Code, Pertaining to Energy Efficiency Standards for Buildings, www.co.maui.hi.us/files/ordinance/LF-Ord3240_etkoujogl.pdf MICROPAS (2007) MICROPAS product website, http://micropas.com Parker, D. (2005) Technical Support for Development of an Attic Simulation Model for the California Energy Commission, Florida Solar Energy Center report FSEC-CR-1526-05, www.fsec.ucf.edu/en/publications/pdf/FSEC-CR-1526-05.pdf Parker, D. (2007) Pers comm from Danny Parker, Florida Solar Energy Center, 13 August Parker, D., J. Huang, S. Konopacki, L. Gartland, J. Sherwin, and L. Gu (1998a) ‘Measured and simulated performance of reflective roofing systems in residential buildings’, ASHRAE Transactions, vol 104, no 1, Atlanta, GA Parker, D., J. Sherwin and J. Sonne (1998b) ‘Measured performance of a reflective roofing system in a Florida commercial building’, ASHRAE Technical Data Bulletin, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA, vol 14, no 2, pp7–12 PG&E (Pacific Gas & Electric) (2007) Pacific Gas & Electric Cool-Roof Rebate Program, www.pge.com/myhome/saveenergymoney/ rebates/remodeling/coolroof ROH (Revised Ordinances of Honolulu) (2001) Revised Ordinances of Honolulu, City and County of Honolulu ROH (2004) Revised Ordinances of Honolulu, City and County of Honolulu ROH (2007) Revised Ordinances of Honolulu, www.honolulu.gov/refs/roh SCE (Southern California Edison) (2007) Southern California Edison Cool-Roof Rebate Program, www.sce.com/Rebatesand Savings/Residential/_Heating+and+Cooling/CoolRoof SMUD (Sacramento Municipal Utility District) (2007) Sacramento Municipal Utility District Residential Cool-Roof Program, www.smud.org/rebates/cool%20roofs Swami, M. (2007) Pers comm from Muthusamy Swami, Florida Solar Energy Commission, Developer of FLA/COM performance compliance software, 14 June Taha, H. (2001) Potential Impacts of Climate Change on Tropospheric Ozone in California: A Preliminary Episodic Modeling Assessment of the Los Angeles Basin and the Sacramento Valley, LBNL-46695, Lawrence Berkeley National Laboratory, Berkeley, CA Wiig, H. (2007) Pers comm from Howard Wiig, Institutional Energy Analyst, Department of Business, Economic Development and Tourism, Hawaii, 14 June Wray, C., H. Akbari, T. T. Xu and R. Levinson (2006) Inclusion of Solar Reflectance and Thermal Emittance Prescriptive Requirements for Residential Roofs in Title 24, www.energy.ca.gov/title24/2008standards/prerulemaking/ documents/2006-05-18_workshop/2006-05-17_RESIDENTIAL_ROOFS.PDF

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 1–32

ABER_CH_02

6/12/08

3:31 PM

Page 33

publishing for a sustainable future

2

A Review of Innovative Daylighting Systems A. Tsangrassoulis

Abstract One of the challenges of architectural design is to bring daylight into the core of a building in an effort to improve visual environment and, at the same time, to increase energy savings. This can be performed through the use of daylighting systems, the operational principles of which are presented in this chapter. Using sunlight/skylight redirection or exclusion these systems can affect a building’s energy behaviour since their adoption is subject to a compromise between shading needs and daylight provision. The choice of the proper daylighting system should be based on a ‘solid’ rating system, which is quite difficult to develop since a large number of parameters are involved. These include the site, climate, openings, operational and integration constraints, and the objectives that the design team has initially set. Although a large number of daylighting systems are available to the building industry today, their use is limited for reasons such as high initial cost, maintenance and variability in their performance parameters.

■ Keywords – daylighting systems; shading; simulation

INTRODUCTION After the 1970s energy crisis, interest in daylighting – among other solar applications – began to be re-examined in an effort to reduce energy consumption. Today, with the advancing problem of global warming, daylighting is considered a very important strategy to ‘replace’ the electric energy demanded by artificial lighting, reducing not only lighting (and cooling) consumption, but efficiently decreasing peak electrical loads. The literature presents a variety of results in relation to energy savings through daylighting, which are difficult to compare since the data refer to a particular place, building and system. Based on simulation or field measurements results, energy savings are somewhere between 20 and 50 per cent of a building’s electricity consumption (Bodart and De Herde, 2002). It is evident that maximization of daylight use can be increased through the development of innovative systems. Although many systems have been tested during major research programmes, such as IEA TASK 21 (2002), market

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56 doi:10.3763/aber.2008.0202 ■ © 2008 Earthscan ■ ISSN 1751-2549 (Print), 1756-2201 (Online) ■ www.earthscanjournals.com

ABER_CH_02

6/12/08

34

3:31 PM

Page 34

A. TSANGRASSOULIS

penetration is very limited for the following reasons: ● high cost due to the precision required in the manufacturing process; ● difficulty in quantifying energy savings due to the lack of relevant data (such as

bidirectional transmittance distribution function (BDTF) or solar heat gain coefficients); ● risk of glare or over-illuminance due to their control.

Given that these drawbacks are corrected, it is quite possible that the system’s efficiency will increase considerably. Furthermore, during the summer months, reducing solar gains and attaining adequate daylighting levels can be difficult, leading to an optimized design procedure in order to satisfy both needs. Traditionally, this role is left to shading systems (e.g. blinds, fabric rollers, etc.), which minimize daylight supply. Daylighting can contribute to an improvement in lighting quality (e.g. sufficient lighting levels and colour rendition) and to better occupant health and comfort since it is a mood motivator and affects productivity due to reduced stress and fatigue. During the past few decades, new issues in daylighting design have been raised in conjunction with new demands by the building industry. One of the challenges is to bring usable daylight deep into the core of a building or into non-daylit areas, which may result in significant savings in energy consumption while creating an attractive visual environment. Conventional fenestration systems can provide adequate daylighting in areas near building openings (perimeter area). The increase of this area can be realized either by using a design strategy in relation to a different building form (i.e. with an atrium) or a system that offers specialized spectral and/or angular selectivity. The latter can be called a daylighting system in the sense that it can extend performance beyond that of conventional solutions. An overview of the main principles involved, providing sketches and short descriptions of the system’s elements, has been presented in KischkoweitLopin (2002). This chapter, instead, deals with the physical processes that define the operation of these systems.

TAXONOMY Depending upon the position of the daylighting systems within the building envelope, two main categories exist (McCluney, 1998): 1 fenestration systems that are built in between glazing or can be attached to the exterior/interior of the window frame; 2 core systems that are used to introduce daylight from a distance from the building envelope. Another classification can be performed that examines the use of diffuse skylight or direct sunlight. According to Matusiak (2004): ● Daylighting systems with shading are systems that: ● rely primarily on diffuse skylight and reject sunlight; ● generally use direct sunlight, reflecting it onto ceilings or to locations above eye

height.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 35

A Review of Innovative Daylighting Systems

35

● Daylighting systems without shading are: ● diffuse light-guiding systems; ● direct light-guiding systems; ● light-scattering or diffusing systems; ● light transport systems.

Of course, in an effort to achieve seasonal adjustments, daylighting systems can be fixed or moveable (e.g. sun-tracking devices). The introduction of daylighting systems to a building is based on the fulfilment of one or more of the following goals: ● ● ● ●

increase in lighting levels in the areas away from the building perimeter; increase in uniformity; improvement of visual comfort; provision of sufficient shading.

FENESTRATION SYSTEMS Fenestration systems can be separated according to their reflective or refractive operating principles. Many daylighting systems restrict the view outside; thus, they have to be placed on clerestory windows away from the horizontal line of sight. The most common shading systems that can be considered as daylighting systems are Venetian blinds. These can be used either internally or externally, causing redistribution of daylight (sunlight or skylight) to ceilings. The efficiency of Venetian blinds in terms of light redistribution is rather low and depends upon the blind section and its material (and the associated reflectance). Only the optimum orientation of these blinds can allow maximum reflection without preventing outward vision. Visual connection to the exterior environment is vital for human beings, providing access to quality daylight. Reflecting blinds are used as an internal multifunctional system that is able to fulfil all complex demands of fenestration systems. The blinds are equipped with a high specular, high reflective upside surface with reflectance above 90 per cent, and with a grey diffuse downside surface. A further improvement is the perforation of the blinds’ surface, a technique able to provide a visual connection with the exterior environment even when the blind system is closed. The optimum use of a blind system can be performed by automatic control depending upon the external climatic conditions and the internal demands. DiBartolomeo et al (1996) have tested such a system with the following control options: ● direct sun block; ● adjustment of blind angle to maintain the design workplace’s illumination; and ● maintenance of the total illuminance from lights and daylight within the design’s

illuminance range. Ordinary blinds have a curved section in order to achieve mechanical stiffness and adjustment precision.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

6/12/08

36

3:31 PM

Page 36

A. TSANGRASSOULIS

Source: Department of Architecture, University of Thessaly

FIGURE 2.1 Simulation of luminance distribution in a room with no shading system (upper) shaded by fabric roller (middle) and shaded by Venetian blinds (lower)

22°

Solar rays cut-off angle

75°

ABER_CH_02

Acceptance angle for diffuse radiation

Source: original material for this chapter

FIGURE 2.2 Schematic representation of two geometrical characteristics: cut-off angle for solar rays and angle of acceptance for diffuse radiation

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 37

A Review of Innovative Daylighting Systems

37

g Hi h n su RO

T RE

Fla

ts

un

RO

RE

INT

TR

O

Source: Koester LichtPlannung (www.koester-lichtplanung.com/pages_gb/product_01.html)

FIGURE 2.3 Retrolux blind

Since it is an expensive solution to continuously and automatically adjust blind positions, a variety of new designs have been produced. The main idea behind these designs is to maintain the blinds, if possible, in a horizontal position, offering an unobstructed view, and to regulate solar gains during heating/cooling periods. Such a system, developed by Koester (2004), is presented in Figure 2.3. Retrolux is a toothed blind that consists of two parts, with the outer used for the sun’s retro-reflection, while the inner part reflects the sun’s radiation by using a very limited number of reflections. As a result, the reflection of the retro-reflection in the window panes does not generate disturbing glare effects in the interior space. It is evident that in maintaining a constant blind position, performance depends upon the position of the sun. Thus, in order to maintain performance, designers should use quite complex optical systems where direct sunlight can be partly reflected depending upon the incident angle. During the summer, the transmittance is normally low, while for low solar altitudes during winter, the transmittance is high. Due to their accurate shape and highly reflective material, these systems are enclosed between a double-glazing insulating unit. These systems reflect direct sunlight backwards and mainly transmit diffuse radiation. An important advantage is that these systems do not affect the colour of daylight and do not require maintenance. Numerous optical designs with various optical properties exist. A typical example of this system is seen in Figure 2.4. Through special surface treatment of three-dimensionally designed metal or plastic optical structures, compound parabolic concentrators (CPCs) can be constructed for the directional selection of radiation. These structures have a strong directional selectivity in their transmittance/reflectance (Kuckelkorn, 2002). CPCs can be created by using two parabolas, as presented in Figure 2.5; its two-dimensional geometry was proposed by Winston (1974). Any ray entering the CPC’s entry aperture with an angle of incidence smaller than the angle (α) will emerge from the exit aperture. In all other cases, rays are reflected back

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

38

3:31 PM

Page 38

A. TSANGRASSOULIS

Source: OKALUX (www.okalux.com)

FIGURE 2.4 Okasolar-W system

α

Source: original material for this chapter

FIGURE 2.5 Creation and operational principle of a compound parabolic concentrator (CPC) using two parabolas and definition of maximum acceptance angle (α)

(depending, of course, upon the surface roughness and local curvatures). For symmetrical CPCs, the concentration ratio (C) is given by: C = 1/sin(α).

[1]

Increase of C means that α has to be decreased, which, in practice, signifies that the collector accepts light from a very small part of the sky. Taking into account the azimuth

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 39

A Review of Innovative Daylighting Systems

39

Source: Siteco (www.siteco.de/de/produkte/tageslichtsysteme.html)

FIGURE 2.6 Construction scheme of the Microlouver (by Siteco)

and slope of the window where the system is to be installed, asymmetric profiles can be designed in order to accept light from a predefined part of the sky (e.g. excluding sunlight during summer). A similar system used for skylights is the Microlouver by Siteco. This system consists of specially shaped and aluminium vapour-deposited louvres that are designed to retro-reflect radiation coming from the southern hemisphere while transmitting light from the northern hemisphere. The solar heat gain coefficient is 14 per cent, with direct transmittance varying from 5 to 55 per cent, depending upon the sun’s position, and diffuse transmittance of 38 per cent. Refraction is also used as a design principle for daylighting systems. The most representative system in this category is the prismatic glazing, which is manufactured from ultraviolet (UV) light- and weather-resistant polymethyl methacrylate (PMMA). The geometry of the prisms is designed to reflect direct sunlight by means of internal total reflection, and to transmit diffuse light. Due to the narrow angle of this operation, the prismatic glazing has to be tilted according to the height of the sun. Typical prism geometry is presented in Figure 2.7. Following the analysis presented by Lorenz (2001), the two prism angles are critical for performance. Angle θ is determined in such way that the following equation applies: ⎛ 1⎞ sin(nG −θ ) = n* sin(arcsin⎜⎜ ⎟⎟⎟ −θ) ⎜⎝ n ⎠

[2]

where θ is the prism angle, n is the refractive index and nG is the angle that defines if transmission or rejection of solar radiation will occur. If the sun’s elevation (on the crosssectional plane of the prism) is larger that nG, then the sun’s rays are to be rejected; otherwise, they are transmitted as shown in Figure 2.8. Prismatic systems with various angles, such as θ and Ω, have been designed, and in some of them one facet of the prism is coated with a reflective layer. On certain occasions,

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

40

3:31 PM

Page 40

A. TSANGRASSOULIS

Source: original material for this chapter

FIGURE 2.7 Prismatic glazing shading operation

θ

nG

θ

Ω

Source: original material for this chapter

FIGURE 2.8 Geometric configuration of a typical prismatic glazing section

prismatic glazing is enclosed in a double-glazing unit for protection. Prismatic glazing can cause visual distortion; thus, in many cases they are placed in clerestory windows. Laser-cut light-deflecting panels (LCPs) (Edmonds, 1993) consist of PMMA panes in which narrow parallel grooves have been cut out using a laser beam. A typical crosssection is presented in Figure 2.9.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 41

A Review of Innovative Daylighting Systems

41

Source: original material for this chapter

FIGURE 2.9 Laser-cut light-deflecting panel (LCP) operational principle

Depending upon the geometry of LCPs and the angle of incidence, light can be deflected upwards (Edmonds et al, 1995). Thus, in the case of a vertical window equipped with this system, light flux is redirected to the ceiling and improves light levels in deep rooms, while at the same time permitting a clear view to the exterior environment. Holographic optical elements (HOEs) use diffraction effects to redirect sunlight and are manufactured by exposing a photographic film to a coherent and monochromatic laser beam. There are two parameters that affect their operation: 1 the period Λ of the photolithographic structure; and 2 the wavelength of the radiation. If Λλ light redistribution is possible. Thus, their use in laminated glazing can offer either shading or redirection of sunlight to the ceiling in an effort to increase light levels in areas away from windows. The angle of diffraction is dependent upon the wavelength and this can be achieved only for a certain angle of incidence. For all other angles of incidence, diffraction does not occur and the HOE is transparent. Although their efficiency is relatively poor (depending upon the wavelength), the technology is sophisticated (IEA, 2000).

CORE SYSTEMS Core systems introduce daylight into the core of a building away from its perimeter zone. These systems usually consist of three functions: ● collecting; ● transmitting; and ● emitting.

A heliostat is, essentially, a mirror that tracks the sun and redirects sunlight onto a specific receiving area. If this area is, for example, a skylight, then direct sunlight can be guided

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

42

3:31 PM

Page 42

A. TSANGRASSOULIS

through the core of the building. In common configurations there are two mirrors employed: one of them is actually the heliostat, while the second one is static and is used to direct sunlight to a predefined direction. Due to the divergence of the sun’s rays (~0.5°), there is a difference between the light flux emitted from the heliostat and the light flux that impinges upon the receiving area. This difference is due to: ● mirror surface roughness; ● local curvature of the mirror surface; and ● tracking error.

The above parameters define the efficiency of the heliostat in transporting solar light flux. It is evident that if the distance between two plane mirrors increases, then the light transportation efficiency decreases. In an effort to overcome this problem, various shapes of concave heliostats have been introduced that are capable of focusing the light beam on the receiving area. A lens can be used to collimate the beam as well, with a further increase in efficiency. Glass and metal are the most common materials used for constructing the mirror surfaces of heliostats, although the potential of stretched reflective films (which can be metal, polymeric or composite construction) on a frame has been exploited by Murphy

Source: BartenBach LichtLabor (www.bartenbach.com)

FIGURE 2.10 Heliostat designed and installed at BartenBach LichtLabor, Innsbruck, Austria

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 43

A Review of Innovative Daylighting Systems

43

Source: original material for this chapter

FIGURE 2.11 Schematic drawing of the inflatable heliostat prototype

(1983). Until recently, heliostat designs based on this principle were extremely limited. The main problem associated with this type of heliostat is the expected lifetime of the material and the rigidity of the construction. During 2002, Sankrithi presented a prototype inflatable heliostat in which a mirrored film is located between two equally sized inflated transparent domes. By adjusting the pressure in these domes, various focal distances can be achieved. The whole system is moved by using two motors and, due to its minimal weight, there are low torque requirements. A crucial element of heliostat efficiency is the proper tracking mechanism. The required accuracy of this mechanism depends upon the receiver acceptance angle. Since the application of heliostats, tracking has been performed with azimuth-elevation axes using two motors. The motors can be controlled directly through a proper controller and through use of irradiance sensors. These sensors help the motors to align the heliostat in the correct direction and can be placed, for example, on the receiver. When clouds block solar radiation, the tracking mechanism should be capable of following the sun’s path using a time-based algorithm; after the sun’s appearance, the sensors take over control again. Because of the level of accuracy required, rotation speed in both axes is extremely low. Such a control mechanism is presented in Figure 2.12. This mechanism consists of two pairs of photo-sensors separated by small fins. The controller is trying to equalize signals from four sensors, in this way achieving perfect alignment with the sun. Transferring daylight with fibres is quite an old concept; Robieux (undated) was the first to suggest the use of concentrating parabolic mirrors associated with solid light guides. There are different types of fibre optics according to the material that has been used for their construction, as presented in a review by Jacobs (2002). Plastic fibres made

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

44

3:31 PM

Page 44

A. TSANGRASSOULIS

Source: BartenBach LichtLabor, system Solux (2004, available from www.bomin-solar.de)

FIGURE 2.12 Mechanism for alignment of a solar tracker

of PMMA have a clear PVC sheath and losses of around 3 per cent per metre. Strands of fibres (0.25 to 3mm in diameter) can be bundled together for the increase of cross-section. However, the section includes not only the core but the cladding and the coating. Even with the coating removed, the cladding and the space between the fibres leaves an unused part of the bundle section. Plastic fibres have a maximum operational temperature of 80°C. Glass fibres are extruded from borosilicate glass with a cladding of lead silicate having losses of > 6 per cent per metre and are produced in small diameters (~0.05mm). A special type of glass fibre is the silica fibre, which comes with a polymer cladding and typical diameters of 1mm. Packing losses are approximately 20 per cent, while maximum operational temperature can reach 350°C. The last category is the liquid fibre, which normally has a large diameter (>13mm) and consists of a pressurized tube with a Teflon cladding, filled with a liquid and plugged on both ends with PMMA solid taps. The liquid that is especially produced to have low infrared (IR) transmission makes this type of fibre suitable for daylighting applications. A critical parameter for the coupling of a collection system with the fibre is the latter’s acceptance angle. Core material has a refracting index n1, while the cladding has a lower refracting index n2. If certain incidence conditions are fulfilled, a ray undergoes total internal reflection on the surface separating the core and its cladding. These conditions occur when the ray’s incidence angle θ is smaller than θA: θA = sin−1 (n12 − n22 ). There are some systems that have been designed to carry daylight through fibres.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

[3]

ABER_CH_02

6/12/08

3:31 PM

Page 45

A Review of Innovative Daylighting Systems

45

θ

n1

n2

Source: original material for this chapter

FIGURE 2.13 Definition of acceptance angle for fibre optics

During 1979, the Himawari system (Himawari, 2004, available from www.himawarinet.co.jp) was developed. This system is actually a heliostat with small hexagonal Fresnel lenses – the number of lenses being 12, 36 or 90, depending upon the size – capable of concentrating sunlight to glass fibres with a diameter of 1mm. The other end of the fibres forms a luminescent spotlight in the interior of the building. Hybrid solar lighting systems (Muhs, 2000) consist of five major elements: 1 2 3 4 5

light sources (both sunlight and electric lamps); sunlight collection and tracking systems; light distribution systems; hybrid lighting control systems; and hybrid ‘luminaires’.

A system of two parabolic mirrors is used to concentrate solar radiation. The primary reflective dish reflects solar radiation on a secondary mirror located at the focal point. Since the secondary mirror is a cold mirror, only visible radiation is reflected to a bundle of optical fibres placed in the centre of the dish. At the other end of the fibres an acrylic rod is attached, with its top hemisphere sandblasted in order to produce a uniform scattering surface, and both the top hemisphere and concave end mirror coated with aluminium. This system is used as a luminaire and its similarity with a linear fluorescent lamp is profound (Earl and Muhs, 2001). The system’s total efficiency is approximately 50 per cent if sunlight is transferred to one floor.

Source: La Foret Engineering, Himawari (2004, available from www.himawari-net.co.jp/e_page-index01.html)

FIGURE 2.14 Himawari system

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

46

3:31 PM

Page 46

A. TSANGRASSOULIS

Plastic Primary Mirror Secondary Mirror

Receiver Module

Tracking Mechanism

Source: Sunlight Direct (www.sunlight-direct.com)

FIGURE 2.15 HSL 3100 hybrid lighting system

The Solux system was developed by Bomin Solar Research (Solux, 2004, available from www.bomin-solar.de) and is a heliostat system in which sunlight is captured by a Fresnel lens (concentration 10,000) and then transmitted to the building interior through the use of a liquid fibre optic. The Fresnel lens has a diameter of 1m with a focal length of 80cm and visible transmittance of 94 per cent. To optimize the coupling efficiency into the liquid fibre, this cone angle (which is defined by lens diameter and focal point) must be equal to or smaller than the acceptance angle of the light fibre. To avoid thermal damage of the PMMA end of the liquid fibre, a closed loop water system with a heat exchanger is used to remove infrared radiation from the concentrating sunlight. The system was tested during the 2002 Universal Fibre Optics Project. The end of the fibre in the building interior was coupled with a flat emitter. The emitter is a sheet of Perspex material with a white dot pattern screen printed on it (Prismex). These dots allow the light that is trapped inside the panel through total internal reflection to be released. By varying the size and distance of the dots, any arbitrary distribution of luminance across the surface can be created. The design of the luminaire is mostly

Source: Universal Fibre Optics Project (2002, www.learn.londonmet.ac.uk/portfolio/1999-2001/ufo.shtml)

FIGURE 2.16 Solux daylighting system installed in the University of Athens during the Universal Fibre Optics research programme

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 47

A Review of Innovative Daylighting Systems

47

influenced by the overall geometry of the Prismex with its printed area, which acts like a large, even emitter. The top emitting surface can be covered with a high specular sheet (ρ = 93 per cent) in an effort to increase luminous intensity. In an attempt to use larger cross-sections and to increase the ability to transfer light flux, Callow and Shao (2003) used a set of ‘light rods’ that are sufficiently small to be fitted into most existing buildings without structural modification. During testing, rods were found to have a transmittance equal to that of a light pipe which had an aspect ratio six times smaller. Rods would collect both direct and diffuse light throughout the year, with high collection efficiency. Light pipes are quite simple daylight devices, which through the use of a highly specular (aluminium-anodized plastic films or polymeric multilayer optical stacks) or refractive (Optical Lighting Film) material can guide sunlight into the core of a building. Light pipes transmit less solar radiation into the building interior, and thus have an advantage over ordinary skylights. A simple formula for the transmittance of a reflective light pipe is: T =R

41tan θ πd

[4]

where R, θ and d are the reflectance, the sun’s elevation and the diameter of the light pipe, respectively. It is evident that the light pipe’s performance depends upon the sun’s elevation; thus, during the winter months, transmittance is extremely low. Some new designs incorporate light redirection systems (such as laser-cut panels) on top in an effort to modify the angle of incidence of light rays. Increase of light pipe performance can be achieved by using sun-tracking devices, which redirect sunlight at a proper angle through the pipe; but their complexity and intense maintenance requirements are really a disadvantage. Following other applications in the field of daylighting and solar concentration (Welford and Winston, 1978), Molteni et al (2001) used the concept of anidolic optics to design an external collector that is characterized by an 80º input acceptance angle. Oakley et al (2000) presented light-level measurements from four buildings in the UK, concluding that light pipes can be effective devices for introducing daylight into buildings provided that they are designed in such a way as to avoid large aspect ratios and bends. A light pipe variant is the prism light guide that is actually a hollow pipe made of transparent acrylic material in which one of its surfaces has longitudinal prisms. Normally, the smooth surface of the transparent material is in the interior of the pipe. This configuration causes the light that enters the pipe’s inner space to be reflected by total reflection at the outside surface, which has a 90º prismatic ridge. Of course, total reflection occurs providing that it is less than 27.5º. During the European research project Arthelio (Kasse et al, 1998–2000), a similar system was examined with the aim of providing daylighting in deeper building interiors. Two prototypes have been installed at the 3M distribution centre in Carpiano, Italy, and at the Semperlux Building in Berlin, Germany. Both prototypes use the hollow light prism guide concept with heliostats as a light collection system on the roof. Rosemann and Kaase (2005) report a high user acceptance of both systems, while the sunshine probability affects the energy balance of the system. Hollow horizontal light

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

48

3:31 PM

Page 48

A. TSANGRASSOULIS

Source: Heliobus AG (www.heliobus.com)

FIGURE 2.17 Heliobus system installed at Graubunden, Switzerland

pipes with polygonal sections can be used to transport daylight from the façade to the building core. Such a system has been examined by Beltrán et al (1997). Their paper mentions some additional parameters that have to be considered: ● The light pipe size should be quite small in order to fit with other building

subsystems within the ceiling plenum. ● The cross-section of the light pipe must be varied in order to study the changes

in illuminance efficiency and distribution. ● The reflector system needs to partially collimate incoming sunlight in order to

minimize inter-reflections within the transport section of the light pipe, and to maximize the efficiency of the system. ● The shape of the light pipe transport cross-section may have to be altered and various reflector options may have to be investigated in order to redirect daylight to the work plane. Such a configuration is depicted in Figure 2.18. Beltrán et al’s (1997) conclusions were that the light pipe can achieve work-plane illuminance levels consistently above 200lx for about seven hours per day throughout the year. Lower, but still useful, levels of daylight (>100lx) are provided for a greater range of sun angles. This is one of the most critical parameters that directly affects light pipe efficiency. Normally, efficiency is quite high as long as the sun’s rays impinge upon the system with a small angle of incidence. Recently, Canziani et al (2004) presented a new type of horizontal light pipe that includes a planar closing element and a sunlight collecting and deflecting device that optimizes the direction of the incoming solar rays as the solar position varies.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 49

A Review of Innovative Daylighting Systems

49

Source: original material for this chapter

FIGURE 2.18 Horizontal light pipe

Source: original material for this chapter

FIGURE 2.19 Horizontal light pipe proposed by Canziani et al (2004)

In a recent research programme initiated and led by Lorne Whitehead at the University of British Columbia, Canada (Whitehead, 2006), a daylighting system was developed that demonstrates both the feasibility and cost-effectiveness of a solar canopy system that guides direct sunlight deep into the interior of a building. The design of the canopy system is compatible with standard building construction techniques, which is critically important for widespread adoption of the technology. This system consists of a number of small moveable mirrors (mirror array) that redirect sunlight into a light pipe.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

50

3:31 PM

Page 50

A. TSANGRASSOULIS

Source: Structured Surface Physics Laboratory, University of British Columbia (www.physics.ubc.ca/ssp/solar_canopy.htm); Whitehead (2006)

FIGURE 2.20 Prototype daylighting system developed by Whitehead (2006)

A similar system is the Aszen (Kinney and Burnett, 2002) collection system, which consists of two systems of highly reflective blinds placed horizontally and vertically and controlled by two motors. During 2002, Scartezzini and Courret presented a horizontal daylight transportation system where anidolic elements had been used in both ends.

Source: original material for this chapter

FIGURE 2.21 Anidolic ceiling

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 51

A Review of Innovative Daylighting Systems

51

Anidolic systems attempt to solve two primary radiation transfer design problems: 1 maximization efficiency; 2 creation of controlled illuminance distribution. A system that is commonly applied to solar collection is the compound parabolic concentrator (CPC), although some designs do not use parabolic profiles and are referred to as CPC type. Anidolic systems have, in principle, the following properties: ● The number of reflections can be minimized. ● Angular selectivity can be achieved though appropriate design.

Consequently, such a system can achieve high efficiency in terms of transmitting light flux. The main advantage of the system is the efficient collection and redistribution of diffuse skylight, achieving an increase in daylight levels in the deeper part of the room equal to 32 per cent, with no external obstructions, while the percentage increases to 45 per cent when the system is installed in an urban environment. Depending upon the climate, some protection against solar gains is needed during the summer months.

PERFORMANCE METRICS Innovative daylighting systems are complex optical devices, and it is important to know their performance parameters through simulation or measurements. Performance can be estimated either for the system itself by measuring/simulating its luminous and radiant properties or for the system combined with a test cell. The latter has the advantage that systems with arbitrary complexity can be measured, while the main disadvantage of this method is the time needed to investigate all possible combinations of solar position and sky luminance distribution. Aizlewood (1993) has used criteria to define the performance of innovative daylighting systems, percentage increase in illuminance and the glare index when the system is compared with clear glass. Moeck’s (1998) work was a continuation of Aizlewood’s paper, proposing not clear glass as a base case, but an ordinary Venetian blind system, which is more likely to be present in an office environment. According to his paper, quantitative and qualitative criteria must be used. Among a number of criteria mentioned the following have been used during his experiments: ● ● ● ● ● ● ● ●

luminous intensity distribution of the system; illuminance on the working plane; uniformity; total light flux entering the space; light flux on the upper back part of the space; luminance of the window; total area of the space that is over-illuminated; spherical illuminance values.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

52

3:31 PM

Page 52

A. TSANGRASSOULIS

Although the above list sufficiently describes the impact of the system to the visual environment, additional criteria should be added, such as: ● ● ● ●

estimation of daylight savings; estimation of glare indices; visual contact with the exterior environment; estimation of radiant properties (such as g-value).

By using a facility that reproduces sky luminance distributions over the hemisphere, a daylighting system in an artificial sky can be examined. Errors associated with sky coverage (due to the limited number of lamps used to cover the entire area of the sky) and with parallax have to be taken into account, while the presence of a solar simulator is essential. A method that uses artificial sky in order to estimate the performance of daylighting systems is the integration of directional coefficients (IDC) method, which has been presented by Papamichael and Beltrán (1993). The method uses a scale model of the system and a space in an artificial sky. Using a collimated beam of light, the directional illuminance coefficients are estimated. These coefficients are defined as follows: C(azimuth, elevation) = Ei(azimuth, elevation)/Ee(azimuth, elevation)

[5]

where Ei(azimuth, elevation) is the interior illuminance due to the collimated beam in the direction specified by the pair of (azimuth, elevation), and Ee(azimuth, elevation) is the exterior illuminance due to, and normal to, the collimated beam. A similar procedure is applied for the calculation of skylighting as well. It is evident that if the above coefficients have been calculated at reference points of interest, then they can be used in an algorithm to determine daylight factors or actual daylight illuminance values under any sun, sky and ground conditions. This method is similar to that proposed by Tregenza and Waters (1983). Daylight coefficients for sunlight have to be calculated separately using a collimated beam of light for various positions. Depending upon the system examined, the number of solar positions should be increased or decreased accordingly. Direct measurement of the performance of a daylighting system is not an easy task since the majority of today’s facilities have a limitation on the sample size. In general, characteristics with hemispherical or conical incidence or collection of light can be measured either using gonio-photometers (measurement of intensity distributions or bidirectional transmittance distribution function: BTDF) or integrating sphere photometers. With the gonio-photometric measurements, the radiation transmitted by the sample into different directions is measured in these directions, while in sphere photometers light flux transmitted by the sample in all directions is collected and spatially integrated by the sphere (Aydinli and Kaase, 2002). As mentioned above, sample size is of critical importance so that the gonio-photometer can measure a system and not a material sample. The bidirectional gonio-photometer of Ecole Polytechnique Fédérale de Lausane (EPFL) (Scartezzini et al, 1997) allows the use of samples with a maximum size of 40 x 40cm, while the same size is used in a similar facility at the Fraunhofer Institute for Solar Energy Systems.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 53

A Review of Innovative Daylighting Systems

53

A measurement device that can calibrate the segment luminous intensity distribution of hollow prism light pipes has been developed and built at the Technical University of Berlin (Kloss et al, 2000; Kloss, 2001). It essentially consists of an arc that can move around a segment of the hollow light guide and a rail system that can linearly drive the hollow light guide through the arc. As a result, all segments of the hollow light guide system can be measured. The maximum length of light pipes that can be measured is 30m. Simulation of BTDF for complicated daylighting systems can be performed using specialized ray-tracing programmes such as ASAP (see www.breault.com) and TracePro (see www.lambdares.com); but the estimated data should be treated with caution since construction inaccuracies can affect the system’s performance considerably. Another issue is that these programmes are not linked – at least directly – to a lighting simulation software in order for luminance distribution along interior surfaces to be estimated on an hourly basis. To examine the influence of a daylighting system in a building, thermal simulation models should be used. In these models, the energy transport through the individual building components is calculated on the basis of hourly values for the external climatic conditions and for the assumed behaviour of the occupants of the building. The problem with these models – in terms of radiation – is the difficulty of estimating the dynamic performance of complex shading/daylighting systems. For today’s models, evaluating the performance of such a system can be performed by using empirical or flux-transfer methods, radiosity-based methods or ray-tracing techniques. Energy+ (available from www.energyplus.gov) supports light pipes and shelves, while complex fenestration systems can be used as long as precalculated BTDFs are given. A different approach is used by Esp-r (available from www.esru.strath.ac.uk/ Programs/ESP-r.htm). In this case the thermal algorithm is linked with Radiance (available from www.radiance-online.org) for daylight calculation. If daylight coefficients are known, they can be used as well. DAYSIM (daylighting analysis software that predicts the annual daylight availability, as well as electric lighting use in a building; see http://irc.nrc-cnrc.gc.ca/ie/light/daysim.html) can be used in conjunction with Radiance to calculate daylight coefficients and then estimate illuminance levels in the interior according to a predefined time step. Although it is possible to simulate daylighting systems, only a limited number of systems (with simple geometric properties) can be simulated with Radiance. According to Jan de Boer (2006) a database with BTDFs of daylighting systems should be created using gonio-photometric measurements and a calculation method that superimposes the external luminance distribution and the BTDFs of the respective system (TALISYS is available from www.talisys.de). This method allows one to determine and evaluate the effects of different daylighting systems on the illuminance distribution.

CONCLUSIONS Innovative daylighting systems – on top of their shading properties – provide higherquality work environments that are more comfortable for occupants and also provide

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

54

3:31 PM

Page 54

A. TSANGRASSOULIS

owners with greater space value. Their main functions are to bring daylight into the building core, to increase uniformity and to reduce glare. Nevertheless, these characteristics are often associated by high illuminance and possible obstruction of views. It is true that adoption of daylighting systems is rather limited because of equipment and installation costs, high maintenance of control systems, reliance on weather and degradation of optical properties due to air pollution. Another issue is that, in many cases, the choice of a daylighting system is tailor made and requires detailed design. An increase in the use of innovative daylighting products can be promoted if there is an associated technology ‘push’ to justify their benefits. These systems are normally used in high-profile projects where cost is irrelevant and they are seen as more or less a ‘marketing’ tool. The outcomes of various European research programmes have identified the barriers that prevent the widespread adoption of these systems and have highlighted the necessity of improving the following issues: ● ● ● ●

data concerning the characterization of daylighting systems; design tools, especially software tools; demonstration and advertisement of existing case studies; and cost reduction through mass production (if possible).

AUTHOR CONTACT DETAILS A. Tsangrassoulis: Department of Architecture, University of Thessaly, Pedion Areos, 383 34 Volos, Greece; [email protected]

REFERENCES Aizlewood, M. E. (1993) ‘Innovative daylighting systems: An experimental evaluation’, Lighting Research Technology, vol 25, no 4, pp141–152 Aydinli, S. and H. Kaase (2002) Measurement of Luminous Characteristics of Daylighting Materials, Subgroup A4 Report IEA Task 21, ECBCS Annex 29, www.iea-shc.org/task21 Beltrán L., E. Lee and S. Selkowitz (1997) ‘Advanced optical daylighting systems: Light shelves and light pipes’, Journal of the Illuminating Engineering Society, Berkley, CA, pp91–106 Bodart, M. and A. De Herde (2002) ‘Global energy savings in offices buildings by the use of daylighting’, Energy and Buildings, vol 34, pp421–429 Callow, J. and L. Shao (2003) ‘Air-clad optical rod daylighting system’, Lighting Research and Technology, vol 35, no 1, pp31–38 Canziani, R., F. Peron and G. Rossi (2004) ‘Daylight and energy performances of a new type of light pipe’, Energy and Buildings, vol 36, pp1163–1176 de Boer, J. (2006) ‘Modelling indoor illuminance by complex fenestration systems based on bidirectional photometric data’, Energy and Buildings, vol 38, pp849–868 DiBartolomeo, D., E. Lee and F. Rubinstein (1996) Developing a Dynamic Envelope/Lighting Control System with Field Measurements, IESNA 4-7/8/1996, Cleveland, US Earl, D. and J. Muhs (2001) ‘Preliminary results on luminaire designs for hybrid solar lighting systems’, in Proceedings of Forum 2001: Solar Energy – The Power to Choose, 21–25 April 2001, Washington, DC

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_02

6/12/08

3:31 PM

Page 55

A Review of Innovative Daylighting Systems

55

Edmonds, I. (1993) ‘Performance of laser cut panels in daylighting applications’, Solar Energy Materials and Solar Cells, vol 29, no 1, pp1–26 Edmonds, I., G. Moore, G. Smith and P. Swift (1995) ‘Daylighting enhancement with light pipes coupled to laser-cut lightdeflecting panels’, Lighting Research and Technology, vol 27, no 1, pp28–35 IEA (2000) IEA Report Task 21: A Source Book on Daylighting Systems and Components, Report LBNL-47493, July IEA (2002) Daylight in Buildings: A Source Book on Daylighting Systems and Components, IEA SHC Task 21 Report, IEA SHC Task 21/ECBCS Annex 29 Jacobs, A. (2002) State of the Art Review of Fibre-Optic Cable Technology, Final report of Universal Fibre Optics project, ERK6-CT1999-00011, www.learn.londonmet.ac.uk/packages/helioptics/results.html Kasse, H., T. Muller, A. Mingozzi, S. Bottiglioni, J. Ejhed, G. Rehm and O. Pelin (1998–2000) Final Report: Project Arthelio – Intelligent and Energy Optimized Lighting Systems Based on the Combination of Daylight and the Artificial Light of Sulfur Lamps, C.n. JOR3-CT97-0177 Kinney, L. and T. Burnett (2002) ‘Innovations in direct beam daylighting systems’, Paper presented at 5th International Conference on Energy-Efficient Lighting, 29–31 May 2002, Nice, France Kischkoweit-Lopin, M. (2002) ‘An overview of daylighting systems’, Solar Energy, vol 73, no 2, pp77–82 Kloss S.-H. (2001) Ein Goniophotometer zur Messung des Lichtstromes und der Lichtstarkeverteilung von hohlen Lichtleitern, PhD thesis, TU Berlin, Germany Kloss, S.-H., T. Muller, A. Rosemann and H. Kaase (2000) Goniophotometrie an Hohllichtleitern: Licht 2000, Tagungsband der 14, Gemeinschaftstagung der Lichttechnischen Gesellschaften in Goslar, Germany Koester, H. (2004) Dynamic Daylighting Architecture: Basics, Systems, Projects, Birkhauser Publishers Kuckelkorn, T. (2002) Nichtabbildende Konzentratoren für den Sonnenschutz in Gebäuden, PhD thesis, Technische Universitat Munchen, Germany Lorenz, W. (2001) ‘A glazing unit for solar control, daylighting and energy conservation’, Solar Energy, vol 70, no 2, pp109–130 Matusiak, B. (2004) Lighting Systems in Smart Energy Efficient Buildings: A State of the Art, Report STF22A04504, Norwegian Research Council, Norway McCluney, R. (1998) ‘Advanced fenestration and daylighting systems’, Paper presented to Daylighting 1998 International Conference on Daylighting Technologies for Energy Efficiency in Buildings, 10–13 May 1998, Ottawa, Canada Moeck, M. (1998) ‘On daylight quantity and quality and its application to advanced daylight systems’, Journal IESNA, vol 27, no 1, pp3–19 Molteni, S. C., G. Courret, B. Paule, L. Michel and J. L. Scartezzini (2001) ‘Design of anidolic zenithal: Lightguides for daylighting of underground spaces’, Solar Energy, vol 69, pp117–129 Muhs, J. (2000) Design and Analysis of Hybrid Solar Lighting and Full-Spectrum Solar Energy Systems, SOLAR2000 Conference, Madison, Wisconsin Murphy L. M. (1983) Technical and Cost Benefits of Lightweight, Starched Membrane Heliostats, SERI/TR-253-1818, Solar Energy Research Institute, Golden Colorado, May Oakley, G., S. B. Riffat and L. Shao (2000) ‘Daylight performance of lightpipes’, Solar Energy, vol 69, no 2, pp89–98 Papamichael, K. and Beltrán, L. (1993) ‘Simulating the daylight performance of fenestration systems and spaces of arbitrary complexity: The IDC method’, Paper presented to 3rd International Conference of the International Building Performance Simulation Association, BS’93, 16–18 August, Adelaide, Australia Robieux, J. (undated) Patent No 75-14582 (INPI) Intern Class, F03G7/02; H02KX/G02B5/10 Rosemann, A. and H. Kaase (2005) Lightpipe applications for daylighting systems’, Solar Energy, vol 78, pp772–780 Sankrithi, M. (2002) Low-Cost Inflatable-Support Heliostats to Enable Cost-Effective Large-Scale Solar Thermal Power, Final technical report, US Department of Energy Inventions and Innovation Grant No DE-FG36-01GO11023, A000, 24 November

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 33–56

ABER_CH_02

6/12/08

56

3:31 PM

Page 56

A. TSANGRASSOULIS

Scartezzini, J.-L., R. Compagnon, C. Roecker, and L. Michel (1997) ‘Bidirectional photogoniometer for advanced glazing material based on digital imaging techniques’, Lighting Research and Technology, vol 29, no 4, pp197–205 Scartezzini, J. L. and G. Courret (2002) ‘Anidolic daylighting systems’, Solar Energy, vol 73, no 2, pp123–135 Tregenza, P. and I. M. Waters (1983) ‘Daylight coefficients’, Lighting, Research and Technology, vol 15, no 2, pp65–71. Universal Fibre Optics Project (2002) Final Report, ERK6-CT-1999-00011, www.learn.londonmet.ac.uk/packages/ helioptics/results.html Welford W. and R. Winston (1978) The Optics of Nonimaging Concentrators: Light and Solar Energy, Academic Press Publications, San Francisco, CA Whitehead, L. (2006) Solar Canopy Illuminance System, www.physics.ubc.ca/ssp/solar_canopy.htm Winston, R. (1974) ‘Principles of solar concentrators of a novel design’, Solar Energy, vol 16, pp89–95

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 8 ■ PAGES 33–56

ABER_CH_03

6/12/08

3:31 PM

Page 57

publishing for a sustainable future

3

Physically Based Modelling of the Material and Gaseous Contaminant Interactions in Buildings: Models, Experimental Data and Future Developments P. Blondeau, A. L. Tiffonnet, F. Allard and F. Haghighat

Abstract Although potentially having a significant influence on indoor air quality (IAQ), interactions between building materials and gaseous contaminants have often been neglected or crudely modelled in IAQ simulation tools. During the past 20 years, empirical source and sink models have progressively given way to physically based models; but confusion still remains on their applicability, as well as on the adequate experimental data to input for the model parameters. Thus, demonstration is first made that models relating macroscopically the room air phase and material concentrations through adsorption and desorption constants are not scientifically sound. Instead, elemental models combining diffusion equations and local sorption equilibria should be used. The compilation of sorption and diffusion data presented in the second part of this chapter underlines the fact that such data cannot be considered independently from the mass transport equations used to fit the measurements. As a result, a thorough analysis of diffusion processes in polymers and porous media is presented in order to define and relate the diffusion coefficients. Finally, the last part of the chapter discusses the way in which existing models could be extended to account for the contributions of temperature, multi-component mixtures, humidity and chemical transformations within materials. Still based on fundamental considerations, the proposed methodology consists of implementing new functionalities to describe the temperature dependence of the model parameters, elemental models representing the interactions between gaseous contaminants and water, as well as kinetic models coming from the fields of atmospheric and surface sciences.

■ Keywords – Indoor air quality; modelling; adsorption; diffusion; surface reactions; building materials; mass transfer

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93 doi:10.3763/aber.2008.0203 ■ © 2008 Earthscan ■ ISSN 1751-2549 (Print), 1756-2201 (Online) ■ www.earthscanjournals.com

ABER_CH_03

6/12/08

58

3:31 PM

Page 58

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

INTRODUCTION Interactions between gaseous contaminants and materials in buildings involve a variety of phenomena, including volatile organic compound (VOC) emissions from new or newly applied covering materials, deposition of reactive species (ozone and sulphur dioxide, as well as some nitrogen oxides) from the room air to the material surfaces, and what is commonly called the reversible sink effect of the walls and furnishings (or adsorption/desorption processes; but this term proves to be confusing and somewhat inappropriate, as will be shown later on). These phenomena have been extensively studied since the mid 1980s. First, the source/sink behaviour of building materials has been characterized from environmental chamber studies. Most of them lead to the proposal of empirical or semi-empirical models obtained by fitting an appropriate mathematical expression to a set of experimental data, usually concentration at the air exhaust of the chamber as a function of time (Dunn and Tichenor, 1988; Chang and Guo, 1992; Colombo and De Bortoli, 1992; Haghighat and Zhang, 1999). Although such experiments are still relevant today, especially for standardization on material emissions, researches have now also clearly been directed toward the development of physically based models to be implemented in indoor air quality simulation tools. This chapter reviews these models and provides a compilation of adequate correlations and experimental data for the model parameters. Indeed, as noticed by Meininghaus and Uhde (2002), there are still remarkably few sets of relevant diffusion, sorption and reactivity data to be found in the literature compared with the vast amount of existing emission and sink data. Therefore, synthesizing the existing information and searching for correlations to generate more data is another key point to strive for, for simulations of realistic configurations of building operation. Finally, the chapter gives some tracks for future developments by presenting the way in which the influence of temperature, humidity or, more generally, mixtures of contaminants could be implemented in the equations describing the contaminant transports within building materials. Since chemical transformations and, especially, secondary emissions from building materials have recently become of great concern in healthrelated indoor air quality studies (Wilkins et al, 2006), the way in which homogeneous and heterogeneous reaction models could be implemented in these equations is also discussed. On no account is the purpose of the models’ description or compilation of data to be exhaustive. Instead, the aim is to show the way in which existing models can be classified based on different approaches, and subsequently to distinguish between what is scientifically sound and what is not. The emphasis is also put on demonstrating that a deeper look at the fundamentals of diffusion and sorption may be helpful in understanding what has been observed experimentally, extending the model abilities, but also simplifying the problems.

DESCRIPTION OF EXISTING MODELS Physically based models aimed at representing the interactions between gaseous contaminants and materials can first be classified as macroscopic and elemental models. Then, in the case of elemental models, distinction is to be made between what will be

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 59

Modelling of Contaminant Interactions in Buildings

59

called one-phase and two-phase models. The physical principles and related equations defining these classes are presented hereafter.

MACROSCOPIC MODELS The most widely used model for predicting either the VOC emission from materials or the adsorption and desorption cycles of airborne contaminants is the one-sink model initially proposed by Tichenor and co-workers (Tichenor et al, 1991, 1993) and subsequently developed by many authors (Jørgensen et al, 1999; Won et al, 2000, 2001; Elkilani et al, 2003). It considers that at any time, the mass flux at the material surface is the difference between the adsorption and the desorption rates of the material. Analogous to Langmuir’s theory, the former is assumed to be proportional to the contaminant concentration in the bulk air of the room, while the latter is assumed to depend upon the amount of contaminant adsorbed at the material surface. Therefore, the governing equations for indoor and surface concentrations of a non-reactive compound, C∞ (mg/m3) and Cs (mg/m2), respectively, are: V

dC∞ = Q (Cinl − C∞ ) − Ak a C∞ + Ak d Csn , dt dCs = k aC∞ − k d Csn , dt

[1]

[2]

where Cinl (mg/m3) is the pollutant concentration in the inlet air, V (m3) the room volume, Q (m3.s–1) the airflow rate through the room, ka (m.s–1) the adsorption rate coefficient, kd (s–1) the desorption rate constant, and n a constant that may account for potential non-linearities in the desorption process but is typically assumed to equal unity. A (m2) is the surface area of the material that is exposed to the room air, which means that increased surface areas due to surface roughness and internal pores, as well as the effects of bulk mass transport and diffusive transport processes, are implicitly ‘lumped’ into the adsorption and desorption rate coefficients.

ELEMENTAL MODELS Unlike the one-sink model described above, and similar models that relate macroscopically the room concentration to a bulk concentration within the material or at the material surface, elemental models can account for the concentration gradient – that is, non-uniformity of the concentrations – within the material. They usually describe the contaminant transport as a serial assembly of two equations, one representing the diffusion transport in the material and the other the contaminant transport at the material surface. Depending upon whether the boundary layer resistance is considered or not, the latter is the boundary layer equation, a simple partitioning equation of the contaminant between the room air and the material, or a combination of the two. Although showing similarities, elemental models can be classified into two broad groups – namely, onephase and two-phase models, depending upon whether the material is seen as a homogeneous solid or porous medium.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

60

3:31 PM

Page 60

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

One-phase models One phase models consider that materials are bulk homogeneous solids where Fick’s law of diffusion applies. As a result, the equation describing the transient diffusion of gases through the material is (Little et al, 1994; Zhang and Xu, 2003): ∂Cmat = Dmat ∇2 Cmat ∂t

[3]

where Cmat (mg.m–3) is the material-phase concentration, t (s) is time and Dmat (m2.s–1) the diffusion coefficient of the species in the material. When coupling Equation 3 in series with the boundary layer equation: ∂C = h∇C ∂t

[4]

where C (mg.m–3) is the air phase concentration and h (m.s–1) the convective mass transfer coefficient, continuity is obtained by assuming that the solid phase and near-surface air phase concentrations constantly remain in equilibrium at the material surface. This equilibrium relation is most often referred to as the linear adsorption isotherm of the pollutant/material system, and is defined as: Cmat = KC at x = 0

[5]

with x (m) being the spatial dimension perpendicular to the surface and K being the partition coefficient of the compound (note that K is used here as a generic term characterizing the contaminant partitioning between the air phase and the solid phase; distinction between what should be KH or Kp will be made later on). It is of interest to point out here that the combination of Equations 3 and 4 is the more general form describing the pollutant transport between the bulk air phase of the room and the material surface. In many situations (Dunn, 1987; Little et al, 1994), one-phase models have been developed and applied based on the assumption that Cmat = KC∞ at abscissa x = 0, which implicitly means that h is infinite and amounts to considering that the boundary layer resistance to mass transfer is negligible.

Two-phase models Two-phase models apply for porous media and thus consider that the diffusing material is made up of a solid phase and an air phase. In most cases diffusion through solid grains can be considered to be negligible compared to porous diffusion. Moreover, unlike simple diffusion, diffusion in porous materials is constrained by adsorption/desorption phenomena all along the diffusion path. Under these conditions, the contaminant transport within the material can be represented by the equation (Axley, 1991; Lee et al, 2005; Li and Niu, 2007): D p ∇2 C = ε

ρ ∂Cs ∂C + mat ∂t ρair ∂t

[6]

where C (kg/kgair) and Cs (kg/kgmat) are the local air phase and sorbed phase concentrations in the pores, respectively. The first term of Equation 6 accounts for the gas diffusion along

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 61

Modelling of Contaminant Interactions in Buildings

61

the spatial dimension x (m), Dp (m2.s–1) representing the diffusion coefficient of the contaminant; the latter has different expressions depending upon the assumptions made on the diffusion process, as will be discussed further on. The second term of Equation 6 accounts for the accumulation of the diffusing species in the air phase of the pores, with ε (m3air/m3mat) being the material porosity. Finally, the last term accounts for the accumulation of the species at the pore surfaces, with ρmat (kg.m–3) and ρair (kg.m–3) being the material and air densities, respectively. Since Cs is related to C through the adsorption isotherm of the gas and material system: Cs = f (C ) or

∂Cs ∂f ∂C = , ∂t ∂C ∂t

[7]

substituting Equation 7 into Equation 6 yields: ⎛ ρ ∂f ⎞⎟ ∂C ⎟⎟ D p ∇2 C = ⎜⎜⎜ε + mat ⎜⎝ ρair ∂C ⎟⎠ ∂t

[8]

In the end, Equation 8 has the form of a Fickian diffusion equation expressed here in terms of the air phase concentration in the pores. The latter can directly be coupled in series with the boundary layer equation, Equation 4, to describe the contaminant transports from the bulk air phase of the room to the depths of materials, and reversibly. It is important to note that the porosity ε sometimes appears in the left-hand side of Equations 6 and 8 to account for the fact that the actual cross-sectional area offered to diffusion is not the total surface area exposed to the room air, but ε times this surface area. Here, ε is considered to be included in the definition of Dp, as will be discussed later on. From a more general point of view, it is also interesting to note that when dedicated empirical models have been developed to represent the sink effect and the VOC emission, physically based models such as the one-phase and two-phase models described above can either serve as sink or source models. The latter will actually only differ from the initial conditions considered. Taking the VOC content of the material as the initial condition of the model will result in a source model; but this model will automatically turn to a sink model as soon as the concentration gradient between the room air and the material is reversed. Details on the way to determine the initial VOC concentrations in building materials are provided further on. Finally, it is important to point out that one-phase and two-phase models were described considering a single slab of material; but they can easily be extended to multilayer walls, as shown by Kumar and Little (2003), Zhang and Niu (2004), Li and Niu, (2007) and others.

SYNTHESIS Macroscopic or elemental models? As previously described, macroscopic models aim to characterize the dynamics of gas and material interactions through the sorption and desorption rate constants ka and kd. However, still following the principles of Langmuir’s theory of adsorption, some authors

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

62

3:31 PM

Page 62

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

have also proposed to define material and air equilibrium partition coefficients as the ratio of the two rate constants: K =

ka kd

[9]

To some extent, this illustrates the ambiguity that arose out of the definition of sorption processes. Obviously, there is great confusion between what is the sorption dynamics (i.e. the sink effect) and the sorption equilibrium (i.e. the sorption isotherm) of a given gas and material system. When developing models, it is of fundamental importance to consider that at the surface of non-porous materials, and at any point within porous materials, the local sorption kinetics are much faster than the diffusion process in the material. Practically, this means that the sorption transports at the solid surfaces can be considered to be instantaneous: the molecules in the bulk air or in the pore space, on the one hand, and the adsorbed molecules, on the other, are in equilibrium with each other; their concentrations are related by the sorption isotherm of the contaminant and material system. However, as diffusion will contribute to establish a concentration gradient between the bulk air phase of the room and the depths of the material, the sorption equilibrium must be considered locally. Especially in the case of porous materials, space-varying pore-phase concentrations will define different equilibrium points within the material. Finally, there is evidence that the sorption parameters to consider in models are the adsorption isotherms (or pseudo-absorption isotherms) of the contaminants. The sorption dynamics of a wall or furnishings do not relate to the sorption transports themselves, but to the kinetics of boundary layer and solid diffusion. Hence, there is no physical sense in modelling the sink effect by macroscopically relating the room concentrations to the corresponding material bulk concentrations through adsorption and desorption constants, or the ratio of the two. Besides, these constants would depend upon the material thickness, instead of being intrinsic properties of the gas and material systems investigated. Such models, as well as the related experimental data, are not scientifically sound and should be disregarded.

One-phase or two-phase models? One-phase models consider the materials as homogeneous solid media and, consequently, hold that adsorption/desorption processes occur at the material surface exposed to the room air. On the other hand, two-phase models consider diffusion within the pores and adsorption/desorption transports occurring locally at the pore surfaces all along the diffusion path. Theoretically, one-phase models should only hold for materials such as polymers that do not have any measurable porosity, but allow for gas diffusion. Indeed, there is fundamentally no good reason to consider that sorption will occur only at the material surface when the pore surface area available is usually much larger; for instance, Blondeau et al (2003) report effective surface areas as high as 500, 1750 and 1600m2.kg–1 for brick, particle board and gypsum, respectively. Moreover, one must keep in mind that

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 63

Modelling of Contaminant Interactions in Buildings

63

adsorption/desorption is a surface phenomenon. Therefore, if the material has no porosity, or is modelled that way, the sorbed phase concentrations should not be defined in units of mass of gas per mass of material, but in units of mass of gas per unit surface area exposed to the bulk air (in the case of homogeneous porous materials, the specific surface area of the material is correlated to its volume; consequently, the sorbed phase concentration can equally be expressed in units of kg/m2 or kg/kgmat). However, one would then have to face the problem of solving the diffusion equation with boundary conditions defined in terms of surface concentrations, which is tricky to say the least. In the end, the physical consistency of one-phase models applied to non-porous polymeric-like building materials (e.g. paints and vinyl flooring) may come from a different interpretation of the contaminant partitioning between the air phase and the solid phase at the material surface. As suggested in many references dealing with the transport of gases in solid polymers (Treybal, 1981; Sun et al, 1997; Rutherford et al, 2005), the mass transports at the material surface should be seen as absorption/desorption phenomena rather than adsorption/desorption processes, and modelled analogously with Henry’s law for dilute liquid solutions. For gas/solid systems, the latter will be expressed as: Cmat = K H C

[10]

where C (g/m3air) and Cmat (g/m3mat) are the equilibrium air phase and material concentrations, respectively, and KH is a temperature-dependent constant that represents the solubility of the species in the material. Similarly to adsorption processes, the kinetics of absorption are likely to be so rapid compared to diffusion transports that they can be treated as instantaneous. On this assumption, the near-surface air phase and material concentrations will constantly remain in equilibrium, as described by Equation 10, but also by Equation 5. Therefore, one-phase models as described by Equations 3 to 5 are most certainly relevant for the modelling of gaseous species transports in polymeric materials; but K should be referred to as the pseudo contaminant solubility in the polymeric material (KH) rather than the partition coefficient of the adsorption isotherm (Kp).

Nodal, zonal or computational fluid dynamics (CFD) models In addition to the criteria used above, source and sink models can be classified as nodal, zonal and computational fluid dynamics (CFD) models based on the level of accuracy used to describe the contaminant transport in the room air. Nodal models consider that the contaminant is instantaneously and perfectly mixed within the room air. They do not require any information about the airflow distribution; but accordingly do not provide any information on the contaminant distribution within the room. Most of the source and sink models were implemented based on this assumption. In the CFD approach, the room and, in some cases, the material is divided into a large number of small control volumes or elements (Yang and Chen, 2001; Murakami et al, 2003; Ataka et al, 2004). Detailed information about the airflow in the bulk air phase of the room, as well as at the material surface, is obtained by solving the equations expressing the conservation of momentum and energy. This may prove to be useful in accurately evaluating the convection mass transfer coefficient, which is of central importance when boundary layer diffusion is the

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

64

3:31 PM

Page 64

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

rate limiting process of the gas and material interaction. Finally, zonal models are intermediate between CFD and nodal models. These models divide the room volume into a small number of cells where the air is assumed to be perfectly mixed. The airflow rate between adjacent cells is given either by the Bernouilli equation or specific empirical equations representing specific driving flows, such as thermal plumes, jets or boundary layer flows. Solving the coupled heat and mass conservation equations within each cell provides a rough overview of the airflow and resulting contaminant concentration distributions within the room. Damian et al (2003) have presented a representative example of how these room models can be coupled with Equations 4 and 8 to account for the simultaneous influence of ventilation and interactions with materials on IAQ.

EXPERIMENTAL DATA As well as knowing which models are relevant or which models are the most suitable for a given application, the problem of modelling the material and contaminant interactions is also to input adequate experimental data for the model parameters – namely, the convection coefficients at the material surface, the adsorption isotherm or the so-called solubilities for polymeric materials, the contaminant diffusivities and the initial concentrations in the materials. Somewhat surprisingly, this proves to be extremely tricky, especially for sorption and diffusion properties. The reasons are explained hereafter, together with a compilation of existing data and correlations for gas and building material systems.

CONVECTIVE MASS TRANSFER COEFFICIENTS The convection average mass transfer coefficient of Equation 4 strongly depends upon the characteristics of the airflow at the material surface. It can be measured directly (Matthews et al, 1987; Haghighat and Zhang, 1999; Morrison et al, 2003) or indirectly using the naphthalene technique, as described by White (1991). Assuming isothermal conditions (i.e. the boundary layer only originates from a concentration gradient) and low enough indoor air velocities to consider that the flow is laminar, h (m.s–1), can be obtained from correlations derived from the heat and mass transfer analogy. These correlations relate the Sherwood number (Sh) to the Rayleigh (Ra) and Schmidt (Sc) numbers in the case of natural convection, and to the Reynolds (Re) and Schmidt (Sc) numbers in the case of forced convection. For a characteristic length L (m), these are defined by: hL Dm

[11]

g βc ∆CL3 νDm

[12]

Re =

U ∞L ν

[13]

Sc =

ν Dm

[14]

Sh =

Ra =

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 65

Modelling of Contaminant Interactions in Buildings

65

where Dm (m2/s) is the molecular diffusivity of the contaminant in air, U∞ (m/s) the air velocity outside the boundary layer, g the gravitational constant (m.s–2), ν (m2.s–1) the kinematic viscosity of air and ∆C (kg/m3) the concentration gradient within the boundary layer. On the assumption that the ideal gas law applies, βc (m3.kg–1), the volumetric expansion coefficient of the species, is given by: βc = −

VM (M − M air ) VM C (M − M air ) + MM air

[15]

where VM (m3/mol) is molar volume, and M and Mair (kg/mol) are the molecular weights of the contaminant and air, respectively. Based on the assumption that the boundary layer airflow is similar to unidirectional flow over a flat plate with a characteristic length scale taken as the length of the wall in the direction of the flow, Bejan (1995) has suggested the following correlations for natural and forced convection with laminar flow. For natural convection: 1

Sh = 0.671Ra L 4 1

[16]

if Sc > 1

1

Sh = 0.8 Ra L 4Sc 4

[17]

if Sc < 1

For forced convection: 1

1

1

1

Sh = 0.664 Re L 2Sc 3 Sh = 1.128 Re L 2Sc 2

if Sc > 1

[18]

if Sc < 1

[19]

In the case of mixed convection, the dominating convection mechanism is decided by the smaller of natural and forced film thicknesses. Bejan (1995) has shown that the problem consists merely in studying the ratio: 1

Ra L

1 4

1

1

1

Re L 2 Sc 4

if Sc > 1

ReL 2 Sc 3

or

RaL

1 4

if Sc < 1.

[20]

If less than unity, the dominating mechanism is natural convection, and Equations 16 and 17 apply. Otherwise, forced convection prevails and h can be derived from Equations 18 and 22. Still assuming isothermal conditions but turbulent flow at the material surface, White (1991) suggests that h be calculated from the equation: 1

4

Sh = 0.037 Sc 3 ReL 5 ,

[21]

which is also derived from the heat and mass analogy for forced convection over a flat plate.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

66

3:31 PM

Page 66

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

SORPTION PARAMETERS The sorption properties of interest for the modelling of both the sink effect and VOC emissions is the contaminant adsorption isotherm for porous materials, and the pseudo absorption constant KH of Henry’s absorption law for non-porous materials such as polymers. When various adsorption isotherm models exist (see Table 3.1), the adsorption equilibrium of gaseous contaminants in buildings has most often been described as a linear correlation between the sorbed phase and air phase concentrations (e.g. see Equation 5). To some extent, this is most certainly a reasonably good assumption in many circumstances since contaminant concentrations in indoor settings typically do not exceed few parts per billion (ppb), on the one hand, and, for physical consistency, all sorption isotherm models should reduce to linearity at low concentration, on the other. However, as noticed by Elkilani et al (2003), the upper limit of what is called low concentrations is not clearly defined. In particular, the linear isotherm assumption may fail in representing the first stages of the VOC emission process. Indeed, newly manufactured materials often exhibit pore-phase VOC concentrations that are several orders of magnitude higher than those commonly found in indoor settings. Within this concentration range, the adsorption isotherm may show non-linearities, as shown by Tiffonnet et al (2002) when dealing with the emission of acetone from particle board. Practically, sorption parameters have been determined in the context of characterizing either the sink or source behaviour of building materials. Many experiments have been carried out and partition or absorption coefficients for many gas and material systems can be found in the literature. Much of this data is, nevertheless, inadequate, uncertain or proves to be tricky for modellers and practitioners. The reasons are fourfold: 1 Many authors have determined K values (Kp or KH) by fitting Equation 5 or other macroscopic models to measured concentration profiles (non-equilibrium

TABLE 3.1 Representative adsorption isotherm models for building applications MODEL Linear Langmuir

CS = f(C)

MODEL PARAMETERS KP : Partition coefficient

Cs = K pC Cs =

Cs0 K LC 1 + K LC n

Freundlich

Cs = K F C

Temkin

Cs = aln (bC )

Brunauer Emmett and Teller (BET)

Cs0: Monolayer sorbent concentration KL : Langmuir constant KF: Freundlich constant nF: Freundlich exponent

F

a, b: Temkin constants

Cs0K BET Cs =

⎛ ⎜⎜1 − C ⎜⎜⎝ C

C Csat

⎞⎟⎛ C ⎟⎞ C ⎟⎟ ⎟⎟⎜⎜1 − + K BET ⎜⎜⎝ ⎟ ⎟ C C ⎠ sat sat sat ⎠

Cs0: Monolayer sorbent concentration KBET: Brunauer, Emmett and Teller (BET) constant Csat: Saturation concentration

Source: Axley (1993)

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 67

Modelling of Contaminant Interactions in Buildings

67

concentrations). K was either defined by the ratio of adsorption and desorption constants, ka /kd, or the ratio of rate coefficients describing the rates of mass transfer among the sinks (Colombo et al, 1993; Jørgensen et al, 1999, 2000; Won et al, 2000, 2001; Singer et al, 2004). However, as discussed in section ‘Synthesis’, such data characterizes the sorption dynamics (which includes the contribution of diffusion) rather than the sorption equilibrium, and is therefore not relevant. The most compelling proof is that the proposed values would prove to depend upon the material thickness when the partition coefficient is a physical property of the gas and material system. 2 In other cases, the partition coefficients have been determined simultaneously with the contaminant diffusivities by fitting one-phase or two-phase models to transient sorption or desorption data (He et al, 2005; Li and Niu, 2005). Such models consider local sorption equilibrium and thus do provide equilibrium sorption data even if drawn from dynamic tests. The problem here is that two parameters are to be determined simultaneously; the best mathematical fitting does not provide any guarantee on the physical consistency of each parameter. 3 The units used for the air phase concentrations, sorbed phase concentrations and partition coefficients add one step of complexity to the problem. As discussed earlier, in the case of porous materials, the surface area available for adsorption is the effective surface area of the material (sometimes called BET surface area) and the sorbed phase concentrations should be expressed accordingly in units of mass of contaminant adsorbed per square metre of pore surface area. As a result, partition coefficients determined from sorbed phase concentrations expressed in units of mass of contaminant per nominal surface area of the material (surface area exposed to the bulk air of the test chamber) have no sense; the proposed values would actually prove to vary with the size of the material sample tested, which is not consistent. On the other hand, since the effective surface area of a porous media depends linearly upon its volume, the sorbed phase concentrations can also be expressed in units of mg/kgmat or mg/m3mat, which is undoubtedly more convenient for modelling purposes, but raises another critical problem regarding the data to input in the model: as defined by Equation 10, or the linear model presented in Table 3.1, the partition coefficient describing the adsorption or absorption equilibrium is defined by the ratio of the sorbed phase and air phase concentrations. Thus, whether these two concentrations are in units of mg/kg or mg/m3, K (Kp or KH) will be dimensionless and, in practice, most often presented that way in the literature. However, it is essential to consider that K will actually have units of kgair/kgmat or m3air/m3mat, so we have: ⎡ m3 ⎤ ⎡ kg ⎤ ρ K ⎢⎢ air ⎥⎥ = air K ⎢⎢ 3air ⎥⎥ ⎣ kgmat ⎦ ρmat ⎢⎣ mmat ⎥⎦

[22]

where ρmat and ρair (kg/m3) are the material and air densities, respectively. Given that ρmat is from two to three orders of magnitude higher than ρair, partition coefficients

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

68

3:31 PM

Page 68

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

determined from volume concentrations will also be from two to three orders of magnitude higher than those determined from mass concentrations. Although being of fundamental importance when proposing or reporting K values, this information is most often omitted or hard to find. Systematically detailing the units of partition coefficient data would definitely be an easy way to avoid confusions and mistakes. 4. Although some experiments have been carried out in dry conditions, most of the existing sorption data were determined at typical relative humidity levels indoors (from about 40 to 60 per cent; see Table 3.2). The problem is that humidity can either decrease or promote the sorption of contaminants due to the competition for the active sites or other complex interaction processes within the material. This will be discussed in section ‘Influence of humidity’. We simply note here that partition coefficients determined in humid conditions may not correctly represent the intrinsic sorption properties of the contaminant species. Tables 3.2 and 3.3 present a non-exhaustive compilation of relevant sorption data based on the previous remarks. A distinction has been made between the data that were fit from one-phase and two-phase models or, similarly, what should actually be considered to be absorption coefficients of Henry’s law (KH) or partition coefficients of the linear adsorption isotherm model (Kp). As one will notice from the first list, some data have been determined from one-phase models when the materials investigated are likely to be porous. Such data can, nevertheless, be input in two-phase models, provided that adequate units are used.

DIFFUSION COEFFICIENTS Diffusion models and definition of diffusivities Diffusion of gases in non-porous building materials such as synthetic floorings, paints and other kinds of polymeric materials can be characterized by a bulk diffusivity that depends both upon the diffusing species and the chain structure of the polymer. Diffusion through polymers requires that a hole or passage of sufficient size be available and, thus, in turn, depends upon the thermal motion of the polymer chains (Treybal, 1981). Practically, this means that small molecules are expected to diffuse faster than larger ones, and that the diffusion rate of any species will increase with temperature, as will be further discussed in section ‘Influence of temperature’. A more detailed description of the mechanisms involved in gas diffusion in polymers and their practical implications on VOC emissions can be found in Cox et al (2001a). Unlike diffusion in polymers, diffusion of gases in porous materials involves a complex variety of processes, including air phase diffusion and surface diffusion within the porous interstices (Treybal, 1981; Ruthven, 1984; Masel, 1996). Air phase diffusion puts into play two distinct diffusion processes depending upon the pore size considered: molecular diffusion occurs in the largest pores where collisions between gaseous molecules are predominant, while Knudsen diffusion occurs in the smallest pores where the mean free path of the diffusing species is limited by the pore dimensions. The mean free path, λ (m), is the average distance travelled by molecules

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 69

Modelling of Contaminant Interactions in Buildings

69

TABLE 3.2 Compilation of sorption and diffusion coefficients: Data fit with one-phase models (K
KH) REFERENCE Little et al (1994)

Bodalal et al (2000)

Cox et al (2001b)

MATERIAL/CONTAMINANT Carpet 1a/styrene Carpet 1b/styrene Carpet 1a/4ethenylcyclohexene Carpet 1a/ethylbenzene and xylenes

K 4200 6500 1400

D (M2/S) 4.1 × 10–12 3.6 × 10–12 5.2 × 10–12

COMMENTS K in units of m3air /m3mat

1500

10.2 × 10–12

Carpet 1b/ethylbenzene and xylenes Carpet 1a/4-phenylcyclohexene Carpet 1b/4-phenylcyclohexene Carpet 3/formaldehyde Carpet 3/acetaldehyde Carpet 3/2,2,4-trimethylpentane Carpet 3/1,2-propanediol Carpet 3/2-ethyl-1-hexanol Carpet 4/styrene Carpet 4/4-ethenylcyclohexene Carpet 4/ethylbenzene and xylenes Carpet 4/4-phenylcyclohexene Carpet backing material/toluene

2400

4.3 × 10–12

The numbers following ‘carpet’ identify different samples

81,000 67,000 11,000 1 59,000 180,000 450,000 5700 1700 5300

0.59 × 10–12 0.50 × 10–12 3.2 × 10–12 6.4 × 10–12 0.060 × 10–12 0.065 × 10–12 0.088 × 10–12 3.1 × 10–12 2.1 × 10–12 1.5 × 10–12

170,000 6171

1.2 × 10–12 43.10 × 10–12

Carpet backing material/nonane Carpet backing material/decane Carpet backing material/undecane Plywood/cyclohexane Plywood/ethylbenzene Plywood/decane Floor tile/ethylbenzene Floor tile/nonane Floor tile/decane

6216 14,617 24,255

28.30 × 10–12 5.42 × 10–12 2.79 × 10–12

348 1636 6948 1920 2142 13,045

155 × 10–12 40.4 × 10–12 12.8 × 10–12 16 × 10–12 14.8 × 10–12 2.09 × 10–12

Floor tile/undecane Vinyl flooring/n-butanol Vinyl flooring/toluene Vinyl flooring/phenol Vinyl flooring/n-decane Vinyl flooring/n-dodecane Vinyl flooring/n-teradecane Vinyl flooring/n-pentadecane

26,647 810 ± 77 980 ± 34 120,000 ± 3000 3000 ± 420 17,000 ± 260 120,000 ± 1300 420,000 ± 38,000

0.85 × 10–12

K and D determined as the best fit of published emission data

K in units of m3air/m3mat T = 23°C ± 1°C RH = 50% ± 5%

T = 25.6 ± 0.3°C Dry conditions

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

70

3:31 PM

Page 70

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

TABLE 3.2 Compilation of sorption and diffusion coefficients: Data fit with one-phase models (K
KH) (Cont’d) D (M2/S)

ACT2/tetradecane

K 1550 10.05 34.90 2.80 at 20°C 2.56 at 25°C 2.53 at 30°C 400

ACT3/hexanal CRP1/4-phenylcyclohexene CRP1/heptane

5524 66,833 300

109 × 10–12 235 × 10–12 609 × 10–12

CRP2/dodecane CRP2/limonene CRP2/styrene

986 1441 268

192 × 10–12 163 × 10–12 364 × 10–12

CRP3/4-phenylcyclohexene CRP3/dodecane CRP3/limonene CRP3/styrene CRP3/tetradecane CRP3/tridecane CRP4/dodecane CRP4/iso-octane CRP4/tridecane CRP5/iso-octane CRP6/tridecane

10,512 3840 1400 275 22,476 7336 42,549 50,000 69,942 84,151 71,560

117 × 10–12 174 × 10–12 325 × 10–12 455 × 10–12 12 × 10–12 81 × 10–12 75 × 10–12 9 × 10–12 72 × 10–12 1 × 10–12 136 × 10–12

Partition coefficients determined simultaneously with diffusivities and initial concentrations by fitting one-phase model to various emission data measured in small chamber tests

GB1/α-pinene GB2/α-pinene GB3/α-pinene OSB1/α-pinene

80,329 80,200 52,744 194,080

10 × 10–12 5 × 10–12 16 × 10–12 0.2 × 10–12

Proposed values are optimal values from the best fit

OSB2/α-pinene OSB2/furan PB0/α-pinene PB0/camphene PB0/hexanal PB0/limonene PB4/α-pinene PB4/camphene

21503 60,256 849 1090 550 334 86 1470

38 × 10–12 6 × 10–12 99 × 10–12 106 × 10–12 74 × 10-–12 126 × 10–12 800 × 10–12 1052 × 10–12

REFERENCE Yang et al (2001)

MATERIAL/CONTAMINANT Gypsum board/ethylbenzene Gypsum board/ethylbenzene Gypsum board/ethylbenzene Tiffonnet et al Acrylic paint/acetone (2002)

He et al (2005)

500 × 10–12

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

COMMENTS K in m3air/m3mat T = 23°C ± 1°C RH = 50% ± 2% K in gair/gmat (original data in units of gair /m2mat) ACT: acoustic ceiling tile CRP: carpet GB: gypsum board OSB: oriented strand board PB : particle board PLY : plywood K in units of m3air/m3mat

of emission profiles The numbers following the material name identify different samples

ABER_CH_03

6/12/08

3:31 PM

Page 71

Modelling of Contaminant Interactions in Buildings

71

TABLE 3.2 Compilation of sorption and diffusion coefficients: Data fit with one-phase models (K
KH) (Cont’d) REFERENCE

Elkilani et al (2003)

MATERIAL/CONTAMINANT PB5/α-pinene PB5/camphene PB5/hexanal PB5/limonene PB6/camphene PB6/hexanal PLY1/3-carene PLY2/3-carene PLY2/α-pinene PLY2/p-cymene PLY3/3-carene PLY3/α-pinene PLY3/limonene PLY3/p-cymene Carpet fibres/toluene

K 492 1570 70,927 1898 564 19,650 37,562 125,632 80,560 29,736 19,676 6749 11,454 15,974 0.86 at 25°C 0.73 at 30°C 0.50 at 35°C 0.34 at 45°C Carpet fibres/1,2-dichlorobenzene 1.60 at 25°C 0.83 at 30°C 0.74 at 35°C 0.36 at 45°C Carpet fibres/1,1,1-trichloroethane 0.18 at 25°C 0.16 at 30°C 0.13 at 35°C 0.11 at 45°C

D (M2/S) 655 × 10–12 1011 × 10–12 16 × 10–12 900 × 10–12 63 × 10–12 66 × 10–12 388 × 10–12 0.4 × 10–12 0.4 × 10–12 0.4 × 10–12 22 × 10–12 0.1 × 10–12 36 × 10–12 44 × 10–12

COMMENTS

Dry conditions Original data in units of m – converted to gair/gmat using the specific surface area of the carpet fibres provided by the authors

between collisions with other particles. Assuming that the ideal gas law applies, it can be determined from the equation (Mulder, 1991): λ=

RT 2πd 2N AP

[23]

where T (K) is temperature, NA the Avogadro’s number, R (J.mol–1.K–1) the universal gas constant, d (m) the diameter of the gas molecules, and P (Pa) the pressure of the diffusing system. In the absence of data for the molecule diameter, the mean free path can also be estimated from the relation (Treybal, 1981): λ=

0 .5 3.2µ ⎛⎜ RT ⎞⎟ ⎟⎟ ⎜⎜ P ⎝ 2πM ⎠

[24]

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

72

3:31 PM

Page 72

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

TABLE 3.3 Compilation of sorption and diffusion coefficients: Data fit with two-phase models (K
Kp) REFERENCE Silberstein (1989) RadulescuBouilly et al (2006) Tiffonnet et al (2002)

MATERIAL/ CONTAMINANT Gypsum/ formaldehyde Gypsum/ tetrachloroethene Gypsum/acetone

D (M2/S)

In units gair/gmat

0.094

Dry air T = 20°C In units gair/gmat Dry air Kp values for acetone in particle board were determined from the low concentration asymptote of nonlinear isotherms Effective diffusivities in the materials were determined independently from mercury intrusion porosimetry tests (see Blondeau et al, 2003) K in units of gair/gmat T = 24 ± 0.5°C RH = 45 ± 3%

1.60 x 10–6

0.70 at 20°C 0.60 at 25°C 0.46 at 30°C 0.25 at 20°C

1.41 x 10–6 1.46 x 10–6 1.51 x 10–6 4.92 x 10–7

Carpet with styrene budadiene rubber (SBR) backing/ n-octane Aerated concrete/ n-octane Solid concrete/n-octane Brick/n-octane

0.381

3.5 × 10–7

0.12

7.6 × 10–7

0.031 0.016

1.0 × 10–7 3.2 × 10–7

Gypsum/n-octane Carpet with SBR backing/ethyl-acetate Aerated concrete/ ethyl-acetate Solid concrete/ ethyl-acetate Brick/ethyl-acetate Gypsum/ethyl-acetate

0.105 0.170

8.4 × 10–7 4.5 × 10–7

> 2.614

> 5.7 × 10–7

> 0.590

> 0.5 × 10–7

0.129 0.133

4.7 × 10–7 11.2 × 10–7

Concrete/ acetone

COMMENTS

5

0.69 at 20°C

Particle board/ acetone

Meininghaus et al (2000)

K

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 73

Modelling of Contaminant Interactions in Buildings

73

where µ (Pa.s) and M (kg.mol–1) are the dynamic viscosity and molecular weight of the diffusing system, respectively. For a given pore radius, r (m), the dominant diffusion mechanism is determined by the ratio of the pore diameter and the mean free path of the gas molecules. In the range 2r/λ from roughly 0.2 to 20, both molecular and Knudsen diffusion have influence. When 2r/λ is much less than unity, diffusion is induced by the bouncing of the molecules to and from the pore walls, and is thus represented by the Knudsen diffusivity, Dk (m2/s), which is defined as: Dk =

2r 3

8RT πM

[25]

On the other hand, molecular diffusion prevails when 2r/λ is much larger than unity. In such conditions, the pore diffusion rate depends upon the contaminant molecular diffusion coefficient in air, Dm (m2/s), which can either be found in tables or evaluated from online tools such as the one developed by the US Environment Protection Agency (EPA, 2006). This molecular diffusion coefficient in air is in the order of 10–5 to 10–4 m2/s for most gases at normal temperature (Do, 1998). Since building materials usually exhibit a wide range of pore sizes (see Blondeau et al, 2003), both molecular and Knudsen diffusions can occur simultaneously within a unit volume of material. Considering, furthermore, that the porous structure is not made up of straight and cylindrical pores, but a series of interconnecting tortuous paths of varying cross-sectional area, the contaminant flux in the air phase of the pores, Jp (kg.m–2.s–1), can be described using Fick’s law by defining an effective diffusion coefficient, De (m2.s–1): J p = ρair De ∇C

[26]

with: De =

D °εσp τ

[27]

In Equation 27, ε, the connected-pore volume fraction, accounts for the real cross-sectional area offered to diffusion, while τ, the tortuosity factor, accounts for the geometric constraints that give a longer diffusion path in the flow direction. σp, the constriction factor, stands for the effects of varying cross-sectional areas within a pore. It is often omitted, either because it is close to unity or because these effects are included in the tortuosity factor. Finally, D° (m2/s) is the mean diffusion coefficient in the pores of the material. The latter can be calculated from the pore-volume distribution and an extension of the well-known Bosanquet equation: ∞ ⎛ 1 1 ⎟⎞ ⎟⎟f (r )dr , D ° = ∫ ⎜⎜⎜ + ⎜⎝ Dm Dk (r ) ⎟⎠ 0

[28]

where f(r)dr is the volume fraction for the pores with radii between r and r + dr. Surface diffusion consists in the jumping of adsorbed gas molecules from one adsorption site to another. It occurs even at very low sorbed phase concentrations and can be modelled using Fickian-like diffusion laws expressed in terms of the sorbed phase

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

74

3:31 PM

Page 74

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

concentration, Cs (kgspecies/kgmat). This way, the contaminant flux in the flow direction x, Js (kg.m–2.s–1) is given by: Js = ρmat Ds ∇Cs ,

or:

Js = ρmat Ds

∂Cs ∇C , ∂C

[29]

[30]

where ρmat (kg.m–3) is the bulk density of the porous material, and Ds (m2.s–1) the surface diffusivity of the contaminant. As surface diffusion occurs in parallel to molecular or Knudsen diffusion processes, the total diffusion flux is the sum of the contributions of pore and surface diffusions: Jtot = J p + Js .

[31]

Therefore, substituting Equations 26 and 30 into equation 31 gives: ⎛ ∂Cs ⎞⎟ Jtot = ⎜⎜ρair De + ρmat Ds ⎟ ∇C . ⎜⎝ ∂C ⎟⎠

[32]

Finally, if further assuming a linear adsorption isotherm: ∂Cs = Kp, ∂C

[33]

Equation 32 turns to the classic form of Fick’s second law of diffusion, written in terms of only the pore gas-phase concentration: Jtot = ρair D p ∇C ,

[34]

with: D p = De +

ρmat Ds K p . ρair

[35]

As defined by Equation 35, Dp (m2.s–1) represents the more general form of the porous diffusion coefficient used in Equation 8. However, care must be taken that if the basic form of Fick’s first law for diffusion: Dapp ∇2C =

∂C ∂t

[36]

is taken as the reference mass transport model in the porous material, the diffusion coefficient of interest is the so-called apparent diffusivity of the contaminant and material system, Dapp (m2.s–1). The latter embeds both the air phase and surface diffusion coefficients and the slope of the adsorption isotherm, Kp: ρmat Ds K p ρair . ρ ε + mat K p ρair

De + Dapp =

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

[37]

ABER_CH_03

6/12/08

3:31 PM

Page 75

Modelling of Contaminant Interactions in Buildings

75

In the field of indoor air-quality modelling, it is common to ignore surface diffusion, either due to the lack on knowledge on the actual surface diffusivities of contaminants in building materials, or because it is thought to be negligible compared to air phase diffusion. The relevance of this assumption can be assessed by comparing the two terms of the right-hand side of Equation 35. After Blondeau et al (2003), the effective diffusivities of VOCs in building materials range from 10–6 to 10–7 m2.s–1. On the other hand, Treybal (1981) indicates that surface diffusivities are typically of the order of 10–7 to 10–9 m2/s at ordinary temperatures for physically adsorbed gases. However, considering that surface diffusivities significantly decrease with decreasing sorbed phase concentrations, and that the latter are usually low compared to industrial applications, Ds values of the order of 10–9 m2/s are more likely. Partition coefficients also lie in a very wide range depending upon the material and contaminant systems considered. Kp values are generally of the order of 10–2 to 10–1; but Silberstein (1989) reported a value of Kp as high as 5.5kg/kgair for formaldehyde in gypsum board (see Table 3.2). As the density of building materials typically ranges from 200kg.m–3 to 3000kg.m–3, one can finally note that ρmatDsKp may be of the same order of magnitude as ρairDe, and will most of the time be from one to three orders of magnitude smaller. Therefore, contrary to what one might think, surface diffusion may be significant and even the dominant diffusion process in some practical situations. This indicates that diffusion in porous materials can be controlled by sorption, which will prove to be of central importance when discussing the order of magnitude of diffusivities, as well as the influence of environmental parameters on the sorption and diffusion properties of gases in building materials. Another noteworthy point coming from the above discussion, and which is of great importance in regard to the following sub-section, is that diffusion coefficient data without any reference to the related diffusion model are worthless and unusable for modelling issues.

Experimental data Most of the contaminant diffusivities in building materials have been determined either from diffusion-cell methods or by fitting one-phase or two-phase models to emission or sink data. The general principles as well as the drawbacks of diffusion-cell methods have been described by Park et al (1996) and, later on, by Haghighat et al (2002). In the frame of IAQ-related studies, two kinds of diffusion-cell methods have been used: the cup method (Kirchner et al, 1999; Hansson and Stymne, 2000; Meininghaus and Uhde, 2002) and twin chamber methods (Bodalal et al, 2000; Meininghaus et al, 2000). The cup method consists of measuring the diffusion capacity of the tested system by weighting the diffusion loss of VOC through a material sample covering a cup that contains the VOC at saturation concentration (temperature and humidity are controlled). In twin chamber methods, the diffusion capacity is assessed by measuring the VOC concentration in the supply and exhaust air of two small-scale chambers placed on each side of the material sample to be tested. In both cases, the introduction of the measured steady-state concentrations and VOC flow rate in Fick’s first law of diffusion provides the intended

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

76

3:31 PM

Page 76

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

diffusion coefficients. These may be solid-phase, effective or apparent diffusivities, depending upon whether the assumption of porous material is considered or not. In all cases, one problem is that the concentrations at the material sample surfaces can hardly be measured. As a result, the concentration gradient in the material is most often taken as the bulk air concentration difference between the primary and secondary side of the cell; it includes the resistance to diffusion in the boundary layers, which leads to an underestimation of the actual diffusivity of the diffusing system (Meininghaus et al, 2000; Blondeau et al, 2003). In the specific case of the cup method, another drawback lies in the material being subjected to VOC saturation concentration, which is unrealistically higher than typical VOC concentrations in indoor settings. As diffusion coefficients have been shown to be concentration dependent at high concentrations, the data may not be representative of actual diffusivities in materials as installed in buildings (Blondeau et al, 2003). Haghighat et al (2002) proposed an extensive compilation of published VOC diffusivities in building materials. As for sorption properties, the data derived from concentration profiles measured in environmental chambers may be classified into two groups: those where the diffusivities were determined simultaneously with other parameters (partition coefficient and, in some cases, convective mass transfer coefficient and initial concentrations) from measured transient concentrations (Bodalal et al 2000; He et al, 2005; Li and Niu, 2005), and those where the sorption properties were determined first by considering the steady state, diffusivity then being defined as the best fit of the mass transport equation to the sorption or desorption concentration profiles (Meininghaus et al, 2000; Cox et al, 2001b). For the reasons explained previously, the former are much more uncertain than the latter. This was confirmed by Li and Niu (2005), as well as He et al (2005), who report potentially large uncertainties on diffusivities due to cross-uncertainties between each parameter. From a practical point of view, one common limitation of all the methods discussed above is that the experiments are costly in time and, perhaps, money. In this context, Blondeau et al (2003) suggested another method based on the analysis of the material porosity first, and then the application of Carniglia’s empirical model to determine the tortuosity factor τ. The effective diffusivity of any gas in the tested material can finally be computed from Equation 27, provided that its molecular diffusivity in air is known. A comparison between diffusivities determined that way, and data coming from traditional diffusion-cell methods, revealed a good agreement and emphasized the potential of the methodology to generate a large number of data with limited experimental effort. Table 3.2 presents diffusion coefficients that were determined simultaneously with partition coefficients and, in some cases, initial concentrations. It shows that measured De (or Dapp) values are of the order of 10–6 to 10–7 m2.s–1, while measured solid-phase diffusivities vary in a wider range from 10–10 to 10–13 m2.s–1. In the latter case, however, distinction has to be made between two kinds of data. True solid-phase diffusivities (i.e. diffusivities in materials such as vinyl flooring or carpet backing) lie in the range from 10–10 to 10–14 m2.s–1, which is in agreement with typical diffusion coefficients in polymers at ordinary temperatures for buildings (Treybal, 1981; Sun et al, 1997). On the other hand, many data have been determined based on the assumption of solid diffusion when the tested materials are likely to be porous, and would thus be better characterized by

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 77

Modelling of Contaminant Interactions in Buildings

77

pore-phase diffusivities. If such data are to be input in two-phase models, comparing Equations 3 and 8 and further assuming a linear adsorption isotherm (with Kp
KH
Kp in units of gair /gmat being the measured partition coefficient) show that measured Dmat values can be converted to Dp values from the relation: ⎛ ρ K⎞ D p = ⎜⎜⎜ε + mat ⎟⎟⎟Dmat . ⎜⎝ ρair ⎟⎠

[38]

Given that ρmat is expected to be two to three orders of magnitude higher than ρair, Dmat measured for porous materials is expected to be one order of magnitude lower than Dp if Kp < 10–3, and several orders of magnitude lower otherwise. Most of the data where Kp and Dmat values were determined simultaneously are consistent with this (see Table 3.2).

INITIAL CONCENTRATIONS Initial concentrations are a key parameter for the modelling of VOC emissions from materials. However, knowing the initial VOC content of building materials faces the problem of industrial secrecy, and no information is to be expected from manufacturers. Besides, this information is usually hard to get experimentally. Methods such as the solvent extraction or high thermal desorption (Tichenor, 1996) have been widely used but carry the risk that the chemical materials are transformed during operation. For dry materials, an interesting alternative has recently been proposed by Cox et al (2001a), who developed a method based on cryogenic grinding and fluidized bed desorption to measure the concentrations of volatile organic compounds in vinyl flooring. Finally, a third way to access the initial concentrations is first to integrate, over time, the emission profiles measured in environmental chambers, and then to divide the estimated total mass emitted by the material mass or volume (Little et al, 1994; Elkilani et al, 2003). This method has the advantage that many emission data exist and thus can provide a substantial amount of information within a reasonable time period. However, it may lead to a great underestimation of the initial concentrations if significant amounts of VOC remain in the material at the end of the experiment. Regarding the modelling issue, it is important to consider that the measured VOC contents, C0tot (kgspecies/m3mat), can directly be used as the boundary condition of one-phase models. On the other hand, the parameter to input in two-phase emission models is the initial concentration in the air phase of the pores, C0 (kg/kgair). The latter can be determined by considering that the total mass of VOC contained in the material is the lump sum of the VOC contained in the gas phase and in the sorbed phase – that is to say: 0 mmat C tot = mair C 0 + mmat Cs0 ,

[39]

where C0s (kgspecies/kg) is the initial sorbed phase concentration, mmat (kg) the mass of the material sample, and mair (kg) the mass of air in the material sample. Since, the last two parameters are correlated through the relation: mair = ρair ε

mmat , ρmat

[40]

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

78

3:31 PM

Page 78

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

substituting Equation 39 into equation 40 and rearranging gives: 0 Ctot =

ερair 0 C + Cs0 ρmat

[41]

Finally, if assuming a linear isotherm, we have: C0 =

0 Ctot

ε

ρair + Kp ρmat

[42]

CORRELATION WITH PHYSICAL PROPERTIES In spite of research efforts, there is evidence that there is insufficient data on the sorption and diffusion properties of contaminants in building materials to model the complexity of real indoor settings. The proposal of standards, or at least validated methods, to get such data would definitely be a first step in progressing towards informative databases. However, experimental methods find their limitations in the huge number of gases and materials encountered in buildings, which results in a prohibitive number of systems to be tested. Identical materials, furthermore, may exhibit different properties depending upon the manufacturer and manufacturing process. For instance, Blondeau et al (2003) noted that even slight differences in the porous structure of gypsum board samples would result in differences in the contaminant diffusivities from one sample to another. Recently, Seo et al (2005) carried out porosity tests on a commercial gypsum board and reported a porosity as high as 68 per cent – that is to say, more than twice the porosity measured by Blondeau et al (2003)! In this context, several authors have searched for correlations between the experimental parameters of interest and the gas properties, such as van Der Waals molar volume, mean diameter of the molecule, saturation vapour pressure/boiling point, or molecular weight (Berens and Hopfenberg, 1982; An et al, 1999; Bouhamra and Elkilani, 1999; Bodalal et al, 2000; Cox et al, 2001b; Meininghaus and Uhde, 2002; Elkilani et al, 2003, He et al, 2005). The general trends are that VOC diffusion coefficients often decrease as the molecular weight of the compound increases, and sorption constants generally increase as the saturation vapour pressure of the compound decreases (Little and Hodgson, 1996). These conclusions are somewhat consistent with the principles of diffusion in polymers: species having a high molecular weight are generally large molecules; they will diffuse slowly through the polymer chains. On the other hand, the relationship between diffusivity and molecular weight is harder to establish analytically for porous materials. Three cases are to be distinguished: if Knudsen diffusion is the controlling mechanism within the pores of the material, Equation 25 indicates that the diffusion coefficient is expected to vary as a function of M–1/2. If molecular diffusion prevails, the Chapman-Enskog equation (Treybal, 1981), which is often used to assess the molecular diffusivity of gases in air, relates Dm to M in a non-explicit way. It would, nevertheless, show that molecular diffusivities also decrease as the molecular weight of the species increase. Finally, if pore phase diffusion is negligible compared to surface diffusion, which has been

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 79

Modelling of Contaminant Interactions in Buildings

79

demonstrated to be the case in some circumstances, Equation 35 indicates that diffusivities are defined by the partition coefficient of the diffusing system. Therefore, Dp will be expected to vary the same way as Kp – that is to say, to increase with molecular weight. Regarding sorption, some authors found a linear increase in KH with the reciprocal of saturation vapour pressure (An et al, 1999; Elkilani et al, 2003), while, in a similar way, others concluded that the logarithm of KH linearly correlates with the logarithm of saturation vapour pressure (Cox et al, 2001b; Won et al, 2001; Weschler, 2002). Practically, these correlations are of great interest in the sense that they provide a convenient means to extrapolate the existing data to other compounds. Since vapour pressure depends upon temperature, they may also serve to estimate KH at temperatures other than the one set experimentally (see section ‘Influence of temperature’). Care should, nevertheless, be taken when extrapolating the data since the correlations do not account for molecular structure or functionality, and thus only apply for compounds belonging to the same chemical group; in other words, when drawing KH versus vapour pressure for different VOCs, the plots will not lie on the same unique straight line (Elkilani et al, 2003). Furthermore, the linear relationships may prove to depend upon temperature. As an example, Elkilani et al (2003) noted that KH in carpet fibres steadily increases with the boiling point of the species (and decreases with their vapour pressure) at 25°C, but levels off in the boiling point range of 100°C to 200°C at temperatures up to 30°C.

MODELLING THE INFLUENCE OF ENVIRONMENTAL PARAMETERS The influence of indoor temperature, relative humidity and, to a lesser extent, airflow and gas mixtures on both the VOC emissions and sink effect has been widely investigated by experimental means during the past 20 years (Haghighat and De Bellis, 1998; Wolkoff, 1998; Fang et al, 1999). Although useful in understanding and interpreting the phenomena, these experiments, nevertheless, cannot fit real configurations of building operation where hundreds of contaminants are present in the air, on the one hand, and environmental parameters are constantly varying over time, on the other. This section discusses the way in which physically based models could be extended to address this issue.

INFLUENCE OF TEMPERATURE Whether they have been carried out in real buildings or in small-scale chambers, most of the experiments dealing with the influence of temperature on gas and material interactions conclude on a lower adsorption of building materials or an increased emission at higher temperatures. For instance, Seifert et al (1989) found that the emission of 4-phenyl cyclohexene from carpets increased fivefold when increasing the temperature from 25°C to 50°C. More recently, when studying the influence of temperature on the emission rate of VOC from two carpets, a PVC flooring and an acrylic paint, Van Der Wal et al (1997) concluded that the general trend was a faster emission with increasing temperature. They also noted that, except for paint, the total emitted mass was higher at 50°C than at 23°C. Numerically, the influence of temperature on the sink effect and VOC emission can be modelled by considering the way in which diffusivities and sorption properties vary with

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

80

3:31 PM

Page 80

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

temperature. In both cases, a distinction has once more to be made between polymeric and porous materials.

Influence of temperature on diffusivities Diffusion coefficients in polymers at different temperatures are often found to be well predicted by: DA = DA 0 e

E − a RT

,

[43]

where DA0 (m2.s–1) is a constant and Ea (J.mol–1) is the energy of activation for diffusion. The influence of temperature on pore diffusivity is somewhat more difficult to assess since the latter may depend upon the sorption properties of the diffusing system, which are themselves strongly influenced by temperature (see below). If sorbed phase concentrations are low enough so that surface diffusion can be neglected (Dp
De), pore diffusivities will diminish to effective diffusivities. The latter have the following temperature-dependent form (Do, 1998): ⎛ T ⎞α De = De 0 ⎜⎜⎜ ⎟⎟⎟ , ⎜⎝T0 ⎟⎠

[44]

where De0 (m2.s–1) is the effective diffusivity at some reference temperature T0 (K). By reference to Equation 25 for Knudsen diffusivity, and the Chapman-Enskog equation for molecular diffusivity, α will actually have the value of 0.5 if Knudsen diffusion is the controlling mechanism within the pores of the material, and about 1.75 if molecular diffusion prevails. α will range from 0.5 to 1.75 in the intermediate region where both processes control the pore diffusion. If surface diffusion in the porous medium is to be considered, the influence of temperature will be stronger as surface diffusivities also follow an Arrhenius-type law for temperature dependency: ⎛⎜ E s ⎞⎟ ⎟ ⎜⎜− RT ⎟⎟⎠

Ds = Ds∞ e⎝

,

[45]

with Ds∞ (m2.s–1) being the surface diffusivity at infinite temperature and Es (J.mol-1) the activation energy for surface diffusion. Practically, Equations 43 to 45 indicate that within typical temperature ranges for buildings, the influence of temperature on gas diffusivities will be significant or almost negligible depending upon the materials and the concentration levels that are put into play. For instance, based on the results of Kumar and Siddaramaiah (2005) and other references, the activation energy for VOC diffusion in polymers at ordinary temperatures may be thought to range from 20 to 30 kJ/mol. Taking these values for Ea in Equation 43 gives that DA will be increased about twofold when the temperature is increased from 20°C to 35°C. On the other hand, Equation 44 indicates that the corresponding increase in De would not exceed 10 per cent in the case of a porous material operating at moderate sorbed phase concentrations.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 81

Modelling of Contaminant Interactions in Buildings

81

After Do (1998), the temperature dependency of the partition coefficient Kp follows an Arrhenius-type law: ⎛⎜ ∆Hads ⎟⎞ ⎟ ⎜⎜− RT ⎟⎟⎠

K p = K p∞ e ⎝

,

[46]

with Kp∞ being the partition coefficient at infinite temperature, ∆Hads (J.mol–1) the heat of adsorption, R (J.mol–1.K–1) the universal gas constant and T (K) the absolute temperature. On the other hand, the temperature dependency of the absorption constant in polymers, KH, follows a Van’t Hoff equation, which by integration yields: ⎛⎜ ∆Habs ⎟⎞ ⎟ ⎜⎜− RT ⎟⎟⎠

K H = K H ∞ e⎝

[47]

Here the pre-exponential factor KH∞ is the absorption constant at infinite temperature and ∆Habs (J.mol-1), the activation energy for absorption (or heat of absorption), is the difference between the internal energies of the adsorbate in the absorbed and gaseous states. In the end, one can note that Equations 46 and 47 have similar shapes. Furthermore, since ∆Hads and ∆Habs have negative values, they both indicate that KH and Kp decrease with increasing temperatures, and the higher the heat of sorption, the greater the temperature sensitivity of the sorption parameters. By way of illustration, Elkilani et al (2003) found ∆Habs absolute values of 36.6, 56.9 and 20.3kJ.mol–1 for absorption of toluene, 1-2-dichlorobenzene and 1,1,1-trichloroethane in carpet fibres, respectively. As shown in Table 3.2, it results in a 41, 53 and 28 per cent decrease in the absorption coefficient when the temperature is increased from 25°C to 35°C. Tiffonnet et al (2002) found a lower heat of absorption of –7.5 kJ/mol for acetone in acrylic paint, reflecting slighter changes in KH with temperature (see Table 3.2). Finally, still based on the results of Tiffonnet et al (2002), the acetone heat of adsorption in particle board would have a value of about –30 kJ/mol – that is to say, the same order of magnitude as typical heats of absorption in polymers. However, there is too little data to draw definitive conclusions on the compared temperature sensitivity of Kp and KH. If contaminant and material interaction models were to be developed in conjunction with heat-transfer models, the influence of temperature on mass transports within porous materials could be automatically accounted for by considering that Cs is a function of both the species air-phase concentration and temperature. Therefore, Equation 8 would translate to: ⎛ ∂C ∂C ρ ∂Cs ∂T ⎞⎟ ∂C D p ∇2C = ε + mat ⎜⎜ s + ⎟ [48] ∂t ρ ⎝⎜ ∂C ∂t ∂T ∂t ⎟⎠ air

Then, assuming a linear isotherm and substituting Equation 46 into Equation 48 would yield: ⎛⎜ ∆Hads ⎞⎟ ⎞ ⎛⎜ ∆Hads ⎞⎟ ⎛ ⎟⎟ ⎜⎜ ⎜⎜− ρmat ∆Hads ⎜⎜⎝− RT ⎟⎟⎟⎠ ∂T ρmat ∂C ⎝ RT ⎟⎟⎠ ⎟ ⎟ ⎜ ⎟⎟ + K p∞ D p ∇ C = ⎜ε + K p∞e e ⎜⎜ ⎟⎟ ∂t ∂t ρair ρair RT 2 ⎜⎝ ⎠⎟ 2

[49]

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

82

3:31 PM

Page 82

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

INFLUENCE OF CONTAMINANT MIXTURES The few experiments that have addressed the problem of multi-component sorption in building materials lead to different conclusions. Cox et al (2001b) examined the effect of simultaneous sorption of two VOCs on vinyl flooring by exposing a vinyl flooring sample to a gas stream containing phenol at 28,000µg.m–3 and n-dodecane at 33,000µg.m–3. Although the tested concentrations were much higher than those typically found indoors, the sorption of one VOC was found to be unaffected by the simultaneous sorption of the other compound. On the other hand, Huang et al (2006) recently showed from a statistical analysis that the uptake of toluene by a ceiling tile as a single compound was higher than as part of a mixture. Finally, Jørgensen and Bjørseth (1999) concluded from dynamic adsorption tests that the sorption of α-pinene and toluene on wool and nylon carpets was enhanced when both compounds were present in the chamber. It seems that the nature of the VOC interactions in materials depends upon the type of material, the type of compound and probably the concentrations that are put into play. Moreover, while the results of the experiments have been analysed solely in terms of potential changes in the sorption properties of the gas and material systems investigated, and will be considered that way here, the observed discrepancies might actually result from complex reactivity phenomena at the material surfaces and should be modelled accordingly (see the section on ‘Modelling chemical transformations within porous materials’). Regarding polymeric materials, Schwope et al (1989) indicated that VOCs in pure polymers generally behave ideally – that is, follow Henry’s law – if the concentration of VOCs in the material phase is lower than 1 per cent by weight. Practically, this means that except in the case of heavily polluted environments or emissive materials containing large amounts of residual VOCs, linear models would still apply for multi-component absorption systems. Moreover, the absorption constants measured for single compounds would be unaffected by the presence of other gases in the air. In the case of porous media, a variety of theories and related multi-component adsorption isotherm models have been proposed. To date, their efficacy has only been investigated for industrial separation applications and for air cleaning applications in industrial applications (Wood, 2002). In general, multi-component adsorption isotherm models state that under isothermal conditions at atmospheric pressure, the equilibrium sorbed phase concentration of a given component i, Csi, will depend upon not only the air phase concentration of that component, but the air phase concentrations of all compounds in the mixture. Whatever the model, the computed sorbed-phase concentrations will always be lower than those predicted from the individual adsorption isotherms of the compounds, thus reflecting the competition between species for a limited number of active sites at the material surface. As a result, experiments showing increased adsorption capacities for contaminant mixtures would tend to suggest that phenomena other than only competition for sorption occur within the material (e.g. absorption in condensed gases and surface reactions). Still based on the premise that contaminant concentrations are usually low indoors, the extended Langmuir equation for component molecules that do not interact at the material surface is likely to be the simplest and most suitable multi-component adsorption

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 83

Modelling of Contaminant Interactions in Buildings

83

model for building applications. The latter relates the sorbed phase concentration of a given species i to the air phase concentration of all other compounds j in the mixture as: Csi =

Csi0 K Li Ci n

, [50]

1 + ∑ K Lj C j j =1

where the model parameters are defined for each component as outlined in Table 3.1. Consequently, one practical constraint of using the extended Langmuir equation is that both Cs0 and KL must be determined for all compounds when only Kp values are available in many situations. It is important to note here that for very low air-phase concentrations and/or a small number of compounds in the mixture,

n

∑ K Lj C j

will be much smaller than 1 and Equation

j =1

50 simplifies to a series of n uncoupled single-component linear isotherms. Therefore, in n

this case, multi-component competition can be ignored. On the other hand, if

∑ K Lj C j j =1

does not prove to be negligible compared to 1, Equation 50 will contribute to couple the equations describing the contaminant transports within porous materials. For instance, if considering a mixture of two gases in air under constant temperature, Equation 8 for the first component becomes: D p1∇ 2C1 = ε

∂C1 ρmat + ρ air ∂t

⎛ ∂Cs1 ∂C ∂Cs1 ∂C 2 ⎞⎟ ⎜⎜ 1 ⎟ ⎜⎜⎝ ∂C ∂t + ∂C ∂t ⎟⎟⎠ 1 2

[51]

Then, substituting the partial derivatives of Equation 50 into Equation 51 gives: ⎡ ρ D p1∇2C1 = ⎢⎢ ε + ρmat air ⎢⎣

⎛ ⎞⎟⎤ ∂C Cs01K L1 Cs01K L21C1 ⎜⎜ ⎟⎥ 1 − ⎜⎜ ⎟⎥ 2⎟ ⎝1 + K L1C1 + K L 2 C 2 (1 + K L1C1 + K L 2C 2 ) ⎟⎠⎥⎦ ∂t

[52]

Cs01K L1K L 2C1 ρ ∂C 2 − ρmat air (1 + K C + K C )2 ∂t L1 1 L2 2 The equation for the second component is obtained by simply interchanging the subscripts 1 and 2 in Equation 52. Another multi-component adsorption isotherm model that may be interesting for indoor air applications is the ideal adsorbed solution (IAS) theory of Myers and Prausnitz (1965). Although it has been developed for applications where porous adsorption occurs in the form of capillary condensation, which is unlikely for most building and pollutant systems, applying this theory with Langmuir equations as individual adsorption isotherm models yields the Langmuir extended equation. Besides, if considering linear adsorption isotherm models for individual compounds, the method will return the same models for the mixture of compounds, which reflects the fact that concentrations are low enough to ignore the competition for sorption. All of these results are consistent and tend to prove

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

84

3:31 PM

Page 84

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

that the IAS theory is probably more general than one usually thinks. Compared to the extended Langmuir equation, it has the advantage that different individual adsorption isotherm models can be used as inputs of the method; this may prove to be useful when dealing with compounds, such as water vapour, with isotherms that exhibit such strong non-linearities that they cannot be approximated by linear or even Langmuir models. On the other hand, the IAS is numerically much more difficult to implement than the extended Langmuir equation. To a lesser extent, the presence of contaminant mixtures in the air may also influence the diffusion of each individual species within materials. Although little is known on this topic, the fundamentals of diffusion presented in the section on ‘Diffusion coefficients’ point out the fact that the question of multi-component diffusion in porous materials cannot be treated independently from the question of the dominant diffusion mechanism: the effect of simultaneous diffusion of dilute compounds in the air can probably be ignored if only pore air-phase diffusion is to be considered. On the other hand, it may really become significant through the above-mentioned competition for adsorption if surface diffusion prevails. In such a case, the diffusion coefficients are expected to be lower than those for single species.

INFLUENCE OF HUMIDITY The experiments dealing with the influence of humidity on the sink (Jørgensen et al, 1999; Won et al, 2000, 2001; Zhang and Niu, 2004) and source (Haghighat and De Bellis, 1998; Wolkoff, 1998; Fang et al, 1999) behaviour of materials have led to different and sometimes divergent conclusions regarding the influence of humidity. A meaningful example is provided by Haghighat and De Bellis (1998): when measuring the VOC emissions from a water-based acrylic paint and a polyurethane plastic finish varnish, the authors came to the conclusion that, during the first stage of the emission process, the total amounts of VOC (TVOC) emitted from the paint were higher at 32 per cent relative humidity (RH) than at 62 per cent RH. On the other hand, the TVOC emissions from the varnish proved to increase with increasing humidity. Nevertheless, in this latter case, the authors also noticed that unlike the TVOC, individual compounds such as toluene, m,p-xylene and 1,3,5-trimethylbenzene showed higher emission rates at 32 per cent RH than at 62 per cent RH. Of all the studies quoted above, few have addressed the question of how humidity physically interacts with contaminants in materials. Recently, Radulescu-Bouilly et al (2006) presented a theory that may serve as a first step in understanding and modelling these complex interactions for porous materials. First, this theory is based on the principle that the transport of water vapour in building materials can be modelled analogously to the transport of any gas, except that a distinction is made between monolayer and multilayer adsorption, on the one hand, and capillary condensation, on the other hand. Next, three elemental interaction mechanisms are considered (see Figure 3.1). First is the competition for adsorption at the surfaces of the largest pores, the latter defined as those where capillary condensation does not occur; the authors suggest that it should be modelled using the extended Langmuir equation: Equation 50. The second and third interaction processes come from the possible capillary condensation of water vapour in

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 85

Modelling of Contaminant Interactions in Buildings

85

Absorption/desorption in condensed water

Competition for sorption at the pore surfaces

Capillary condensation of water in the thinnest pores Source: adapted from Radulescu-Bouilly et al (2006)

FIGURE 3.1 Elemental interaction phenomena between gaseous contaminants and water vapour in poly-dispersed porous materials

the thinnest pores of the material. As gases are soluble in liquids to a greater or lesser degree (Sander, 1999), the pores that are filled with liquid water will give rise to absorption/desorption mass transports inside the material, with the result of increasing its mass capacity. In addition, the pores that are filled with liquid water no longer contribute, or contribute differently, to the porous diffusion of gases in the material. Therefore, the diffusivity of VOCs may be decreased significantly when humidity increases. Using, first, published data for the pore-size distribution of some materials, on the one hand, and, second, the Laplace equation to determine the critical pore radius below which the pores will be filled with liquid water, and the resulting water content of the materials, and, third, Henry’s law to represent the absorption/desorption of gases in water, Radulescu-Bouilly et al (2006) show that increasing the air humidity can contribute to change from one dominant interaction phenomenon to another. This may explain the non-linearities in the observed sink or source behaviour at various humidity levels, as well as the differences in the influence of humidity depending upon the contaminant and material systems investigated. With regard to polymeric materials, Cox et al (2001b) suggested various possible mechanisms to explain the observed changes in the adsorption capacity of vinyl flooring in humid conditions. Some relate to direct interactions between sorbed VOC and water vapour within the material, while others suggest that sorbed water molecules change the

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

86

3:31 PM

Page 86

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

structural properties of the material and, subsequently, the sorption properties of the VOC and vinyl flooring system. In all cases, these processes nevertheless remain very difficult to represent analytically.

MODELLING CHEMICAL TRANSFORMATIONS WITHIN POROUS MATERIALS During the past 20 years, many studies have emphasized the potentially strong influence of homogeneous chemistry on the concentration levels of some hazardous contaminants (Nazaroff and Cass, 1986; Weschler and Shields, 1997; Blondeau et al, 1998, Sarwar et al, 2002). All have focused on chemical reactions occurring in the bulk air phase of the rooms. However, the kinetic coefficients reported for those reactions, which are most of the time directly transposed from atmospheric chemistry mechanisms, show that chemical reactions are likely to proceed much faster than diffusion in materials. Therefore, there is evidence that organic or non-organic species may also be transformed through homogeneous processes when transported in the pores of porous building materials. Similarly, surface chemical transformations (also called heterogeneous processes or chemisorption, in contrast to the reversible physical adsorption considered previously) are likely to be important for a large number of species, including nitrogen oxides and their radicals; oxygen; ozone and oxygen radicals; sulphur dioxide and sulphate ions; organic and inorganic acids; and a large number of volatile organic compounds (Axley, 1991). They may be considered even more important for porous materials where connected pores offer huge surface areas for surface reactions. In the field of indoor air analysis, surface chemistry has most often been modelled using the highly simplified deposition velocity approach (Nazaroff and Cass, 1989). The latter is based on the assumption that the removal rates of contaminants, r (kg.s–1), are related directly to the bulk air-phase species concentration, C: r = − ρairVd AsC ,

[53]

where Vd (m.s–1) is the deposition velocity of the compound and As the nominal surface area of the material (i.e. surface area exposed to the room air). From the computational point of view, the deposition velocity model has the great advantage that r can be included in the equations describing the contaminant mass conservation in the bulk air phase of the room: no equations and unknowns need to be added in the problem formulation to account for the contribution of surface reactions. Furthermore, the model is linear and thus does not call for advanced solving methods. Physically, the use of the deposition velocity model is much more questionable. First, this model does not account for boundary layer resistance: the removal rate should be defined as a function of the contaminant near-surface concentration, rather than its concentration in the bulk air. Practically, this means that the deposition velocity of a given compound and material system will also depend upon the characteristics of the airflow at the material surface, which is not fully satisfactory from a physical point of view. Another drawback of the model is that it can only handle unimolecular decomposition reactions although other

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 87

Modelling of Contaminant Interactions in Buildings

87

reactivity processes may exist (see below); moreover, it does not consider the yield of product species that may be important from an indoor air quality point of view. Finally, the deposition velocity model is not suitable for porous materials where, as mentioned above, the surface area available for surface chemistry is much higher than the material surface area that is exposed to the room air. Considering that homogeneous and heterogeneous reactions will occur in parallel in porous materials, and will subsequently be constrained by the diffusion and sorption processes that determine the amounts of reagents locally available, Axley (1991) suggested that the two phenomena may be modelled analogously by extending Equation 8 to: D p ∇2C +

1 ρair

⎛ ρ ∂f ⎞⎟ ∂C ⎟⎟ , r = ⎜⎜⎜ε + mat ⎜⎝ ρair ∂C ⎟⎠ ∂t

[54]

where r, here in units of kg.m–3.s–1, is the local removal or yield rate of the compound due to any kind of transformation within the material. To be more precise, r will actually be the sum of the contributions of each homogeneous or surface reaction j involving the compound as a reagent or as a product: r = ∑ rair , j + ∑ rsurf , j . j

j

[55]

Based on the knowledge and theories developed in the field of atmospheric chemistry, rair,j (kg.m–3.s–1) will be defined as (Seinfeld, 1986): rair , j = ± ερair xk j

nj

∏ Ck ,

k =1

[56]

where x is the stoichiometric coefficient of the compound in reaction j, kj (s–1) the kinetic coefficient of the reaction, nj the number of reagents involved in reaction j (i.e. 1, 2 or 3 for unimolecular, bimolecular or termolecular reactions, respectively) and Ck (kg/kgair) represents the local pore air-phase concentration of these reagents. Finally, rair,j is negative if the compound is involved as a reagent in the reaction, and positive if acting as a reaction product. In the fields of catalysis and surface science, surface reactions are frequently described as unimolecular decomposition or bimolecular combination reactions, although more complex interactions may exist (Masel, 1996). By analogy with the modelling of homogeneous reactions, the local rate of removal of reagents or yield of product species resulting from reaction j is in both cases given by the kinetic coefficient of the reaction, kj (in units of s–1 for a unimolecular reaction and s–1.kg–1 for a bimolecular reaction), multiplied by the stoichiometric coefficient of the compound and the mass(es) of reagent(s) adsorbed – that is: rsurf , j = ± ρmat xk j

nj

∏ Csk .

k =1

[57]

As for homogeneous reactions, rsurf,j will be negative if the compound is involved as a reagent in reaction j, and positive otherwise. For a bimolecular reaction involving species

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

88

3:31 PM

Page 88

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

A and B as reagents, substituting Equation 50 into Equation 57 then leads to the so-called Langmuir-Hinshelwood model: rsurf , j = ±

0 0 K LA C A CsB K LB CB ρmat xk j CsA

(1 + K LA C A + K LB CB )2

.

[58]

To date, the implementation of surface chemistry models is of limited interest since not one relevant surface reaction on building materials has been elucidated to the point that the mechanism is understood and the kinetic constant is known. Only Spicer et al (1986) suggested a catalytic conversion of nitrogen dioxide to nitric oxide on some surfaces and Ryan and Koutrakis (1993) indicated that the decomposition of ozone on latex paints may be governed by a bimolecular reaction with sorbed water. Experiments are needed in order to gain more knowledge and to support research in this area. Then, considering that most reagent species will follow a Langmuir adsorption isotherm model in porous building materials, and that Langmuir-Hinshelwood type rate equations often work under conditions where the adsorption of the reagents does not follow a Langmuir adsorption isotherm (Masel, 1996), Equation 58 definitely comes in as a good candidate for numerical developments in this field. In the end, it is important to note that from a numerical point of view, the implementation of homogeneous and/or heterogeneous chemistry models in the equations describing the contaminant transport in building materials will lead to systems of non-linear and coupled equations. Moreover, not only the contaminant of interest for indoor air-quality analysis, but all the species involved as reagents in the transformations, must be considered in the model. This may result in large systems to be solved and, possibly, prohibitive calculation times.

CONCLUSION Although potentially having a great influence on predicted contaminant concentrations, modelling homogeneous and heterogeneous processes that occur indoors still remains the drawback of most indoor air-quality simulation tools. The review of existing gas/material interaction models presented here shows that research in this field has suffered confusion regarding which physical phenomena to consider and the way in which to model data. Frequently, misinterpretation occurs between sorption dynamics and sorption equilibrium. From the knowledge developed in the related fields of surface science and catalysis, it seems obvious that the former results from the dynamics of diffusion within the materials. Therefore, elemental models that couple local equilibrium sorption models with diffusion equations are most probably a more realistic fit than what has been referenced as macroscopic models. Another confusing problem comes from the nature of the sorption phenomena involved depending upon the materials considered. In most studies, the contaminant transfers from the bulk air to the material (and vice versa) are referred to as adsorption/ desorption, and are modelled accordingly, when they actually consist of absorption/ desorption processes in the case of non-porous polymeric materials. Practically, this is of no consequence since Henry’s law for ideal absorption in solids or liquids is analogous to a linear adsorption isotherm model. However, from a general viewpoint, it underlines the

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 89

Modelling of Contaminant Interactions in Buildings

89

central importance of the porous features of building materials for modelling purposes: knowing whether the material is porous or not determines the kind of mass-transport model to be used. Furthermore, having information on the pore-size distribution of porous materials may help one to assess the contaminant diffusivities in the air phase of the pores and, subsequently, to determine whether surface diffusion is to be considered or not in the model. Without going into philosophical considerations, physically based models somewhat imply universality or, at least, generality. As a result, this chapter placed particular emphasis on demonstrating that, depending upon the initial conditions and the parameters set, elemental models can either represent the emission or the sink effect of any material, predict the transition from one phenomenon to another, and possibly account for the decomposition of reactive species (what is referred to as deposition) by defining surface reactions. To date, very little is known about the latter point. Developing knowledge on this topic and, especially, combination reactions (e.g. oxidation, reduction or hydrolysis of contaminants at the material surfaces) comprise a real challenge for the next years. Implementing multi-component facilities and coupling the heat and mass transfer equations are undoubtedly other steps to overcome before accurately modelling the variety and great complexity of all indoor gas and material interactions.

AUTHOR CONTACT DETAILS P. Blondeau: (corresponding author) LEPTAB, University of La Rochelle, France; [email protected] A. L. Tiffonnet: LUSAC, University of Caen, France F. Allard: LEPTAB, University of La Rochelle, France F. Haghighat: Centre for Building Studies, Concordia University, Montreal, Canada

REFERENCES An, Y., J. S. Zhang and C. Y Shaw (1999) ‘Measurements of VOC adsorption/desorption characteristics of typical interior building materials’, HVAC Research, vol 5, pp297–316 Ataka, Y., S. Kato, S. Murakami, Q. Zhu, K. Ito and T. Yokota (2004) ‘Study of effect of adsorptive building material on formaldehyde concentrations: Development of measuring methods and modeling of adsorption phenomena’, Indoor Air, vol 14, pp51–64 Axley, J. W. (1991) ‘Adsorption modelling for building contaminant dispersal analysis’, Indoor Air, vol 2, pp147–171 Axley, J. W. (1993) ‘Modeling sorption transport in rooms and sorption filtration systems for building air quality analysis’, Indoor Air, vol 3, pp298–309 Bejan, A. (1995) Convection Heat Transfer, 2nd edition, Wiley and Sons, New York Berens, A. R. and H. B. Hopfenberg (1982) ‘Diffusion of organic vapors at low concentrations in glassy PVC, polystyrene and PMMA’, Journal of Membrane Science, vol 10, pp283–303 Blondeau, P., F. Allard, M. Spérandio and F. Haghighat (1998) ‘Detailed modeling of gas-phase chemistry mechanisms in IAQ simulation’, ASHRAE Transactions, vol 104, no 2, pp1309–1317 Blondeau, P., A. L. Tiffonnet, A. Damian, O. Amiri and J. L. Molina (2003) ‘Assessment of contaminant diffusivities in building materials from porosimetry tests’, Indoor Air, vol 13, pp310–318

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

90

3:31 PM

Page 90

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

Bodalal, A., J. S. Zhang and E. G. Plett (2000) ‘A method for measuring internal diffusion and equilibrium partition coefficients of volatile organic compounds for building materials’, Building and Environment, vol 35, pp101–110 Bouhamra, W. S. and A. S. Elkilani (1999) ‘Investigation and modeling of surface sorption/desorption behaviour of volatile organic compounds for indoor air quality analysis’, Environmental Science and Technology, vol 20, pp531–545 Chang, C. S. and Z. Guo (1992) ‘Characterization of organic emissions from a wood finishing product – wood stain’, Indoor Air, vol 2, pp146–153 Colombo, A. and M. De Bortoli (1992) ‘Comparison of models used to estimate parameters of organic emissions from materials tested in small environmental chambers’, Indoor Air, vol 2, pp49–57 Colombo, A., M. De Bortoli, H. Knöppel, E. Pecchio and H. Vissers (1993) ‘Adsorption of selected volatile organic compounds on a carpet, a wall coating, and a gypsum board in a test chamber’, Indoor Air, vol 3, pp276–282 Cox, S. S., A. T. Hodgson and J. C. Little (2001a) ‘Measuring concentrations of volatile organic compounds in vinyl flooring’, Journal of the Air and Waste Management Association, vol 51, no 8, pp174–185 Cox, S. S., D. Zhao and J. C. Little (2001b) ‘Measuring partition and diffusion coefficients for volatile organic compounds in vinyl flooring’, Atmospheric Environment, vol 35, pp3823–3830 Damian, A., P. Blondeau and F. Allard (2003) ‘Distribution of room air contaminant concentrations as a function of ventilation and air cooling’, Proceedings of the 7th International Conference Healthy Building 2003, Singapore, vol 2, pp410–415 Do, D. D. (1998) Adsorption Analysis: Equilibria and Kinetics, Imperial College Press, London Dunn, J. E. (1987) ‘Models and statistical methods for gaseous emission testing of finite source in well-mixed chambers’, Atmospheric Environment, vol 21, pp425–430 Dunn, J. E. and B. Tichenor (1988) ‘Compensating for sink effects in emissions test chambers by mathematical modeling’, Atmospheric Environment, vol 22, pp885–894 Elkilani, A. S., C. G. J. Baker, Q. H. Al-Shammari and W. S. Bouhamra (2003) ‘Sorption of volatile organic compounds on typical carpet fibers’, Environment International, vol 29, pp575–585 EPA (US Environmental Protection Agency) (2006) EPA On-line Tools for Site Assessment Calculation – Diffusion Coefficient Estimation, www.epa.gov/athens/learn2model/part-two/onsite/estdiffusion.htm Fang L., G. Clausen and O. Fanger (1999) ‘Impact of temperature and humidity on chemical and sensory emission from buildings materials’, Indoor Air, vol 9, pp193–301 Haghighat, F. and L. De Bellis (1998) ‘Material emission rates: Literature review, and the impact of indoor air temperature and relative humidity’, Building and Environment, vol 33, pp261–277 Haghighat, F. and Y. Zhang (1999) ‘Modelling of emission of volatile organic compounds from building materials – estimation of gas-phase mass transfer coefficient’, Building and Environment, vol 34, pp377–389 Haghighat, F., C.-S. Lee and W. S. Ghaly (2002) ‘Measurement of diffusion coefficients of VOCs for building materials: Review and development of a calculation procedure’, Indoor Air, vol 12, pp81–91 Hansson, P. and H. Stymne (2000) ‘VOC diffusion and absorption properties of indoor materials – consequences for indoor air quality’, Proceedings of Healthy Building 2000, Espoo, Finland, vol 4, pp151–156 He, G., X. Yang and C. Y. Shaw (2005) ‘Material emission parameters obtained through regression’, Indoor and Built Environment, vol 14, pp59–68 Huang, H., F. Haghighat and P. Blondeau (2006) ‘Volatile organic compound (VOC) adsorption on material: influence of gas phase concentration, relative humidity and VOC type’, Indoor Air, vol 16, pp236–247 Jørgensen, R. B. and O. Bjørseth (1999) ‘Sorption behaviour of volatile organic compounds on material surfaces – the influence of combinations of compounds and materials compared to sorption of single compounds on single materials’, Environment International, vol 25, pp17–27 Jørgensen, R. B., O. Bjørseth and B. Malvik (1999) ‘Chamber testing of adsorption of volatile organic compounds (VOCs) on material surfaces’, Indoor Air, vol 9, pp2–9

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 91

Modelling of Contaminant Interactions in Buildings

91

Jørgensen, R. B., T. H. Dokka and O. Bjørseth (2000) ‘Introduction of a sink-diffusion model to describe the interaction between volatile organic compounds (VOCs) and material surfaces’, Indoor Air, vol 10, pp27–38 Kephalopoulos, S., H. Knoeppel and M. De Bortoli (1996) ‘Testing of sorption and desorption of selected VOCs on/from indoor materials’, Proceedings of the 7th International Conference on Indoor Air Quality and Climate, Indoor Air 1996, Nagoya, pp61–66 Kirchner, S., J. R. Badey, H. N. Knudsen, R. Meininghaus, D. Quenard, K. Saarela, H. Sallee and A. Saarinen (1999) ‘Sorption capacities and diffusion coefficients of indoor surface materials exposed to VOCs: Proposal of new test procedures’, in Proceedings of the 8th International Conference on Indoor Air Quality and Climate, Indoor Air 1999, Edinburgh, vol 1, pp430–435 Kumar, H. and J. C. Little (2003) ‘Characterizing the source/sink behavior of double-layer building materials’, Atmospheric Environment, vol 37, pp5529–5537 Kumar, H. and Siddaramaiah (2005) ‘A study of sorption/desorption and diffusion of substituted aromatic probe molecules into semi interpenetrating polymer network of polyurethane/polymethyl methacrylate’, Polymer, vol 46, pp7140–7155 Lee, C. S., F. Haghighat and W. S. Ghaly (2005) ‘A study on VOC source and sink behaviour in porous building materials – analytical model development and assessment’, Indoor Air, vol 15, pp183–196 Li, F. and J. Niu (2005) ‘Simultaneous estimation of VOCs diffusion and partition coefficients in building materials via inverse analysis’, Building and Environment, vol 40, pp1366–1374 Li, F. and J. Niu (2007) ‘Control of volatile organic compounds indoors – development of an integrated mass-transfer-based model and its application’, Atmospheric Environment, vol 41, pp2344–2354 Little, J. C. and A. T. Hodgson (1996) A Strategy for Characterizing Homogeneous, Diffusion-Controlled, Indoor Sources and Sinks, Standard Technical Publication 1287, American Society for Testing and Materials, pp294–304 Little, J. C., A. T. Hodgson and A. J. Gadgil (1994) ‘Modeling emissions of volatile organic compounds from new carpets’, Atmospheric Environment, vol 28, pp227–234 Masel, R. I. (1996) Principles of Adsorption and Reaction on Solid Surfaces, Wiley and Sons, New York Matthews, T. G., A. R. Hawthorne and C. V. Thompson (1987) ‘Formaldehyde sorption and desorption characteristics of gypsum wallboard’, Environmental Science and Technology, vol 21, pp629–634 Meininghaus, R. and E. Uhde (2002) ‘Diffusion studies of VOC mixtures in a building material’, Indoor Air, vol 12, pp215–222 Meininghaus, R., L. Gunnarsen and H. N. Knudsen (2000) ‘Diffusion and sorption of volatile organic compounds in building materials – impact on indoor air quality’, Environmental Science and Technology, vol 34, pp3101–3108 Morrison, G. C., Z. Ping, D. J. Wiseman, M. Ongwandee, H. Chang, J. Portman and S. Regmi (2003) ‘Rapid measurement of indoor mass-transfer coefficients’, Atmospheric Environment, vol 37, pp5611–5619 Mulder, M. (1991) Basic Principles of Membrane Technology, Kluwer Academic Publishers, Dordrecht, The Netherlands Murakami, S., S. Kato, K. Ito and Q. Zhu (2003) ‘Modeling and CFD prediction for diffusion and adsorption within room with various adsorption isotherms’, Indoor Air, vol 13, pp20–27 Myers, A. L. and J. M. Prausnitz (1965) ‘Thermodynamics of mixed-gas adsorption’, American Institute of Chemical Engineers Journal, vol 11, pp121–127 Nazaroff, W. W. and G. R. Cass (1986) ‘Mathematical modeling of chemically reactive pollutants in indoor air’, Environmental Science and Technology, vol 20, pp924–934 Nazaroff, W. W. and G. R. Cass (1989) ‘Mass-transport aspects of pollutant removal at indoor surfaces’, Environment International, vol 15, pp567–584 Park, I., D. D. Do and A. E. Rodrigues (1996) ‘Measurement of the effective diffusivity in porous media by the diffusion cell method’, Catalysis Reviews: Science and Engineering, vol 38, pp189–247

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

92

3:31 PM

Page 92

P. BLONDEAU, A. L. TIFFONNET, F. ALLARD AND F. HAGHIGHAT

Radulescu-Bouilly, C., F. Allard, P. Blondeau, B. Collignan, R. Popescu and A. Sjöberg (2006) ‘A physically-based analysis of the interactions between humidity and VOCs in building materials’, Proceedings of the Healthy Building 2006 International Conference, Lisboa, vol IV, pp89–94 Rutherford, S. W., R. E. Kurtz, M. G. Smith, K. G. Honnell and J. E. Coons (2005) ‘Measurement and correlation of sorption and transport properties in polystyrene – POSS membranes’, Journal of Membrane Science, vol 263, pp57–65 Ruthven, D. M. (1984) Principles of Adsorption and Adsorption Processes, Wiley and Sons, New York Ryan, P. B. and P. Koutrakis (1993) ‘The reactive chemistry of ozone in indoor environments’, CIAR Currents: The Newsletter of the Center for Indoor Air Research, Linthicum, MD, pp1–3 Sander, R. (1999) Compilation of Henry’s Law Constants for Inorganic and Organic Species of Potential Importance in Environmental Chemistry (Version 3), www.henrys-law.org Sarwar, G., R. Corsi, Y. Kimura, D. Allen and C. J. Weschler (2002) ‘Hydroxyl radicals in indoor environments’, Atmospheric Environment, vol 36, pp3973–3988 Schwope, A. D., W. J. Lyman and R. C. Reid (1989) Methods for Assessing Exposure to Chemical Substances, US EPA Report (EPA 560/5–85–015), vol 11 Seo, J., S. Kato, Y. Ataka and Q. Zhu (2005) ‘Evaluation of effective diffusion coefficient in various building materials and absorbents by mercury intrusion porosimetry’, Proceedings of the 10th International Conference on Indoor Air Quality and Climate, Indoor Air 2005, Beijing Seifert, B., D. Ullrich and R. Nagel (1989) ‘Volatile organic compounds from carpeting’, in Proceedings of the 8th World Clean Air Congress, The Hague, vol 1, pp253–258 Seinfeld, J. H. (1986) Atmospheric Chemistry and Physics of Air Pollution, Wiley and Sons, New York Silberstein, S. (1989) A Gypsum Wallboard Formaldehyde Sorption Model, National Institute for Standards and Technology (NIST), Document no NISTDIR 89-4028, Gaithersbourg, US Singer, B. C., K. L. Revzan, T. Hotchi, A. T. Hodgson and N. J. Brown (2004) ‘Sorption of organic gases in a furnished room’, Atmospheric Environment, vol 38, pp2483–2494 Spicer, C. W., R. W. Coutant and G. F. Ward (1986) Investigation of Nitrogen Dioxide Removal from Indoor Air, Gas Research Institute, 1986, Document no GRI–86/0303, Chicago, US Sun, Y. M., C. L. Lin, Y. K. Chan, and C. H. Wu (1997) ‘Sorption and diffusion of organic vapors in poly[bis(trifluoroethoxy) phosphazene] and poly[bis(phenoxy)phosphazene] membranes’, Journal of Membranes Sciences, vol 134, pp117–126 Tichenor, B. A., Z. Guo, J. E. Dunn, L. E. Sparks and M. A. Mason (1991) ‘The interaction of vapour phase organic compounds with indoor sinks’, Indoor Air, vol 1, pp23–35 Tichenor, B. A., Z. Guo and L. E. Sparks (1993) ‘Fundamental mass transfer model for indoor air emissions from surface coatings’, Indoor Air, vol 3, pp263–268 Tichenor, B. A. (1996) ‘Overview of source/sink characterization methods’, ASTM STP, vol 1287, pp9–19 Tiffonnet, A. L., P. Blondeau, F. Allard and F. Haghighat (2002) ‘Sorption isotherms of acetone on various building materials’, Indoor and Built Environment, vol 11, pp95–104 Topp, C., P. V. Nielsen and P. Heiselberg (2001) ‘Influence of local airflow on the pollutant emission from indoor building surfaces’, Indoor Air, vol 11, pp162–170 Treybal, R. E. (1981) Mass Transfer Operations, McGraw Hill International, Boston, US Van Der Val, J. F., A. W. Hoogeveen and P. Wouda (1997) ‘The influence of temperature on the emission of volatile organic compounds from PVC flooring, carpet and paint’, Indoor Air, vol 7, pp215–221 Weschler, C. J. (2002) ‘Connections: particles, sensory offending filters, the “sink effect” and nasal pungency thresholds’, Proceedings of Indoor Air 2002, Monterey, US

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 93

Modelling of Contaminant Interactions in Buildings

93

Weschler, C. J. and H. C. Shields (1997) ‘Potential reactions among indoor pollutants’, Atmospheric Environment, vol 31, pp3487–3495 White, F. M. (1991) Heat and Mass Transfer, Addison-Wesley, New York Wilkins, C. K., P. Wolkoff, P. A. Clausen, M. Hammer, S. T. Larsen and G. D. Nielsen (2006) ‘Effects of repeated exposure to limonene oxidation products in a mouse bioassay’, Proceedings of the Healthy Building 2006 International Conference, Lisboa, Portugal Wolkoff, P. (1998) ‘Impact of air velocity, temperature, humidity and air on long-term VOC emission from building products’, Atmospheric Environment, vol 32, pp2659–2668 Won, D., R. L. Corsi and M. Rynes (2000) ‘New indoor carpet as an adsorptive reservoir for volatile organic compounds’, Environmental Science and Technology, vol 34, pp4193–4198 Won, D., R. Corsi and M. Rynes (2001) ‘Sorptive interactions between VOCs and indoor materials’, Indoor Air, vol 11, pp246–256 Wood, G. O. (2002) ‘Review and comparisons of D/R models of equilibrium adsorption of binary mixtures of organic vapors on activated carbons’, Carbon, vol 40, pp231–239 Yang, X. and Q. Chen (2001) ‘A coupled airflow and source/sink model for simulating indoor VOC exposures’, Indoor Air, vol 11, pp257–269 Yang, X., Q. Chen, J. S. Zhang, Y. An, J. Zeng and C. Y. Shaw (2001) ‘A mass transfer model for simulating VOC sorption on building materials’, Atmospheric Environment, vol 35, pp1291–1299 Zhang, L. Z. and J. L. Niu (2004) ‘Modeling VOCs emissions in a room with a single-zone multi-component multi-layer technique’, Building and Environment, vol 39, pp523–531 Zhang, Y. and Y. Xu (2003) ‘Characteristics and correlations of VOC emissions from building materials’, International Journal of Heat and Mass Transfer, vol 46, pp4877–4883

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 57–93

ABER_CH_03

6/12/08

3:31 PM

Page 94

ABER_CH_04

6/12/08

3:32 PM

Page 95

publishing for a sustainable future

4

The Application of Urban Climate Research in the Design of Cities Evyatar Erell

Abstract In spite of the growing body of research on urban climatology and the increasing demand for architects and urban planners to practise climate-conscious design, there is too little evidence of the application of urban climatology in practice. This chapter explores the relationship between climatologists and urban planners, seeking to establish some of the reasons for this state of affairs. It then sets out a methodological framework for the application of climatology in the planning process, outlining possible goals for such intervention, as well as its limitations. The chapter then attempts to establish the effects on the urban microclimate of a broad range of decisions taken routinely by architects and planners, based on an extensive survey of applied research in the field.

■ Keywords – urban microclimate; urban planning; architectural design

INTRODUCTION The quality of life of millions of people living in cities can be improved if the factors that affect the urban microclimate are understood and the form of the city responds to them in a manner that is appropriate to its environment. Underlying this approach is the idea that climatically responsive urban design is vital to any future notion of sustainability: it enables individual buildings to make use of ‘natural’ energy, it enhances pedestrian comfort and activity in outdoor spaces, and it may encourage city dwellers to moderate their dependence upon air conditioning in buildings and upon private vehicles. It has been known for well over a century that cities generate their own ‘climate’ (Howard, 1833). However, urban climate research has evolved as a specialist discipline within meteorology and climatology only in the past 50 years or so. During this period, the focus of research has shifted from descriptive studies of the properties of the urban wind field, temperature and humidity, to an experimental approach designed to investigate the physical processes responsible for the unique meteorological conditions found in cities. Breaking down complex urban forms into basic components, notably the urban canyon, has allowed researchers to isolate the effects of individual factors. Comprehensive

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121 doi:10.3763/aber.2008.0204 ■ © 2008 Earthscan ■ ISSN 1751-2549 (Print), 1756-2201 (Online) ■ www.earthscanjournals.com

ABER_CH_04

6/12/08

96

3:32 PM

Page 96

EVYATAR ERELL

reviews of progress in research on urban climatology, in general, may be found in Arnfield (2003) and Grimmond (2006). Architecture (the design of buildings) and urban planning (the design of the urban context within which individual buildings are constructed) have traditionally responded to climatic conditions, as, indeed, they have to other aspects of the location of towns. The importance of environmental factors in the design of settlements was recognized at least as early as Roman times, as is recorded in the writings of Vitruvius (1999), who warned planners, for example, that ‘if the streets run straight in the direction of the winds then their constant blasts rush in and sweep the streets with great violence’. More recently, there have been numerous manuals describing modern methods of bio-climatic architecture, many of which owe much to the pioneering work of Olgyay (1963). Two modern inventions have diverted the attention of architects and allowed them, for a time, to ignore the effects of climate: first, the automobile, which replaced human beings as the focus of attention in the design of outdoor space; and, second, advanced space heating and cooling systems, which provided comfortable and stable interiors regardless of the vagaries of climate, and allowed architects to avoid some of the consequences of designing buildings that showed no regard to local conditions. Climatic considerations have nonetheless had a major effect on the plans of a number of new neighbourhoods or towns during recent years (Gotz, 1982; Etzion, 1990; Evans and de Schiller, 1990–1991; El-Shakhs, 1994). In all of these cases, planning decisions were made primarily on the basis of an intuitive understanding of local conditions, and less on the basis of scientific analysis of the meteorological conditions and likely urban effects. The only variable treated in a quantitative manner in some modern neighbourhood plans is solar access rights, which are based on geometric rules derived from apparent solar position. Unfortunately, most publications on the application of urban climatology to the planning process fall into one of two categories: they are either cases studies of conditions in existing settlements (Potchter, 1988); or they provide only general recommendations, but not detailed design tools (Landsberg, 1968; Golany, 1996). Borve (1982), Westerberg and Glaumann (1990/1991) and Pressman (1996) offered general guidelines for so-called ‘winter cities’; Aynsley and Gulson (1999) proposed design strategies for humid tropical cities, in general; while Emmanuel (1995) suggested strategies specific to Colombo, Sri Lanka. Herz (1988) drew up recommendations for the Sahel region. The awareness of the importance of climatological inputs in the process of urban planning is growing in recent years. However, even where environmental concerns are the subject of much public debate and where urban planners are interested in climatic aspects of design, the use of climatic information is not systematic and climatology has little impact on the planning process (Eliasson, 2000). The reasons for this are not only conceptual and knowledge based, but are also related to technical matters, policy issues, organizational aspects and the market.

ON THE DIALOGUE BETWEEN URBAN PLANNERS AND URBAN CLIMATOLOGISTS Mills (1999) has suggested that the difficulty in applying research in urban climatology to architectural design problems may be explained by the fact that ‘despite the common

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 97

The Application of Urban Climate Research in the Design of Cities

97

interest in the urban climate, these fields pursue different research interests, employ contrasting methodologies and present results differently’. Each of these contrasts deserves further elaboration. Urban climatology now encompasses issues as diverse as the effect of street design on pollution dispersion and the effect of cities on the hydrological balance; but a large proportion of the studies has been devoted to the phenomenon of the urban heat island. Although it has been demonstrated that the urban heat island occurs mostly on clear nights with no wind – conditions that are not necessarily frequent in many cities – it has been the focus of considerable research since it was proposed that its maximum intensity was related to the population size of cities (Oke, 1973). The success of models devoted to the urban street canyon (Nunez and Oke, 1977; Oke, 1981; Nakamura and Oke, 1988) has led to its widespread use as the preferred unit of study – irrespective of the fact that most urban spaces do not conform to this simplified generic form. Architectural research, meanwhile, has tended to focus mostly on daytime phenomena and on issues of human thermal comfort. Research by architects on urban climate is typically concerned with observing the climatic behaviour of urban spaces, with the underlying assumption that ‘successful’ examples may then be examined to elucidate the fundamental physical characteristics that are most responsible for creating the desirable conditions. It often relies on studies of vernacular architecture, seeking to apply the distilled experience of previous generations to modern-day situations (see, for example, Krishan, 1996). The emphasis on physical form and material is in marked contrast to the interest of climatologists in studying the processes and the fundamental principles that drive them, which may require isolation of a process and its presentation in abstract terms. Meteorological models necessarily involve simplification of the real world, and applying the insights gained by such methods to planning in the complex reality of a city may therefore be quite difficult. Climatic research by architects generally has a strong focus on practical application. As Mills (1999) noted, the results of architectural research are often formulated in terms of guidelines or methodologies for other designers. These are frequently demonstrated with examples from the ‘real world’, which are presented as proof that the underlying principles discovered through the research may be applied in practice.

A PARADIGM FOR CLIMATE-CONSCIOUS URBAN PLANNING DESIGN GOALS Oke (1988a) with mid-latitude cities in mind, proposed the following objectives for urban planners considering an appropriate response to climate: ● ● ● ●

maximize maximize maximize maximize

shelter; dispersal of pollutants; urban warmth; solar access.

Considering the aspect ratio of streets as the only parameter to be modified in order to achieve these objectives, Oke (1988a) suggested that a height-to-width ratio (H/W) of 0.4

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

98

3:32 PM

Page 98

EVYATAR ERELL

was a compromise that would lead to satisfactory performance with respect to all of them. This implies a very low density compared to what is found in many existing urban centres. However, several studies (Manins et al, 1998; Mills, 2006) have shown that a compact city performs best on a number of measures, especially reduction of energy use in transport – implying a much higher density is desirable. Oke’s (1998a) objectives were framed in a very narrow and specific framework. Page (1972) discussed the effects of microclimate climate on a very broad range of issues encompassed in the field of urban planning and design. These include optimization of land-use patterns in relation to different activities to be carried out in the town; identification and development of suitable microclimates for various activities, such as parks or recreation; identification of adverse microclimatic factors likely to affect the detailed design of urban systems, such as high local winds; optimization of building form in relation to external climatic inputs, such as solar radiation and wind; optimization of building form in relation to microclimatic modification of the immediate exterior domain of the building, such as the high winds induced near ground level by tall buildings; constructional safety, especially with respect to high winds; selection of appropriate building materials; planning of the construction process itself in view of climatic constraints; control of water runoff; assessment of building running costs (heating, ventilation and air conditioning (HVAC), lighting, etc.) in advance of construction; optimization of the operating environment of transport systems (e.g. avoidance of ice hazards); and control of the environmental impact of a transport system on its adjacent urban systems (e.g. with respect to air pollution by vehicles). As Page (1968) demonstrated, urban microclimate may affect our lives in diverse ways. To summarize this section, it may be useful to organize these effects into two main categories: 1 The effect of microclimate on human activity, especially pedestrian, in the spaces between buildings. The urban fabric consists of buildings and open space, which may, in turn, be classified according to intended patterns of use. Where pedestrian access is considered valuable, design of the outdoor spaces intended for humans should provide optimal conditions, as appropriate in the local climatic conditions. 2 The effect of microclimate on the performance of buildings, especially with respect to energy conservation. The magnitude of modifications to the microclimate resulting from the effects of the urban fabric has drawn the attention of researchers to the need for tools to predict them and to devise design strategies to respond to them (Taha, 1978; El-Sioufi, 1987; El Nahas, 1996). The impetus for some of the research has been the proliferation of computer software for building thermal analysis, which relies on meteorological data to predict interior conditions. While many building simulation codes are now considered to be quite accurate, significant errors may be introduced to the simulations as a result of weather data, which is based on regional averages but which may not be representative of site-specific conditions (Even and Williamson, 2006).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 99

The Application of Urban Climate Research in the Design of Cities

99

SUBSIDIARITY Architects and planners must deal with a multitude of factors. Often, the demands of different consultant experts introduce conflicting requirements upon the architect, so the design of urban space frequently involves a process of optimization. It is thus of great value to be able to establish the benefits of a particular approach in general terms without resorting to a unique policy required to achieve the desired goal. Furthermore, if there is more than one solution that may yield the required result, the preferred solution is one that may be applied as late as possible in the planning process, and which thus has the least impact upon other aspects of the design. This approach, of seeking the solution for a particular issue at the lowest possible level of the planning process, may be termed ‘subsidiarity’. The following example may serve to illustrate this principle. Solar access, especially with respect to direct radiation, has generated much research interest (Gupta, 1984; Littlefair, 1998; Bourbia and Awbi, 2004; Robinson and Stone, 2004; Bozonnet et al, 2005; Ali-Toudert and Mayer, 2006). The reasons for the profusion of studies into this particular aspect of the urban microclimate are self-evident: solar radiation is the driving force of all climatic systems; it may be studied through the application of simple rules of geometry; and its recommendations are likewise formulated in terms of geometrical restrictions on building volumes or the proportions of streets. Gupta’s (1984) work is instructive: using the rather basic computing tools of the time, he produced an excellent analysis of the effects of solar radiation on urban geometry in hot climates. Yet, while acknowledging the fact that external shading devices such as pergolas, awnings, etc. may be used to limit the solar exposure of building openings – thus undermining much of the rationale for the analysis – the study then proceeds to study the effects of structural shading (i.e. the shadows cast by the basic form of the building) as the primary criterion for assessing the relative merits of the different options. These include the choice of building form, building height, street orientation and street height-to-width ratios – all of which could be decided upon prior to the study of window shading.

COMPLEXITY The previous example illustrates another common shortcoming of some attempts to apply the scientific method to urban planning. In order to analyse a particular question, researchers often simplify the issue, studying its effects in isolation from other factors that may be involved. This has clear advantages as far as analysis of the physical processes is concerned; yet, great care must be used in the synthesis of research results into an overall planning strategy that may be applied to a specific urban location. Thus, deriving an optimal urban form on the basis of exposure to solar radiation risks overlooking the effects of other factors, such as energy emitted by long-wave radiation or heat exchange by convection. The formulation of design recommendations on the basis of such research must also take care to very carefully define the goals to be achieved: design for pedestrian comfort and for building energy savings may lead to contradictory requirements. The role of climatologists in a real-life design team also includes responsibility for analysing climatic conditions in the urban area in question in order to identify the critical

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

100

3:32 PM

Page 100

EVYATAR ERELL

issues to be addressed by the proposed plan. Some cities may have only one clearly defined period in which meteorological conditions introduce significant stress that deserves the planners’ attention. Yet, many cities, such as New York, for instance, experience cold winters and hot-humid summers, both conditions requiring careful analysis with respect to possible planning intervention. Givoni (1989) presents such an analysis for a number of different climates. The report deals with the definitions of human thermal comfort, provides a general description of the relevant characteristics of the urban climate, and discusses the effects of various planning features, housing types and vegetation on the urban microclimate. Its broad scope is an indication of the complex nature of the issue and the multifaceted response required in order to be of value in the context of city planning.

ECONOMIC VIABILITY The recommendations of urban climatologists with respect to city planning may often have significant economic implications. Urban development is driven to a great degree by economic considerations, and zoning regulations often reflect the desire of city planners to attract investment by real-estate developers. Street width, for instance, is generally determined by the requirements of vehicular access, while building height reflects the desire to maximize the value of land. Thus, any recommendation concerning the heightto-width ratio of streets, which has a major influence on the canopy layer climate of cities, may also have considerable economic implications. Any explanation for the relative lack of success in implementing climate-related strategies in urban planning must therefore consider the lack of a practical framework to assess their economic effects too.

CLEAR AND IMMEDIATE BENEFITS In order for a particular recommendation based on urban climate considerations to be adopted by planners, the proposed strategy should have clear and immediate benefits. For example, the effect of urban climate modification on building energy consumption could be estimated, taking into account the urban heat island, as has been demonstrated from measured temperature data for Athens (Santamouris et al, 2001) and London (Kolokotroni et al, 2006). To realize this aim, urban climatology is required to develop sufficiently accurate and reliable predictive tools. In the absence of quantitative studies on the effect of proposed designs upon climate, and on the basis of well-documented evidence from other planning professions, decision-makers, in general, tend to downgrade the importance of climatic considerations in urban planning.

COMPREHENSIVE APPROACH TO PROBLEM-SOLVING In order to apply urban climatology effectively in the process of town planning, a comprehensive approach must be adopted. Recommendations must be based on analysis of all factors influencing the urban microclimate. Conclusions based upon the study of one or even several factors in conjunction may be misleading if they fail to take into account the effect of other significant processes. With the increase in computing power, computerized modelling may be capable of providing accurate and comprehensive analysis of the urban microclimate. Once such models become reliable and accurate

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 101

The Application of Urban Climate Research in the Design of Cities

101

enough, they should be applied wherever possible to inform the decision-makers of the microclimatic implications of urban planning strategies under consideration. For the models to be useful, they must allow the study of the particular issues that are foremost in the architect’s mind. In other words, they must be formulated so that the inputs include parameters related directly to the architect’s decision-making process. The following section illustrates the effect of planning decisions on the resulting microclimate.

THE EFFECT OF DESIGN DECISIONS ON THE URBAN MICROCLIMATE Modern urban planning is a complex process, generally carried out by teams of professionals from a number of fields. While the architect or town planner may have overall responsibility for producing the plan and for coordinating the project, crucial inputs are provided by consultants from other disciplines. The process is generally driven primarily by economic forces in response to market demand for housing, retail space, etc. The input of the urban climatologist is therefore but one of many that compete not only for the architect’s attention, but also for appropriate consideration in the process of optimization that takes place in urban planning. Mills (2006) noted that ‘while the meteorologically ideal settlement serves a useful pedagogical purpose, it does not recognize planning realities where climate issues are rarely a dominant concern’. So, while a statement of desirable outcomes and the means of achieving them may be a logical means of applying urban climatology, climatologists seeking to inform the decision-making process are rarely in the position of generating a plan. Rather, they may be required to respond to proposals made by other members of the design team. In order to do so effectively, urban climatologists must be aware of the effect that climatic considerations may have on the issues that the planners must resolve when preparing a town plan. This section is therefore structured to deal with the effects of various features of the urban form on microclimate from the perspective of an urban planner.

URBAN DENSITY The density of a city is generally determined by economic considerations, reflected in the price and availability of land. It is also influenced, and in turn has an effect upon, the overall form of the town. Architects and planners typically measure urban density by means of the number of dwellings per unit area of the site (Knowles, 2003) or by the ratio of the total built floor area to the area of the site, an index more suited to non-residential development. Urban climatologists, on the other hand, refer to density by different measures: the plan area density, which is the ratio of the building’s footprint to the total area of the site; or the frontal area density, which is the ratio of its (windward) elevation to the site area (Grimmond and Oke, 1999). Density has a direct effect on the exposure of urban surfaces to direct solar radiation. Such exposure may be considered beneficial in cold climates or detrimental in hot conditions, where shade is considered desirable. However, whereas the latter objective is fairly easy to achieve – for example, by the addition of specialized shading elements such as pergolas and blinds – the former imposes stringent limitations on the overall built volume that can be constructed in a given site, as well as upon its geometry. The so-called

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

102

3:32 PM

Page 102

EVYATAR ERELL

solar envelope has been the subject of numerous studies by architects, with the aim of ensuring appropriate solar exposure of buildings (Knowles, 1981) or of open space (Yezioro et al, 2006). Solar exposure of buildings has been a primary concern of many planners in climates with cold winters to allow for passive solar heating. It has sometimes been investigated in isolation from other aspects of urban climatology because it only requires knowledge of geometry. Littlefair (1998) surveyed a variety of graphic methods to establish solar exposure, though a substantial number of computer-aided design (CAD) programmes now perform the task automatically, given a geographical location and time of year. However, the study raises the intriguing question of how to establish the criteria for solar exposure. It proposes that rather than mandate solar access on the winter solstice, when the sun is lowest, it may be preferable to define a heating season and aim to maximize gains over the whole period. Meeting the first requirement in high-latitude locations requires exceedingly large distances between adjacent buildings in order to provide what may be marginal benefits because sunshine hours are short and insolation is limited. Guaranteeing exposure when insolation levels are higher may fill a proportion of heating requirements that is only slightly smaller, yet requires far less stringent geometric limitations on building height. In tropical climates, exposure to solar radiation is generally undesirable. Planners in such locations are therefore concerned with creating urban geometries that maximize shade. Narrow streets and a dense urban matrix sometimes are often recommended for desert locations (Mazouz and Zerouala, 1998), although they restrict ventilation; but designs for warm-humid locations must maximize airflow while providing shade in public spaces too (Emmanuel, 1993), so tropical cities may be less dense. In temperate climates, design for solar access reflects a desire to accommodate sometimes contradicting requirements in response to seasonal weather and solar exposure. Swaid (1992) proposed that operable screens be installed on building rooftops to restrict solar access into street canyons when radiation is excessive, yet which are capable of being folded away when full exposure is desirable. He simulated the effect of adjustable screens on canyon air temperature, and reported significant differences among the various configurations tested. Knowles (2003), too, suggested that the solar envelope of a building should be adjustable, terming the zone that lies between the summer and winter extremes the ‘interstitium’. The effect of urban thermal mass on air temperature may be seen as being analogous to the effect of thermal mass on interior temperatures of buildings (Ratti et al, 2003). Erell and Williamson (2007) suggest that since the three-dimensional geometry of the city results in much larger surface area compared with a flat rural site with the same plan area, effective thermal mass is greater in the former, which tends to reduce the diurnal temperature swing. Field experiments carried out in Adelaide, Australia, have shown that the diurnal amplitude of air temperature in a deep street canyon was 60 to 70 per cent of the rural one, and was the result not only of elevated night-time minima, but also of reduced daytime maxima. This suggests that the dense urban structure often found in traditional quarters of many desert cities reduces the diurnal temperature extremes characteristic of such climates.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 103

The Application of Urban Climate Research in the Design of Cities

103

STREET ORIENTATION The planning of streets is often the primary input in the master plan of a city. While it is true that it is the architecture of the buildings that give a city character, it is the street layout that determines its structure. In modern town planning, streets are generally planned in response to the requirements of transport systems. Transport planners draw up the appropriate road links in response to the density and type of expected land use in accordance with national standards or accepted best practice. The orientation of streets with respect to the path of the sun or to the prevailing winds is now rarely considered during the design process, although its effect on microclimate was recognized even in historical times: Vitruvius (1999), discussing orientation in approximately 20 BC, wrote: . . . the orientation of streets and lanes according to the regions of the heavens. This process will be properly accomplished if, with foresight, the lanes are kept from facing into the path of the prevailing winds. For if the winds are cold, they injure; if hot, they corrupt; if moist, they are noxious. Kenworthy (1985), discussing the regular street pattern of some ancient cities such as Miletus (in Asia Minor), proposed that the opposite was true: promoting exposure to regional winds on an urban scale has been an aim of city planners from ancient times. What, then, is the orientation (relative to wind direction) that best promotes ventilation? Kenworthy (1985) found that while streets oriented parallel to the prevailing winds would appear to offer the least aerodynamic resistance, scale-model tests of orthogonal block grids showed that the maximum wind speed at street level was, in fact, measured when the wind blew at a small angle to the main axis. Bottema (1999) also found that parallel flow resulted in lower roughness (z0) than normal flow, but noted that long buildings aligned with the wind created ‘flow channelling’ that actually impaired ventilation and removal of pollutants because of reduced vertical mixing. If the street width is less than twice building height, shelter is enhanced, but ventilation is reduced. Several studies on the orientation of streets, usually in a grid scheme, have recommended various orientations on the basis of exposure of buildings or of street surfaces to direct solar radiation. Gupta (1984) found that in composite climates, an east–west street with continuous wall surfaces was the optimum configuration (from building energy considerations), but in low-latitude locations, a north–south street axis gave buildings equal solar protection to an east–west oriented street. A similar study (Mills, 1997) compared the effect of building group configuration on the thermal stresses affecting individual buildings on the basis of two measures: solar exposure (controlling heating) and sky view factor (controlling cooling by long-wave radiation). The study provides a useful insight into the effects of urban geometry; but recommendations for different climate zones (defined by latitude) are, of necessity, too simplistic, not least since they ignore the effects of convection on building energy needs. Considering solar exposure of streets, rather than buildings, Knowles (1981) arrived at different conclusions: a ‘Spanish grid’ (in which streets are oriented at 45° to the cardinal points of the compass) was found to be preferable to the so-called ‘Jeffersonian grid’

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

104

3:32 PM

Page 104

EVYATAR ERELL

(in which streets are oriented east–west and north–south). This is because in the Spanish grid, part of the street is shaded while the opposite sidewalk is exposed to the sun all year round, allowing pedestrians to choose between different conditions. The Jeffersonian grid, in contrast, maximizes midday exposure on the north–south streets in both summer and winter, while east–west streets will be in shadow most of the time in winter, unless they are very wide, and exposed to the sun during summer. The effects of street orientation on solar access in low-latitude desert cities were also modelled by Bourbia and Awbi (2004) and Ali-Toudert and Mayer (2006), the latter emphasizing that the effects of street orientation on thermal comfort should be considered with respect to the daily patterns of use by pedestrians since solar access varies with time of day. Shashua-Bar and Hoffman (2003) pointed out that extensive planting of trees minimizes the effect of street orientation with respect to solar access. However, the question of street orientation should be considered not only in relation to solar access (or protection), but also with respect to wind direction, especially where lack of ventilation is a major problem (Ahmed, 2000).

STREET ASPECT RATIO Practising architects typically refer to the height of building façades facing a street or to the width of the street; but rarely, if at all, do they refer to the ratio between them, except in a qualitative sense. Climatologists, on the other hand, have found it useful to study the city by means of a simplified form referred to as an ‘urban canyon’: a semi-infinite street with a rectangular cross-section bounded on both sides by continuous buildings of equal height. This prototypical urban space is characterized by means of its ‘aspect ratio’, or the ratio of building height (H) to the width of the street (W). The aspect ratio of a street is one of the most important controls on microclimatic conditions in the street, affecting the transfer of energy, mass and momentum occurring in the space between buildings.

Radiant exchange The canyon aspect ratio affects radiant exchange in different ways during the daytime and at night. Deep streets have a small sky view factor and therefore lose less heat at night by long-wave radiation than shallower ones. Oke (1981) demonstrated how canyon geometry could explain the formation of the nocturnal urban heat island (UHI), previously thought to be linked to the population size of the city (Oke, 1973). Oke’s (1981) experiment is supported by numerous field studies (Barring et al, 1985; Yamashita et al, 1986; Runnalls and Oke, 1998; Goh and Chang, 1999; Livada et al, 2002; Chow and Roth, 2006; Erell and Williamson, 2006) and has formed the basis for much subsequent research on the UHI. The effect of the canyon aspect ratio on air temperature at night is well understood: the city cools down more slowly than rural areas and has higher minimum temperatures. However, the processes that occur during the daytime are more complex. Deep street canyons restrict access to direct solar radiation and provide shade at street level, but they also act as ‘radiation traps’: solar radiation that might otherwise have been reflected back to the atmosphere from high-albedo surfaces experiences multiple-reflections among canyon surfaces. The overall balance of these opposing tendencies depends inter alia on

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 105

The Application of Urban Climate Research in the Design of Cities

105

the exact proportions of the street and on the thermal properties of the surfaces. Several studies of air temperature in urban streets show that they are warmer than rural areas in the daytime as well as at night (Santamouris, 1998; Giridharan et al, 2004), while others demonstrate that air temperature in city streets may, in fact, be slightly cooler during the day (Steinecke, 1999; Runnalls and Oke, 2000; Erell and Williamson, 2007). The canyon aspect ratio also has an effect on the overall albedo of a city with a regular geometry. Aida and Gotoh (1982), using a numerical model, found that the maximum solar absorption in a mid-latitude city such as Tokyo occurs when the canyon width is approximately twice the block width. Soler and Ruiz (1994) confirmed this finding in an empirical study comparing the intensity of reflected radiation in satellite images of urban and rural areas near Barcelona with terrestrial measurements. It should be noted that existing high-albedo cities are characterized by either high density (where roof albedo dominates) or very low density (where the albedo of the ground is most important). In medium-density cities, multiple reflections among canyon facets emphasize the contribution of walls in addition to the horizontal surfaces. The effect of street canyon geometry on radiant exchange may have a great effect on human thermal comfort, an issue often overlooked in modern street design. In hot climates with high radiant loads, net radiant balance may be more important than convective exchange, and the benefit to pedestrians of shade outweighs minor modifications to air temperature that might be measured in the street (Pearlmutter et al, 1999, 2006).

Airflow Although airflow in cities takes place in open spaces of all types and sizes – for example, backyards, parks and plazas, in addition to streets – it is streets that have received the most attention from both climatologists and designers addressing this issue (although for different reasons). Climatologists have found the two-dimensional street canyon a useful generic urban form to model airflow, using the aspect ratio (H/W) as one of its primary geometric descriptors. Oke (1987) classified urban airflow into three basic regimes – isolated roughness flow, wake interference flow and skimming flow – identifying flow patterns associated with certain ranges of street aspect ratios. Airflow in street canyons with uniform building heights and an aspect ratio of approximately unity is characterized by a lee vortex generated by the above-roof wind, and may display a corkscrew pattern as the cross-canyon circular motion is combined with an along-canyon component (Hotchkiss and Harlow, 1973; Yamartino and Wiegand, 1986). The effect of the canyon aspect ratio on airflow patterns in street canyons that are either much shallower or much deeper, or which have buildings of unequal height on either side (step up or step down with respect to the wind), has been the subject of several recent studies (Baik and Kim, 1999; Baik et al, 2000). In deep canyons, two major counter-rotating vortices may occur in some conditions, the lower one driven by the upper (So et al, 2005). The thermal flows generated by unequal heating of canyon surfaces further complicate the picture (Kim and Baik, 2001; Xie et al, 2005, 2006). Assimakopoulos et al (2006) report on field studies that show the existence of intermittent vortices that cannot be modelled by quasi-steady state models.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

106

3:32 PM

Page 106

EVYATAR ERELL

They note that existing computational fluid dynamic (CFD) models are still incapable of fully reproducing the turbulent nature of airflow in reality, and cannot be used to predict the evolution of airflow at a particular urban location for any length of time. The agreement between predicted and actual airflow is especially poor when wind speed is low – less than 2m/s – when ventilation might be difficult to provide (Dixon et al, 2006). Large eddy simulation (LES), requiring very substantial computing resources, though much more powerful than CFD, is also still not sufficiently developed to be applied as a tool for generating design strategies for even simplified generic forms of the urban structure. The geometry of street canyons is of primary importance in the study of pollutant dispersion, especially automobile exhaust fumes. Studies have been carried out using scale models in wind tunnels (Cermak, 1995; Kastner-Klein and Plate, 1999; Al-Sallal et al, 2001), as well as numerical simulation, once computing power had developed sufficiently to allow numerical simulation of airflow in the presence of arrays of bluff objects representing buildings (Kim and Baik, 1998, 2004; Chan et al, 2001, 2003). The flow regime in a canyon (and, hence, the effectiveness of flushing out pollutants) depends not only upon the wind speed but also, to a great degree, upon the canyon aspect ratio (So et al, 2005). As might be expected, deeper canyons are more likely to suffer from high concentrations of pollutants than shallower ones, while irregular building height promotes turbulence and therefore improves pollutant removal. The effect of roof geometry is also important (Rafailidis, 1997; Kastner-Klein and Plate, 1999). Much of the pollutant removal occurs at intersections, which are characterized by unique flow regimes (Chan et al, 2003; Dobre et al, 2005).

NEIGHBOURHOOD AND BUILDING TYPOLOGY A feature of urban development clearly visible to all city dwellers is building typology. Many streets conform to the ‘urban canyon’ model that has proved such a useful tool for urban climatologists. Yet, many more do not. Gupta (1984) identified three basic building forms, which may translate into quite different urban morphologies: pavilion, court and street. Steemers et al (1997) found six generic urban forms: pavilions; slabs; terraces; terrace courts; pavilion courts; and courts – different combinations of which produced distinct wind patterns in wind tunnel experiments. The complexity of real cities, as opposed to simplified conceptual models composed of generic building types and regular street patterns, makes analysis of the effect of urban form on micro-scale environmental behaviour very difficult. Computerized imageprocessing techniques offer a method of dealing with such conditions by analysing urban texture using a digital elevation model (DEM) of the site in question (Ratti and Richens, 1999, 2004). The DEM is a compact way of storing urban 3D information using a 2D matrix of elevation values: each pixel represents the height of a building (or part of it) and can be displayed in shades of grey as a digital image. The computer model may then be analysed for any geometric characteristic desired, such as the sky view factor from the ground at any point, the extent of shadowing or sunlight availability. The technique was used in an analysis of building energy consumption and urban texture (Ratti et al, 2005) using the LT model1 to calculate energy requirements and the DEM to differentiate

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 107

The Application of Urban Climate Research in the Design of Cities

107

between core sections of buildings, which require more energy to service, and perimeter areas, which are more amenable to daylighting, ventilation and passive heating. A further study using this technique suggested that in a hot-arid climate, courtyard buildings (of certain proportions) were a more appropriate building type than the ‘pavilion’ type on the basis of their surface-to-volume ratio, shadow density, daylight distribution and sky view factor (Ratti et al, 2003). The same study also suggests that courtyard buildings would be less suitable for tropical climates, where a narrow temperature variation would not benefit from the high thermal mass and reduced ventilation characteristic of this configuration. Texture analysis with image processing can also be used in the study of winds and dispersal of pollutants in the city (Ratti et al, 2006). The variation of wind speed above the ground is affected by the morphology of the surface, which can be described by means of the roughness length (z0) and the displacement height (zd). The roughness length is, in turn, a function of the average height of the buildings and a measure of their density, such as the frontal area density (for a given wind direction) or the plan area density (Grimmond and Oke, 1999). These can be obtained easily from a DEM and the aerodynamic roughness derived for the site being studied. Using this tool, analysis of hypothetical arrangements of long buildings arranged in parallel rows showed that counter to design intuition and common sense, aligning the streets parallel to the wind resulted in poorer ventilation and slower pollutant removal than aligning them perpendicular to the wind.

SIZE, TYPE AND LOCATION OF URBAN PARKS Vegetation affects conditions in the city in a variety of ways. For instance, several studies found that vegetation dry-precipitates and adsorbs pollutants and, by doing so, decreases the mass of airborne gases and particulate matter (Raza et al, 1990–1991; Taha et al, 1997). However, the following section deals only with the effects of vegetation on the energy balance and upon air temperature. The effect of local parks in a non-homogeneous urban area has been the subject of intense study, especially once it became clear that the microclimate of built-up areas differed substantially from that of rural areas. The so-called park cool island (PCI), a manifestation of the more general oasis effect, is the converse of the urban heat island (UHI): empirical findings show that air temperature in moderate- to large-sized parks may be substantially lower than temperature in surrounding built-up areas (Jauregi, 1990–1991; Kanda and Moriwaki, 1998), although there are significant variations among different types of parks. Landscaping, specifically the incorporation of planted areas within the urban fabric, may reduce differences between natural terrain and the urban surface. Bonan (2000) demonstrated in an empirical study of a town that the availability of water and the resulting increase in evaporation was the main factor responsible for lower surface temperature and air temperature above the lawns, rather than the mere presence of plants. This helps to explain the finding that increasing the proportion of planted areas in a city tends to reduce daytime maximum temperatures, but may often have little or no effect on night-time minima (Urano et al, 1995). Emmanuel (1997) noted that the main effect of trees is to reduce radiant exchange at the ground surface: this may decrease

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

108

3:32 PM

Page 108

EVYATAR ERELL

daytime maximum temperature, but would also restrict nocturnal cooling, actually leading to higher minima.

Classification of urban parks Spronken-Smith and Oke (1998) distinguished between park cool islands that are defined by surface temperature, which may be quite large, and cool islands defined by air temperature, where the effects of surface temperature variation are diluted by nearsurface turbulent mixing and advection by wind. During daytime, surface temperatures are affected by the presence or absence of shade, by surface albedo, by water availability and by the thermal properties of the soil. These properties govern the receipt of solar radiation, its absorption and the role of evaporative cooling. At night the thermal properties of surfaces and the radiative geometry are the major controls on cooling. Urban parks vary substantially with respect to the above factors, and may be classified according to the arrangement of vegetation (Spronken-Smith and Oke, 1998): grass; grass with tree border; savannah (grass with isolated trees); garden; forest; and multi-use. Park cool islands may develop either during the daytime or at night. However, a given urban park will display a regular diurnal pattern, indicating that the formation of PCIs may be the result of a number of mutually exclusive factors. Spronken-Smith and Oke (1999) found that daytime PCIs formed as a result of the combined effects of soil moisture and shading: trees shade the surface, while grass is typically cooler than most solid surfaces during the daytime if it is well irrigated. The relative coolness of irrigated parks therefore peaks in the afternoon (forest type) or early evening (garden, savannah and multi-use types). However, trees also inhibit nocturnal long-wave radiative cooling by blocking off part of the sky, while excess moisture increases the thermal capacity of the soil and slows down surface cooling. Night-time PCIs therefore typically form in relatively dry urban parks with a sparse tree cover. They are driven by long-wave radiative cooling (since the sky view factor is close to unity), and since evaporative fluxes are generally weak at night, evaporation does not play a significant role in the formation of this type of PCI. In such parks, daytime temperatures may sometimes be higher than in neighbouring urban areas. However, an ‘edge effect’ exists that applies within distances of about 2.2 to 3.5 times the height of the park border, and which results in weaker radiative cooling where the sky view factor is reduced by the obstructing features, such as perimeter trees or buildings.

Effects of size The extent of vegetated area required to produce measurable effects on air temperature is of great interest to urban planners. Upmanis et al (1998) found that the magnitude of the intra-urban temperature difference between parks and their urban surroundings at night increased with increasing park size, although large differences were also found within the parks and in the built-up areas. These were attributed to the degree of exposure to the sky and, hence, to the intensity of radiant cooling. Saito et al (1990–1991), in contrast, reported cool islands in clumps of vegetation less than 200m across, although the effect of vegetation was limited to the planted area itself and was not felt at distances as little as 20m away from the park edge. Ca et al (1998) reported on the basis of measured data that an urban park 0.6km2 in area can reduce air temperature in a commercial

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 109

The Application of Urban Climate Research in the Design of Cities

109

area 1km downwind by as much as 1.5°C; but Shashua-Bar and Hoffman (2000), who reported differences of up to 3°C between air temperature in tree-shaded urban avenues and nearby non-shaded reference points when wind speed was very low, noted that the cooling effect declined at an exponential rate with increasing distance from the border of the planted area, and vanished less than 100m away from the sites studied. Numerical modelling (Bruse and Fleer, 1998) indicates that small parks of only tens of metres across may create temperature differentials of 2°C or more. However, the horizontal gradients of air temperature represented by the models may be quite large, and their spatial patterns shift constantly with wind speed and direction, as demonstrated by Upmanis and Chen (1999) in Goteborg. A pronounced difference in air temperature is almost always observed very close to the edge of the park, a phenomenon also noted by Santamouris (2001) in a study carried out in Athens.

The effect of trees on urban air temperature The presence of trees in the urban matrix may affect air temperature at a variety of spatial scales, from individual streets (Shashua-Bar and Hoffman, 2000) to city-scale modifications (Huang et al, 1987). However, the magnitude of this effect may depend upon a variety of factors because the interaction between trees and other constituents of the urban environment is so complex (Oke, 1989). Trees intercept not only solar radiation, but also long-wave radiation from the ground, building surfaces and the sky. The dissipation of this heat load depends upon the water balance and wind climate of the tree. In the presence of unrestricted water, transpiration will cause substantial cooling. However, water supply to the root system may be restricted; stomata may be physically blocked by particulates; or the heat load may be excessively high, leading to closure of the stomata. Furthermore, at night (in the absence of sunlight) there is no photosynthesis, so the stomata are closed and the tree is not cooled by transpiration. The response of trees to increased energy loading, which may occur when individual trees are planted in extensive paved areas such as parking lots, for example, will vary with species, humidity of the atmosphere and how much of the crown is exposed (Kjelgren and Montague, 1998). Species from hot or arid habitats may be tolerant of high temperatures or able to dissipate heat with small leaves – but the evapotranspiration rates from such trees may accordingly be lower than those of broadleaf trees, and thus they have a smaller effect on air temperature in their surroundings. Grimmond et al (1996) found that in a neighbourhood with a relatively dense tree cover, the latent heat flux increased as a fraction of available energy compared to an otherwise similar neighbourhood with a sparse tree cover, as did the storage flux, whereas the sensible heat flux decreased. However, in absolute terms, all fluxes, including the sensible heat flux, were enhanced at this neighbourhood. Trees and shrubs lowered the albedo and surface temperatures, thereby reducing the loss of solar and long-wave radiation, respectively – resulting in an increase in the overall amount of energy to be dissipated compared to the sparsely planted neighbourhood. As a result of the difference in fluxes, the maximum air temperature above the canopy was about 1°C higher in the densely planted neighbourhood, while temperatures below the trees at noon and in the early afternoon were similar or up to 0.5°C higher than in the sparsely planted neighbourhood.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

110

3:32 PM

Page 110

EVYATAR ERELL

In contrast to Grimmond et al (1996), most researchers report that trees reduce urban air temperature. This is usually attributed to evapotranspiration; but Shashua-Bar and Hoffman (2003) suggested that the cooling is due almost entirely to shading, which more than offsets the exchange of sensible heat between the tree canopy and the air. The actual reduction in air temperature might be overstated in many cases due to the difficulty of measuring air temperature accurately in the presence of strong radiant loads (Erell et al, 2005). A model specifically designed to predict the general microclimatic effect on energy consumption of augmenting urban vegetation was proposed by Sailor (1998). The meso-scale model (2 x 2km grid) predicts that by increasing the vegetated fraction of the core of a hypothetical city by 0.065, annual cooling loads could be reduced by 3 to 5 per cent simply by lowering the background air temperature. Shashua-Bar and Hoffman (2002) proposed a quantitative model for predicting the cooling effect of trees in an urban setting; but their method incorporates site-specific empirical factors to account for convective exchange, the magnitude of which is not systematic and for which there is no method of calculation. The method is, nonetheless, demonstrated in an empirical study of the effect of trees on air temperature in the streets of Tel Aviv, Israel, which found that the overall cooling effect of trees could be as high as 3K, depending, in addition to the shade coverage, upon the depth of the urban canyon, the albedo of canyon walls and street orientation (Shashua-Bar and Hoffman, 2004).

The effects of vegetation on building energy consumption Landscaping and careful planting of vegetation near buildings have been credited with energy savings of anything up to 80 per cent in hot, dry climates (Meier, 1990–1991). Several studies have been carried out to quantify the effect of vegetation, especially trees, on the energy consumption of buildings (Simpson and McPherson, 1998; Simpson, 2002); but much of the evidence remains anecdotal. The mechanisms by which vegetation affects the energy exchange between buildings and the environment may be summarized as follows: ● Vegetation can reduce energy consumption in buildings in hot climates if air

temperature is reduced near the planted area. However, it should be noted in this context that heat transfer through building walls is driven by differences in surface temperature, rather than by air temperature. Furthermore, the reduction in air temperature resulting from evapotranspiration is accompanied by an increase in the vapour content of the air. Therefore, the air-conditioning system must deal with an increased latent heat load, offsetting, to some extent, any gains from a lower sensible heat load. ● Plants may shade building surfaces, reducing the radiant load on the envelope. This may be beneficial in hot climates, but detrimental in cold ones. In temperate climates with distinct heating and cooling seasons, deciduous trees are often planted, and vine-covered trellises are common in many Mediterranean areas. However, the timing of defoliation and the permeability of the bare trees vary widely from species to species (Canton et al, 1994) and may not match the desired pattern of exposure to

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 111

The Application of Urban Climate Research in the Design of Cities

111

the sun. McPherson et al (1988) found that in the middle latitudes, cooling loads were most sensitive to shading on the roof and on the west wall, while heating loads were affected most by exposure of the south (equator-facing) and east walls. The direct effect of shading building surfaces by plants was studied by Papadakis et al (2001). Thick foliage producing a full shade effect resulted in a reduction of the surface temperature of a light-coloured concrete wall by up to 8°C, with concomitant reductions in heat flux through the surface. When wind speed was negligible, air moisture content in the planted canopy was up to 2g/m3 greater than in the surrounding air. Likewise, a dense growth of ivy can block radiant exchange at the wall surface almost entirely (Hoyano, 1988). However, several studies show that where the shaded area has a limited spatial dimension – for instance, beneath a pergola (Hoyano, 1988) or in the shade of a liman (small clump of trees in an artificial floodplain in the desert) (Schiller and Karschon, 1974) – the effect on air temperature at a height of 1m above the ground is negligible. ● Plants may reduce wind speed near buildings, limiting unwanted infiltration, but also restricting ventilation and reducing convective exchange at building surfaces. The first two mechanisms are self-explanatory; but the third has less well-known consequences. For instance, in hot climates, wind is an asset for unshaded houses because it helps to remove radiant heat at the external building surfaces (McPherson et al, 1988), reducing temperature differentials between the interior and exterior. However, in poorly insulated houses, especially in cold climates, increased convective exchange at the building envelope results in increased loads on building heating or cooling systems. ● Plants in warm climates may reduce temperatures of ground surfaces through evapotranspiration (although planted surfaces may be warmer than bare soil in cold climates), with two effects: cooler surfaces emit less infrared radiation, thus reducing the radiant load on building surfaces; and they release less sensible heat to the adjacent air, so that buildings are exposed to cooler ambient air. ● Roof gardens (or planted roofs) are perhaps the most obvious example of the use of plants to control building energy performance, and are sometimes credited with improving the urban microclimate as well (Wong et al, 2003). The shading and evapotranspiration of the plants contribute to lower surface temperatures and thus to lower heat gains through the roof. Experiments show that the surface temperature of an exposed roof can be reduced substantially by the addition of an irrigated lawn on a fabric matrix (Onmura et al, 2001). A complete thermal model of a planted roof (Palomo Del Barrio, 1998) showed that the contribution of the planting to the thermal performance of the roof depends mainly upon the density of the foliage, the composition, density and thickness of the substrate, and its water content. However, unless the thermal conductivity of the soil is particularly low or the thickness of the substrate is considerable, the thermal resistance provided by the planting and substrate is usually insufficient during the cold season, even in mid-latitude countries with relatively mild winters. However, much of the work in this field remains empirical and lacks a comprehensive theoretical framework at the urban scale (Niachou et al, 2001).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

112

3:32 PM

Page 112

EVYATAR ERELL

BUILDING AND PAVING MATERIALS Roofing materials, wall finishes and paving blocks are typically specified by architects for a variety of reasons, such as cost, durability and appearance. Their thermal properties, especially insulation, are usually assessed in the context of the energy budget of the building. However, the properties of the materials that form the surfaces that make up the urban canopy also have an effect on urban microclimate, which may be modified through the selection of suitable finishes. The role of materials in the formation of the urban heat island has been the subject of some conjecture. Oke (1981) demonstrated that the nocturnal urban heat island could be explained not only by the effects of street canyon geometry on long-wave radiant exchange, but also by differences in the thermal properties of urban and rural surfaces. However, he also noted that, in reality, differences in thermal admittance between urban and rural materials were too small and were, in any case, affected to a great degree by rural soil moisture. A later study (Oke, 1982) suggested that differences in thermal inertia between rural and urban sites could be accounted for by considering the role of increased surface area in the city, as well as differences in moisture availability. However, the difficulty of measuring the storage flux directly in situ on a neighbourhood scale has thus far limited efforts to explain this mechanism satisfactorily. The effect of the thermal properties of the substrate on the surface temperature and on the magnitude of the energy fluxes was investigated by Asaeda et al (1996), who demonstrated a substantial difference in the behaviour of asphalt and concrete (painted black to create similar albedo). The former has a lower thermal conductivity, and the radiant energy it absorbs at the surface during the daytime results in a higher temperature and therefore greater emission of long-wave radiation and higher rate of sensible heat loss. Conversely, an asphalt surface also cools down more quickly than concrete at night. Buildings in the warmer parts of the Mediterranean are often whitewashed to reduce the external radiant load on the buildings and, hence, their surface temperature during the daytime. The same strategy has been proposed to reduce convective heat transfer from pavements and buildings to the air, although the temperature of air near the surface is also affected by several additional factors. Paving materials with a variety of colours and textures have been investigated by several researchers to evaluate the effect of exposure to intense sunlight on surface temperature (Bretz et al, 1998; Doulos et al, 2004; Synnefa et al, 2006). The studies found that during the daytime, all paving materials had a mean temperature higher than that of the ambient air, while at night all surfaces were cooler, in some cases by almost 6°C (Synnefa et al, 2006). However, differences among the samples are illuminating: compared to an exposed concrete tile serving as the reference, Synnefa et al (2006) found that some of the coatings tested resulted in mean daytime temperatures lower by approximately 4°C to 5°C, with a maximum difference in excess of 7°C, while Doulos et al (2004) found that the difference between the coolest surface (white marble) and the warmest one (asphalt) was over 22°C. The albedo of roof surfaces may have less effect on air temperature in the canopy layer than the properties of the pavement or the ground because it affects heat transfer above roof level. However, since it affects the energy performance of a house directly, it has been the subject of much research. The obvious strategy with respect to roofing

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 113

The Application of Urban Climate Research in the Design of Cities

113

materials in hot climates is to employ light-coloured materials. However, progress in the production of wavelength-selective paints now allows other colours to be specified, which although visually fairly dark, nevertheless have relatively high albedo (Levinson et al, 2007; Synnefa et al, 2007). The so-called ‘cool coatings’ can reduce the temperature of roof tiles having the same appearance (i.e. colour) by between 1.5°C and 10°C (for green and black tiles, respectively). Some of the coatings have been tested to evaluate the long-term effects of exposure to ultraviolet (UV) radiation and mechanical damage, and show relatively little deterioration over extended periods (Bretz and Akbari, 1997; Synnefa et al, 2007). Some of the reduction in reflectance is attributed to soiling and can therefore be reversed by rinsing. High surface temperatures contribute directly to pedestrian discomfort by imposing a long-wave radiant load (Pearlmutter et al, 2006, 2007), and may also affect air temperature in close proximity. However, the thermal properties of urban surfaces may also have an effect on the climate of a city as a whole: if roof albedo is modified on a large proportion of city roofs, the cumulative effect on the climate of a city may be significant. Using a onepoint model of Sacramento, California, Taha et al (1988) suggested that maximum air temperature (in the early afternoon on a summer day) could be reduced by about 4°C if urban albedo was increased from 0.25 to 0.4, while reducing urban albedo to only 0.1 would result in an increase of more than 4°C relative to the base case. Taha et al (1997) found, using the Colorado State University Mesoscale Model (CSUMM), that feasible albedo changes could lower daytime maximum temperatures in central Los Angeles by up to 2°C, with smaller reductions predicted for the suburbs. The actual albedo of most urban areas varies from about 0.1 to about 0.2 (Taha, 1997; Sailor and Fan, 2002), with a mean value of about 0.14 for urban centres (Oke, 1988b), so there appears to be substantial potential for cooling if the large-scale application of high-albedo surfaces can, in fact, be attained (Bretz et al, 1998). The scale of modification possible depends upon the area of solid surfaces, such as rooftops, streets and other paved areas as a proportion of the total urban area. However, the micro-scale intra-urban variation of air temperature in the urban canopy, observed in several field studies (Eliasson, 1996; Erell and Williamson, 2007), means that the benefits of such a strategy may not be felt uniformly over the entire city. Extensive glazed areas also affect radiant exchange in a street canyon. While the use of mirror glass results in extremely high albedo, sunlight transmitted to building interiors through transparent windows creates urban surfaces that have extremely low albedo. Tsangrassoulis and Santamouris (2003) showed that increasing the window-to-wall ratio resulted in a lower canyon albedo because sunlight is transmitted by the glass to building interiors. In very deep canyons (H/W > 2), the predicted albedo for all sun angles as well as for uniform diffuse radiation was between 0.01 to 0.03 – meaning that practically all of the incident solar radiation is absorbed. Deep streets with extensive glazed areas in adjacent buildings may therefore be viewed as a very efficient trap for solar radiation.

DISCUSSION AND CONCLUSIONS As the above survey has demonstrated, there is a substantial body of knowledge on the effects upon microclimate of a wide variety of planning decisions. However, it is equally

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

114

3:32 PM

Page 114

EVYATAR ERELL

evident that very little of this knowledge is being applied in practice, except in a haphazard, piecemeal way. Page (1968) identified three reasons why scientific information available to researchers might, nonetheless, be rejected by practitioners: 1 The potential user considers the information irrelevant. 2 The user considers the information provided inapplicable in the form presented. 3 The user considers the information provided incomprehensible. Page (1968) also noted that in order to be of use to urban planners, urban climatology must first be predictive – descriptive science is not in itself sufficient. Furthermore, the problem is not to produce an idealized climatological plan, but to produce a workable evolutionary plan that is economically viable and accepts that the planner must consider other factors, such as the requirements of transportation systems. Bitan (1988) noted that climatology affected urban design at a variety of spatial scales, and that failure to address the effects at each of them, in turn, would necessarily restrict the effectiveness of any intervention. However, in reality, there are very few projects where climatologists take part in ongoing design processes over many years, beginning with environmental impact assessments at the stage of site selection for large urban developments and continuing throughout the entire planning process. The number and hierarchy of authorities that are involved in the planning and regulatory approval of such projects require a commitment that is rarely found in practice. Oke (1984) suggested that what climatologists require is an ‘ability to demonstrate the importance of climatic information in the design of a settlement, and the predictive power to foretell the climatic impact of alternative design strategies’. According to Oke (1984), several shortcomings in microclimate research at that time prevented it from becoming an applied science, including a lack of quantitative techniques and relationships; lack of standardization, generality and transferability; and the absence of clear guidelines for those wishing to learn and use climatological principles in settlement planning. Advances in addressing any of the following issues would represent real progress in applying urban climatology in practical urban design: ● What are the urban forms that will provide the best microclimate for any given

density, taking into consideration meso-scale climate, latitude, etc.? ● Can street orientation be used as a policy guideline to achieve desirable microclimatic

conditions in a city? ● Can the aspect ratio of streets be used as a policy guideline to achieve desirable

microclimatic conditions in a city? What are the quantitative limits that may be usefully applied in realistic planning situations in high-density cities with regard to specific goals of solar access and airflow? ● What are the most appropriate neighbourhood layouts for every climate type? Are some building typologies better at providing good microclimates than others? ● What are the real effects of vegetation in various urban contexts? There seems to be a consensus that plants ‘improve’ the microclimate of cities; but numerous

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 115

The Application of Urban Climate Research in the Design of Cities

115

quantitative studies have produced a wide range of estimates regarding the magnitude and spatial extent of the effect of urban parks. ● What are the city-scale modifications to microclimate that can be achieved by implementing particular types of building and paving materials? The benefits to individual buildings may be clear; but although models suggest that the compound effect of enforcing their use all across an urban area may be substantial, the magnitude of even local-scale effects is still open to some question. ● Can urban climatology provide detailed site-specific weather data modified from regional weather stations to allow more accurate simulation of building energy performance, including not only air temperature, but also humidity and wind characteristics? The answers to these questions will probably only come about through research collaboration between planners and urban climatologists.

AUTHOR CONTACT DETAILS Associate Professor Evyatar Erell: The Desert Architecture and Urban Planning Group, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Israel; tel +972-8-6596875; www.bgu.ac.il/ CDAUP; E-mail: [email protected]

NOTE 1 LT is the name Baker and Steemers (2000) gave to a computer model dealing with lighting and thermal properties of buildings.

REFERENCES Ahmed, K. S. (2000) ‘Problems and prospects of bioclimatic urban design: An analytical study of shading and ventilation needs of the urban spaces in the humid tropics’, PLEA 2000: Architecture – City – Environment, Cambridge, UK Aida, M. and K. Gotoh (1982) ‘Urban albedo as a function of the urban structure – a two-dimensional numerical simulation’, Boundary Layer Meteorology, vol 23, pp415–424 Al-Sallal, K. A., M. M. AboulNaga, and A. M. Alteraifi (2001) ‘Impact of urban spaces and building height on airflow distribution: Wind tunnel testing of an urban setting prototype in Abu Dhabi City’, Architectural Science Review vol 44, no 3, pp227–232 Ali-Toudert, F. and H. Mayer (2006) ‘Numerical study on the effects of aspect ratio and orientation of an urban street canyon on outdoor thermal comfort in hot and dry climate’, Building and Environment, vol 41, pp94–108 Arnfield, J. (2003) ‘Two decades of urban climate research: A review of turbulence, exchanges of energy and water, and the urban heat island’, International Journal of Climatology, vol 23, pp1–26 Asaeda, T., V. T. Ca and A. Wake (1996) ‘Heat storage of pavement and its effect on the lower atmosphere’, Atmospheric Environment, vol 30, no 3, pp413–427 Assimakopoulos, D. N., C. Georgakis and M. Santamouris (2006) ‘Experimental validation of a computational fluid dynamics code to predict the wind speed in street canyons for passive cooling purposes’, Solar Energy, vol 80, pp423–434

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

116

3:32 PM

Page 116

EVYATAR ERELL

Aynsley, R. and L. Gulson (1999) ‘Microclimate and urban planning in the humid tropics’, Planning in the Hothouse – 27th National Congress of the Royal Australian Planning Institute, Darwin, Australia Baik, J. and J. Kim (1999) ‘A numerical study of flow and pollutant dispersion characteristics in urban street canyons’, Journal of Applied Meteorology, vol 38, no 11, pp1576–1589 Baik, J., R, Park, H. Chun and J. Kim, J. (2000) ‘A laboratory model of urban street-canyon flows’, Journal of Applied Meteorology, vol 39, no 9, pp1592–1600 Baker, N. and Steemers, K. (2000) Energy and Environment in Architecture: A Technical Design Guide, E. and F. Spon, London Barring, L., J. O. Mattsson and S. Lindqvist (1985) ‘Canyon geometry, street temperatures and urban heat island in Malmo, Sweden’, Journal of Climatology, vol 5, pp433–444 Bitan, A. (1988) ‘The methodology of applied climatology in planning and building’, Energy and Buildings, vol 11, pp1–11 Bonan, G. B. (2000) ‘The microclimates of a suburban Colorado (USA) landscape and implications for planning and design’, Landscape and Urban Planning, vol 49, pp97–114 Borve, A. B. (1982) ‘Settlement planning under Arctic conditions in northern Norway’, Energy and Buildings, vol 4, pp67–70 Bottema, M. (1999) ‘Towards rules of thumb for wind comfort and air quality’, Atmospheric Environment, vol 33, pp4009–4017 Bourbia, F. and H. B. Awbi (2004) ‘Building cluster and shading in urban canyon for hot dry climate, Part 2: Shading simulations’, Renewable Energy, vol 29, pp291–301 Bozonnet, E., R. Belarbi and F. Allard (2005) ‘Modelling solar effects on the heat and mass transfer in a street canyon, a simplified approach’, Solar Energy, vol 79, pp10–24 Bretz, S. and H. Akbari (1997) ‘Long-term performance of high-albedo roof coatings’, Energy and Buildings, vol 25, pp159–167 Bretz, S., H. Akbari and A. Rosenfeld (1998) ‘Practical issues for using solar-reflective materials to mitigate urban heat islands’, Atmospheric Environment, vol 32, no 1, pp95–101 Bruse, M. and H. Fleer (1998) ‘Simulating surface–plant–air interactions inside urban environments with a three dimensional numerical model’, Environmental Modelling and Software, vol 13, pp373–384 Ca, V. T., T. Asaeda and E. M. Abu (1998) ‘Reductions in air conditioning energy caused by a nearby park’, Energy and Buildings, vol 29, pp83–92 Canton, M. A., J. L. Cortegoso and C. de Rosa (1994) ‘Solar permeability of urban trees in cities of western Argentina’, Energy and Buildings, vol 20, pp219–230 Cermak, J. E. (1995) ‘Physical modelling of flow and dispersion over urban areas’, in Cermak, J. E., Davenport, A. G., Plate, E. J. and Viegas, D. X. (eds) Wind Climate in Cities, Kluwer Academic Publishers, Dordrecht/Boston/London, pp383–403 Chan, A., E. So and S. Samad (2001) ‘Strategic guidelines for street canyon geometry to achieve sustainable street air quality’, Atmospheric Environment, vol 35, pp5681–5691 Chan, A., W. Au and E. So (2003) ‘Strategic guidelines for street canyon geometry to achieve sustainable street air quality – part II: Multiple canopies and canyons’, Atmospheric Environment, vol 37, pp2761–2772 Chow, W. and M. Roth (2006) ‘Temporal dynamics of the urban heat island of Singapore’, International Journal of Climatology, vol 26, pp2243–2260 Dixon, N. S., J. W. D. Boddy, R. J. Smalley and A. S. Tomlin (2006) ‘Evaluation of a turbulent flow and dispersion model in a typical street canyon in York, UK’, Atmospheric Environment, vol 40, pp958–972 Dobre, A., S. Arnold, R. J. Smalley, J. W. D. Boddy, J. Barlow, A. S. Tomlin and S. Belcher (2005) ‘Flow field measurements in the proximity of an urban intersection in London, UK’, Atmospheric Environment, vol 39, pp4647–4657 Doulos, L., M. Santamouris and I. Livada (2004) ‘Passive cooling of outdoor spaces: The role of materials’, Solar Energy, vol 77, pp231–249 El Nahas, M. M. (1996) Energy Saving Through Urban Design – A Microclimatic Approach, PhD thesis, Department of Architecture, University of Adelaide, Adelaide

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 117

The Application of Urban Climate Research in the Design of Cities

117

El-Shakhs, S. (1994) ‘Sadat City, Egypt, and the role of new town planning in the developing world’, Journal of Architectural and Planning Research, vol 11, no 3, pp239–259 El-Sioufi, M. (1987) Urban Patterns for Improved Thermal Performance, PhD thesis, Department of Architecture, University of Michigan, MI Eliasson, I. (1996) ‘Intra-urban nocturnal temperature differences: A multivariate approach’, Climate Research, vol 7, pp21–30 Eliasson, I. (2000) ‘The use of climate knowledge in urban planning’, Landscape and Urban Planning, vol 48, pp31–44 Emmanuel, R. (1993) ‘A hypothetical “shadow umbrella” for thermal comfort enhancement in the equatorial urban outdoors’, Architectural Science Review, vol 36, pp173–184 Emmanuel, R. (1995) ‘Energy efficient urban design guidelines for warm-humid cities: Strategies for Colombo, Sri Lanka’, Journal of Architectural and Planning Research, vol 12, no 1, pp58–79 Emmanuel, R. (1997) ‘Summertime urban heat island mitigation: Propositions based on an investigation of intra-urban air temperature variations’, Architectural Science Review, vol 40, no 4, pp155–164 Erell, E. and T. Williamson (2006) ‘Simulating air temperature in an urban street canyon in all weather conditions using measured data from a reference meteorological station’, International Journal of Climatology, vol 26, no 12, pp1671–1694 Erell, E. and Williamson, T. (2007) ‘Intra-urban differences in canopy layer air temperature at a mid-latitude city’, International Journal of Climatology, vol 27, pp1243–1255 Erell, E., V. Leal and E. Maldonado (2005) ‘Measurement of air temperature in the presence of a large radiant flux: An assessment of passively ventilated thermometer screens’, Boundary Layer Meteorology, vol 114, pp205–231 Etzion, Y. (1990) ‘A desert solar neighbourhood in Sde-Boqer, Israel’, Architectural Science Review, vol 33, pp105–111 Evans, J. M. and S. de Schiller (1990–1991) ‘Climate and urban planning: The example of the planning code for Vicente Lopez, Buenos Aires’, Energy and Buildings, vol 15–16, pp35–41 Giridharan, R., S. Ganesan and S. Lau (2004) ‘Daytime urban heat island effect in high-rise and high-density residential developments in Hong Kong’, Energy and Buildings, vol 36, pp525–534 Givoni, B. (1989) Urban Design in Different Climates, WMO/TD, no 346 Goh, K. and C. Chang (1999) ‘The relationship between height to width ratios and the heat island intensity at 22:00 for Singapore’, International Journal of Climatology, vol 19, pp1011–1023 Golany, G. S. (1996) ‘Urban design morphology and thermal performance’, Atmospheric Environment, vol 30, no 3, pp455–465 Gotz, L. (1982) ‘Integration of climate in planning and building illustrated in a case of extreme climatic conditions’, Energy and Buildings, vol 4, pp51–65 Grimmond, C. S. B. (2006) ‘Progress in measuring and observing the urban atmosphere’, Theoretical and Applied Climatology, vol 84, pp3–22 Grimmond, C. S. B. and T. R. Oke (1999) ‘Aerodynamic properties of urban areas derived from analysis of surface form’, Journal of Applied Meteorology, vol 38, pp1262–1292 Grimmond, C. S. B., C. Souch and M. D. Hubble (1996) ‘Influence of tree cover on summertime surface energy balance fluxes, San Gabriel Valley, Los Angeles’, Climate Research, vol 6, pp45–57 Gupta, V. K. (1984) ‘Solar radiation and urban design for hot climates’, Environment and Planning B: Planning and Design, vol 11, pp435–454 Herz, R. K. (1988) ‘Considering climatic factors for urban land-use planning in the Sahelian zone’, Energy and Buildings, vol 11, pp91–102 Hotchkiss, R. S. and F. H. Harlow (1973) Air Pollution Transport in Street Canyons, US EPA Report (EPA–R4–73–029) Howard, L. (1833) The Climate of London, Harvey and Darton, London Hoyano, A. (1988) ‘Climatological uses of plants for solar control and the effects on the thermal environment of a building’, Energy and Buildings, vol 11, pp181–200

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

118

3:32 PM

Page 118

EVYATAR ERELL

Huang, Y. J., H. Akbari, H. Taha and A. H. Rosenfeld (1987) ‘The potential of vegetation in reducing summer cooling loads in residential buildings’, Journal of Climate and Applied Meteorology, vol 26, pp1103–1116 Jauregi, E. (1990–1991) ‘Influence of a large urban park on temperature and convective precipitation in a tropical city’, Energy and Buildings, vol 15–16, pp457–463 Kanda, M. and R. Moriwaki (1998) ‘Environmental effect of Meiji-shrine forest as a sink of energy and pollutants’, Second Urban Climate Symposium, Albuquerque, New Mexico Kastner-Klein, P. and E. J. Plate (1999) ‘Wind-tunnel study of concentration fields in street canyons’, Atmospheric Environment, vol 33, pp3973–3979 Kenworthy, A. T. (1985) ‘Wind as an influential factor in the orientation of the orthogonal street grid’, Building and Environment, vol 20, no 1, pp33–38 Kim, J. and J. Baik (1998) ‘A numerical study of thermal effects on flow and pollutant dispersion in urban street canyons’, Second Urban Environment Symposium, Albuquerque, New Mexico Kim, J. and J. Baik (2001) ‘Urban street canyon flows with bottom heating’, Atmospheric Environment, vol 35, pp3395–3404 Kim, J. and J. Baik (2004) ‘A numerical study of the effects of ambient wind direction on flow and dispersion in urban street canyons using the RNG k-ε turbulence model’, Atmospheric Environment, vol 38, pp3039–3048 Kjelgren, R. and T. Montague (1998) ‘Urban tree transpiration over turf and asphalt surfaces’, Atmospheric Environment, vol 32, no 1, pp35–41 Knowles, R. (1981) Sun, Rhythm, Form, The MIT Press, Cambridge, MA Knowles, R. (2003) ‘The solar envelope: Its meaning for energy and buildings’, Energy and Buildings, vol 35, pp15–25 Kolokotroni, M., I. Giannitsaris and R. Watkins (2006) ‘The effect of the London urban heat island on building summer cooling demand and night ventilation strategies’, Solar Energy, vol 80, pp383–392 Krishan, A. (1996) ‘The habitat of two deserts in India: Hot-dry desert of Jaisalmer (Rajasthan) and the cold-dry high altitude mountainous desert of Leh (Ladakh)’, Energy and Buildings, vol 23, no 3, pp217–229 Landsberg, H. E. (1968) ‘Climates and urban planning’, Symposium on Urban Climates and Building Climatology, Brussels, Belgium Levinson, R., H. Akbari and J. Reilly (2007) ‘Cooler tile-roofed buildings with near-infrared reflective non-white coatings’, Building and Environment, vol 42, pp2591–2605 Littlefair, P. (1998) ‘Passive solar urban design: Ensuring the penetration of solar energy into the city’, Renewable and Sustainable Energy Reviews, vol 2, pp303–326 Livada, I., M. Santamouris, K. Niachou, N. Papanikolaou and G. Mihalakakou (2002) ‘Determination of places in the great Athens area where the heat island effect is observed’, Theoretical and Applied Climatology, vol 71, pp219–230 Manins, P. C., M. E. Cope, P. J. Hurley, P. W. Newton, N. C. Smith and M. Lo (1998) ‘The impact of urban development on air quality and energy use’, 14th International Clean Air & Environment Conference, Melbourne, Australia, 18–22 October 1998 Mazouz, S. and M. S. Zerouala (1998) ‘Shading as a modulator for the design of urban layouts based on vernacular experiences’, Energy and Buildings, vol 29, pp11–15 McPherson, E. G., L. P. Herrington and G. M. Heisler (1988) ‘Impacts of vegetation on residential heating and cooling’, Energy and Buildings, vol 12, pp41–51 Meier, A. K. (1990–1991) ‘Strategic landscaping and air-conditioning savings: A literature review’, Energy and Buildings, vol 15–16, pp479–486 Mills, G. (1997) ‘The radiative effects of building groups on single structures’, Energy and Buildings, vol 25, pp51–61 Mills, G. (1999) ‘Urban climatology and urban design’, 15th International Congress of Biometeorology and International Conference on Urban Climatology, Sydney, Australia Mills, G. (2006) ‘Progress toward sustainable settlements: A role for urban climatology’, Theoretical and Applied Climatology, vol 84, pp69–76

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 119

The Application of Urban Climate Research in the Design of Cities

119

Nakamura, Y. and T. R. Oke (1988) ‘Wind, temperature and stability conditions in an east–west oriented urban canyon’, Atmospheric Environment, vol 22, no 12, pp2691–2700 Niachou, A., K. Papakonstantinou, M. Santamouris, A. Tsangrassoulis and G. Mihalakakou (2001) ‘Analysis of the green roof thermal properties and investigation of its energy performance’, Energy and Buildings, vol 33, pp719–729 Nunez, M. and T. R. Oke (1977) ‘The energy balance of an urban canyon’, Journal of Applied Meteorology, vol 16, pp11–19 Oke, T. R. (1973) ‘City size and the urban heat island’, Atmospheric Environment, vol 7, no 8, pp769–779 Oke, T. R. (1981) ‘Canyon geometry and the nocturnal urban heat island: Comparison of scale model and field observations’, Journal of Climatology, vol 1, no 3, pp237–254 Oke, T. R. (1982) ‘The energetic basis of the urban heat island’, Quarterly Journal of the Royal Meteorological Society, vol 108, no 455, pp1–24 Oke, T. R. (1984) ‘Towards a prescription for the greater use of climatic principles in settlement planning’, Energy and Buildings, vol 7, pp1–10 Oke, T. R. (1987) Boundary Layer Climates, Methuen, London and New York Oke, T. R. (1988a) ‘Street design and urban canopy layer climate’, Energy and Buildings, vol 11, pp103–113 Oke, T. R. (1988b) ‘The urban energy balance’, Progress in Physical Geography, vol 12, no 4, pp471–508 Oke, T. R. (1989) ‘The micrometeorology of the urban forest’, Philosophical Transactions of the Royal Society London B, vol 324, pp335–349 Olgyay, V. (1963) Design with Climate, Princeton University Press, Princeton, NJ Onmura, S., M. Matsumoto and S. Hokoi (2001) ‘Study on the evaporative cooling effect of roof lawn gardens’, Energy and Buildings, vol 33, pp653–666 Page, J. K. (1968) ‘The fundamental problems of building climatology considered from the point of view of decision-making by the architect and urban designer’, Symposium on Urban Climates and Building Climatology, Brussels, Belgium Page, J. K. (1972) ‘The problem of forecasting the properties of the built environment from the climatological properties of the green-field site’, in Taylor, J. A. (ed) Weather Forecasting for Agriculture and Industry, David & Charles, London Palomo Del Barrio, E. (1998) ‘Analysis of the green roofs cooling potential in buildings’, Energy and Buildings, vol 27, pp179–193 Papadakis, G., P. Tsamis and S. Kyritsis (2001) ‘An experimental investigation of the effect of shading with plants for solar control of buildings’, Energy and Buildings, vol 33, pp831–836 Pearlmutter, D., A. Bitan and P. Berliner (1999) ‘Microclimatic analysis of “compact” urban canyons in an arid zone’, Atmospheric Environment, vol 33, pp4143–4150 Pearlmutter, D., P. Berliner and E. Shaviv (2006) ‘Physical modeling of the pedestrian energy exchange within the urban canopy’, Building and Environment, vol 41, pp783–795 Pearlmutter, D., P. Berliner and E. Shaviv (2007) ‘Integrated modeling of pedestrian energy exchange and thermal comfort in urban street canyons’, Building and Environment, vol 42, pp2396–2409 Potchter, O. (1988) ‘Climatic aspects in rural settlement development in hot, arid zones: A case study of the Central Jordan Valley’, Energy and Buildings, vol 11, pp73–90 Pressman, N. (1996) ‘Sustainable winter cities: Future directions for planning, policy and design’, Atmospheric Environment, vol 30, no 3, pp521–529 Rafailidis, S. (1997) ‘Influence of building areal density and roof shape on the wind characteristics above a town’, Boundary Layer Meteorology, vol 85, pp225–271 Ratti, C. and P. Richens (1999) ‘Urban texture analysis with image processing techniques’, Computers in Building: Proceedings of the CAADfutures 1999 Conference, Atlanta, Georgia, 7–8 June 1999 Ratti, C. and P. Richens (2004) ‘Raster analysis of urban form’, Environment and Planning B: Planning and Design, vol 31, pp297–309

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

120

3:32 PM

Page 120

EVYATAR ERELL

Ratti, C., D. Raydan and K. Steemers (2003) ‘Building form and environmental performance: Archetypes, analysis and an arid climate’, Energy and Buildings, vol 35, pp49–59 Ratti, C., N. Baker and K. Steemers (2005) ‘Energy consumption and urban texture’, Energy and Buildings, vol 37, pp762–776 Ratti, C., S. Di Sabatino and R. Britter (2006) ‘Urban texture analysis with image processing techniques: Winds and dispersion’, Theoretical and Applied Climatology, vol 84, pp77–90 Raza, S. H., M. S. R. Murthy, O. B. Lakshmi and G. Shylaja (1990–1991) ‘Effect of vegetation on urban climate and healthy urban colonies’, Energy and Buildings, vol 15–16, pp487–491 Robinson, D. and A. Stone (2004) ‘Solar radiation modelling in the urban context’, Solar Energy, vol 77, pp295–309 Runnalls, K. and T. R. Oke (1998) ‘The urban heat island of Vancouver, BC’, Second Urban Environment Symposium, Albuquerque, New Mexico, 2–6 November 1998 Runnalls, K. and T. R. Oke (2000) ‘Dynamics and controls of the near-surface heat island of Vancouver, British Columbia’, Physical Geography, vol 21, no 4, pp283–304 Sailor, D. J. (1998) ‘Simulations of annual degree day impacts of urban vegetative augmentation’, Atmospheric Environment, vol 32, no 1, pp43–52 Sailor, D. and H. Fan (2002) ‘Modeling the diurnal variability of effective albedo for cities’, Atmospheric Environment, vol 36, pp713–725 Saito, I., O. Ishihara and T. Katayama (1990–1991) ‘Study of the effect of green areas on the thermal environment in an urban area’, Energy and Buildings, vol 15–16, pp493–498 Santamouris, M. (1998) ‘The Athens urban climate experiment’, PLEA 1998: Environmentally Friendly Cities, Lisbon, Portugal, 1–4 June 1998 Santamouris, M. (ed) (2001) Energy and Climate in the Urban Built Environment, James & James, London Santamouris, M., N. Papanikolaou, I. Livada, I. Koronakis, C. Georgakis, A. Argiriou and D. N. Assimakopoulos (2001) ‘On the impact of urban climate on the energy consumption of buildings’, Solar Energy, vol 70, no 3, pp201–216 Schiller, G. and R. Karschon (1974) ‘Microclimate and recreational value of tree plantings in deserts’, Landscape Planning, vol 1, pp329–337 Shashua-Bar, L. and M. Hoffman (2000) ‘Vegetation as a climatic component in the design of an urban street: An empirical model for predicting the cooling effect of urban green areas with trees’, Energy and Buildings, vol 31, no 3, pp221–235 Shashua-Bar, L. and M. Hoffman (2002) ‘The Green CTTC model for predicting the air temperature in small urban wooded sites’, Building and Environment, vol 37, pp1279–1288 Shashua-Bar, L. and M. Hoffman (2003) ‘Geometry and orientation aspects in passive cooling of canyon streets with trees’, Energy and Buildings, vol 35, pp61–68 Shashua-Bar, L. and M. Hoffman (2004) ‘Quantitative evaluation of passive cooling of the UCL microclimate in hot regions in summer, case study: Urban streets and courtyards with trees’, Building and Environment, vol 39, pp1087–1099 Simpson, J. R. (2002) ‘Improved estimates of tree-shade effects on residential energy use’, Energy and Buildings, vol 34, no 10, pp1067–1076 Simpson, J. R. and E. G. McPherson (1998) ‘Simulation of tree shade impacts on residential energy use for space conditioning in Sacramento’, Atmospheric Environment, vol 32, no 1, pp69–74 So, E., A. Chan and A. Wong (2005) ‘Large-eddy simulations of wind flow and pollutant dispersion in a street canyon’, Atmospheric Environment, vol 39, pp3573–3582 Soler, M. R. and C. Ruiz (1994) ‘Urban albedo derived from direct measurements and LANDSAT 4 TM satellite data’, International Journal of Climatology, vol 14, pp925–931 Spronken-Smith, R. A. and T. R. Oke (1998) ‘The thermal regime of urban parks in two cities with different summer climates’, International Journal of Remote Sensing, vol 19, no 11, pp2085–2104

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 121

The Application of Urban Climate Research in the Design of Cities

121

Spronken-Smith, R. and T. R. Oke (1999) ‘Scale modelling of nocturnal cooling in urban parks’, Boundary Layer Meteorology, vol 93, pp287–312 Steemers, K., N. Baker, D. Crowther, J. Dubiel, M.-H. Nikolopoulou and C. Ratti (1997) ‘City texture and microclimate’, Urban Design Studies, vol 3, pp25–50 Steinecke, K. (1999) ‘Urban climatological studies in the Reykjavik subarctic environment, Iceland’, Atmospheric Environment, vol 33, pp4157–4162 Swaid, H. (1992) ‘Intelligent urban forms (IUF): A new climate-concerned urban planning strategy’, Theoretical and Applied Climatology, vol 46, pp179–191 Synnefa, A., M. Santamouris and I. Livada (2006) ‘A study of the thermal performance of reflective coatings for the urban environment’, Solar Energy, vol 80, pp968–981 Synnefa, A., M. Santamouris and K. Apostolakis (2007) ‘On the development, optical properties and thermal performance of cool colored coatings for the urban environment’, Solar Energy, vol 81, no 4, pp488–497 Taha, H. G. (1978) An Urban Micro-Climate Model for Site-Specific Building Energy Simulation: Boundary Layers, Urban Canyon and Building Conditions, PhD thesis, Department of Architecture, University of California, Berkeley, CA Taha, H. (1997) ‘Urban climates and heat islands: Albedo, evapotranspiration and anthropogenic heat’, Energy and Buildings, vol 25, pp99–103 Taha, H., H. Akbari, A. Rosenfeld and J. Huang (1988) ‘Residential cooling loads and the urban heat island – the effects of albedo’, Building and Environment, vol 23, no 4, pp271–283 Taha, H., S. Douglas and J. Haney (1997) ‘Mesoscale meteorological and air quality impacts of increased urban albedo and vegetation’, Energy and Buildings, vol 25, pp169–177 Tsangrassoulis, A. and M. Santamouris (2003) ‘Numerical estimation of street canyon albedo consisting of vertical coated glazed facades’, Energy and Buildings, vol 35, pp527–531 Upmanis, H. and D. Chen (1999) ‘Influence of geographical factors and meteorological influences on nocturnal urban-park temperature differences – a case study of summer 1995 in Goteborg, Sweden’, Climate Research, vol 13, pp125–139 Upmanis, H., I. Eliasson and S. Lindqvist (1998) ‘The influence of green areas on nocturnal temperatures in a high latitude city (Goteborg, Sweden)’, International Journal of Climatology, vol 18, pp681–700 Urano, A., Y. Morikawa and M. Nishimura (1995) ‘Urban environment analysis of the Kanto Plain with 1-dimensional and 3-dimensional numerical simulations’, Pan Pacific Symposium on Building and Urban Environmental Conditioning in Asia, Nagoya, Japan, March 1995 Vitruvius (1999) Ten Books on Architecture, Cambridge University Press, Cambridge, UK Westerberg, U. and M. Glaumann (1990/1991) ‘Design criteria for solar access and wind shelter in the outdoor environment’, Energy and Buildings, vol 15–16, pp425–431 Wong, N. H., Y. Chen, C. L. Ong and A. Sia (2003) ‘Investigation of the thermal benefits of rooftop garden in the tropical environment’, Building and Environment, vol 38, pp261–270 Xie, X., Z. Huang, J. Wang and Z. Xie (2005) ‘Thermal effects on vehicle emission dispersion in an urban street’, Transportation Research Part D, vol 10, pp197–212 Xie, X., C. Liu, D. Leung and M. Leung (2006) ‘Characteristics of air exchange in a street canyon with ground heating’, Atmospheric Environment, vol 40, pp6396–6409 Yamartino, R. J. and G. Wiegand (1986) ‘Development and evaluation of simple models for the flow, turbulence and pollutant concentration fields within an urban street canyon’, Atmospheric Environment, vol 20, pp2137–2156 Yamashita, S., K. Sekine, M. Shoda, K. Yamashita and Y. Hara (1986) ‘On relationships between heat island and sky view factor in the cities of Tama River basin, Japan’, Atmospheric Environment, vol 20, no 4, pp681–686 Yezioro, A., I. G. Capeluto and E. Shaviv (2006) ‘Design guidelines for appropriate insolation of urban squares’, Renewable Energy, vol 31, pp1011–1023

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 95–121

ABER_CH_04

6/12/08

3:32 PM

Page 122

ABER_CH_05

6/12/08

3:32 PM

Page 123

publishing for a sustainable future

5

Solar Air Conditioning: A Review of Technological and Market Perspectives S. Oxizidis and A. M. Papadopoulos

Abstract The close relation between maximum insolation, and a building’s cooling loads and peak electricity demand indicates that solar-assisted cooling (SAC) is an interesting method of successfully handling this issue. Furthermore, from an environmental viewpoint, SAC arises as a promising alternative to conventional, electrically driven air conditioning since it results in decreased carbon dioxide (CO2) emissions and does not use of ozonedepleting or greenhouse gas (GHG) refrigerants. The utilization of solar energy in air conditioning may in this sense be the challenge for the next major growth of the solar thermal market. Nevertheless, there is a long way to go, and SAC technologies must prove that they are competitive. This chapter discusses the principles and the state of the art of SAC. Technical and non-technical advantages and barriers are also presented, as well as the possibility of using active SAC technologies, and technical issues raised by the use of electrically driven conventional air-conditioning systems are identified. Finally, by utilizing strengths, weaknesses, opportunities, threats (SWOT) analysis, conclusions are drawn with respect to measures and strategies for enhancing SAC technologies.

■ Keywords – cooling; solar systems; urban environment; SWOT analysis

INTRODUCTION Solar-assisted cooling appears to be a promising alternative to conventional electrically driven air-conditioning units from an environmental point of view since it results in decreased CO2 emissions and, in the case of prevailing solar-cooling technologies, the elimination of chlorofluorocarbons (CFCs) and hydro-chlorofluorocarbons (HCFCs). In addition, SAC can contribute significantly to dealing with continuously increasing peak electricity loads during summer, which burden both the generation and transfer capacities of electricity utility companies, as well as consumers. The heat-driven, solar-assisted chillers seem, in this sense, likely to have an excellent potential in the space

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157 doi:10.3763/aber.2008.0205 ■ © 2008 Earthscan ■ ISSN 1751-2549 (Print), 1756-2201 (Online) ■ www.earthscanjournals.com

ABER_CH_05

6/12/08

124

3:32 PM

Page 124

S. OXIZIDIS AND A. M. PAPADOPOULOS

NOMENCLATURE

Q T H S W COP cp

heat (kJ) temperature (°C) enthalpy (kJ) entropy (kJ/K) work (kJ) coefficient of performance specific heat (kJ/kgK)

GREEK

∆ η

change in quantity efficiency

SUBSCRIPTS

des sor ev con dec syn M L H g c l s HE HP R

desorption sorption evaporation condenser decomposition synthesis medium low high gas complex liquid solid heat engine heat pump refrigerator

air-conditioning businesses, although so far they are not really a competitive alternative, in market terms, to conventional vapour compression cooling (VCC) systems. Increased environmental concern and sensitivity during recent years, combined with technological development in solar energy use and continuous urbanization, have given rise to new priorities for the building sector (Santamouris, 2001). The rapid industrialization and urbanization of the second half of the 20th century led to a dramatic increase in the size and density of cities worldwide, leading to a corresponding upsurge in energy consumption. Thus, energy efficiency within buildings must take into account the particular requirements and constraints of urban buildings themselves, as well as the urban microclimate and environment. In this sense, there are two major concerns that occupy the interest of researchers and decision-makers: first, the urban environment that alters the microclimate, putting

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 125

Solar Air Conditioning: A Review

125

a heavy burden on cooling loads and increasing demand for air conditioning in buildings, which is reflected in the sale of air conditioners; and, second, the fact that this burden can only marginally be covered by existing generation plants or the capacity of transport/distribution electricity grids. There are several reasons for the boost in the cooling demand of urban buildings: ● Climatic aspects: ● Urban heat island effect. Higher air temperatures experienced in cities due to the

urban heat island effect are a significant factor of increased demand for cooling (Papadopoulos, 2001; Santamouris et al, 2001). ● Urban microclimate, which reduces the potential of natural cooling. In densely built cities with buildings that create street canyons, wind velocities are very low and vortices lead to the entrapment of hot air masses, thus reducing natural cooling options (Alvarez and Molina, 2003; Geros et al, 2005). ● Global warming. The major feature of climate change is higher temperatures due to global warming. In addition, the frequency and intensity of heatwaves are expected to increase (Cartalis et al, 2001). ● Architectural aspects: ● Buildings oriented to maximize solar gains. After the first oil crisis of the 1970s, efforts focused on the construction of buildings with low energy consumption for heating, disregarding the effect this would have on cooling loads. ● Current trend in modern architecture. The vast majority of office buildings have been, and are being, constructed by making extended use of transparent surfaces, with the consequence that interior solar loads are enormous (Cheung et al, 2005). ● Occupancy and regulatory trends: ● Continuously increasing capacities of electric equipment used in buildings. The excessive use of electric and electronic office and house equipment has a proportional effect to internal energy gains (Katsakos, 2007). ● Rise of living standards – increase in thermal comfort demands. Today’s buildings’ occupants are placing higher demands on thermal comfort, a fact that, combined with stricter regulations for indoor air quality (IAQ), increases the demand for ventilation and, hence, for cooling fresh air (Chatzidimoula, 2004; Avgelis et al, 2005). ● Change of work schedules. Traditionally, in areas with hot summers such as the Mediterranean belt, there used to be a break during noon and work continued in the afternoon. Lately, even in these countries the working hours follow the Western and Northern European profile of a continuous schedule. ● Availability of HVAC systems: ● Diminishing prices for air conditioners. The price of room air conditioners has fallen dramatically, while cheap air conditioners introduced to the market by lesser-known manufactures present low performance coefficients (BSRIA, 2000; Dethman, 2004).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

126

3:32 PM

Page 126

S. OXIZIDIS AND A. M. PAPADOPOULOS

This increased demand for cooling results in a series of problems, which can be divided into problems of global and regional/local scale: ● Global scale: ● Conventional refrigerants present a high ozone-depletion potential (ODP) and

global warming impact (GWI). The hydrofluorocarbon (HFC) and polyfluorocarbon (PFC) refrigerants that replaced HCFCs and CFCs in new chillers may be less drastic, but are still GHGs and are therefore considered responsible for climate change. ● Increased primary energy consumption leads to higher emissions of pollutants from power plants. ● Regional/local scale: ● Summer peak electricity demand cannot be covered by installed plant capacity due to the high cost of new plants and transmission and distribution networks, which private power utilities in the liberalized electricity market are less and less willing to undertake. In addition, environmental restrictions (e.g. the Kyoto Protocol) do not allow for the installation of large fossil fuel-driven generation plants. ● Any active cooling technology – electrically or thermally driven – rejects high quantities of heat enforcing the urban heat island (UHI), therefore adding to the vicious cycle. From the above, one can deduce that refrigeration is one of the major technological challenges that the developed world has to face. In order for this to happen, a reduction in demand is almost mandatory since on the production side (electricity generation) the options are limited. This will lead to a replacement of a vast quantity of vapour compression by alternative cooling technologies. There are several solutions that can be considered as technically viable alternatives, the most important of which are: ● Methods that deal with the issue of increased primary energy consumption: ● energy-conscious design of buildings, featuring sun protection, enhanced thermal

insulation and utilizing the thermal storage properties of the building’s envelope; ● applying low-energy cooling techniques, such as evaporative, radiative and

ground cooling and ventilation; ● active cooling techniques (mainly solar technologies). ● Alternative, but still conventional, technologies and economies of scale: ● cool storage in the building; ● district cooling systems; ● distributed energy generation systems, preferably combined heat and power

systems; ● natural gas-fired cooling systems.

All of these technologies are suitable to address the problem of reducing high peak electricity loads; but the only ones with a true positive environmental impact are the active solar and the low-energy cooling technologies, which result in a lower primary energy

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 127

Solar Air Conditioning: A Review

127

ratio (PER) for the cooling energy to be delivered to the building. However, in an urban context, it is very difficult to utilize low-energy cooling technologies due to the lack of appropriate heat sinks and the constraints imposed by the built environment (e.g. heavily built sites and canyon geometry).

THERMALLY DRIVEN ACTIVE SOLAR COOLING TECHNOLOGIES A heat pump (see Figure 5.1) can operate either by consuming electricity (with an electric compressor) or fossil fuel (with a gas engine) during the production and use of mechanical work (through compression) either through vapour compression heat pumps or by consuming thermal energy through thermally driven heat pumps. Thermally driven pumps differ fundamentally from the vapour compression heat pumps because they convert heat of a given temperature to heat of another temperature without any intermediate use of mechanical work through an electric compressor or gas engine. The term thermal compression is often used to point out the feature of compression by mechanical means. Thermally driven heat pumps utilize sorption phenomenon. Sorption heat pumps can be operated with different working pairs (e.g. lithium bromide/water, zeolite/water, salt/ ammonia or metal/hydrogen) and employ different operating principles, which may be absorption, adsorption or chemisorption (chemical reaction). Although the physical principles of these systems may vary, all sorption systems can be evaluated in a similar way. Sorption cooling technologies are the prevailing option when using solar energy for air-conditioning purposes. Figure 5.2 provides a schematic representation of a simple ideal sorption heat pump. The system is driven by heat input Qdes (heat of desorption or heat of generating vapour) at temperature Tdes (high temperature), releases heat Qsor (heat of sorption) at Tsor (medium High Temperature Heat Sink T

Heat Sink T

QH

QH Mechanical Energy Heat Pump W QC

Medium Temperature Heat Sink T

Thermal Energy Q

QM Heat Pump

QC

Heat Source T

Heat Source T (a)

(b)

Source: original material for this chapter

FIGURE 5.1 Operational principles of heat pumps: (a) mechanically driven; (b) thermally driven

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

128

3:32 PM

Page 128

S. OXIZIDIS AND A. M. PAPADOPOULOS

HIGH TEMPERATURE HEAT SOURCE (SOLAR COLLECTORS) THERMAL COMPRESSOR

Tdes Qdes MEDIUM TEMPERATURE Tcon HEAT SINK (AMBIENT) Qcon

MEDIUM TEMPERATURE HEAT SINK (AMBIENT)

DESORBER

CONDENSER

Tdes Desorption temperature Tsor Sorption temperature Tev Evaporation temperature Tcon Temperature of condensation

Qsor

SORBER

EVAPORATOR

Qdes Qsor Qev Qcon

Tsor

COOLING WATER

Tev

CHILLED WATER

Qev

Heat of desorption LOW TEMPERATURE HEAT SOURCE (BUILDING COOLING SYSTEM) Heat of sorption Heat of evaporation Heat of condensation

Source: original material for this chapter

FIGURE 5.2 Schematic representation of sorption refrigerator

temperature) and Qcon at Tcon (medium temperature), and produces cold Qev at Tev (low temperature). As a result, the temperature profile of the process is Tev < Tcon < Tsor < Tdes, though often when the heat pump works on a cooling mode the extraction of heat during sorption and condensation is often assumed to take place under the same temperature (Tcon = Tsor = Tm) because for both processes, the heat is transferred to the same environmental heat sink.

CLASSIFICATION OF SORPTION HEAT PUMPS In their capacity of driving heat pumps, sorption phenomena can be categorized as gas absorption by a liquid, gas adsorption on a micro-porous solid, or chemical reaction between a gas and a solid. The most important difference between these three processes concerns the nature of the process itself. Adsorption involves the use of solids for removing substances from gaseous solutions. The process of adsorption concerns separation of a substance from one phase, accompanied by its accumulation or concentration at the surface of another. On the other hand, absorption is the process in which material transferred from one phase to another (e.g. liquid) infiltrates the second phase to form a solution. In the case of adsorption, physical adsorption on the surface occurs without modifying the solid skeleton itself. At the surface of solids, unsaturated and relatively weak intermolecular forces of the van der Waals type, as well as electrostatic forces, are in play. Hence, when a solid is exposed to a gas, the gas molecules will form bonds with it and become attached – though not chemically bonded. The inverse process – the release of gas molecules from the solid surface – is termed desorption. When the gas molecules form a chemical bond with the surface of the solid, the phenomenon is called chemisorption (Schawe, 2001). In this case, chemical changes occur that induce modification of the solid itself. This difference is a result of the variance of the equilibrium of each process (see Figure 5.3):

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 129

Solar Air Conditioning: A Review

P

Sorbate Concentration refrigerant (sorbate) liquid / vapor equilibrium line

Qcon Pcon

129

Qgen

Qev

Pev refrigerant vapour (absorbate or absorbate) / liquid (absorbent) or solid (absorbent) desiccant isosteric network

Qabs

(a)

Tev

Tcon

T

Tgen

Tabs /Tabs

P

refrigerant (sorbate) liquid / vapor equilibrium line Qdec

Qcon

reaction between sorbent – sorbate equilibrium line

Pcon Pev

Qsyn

Qev

(b)

Tev

Tcon

Tsyn

Tdec

T

Source: original material for this chapter

FIGURE 5.3 Basic cycle for adsorption and absorption in a (lnP, –1/T) diagram: (a) closed systems; (b) chemical reaction systems ● Absorption is bi-variant and induces no volume modification. ● Adsorption is also bi-variant and induces no volume modification of the solid. ● Chemisorption is mono-variant and induces volume modification of the solid.

When using the sorption phenomena (heating, cooling and heat storage), a whole range of refrigerant fluids is available as working pairs (sorbate/sorbent) that can be useful for a wide range of applications, from cryogenics to high-temperature heat storage. The pairs of interest in the field of sorption for building applications are: ● adsorption: water/zeolite, water/silica gel, methanol/activated carbon and

ammonia/activated carbon; ● absorption: water/lithium bromide and ammonia/water; ● chemical reaction: ammonia/inorganic salts.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

130

3:32 PM

Page 130

S. OXIZIDIS AND A. M. PAPADOPOULOS

All three processes can be utilized in open systems (desiccant cooling), where the elements comprise the sorbent, the sorbate (usually water) and air for heat and mass transport, while in closed systems, there are the sorbent and the sorbate (usually water) in an evacuated system. In open systems, air is the heat and water vapour the transporting fluid, while in closed evacuated systems no fluid supports the heat transfer and either special heat exchangers have to be installed or the heat exchangers have to be coated by a sorption material.

Absorption heat pumps The two main types of absorption systems – with respect to the working fluid – are lithium bromide (LiBr) (absorber)/water (refrigerant) and water (absorber)/ammonia (NH3) (refrigerant). Since both systems use water, their application range is limited to temperatures above 4°C, which is not a problem with regard to air-conditioning applications. On the other hand, concerning the advantages of LiBr/water over water/ammonia systems, the LiBr/water solution is a non-volatile absorbent of LiBr, which eliminates the need for a rectifier, such as in the water/NH3 system. In addition, water, as a refrigerant, has a high vaporization heat. Furthermore, LiBr/water systems present a lower generator operating temperature, which makes them more suitable for solar utilization and, in principle, achieves a higher coefficient of performance (COP), compared to the water/ammonia systems. Two rather minor problems for the LiBr/water solution arise as a result of its proclivity for crystallization – this problem has substantially been reduced through the use of microprocessor and direct digital control (DDC) (Dorgan et al, 1995) – and its corrosiveness to some metals. Several additives can be added to the solution as corrosion inhibitors or in order to improve the heat mass-transfer performance (Srikhirim et al, 2001). LiBr/water systems are categorized by the number of times that the solution is heated to produce refrigerant vapours. This is referred to as the number of effects. A single-effect system uses the heat input once, while the double-effect system uses the heat input for one desorption effect and uses the warm refrigerant vapours as the heat source for the second effect (Dorgan et al, 1995). Triple-effect systems are under research; but no tripleeffect machine is yet available commercially. The water/ammonia solution has been widely used since the invention of the absorption cycle systems for both cooling and heating purposes, although after the introduction of the LiBr/water solution, its use became limited to industrial applications for refrigeration purposes. Both NH3 (refrigerant) and water (absorbent) are highly stable over a wide range of operating temperatures and pressures. Ammonia has a high latent heat of vaporization, which is necessary for the efficient performance of the system. It can be used for low-temperature applications (as low as –50°C) as the freezing point of NH3 is –77°C. Since both NH3 and water are volatile, the cycle requires a rectifier to strip away water that normally evaporates with NH3. Without a rectifier, the water will accumulate in the evaporator and reduce the system’s performance, since it will freeze in the low operating temperatures. There are other disadvantages such as its high pressure, toxicity and corrosive action to copper and copper alloy. However, the water/NH3 system is environmentally friendly and inexpensive (Srikhirim et al, 2001).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 131

Solar Air Conditioning: A Review

131

Water/ammonia systems have several disadvantages compared to LiBr/water systems: their COP, for instance, is lower. Since they require a higher generator inlet temperature, they operate at a 10 to 15 per cent lower solar fraction than LiBr/water systems. Furthermore, they require higher pressures, resulting in higher pumping power. A more complex system requiring a rectifier to separate ammonia and water vapour at the generator outlet is necessary, otherwise captured water will evaporate within the absorption cycle, affecting temperature and pressure. Finally, there are restrictions on in-building applications of water/ammonia cooling units because of the toxicity hazards associated with the use of ammonia (Li and Sumathy, 2000). A common classification for the water/ammonia systems refers to the number of evaporator stages involved. The number of stages describes the number of basic absorption cycles that are combined to create a particular configuration. The number of effects, furthermore, describes how often a major portion of the driving heat is reused within the system in order to provide cooling. As a rule, increasing the number of effects increases efficiency, and increasing the number of stages requires more hardware. However, for a given number of stages, there are only limited numbers of effects possible. For a cycle with n stages, the maximum number of effects is 2n – 1 (Cheung et al, 1996). The most common systems are the single-stage ones, which feature one stage of evaporation/absorption, and the double-stage ones with two stages of evaporation/ absorption.

Adsorption heat pumps The main adsorbents used for air-conditioning purposes are silica gel and zeolite, which use water as a refrigerant, while the pair-activated carbon/methanol, which has been used since 1930, can be employed for refrigeration purposes as well since methanol can be cooled below 0°C, although methanol’s latent heat of vaporization is smaller than water’s. On the other hand, zeolite/water systems demand generation temperatures of 170oC, while the silica gel/water and, to a lesser degree, the active carbon/methanol systems can utilize heat in temperatures below 100°C. Therefore, the most common pair for airconditioning applications is the silica gel/water system. When fixed adsorbent beds are employed, which is the common practice, these cycles can be operational without moving parts other than magnetic valves. This results in low vibrations, mechanical simplicity, high reliability and availability and, consequently, a longer lifetime. The use of fixed beds also results in an intermittent cycle operation, with adsorbent beds changing between adsorption and desorption stages. Hence, if a constant flow of vapour from the evaporator is required in order to provide continuous cooling, two or more adsorbent beds must be operated out of phase (Critoph, 1999). Compared to absorption, adsorption chillers are more expensive and their commercial availability is still limited. In principle, their COP is lower than those of absorption chillers; but they can utilize lower heat temperatures and thus can be driven by flat-plate solar collectors more efficiently.

Chemical reaction (thermochemical) heat pumps (TCHPs) In general, the operational principle of a chemical reaction heat pump is quite similar to that of an adsorption heat pump. The temperature profile of a thermochemical heat

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

132

3:32 PM

Page 132

S. OXIZIDIS AND A. M. PAPADOPOULOS

Source: original material for this chapter

FIGURE 5.4 Temperature profile of a TCHP cycle

pump’s complete cycle is depicted in Figure 5.4, along with the heat fluxes at every step of the cycle: 1 At high temperature Tdec, the decomposition (endothermic reaction) of the solid material S.nG takes place by utilizing high-temperature heat. The reaction leads to the products S.(n-m)G (solid or liquid) and MG (gas). 2 After the products are separated, the solid loses sensible heat to the environment. 3 Concurrently with step 2, the gas loses sensible heat to the environment. 4 The gas condenses at temperature Tcon, releasing heat either to the environmental heat sink or to useful heating applications. 5 The forming liquid loses more sensible heat to the environment. 6 The liquid extracts heat from the environment at temperature Tev to evaporate and increase its enthalpy. 7 The formed gas increases its enthalpy through the addition of more sensible heat. 8 The products of the initial reaction are brought together so that the reverse exothermic reaction can take place, delivering the useful heat (stored as chemical energy) at medium temperature Tsyn. 9 The addition of heat to the solid brings it back to initial conditions, ready to decompose and repeat the cycle.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 133

Solar Air Conditioning: A Review

133

THERMODYNAMICS OF SORPTION HEAT PUMPS As already mentioned, the temperature of sorption and condensation can be the same; thus, Tcon = Tsor = Tm. Adopting this assumption and in relation to the phase diagram of Figure 5.3, an enthalpy–temperature (H–T) relation is depicted in Figure 5.5. In Figure 5.5, the dotted line corresponds to the working cycle of the refrigerant, along with the following processes: (a) Desorption of the complex: the heat of desorption ∆HH(c) is supplied by the hightemperature heat source (solar collectors).

Source: adjusted from Benach (1984)

FIGURE 5.5 Working media paths in a sorption system

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

134

3:32 PM

Page 134

S. OXIZIDIS AND A. M. PAPADOPOULOS

(b) Cooling of sorbent and gaseous working fluid: sensible heat [cp(s)(TH – TM) + cp(g)(TH – TM)] is released to ambient air. (c) Sorption of the complex: the heat of sorption ∆HM(c) is disposed to ambient air. (d) Heating of the complex: sensible heat [cp(c)(TH – TM)] is supplied from a hightemperature heat source (solar collectors). (e) Vaporization of working fluid: latent heat ∆HL(l) is supplied from a low-temperature environmental heat source (building cooling system). (f) Heating of gaseous working fluid: sensible heat [cp(g)(TM – TL)] is supplied. (g) Condensation of working fluid: the heat of condensation ∆HM(g) is disposed to ambient air. (h) Cooling of liquid working fluid: sensible heat [cp(l)(TM – TL)] is disposed. Thus, the calculation of the minimum entropy production for the system is: ⎡ c p (l ) c p (g ) ⎤ ⎡c + c p (s ) c p (c ) ⎤ ⎥ (TH − TM ). ⎥ (TM − TL ) + ⎢ p (g) ∆S = ⎢⎢ − − ⎥ ⎢ TM ⎥⎦ TM TH ⎥⎥⎦ ⎢⎣ ⎢⎣ TL

[1]

The entropy–temperature (S–T) diagrams of Figure 5.6, for a discontinuous process, and Figure 5.7, for a continuous one, depict the work related to the low- and high-temperature cycles of the process presented in Figure 5.5. The Carnot diagram equivalent to the overall cycle of Figure 5.6 is given in Figure 5.8. However, two different modes are distinguished: they are coupled together and assumed to run simultaneously. These two modes can be described as follows: 1 a high-temperature Carnot heat engine, operating between desorption and condensation temperature; 2 a low-temperature heat pump, operating between evaporation and sorption temperature. Thus, relationships for the Carnot efficiencies can be established as follows. For the heat engine cycle: ηHE ≡

T − Tcon WHE = des . Qdes Tdes

[2]

Qsor Tsor = . WHP Tsor − Tev

[3]

For the heat pump cycle: ηHP ≡ For the refrigeration cycle: ηR ≡

Qev Tev = . WHP Tsor − Tev

[4]

And the overall efficiencies (coefficients of performance) derived from the Carnot efficiencies and Figure 5.8 are as follows.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 135

Solar Air Conditioning: A Review

135

Source: original material for this chapter

FIGURE 5.6 S–T diagram for a discontinuous sorption system process

COP for heating mode, assuming that the heat of condensation is not used in heating applications but is disposed to an environmental heat sink: COPHEATING =

− Tcon ) T (T Qsor = sor des . Qdes Tdes (Tsor − Tev )

[5]

COP for heating mode, assuming that the heat of condensation is used for heating applications: COPHEATING =

Qsor + Qcon TsorTdes − TconTev = . Qdes Tdes (Tsor − Tev )

[6]

− Tcon ) T (T Qev = ev des . Qdes Tdes (Tsor − Tev )

[7]

COP for cooling mode: COPCOOLING =

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

136

3:32 PM

Page 136

S. OXIZIDIS AND A. M. PAPADOPOULOS

Source: original material for this chapter

FIGURE 5.7 Entropy–temperature (S–T) diagram for a sorption heat pump

It should be noted again that the above relationships assume temperature-independent heats of sorption and desorption and do not take into account system-related sensible heat. Specifically, the basic factors that contribute to less-than-ideal system performance are as follows (Benach, 1984): ● heat of formation in the sorption process; ● heat capacities of the working substances; ● heat capacities of system construction materials.

The most common thermodynamic requirements related to sorption heat pumps can therefore be summarized as follows: ● The heat output per mass of the working medium should be as high as possible. ● The working medium should remain isochoric during the entire cycle. ● The working medium should preserve its characteristic properties during many

recycling periods. ● Simple regeneration of the working medium should be possible.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 137

Solar Air Conditioning: A Review

DESORBER

Qdes

137

Tdec

High temperature cycle A Carnot Engine (CE) that supplies the work to drive the low temperature cycle

WHE = - WHP

Qcon = Qdes - WHE Tcon

CONDENSER

TM SORBER Qsor = Qev - WHP (Qsor = Qev + WHE)

Tsor Low temperature cycle A Carnot Heat Pump (CHP) or A Carnot Refrigeration (CR) cycle

Qev

EVAPORATOR

Tev

Source: original material for this chapter

FIGURE 5.8 Equivalent Carnot cycles of a sorption heat pump

TECHNICAL AND MARKET CHARACTERISTICS OF SOLAR-ASSISTED COOLING (SAC) SYSTEMS In this section, commercial technologies are considered, even those which have only had a significant number of demonstration projects or have limited applications in niche markets. These technologies are absorption, adsorption and desiccant cooling. A review of applied systems based on these technologies can be found in Sayigh and McVeigh (1992); Papadopoulos et al (2003); Henning (2004); Balaras et al (2007) and Henning (2007b). The main features of these technologies, with respect to their coefficient of performance, application, input required and stage of availability are presented in Table 5.1. Combining the current trends of the available solar collectors and those of refrigeration technologies, a correlation between the two areas can be produced, which is also depicted in Table 5.1. The most important features of SAC technologies with respect to their market applicability are discussed in the following sub-sections.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

TABLE 5.1 Main features of possible solar sorption cooling technologies TECHNOLOGY COP APPLICATION Lithium Single effect 0.5–0.75 AC bromide/water Double effect 0.8–1.2 AC Triple effect 1.3–1.7 AC GAX AC Water/ Single stage 0.5 AC/R ammonia Two stage 1.2–1.3 AC/R GAX 1 AC/R Half effect 0.2–0.3 AC/R Diffusion 0.05–0.2 R One stage (silica gel/water) 0.3–0.7 AC Two stage (silica gel/water) 0.35 AC Two-bed system 0.4–0.6 AC (AC/methanol) Adsorber/collector 0.1–0.35 R Chemical reaction systems 0.3–0.7 AC/R Liquid >1 D Rotary 0.5–1 D Solid FIX BED D

Absorption

C

60

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

Source: original material for this chapter

FP = flat plate; L = laboratory; R = refrigeration; VT = vacuum tube.

Notes: AC = air conditioning; AFC = air flat collectors; C = commercial; CC = concentrating collectors; D = dehumidification;

C L L C 60–90 45–70 45–95

140–200 60–90 50–75

C L L C C L L

C L L C

100–160 160–230 120–150

AVAILABILITY C

INPUT (OC) 80–100

FP

FP FP, AFC FP, AFC

FP

FP, VT

VT VT, CC

COLLECTORS FP

20–350

50–1050

20–5000

CAPACITY (KW)

138

3:32 PM

Adsorption

6/12/08

Desiccant

ABER_CH_05 Page 138

S. OXIZIDIS AND A. M. PAPADOPOULOS

ABER_CH_05

6/12/08

3:32 PM

Page 139

Solar Air Conditioning: A Review

139

SIZE OF SYSTEMS Currently, the size of most systems is quite large, resulting in limited suitability for domestic applications. Nevertheless, apart from the existence of this ‘smallest capacity available’ limit, there is no respective ‘upper capacity’ limit. Solar-assisted cooling (SAC) systems can therefore be installed in the majority of large public buildings (e.g. offices, complexes, big hotels, etc; see Table 5.1 for the range of capacity for cooling systems). Hence, in order for solar-assisted chillers to take advantage of the already wide and mature market of solar collectors, domestic-sized units should be developed. This is the current trend, where a number of manufactures are introducing or are about to introduce low-power systems of less than 20kW capacity (Mugnier, 2006; Henning, 2007a).

TYPES OF WORKING MEDIUM PAIRS USED (SORBENT/REFRIGERANT) There are many parameters that influence the selection of the appropriate working medium pair: the most important are recognized to be temperature and pressure range; thermodynamic efficiency; the history of applications; the complexity in handling the system; safety; and environmental concerns. Despite the plethora of different possibilities, two pairs have until now dominated the market and can be considered as state of the art: H2O/LiBr for air conditioning and NH3/H2O for refrigeration. Both pairs are utilized in absorption technology. With regard to further options, research has focused on improving the characteristics of these well-known pairs in order to overcome their associated problems, and on the use of new working-medium pairs. According to Ziegler (2002) within the first line of approach, several salt mixtures, based on the LiBr system have been proposed. It seems that until now the problem of corrosiveness and poor heat transfer has not yet been solved. The same holds for the change from salt solutions to hydroxides or acids. Within the second line of approach, there are two alternatives: first, organic pairs are often suggested – a good choice if, for any reason, the standard pairs cannot be applied. Nevertheless, they still do not play a significant role. A more radical solution is the switch to solid sorbents (adsorption), where zeolite attracts much attention, although silica-gel chillers are already commercially available. Moreover, the basic research seems to be more active in the field of silica gel, as well as in the field of ammonia salts. Another alternative considered is the use of chemical heat pumps (chemisorption instead of physisorption for adsorption), a perspective that gains more attention in the effort to use thermal energy for cooling provision. An extended review of the whole range of working pairs used in chemical heat pumps and their applications can be found in Wongsuwan et al (2001).

PROSPECTIVE MARKETS The most attractive potential markets for SAC are areas with hot, sunny climatic conditions. However, taking into consideration the trend monitored during the last few years all over Europe – namely, the rapid increase in summer air-conditioning loads – one should not overlook Northern European areas as prospective markets. In addition, increasing environmental sensitivity in these countries, along with the growth of the solar collectors market and the comparatively mild cooling load, has encouraged the installation and use of SAC.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

140

3:32 PM

Page 140

S. OXIZIDIS AND A. M. PAPADOPOULOS

There are various markets for solar air conditioning in developed and developing countries. Concerning developed countries where air-conditioning loads are steadily increasing, SAC has a good potential since it can successfully meet peak electricity loads during the summer, while also fulfilling environmental regulations – for example, concerning ozone-depletion potential (ODP) and global warming potential (GWP). On the other hand, in developing or underdeveloped countries, the lack of an extended electricity grid gives SAC an opportunity to cover current and future air-conditioning loads in a sustainable way. Another market for SAC could be areas with specific problems in electricity transmission systems, such as islands or isolated rural areas. One can also easily deduce that countries with a high number of installed solar collectors, such as Austria, Cyprus, Germany, Greece and Israel, are very likely to embrace SAC since solar energy is already a seasoned alternative to the use of fossil fuels. The same holds for countries such as Japan where sorption refrigeration controls an essential portion of the air-conditioning market.

COMMERCIAL REQUIREMENTS AND CRITERIA As for most heating, ventilation and air conditioning (HVAC) systems, the most important commercial requirements and criteria for SAC systems are considered to be: ● ● ● ● ● ● ● ● ● ● ●

efficiency and partial load characteristics; useful lifetime; turnkey costs; operation and maintenance costs; payback period of the investment; technical availability and reliability; footprint, weight and volume; noise; technical support; safety; environmental performance.

The following sub-sections analyse these requirements for SAC systems, while (wherever possible) a short comparison is provided with vapour compression cooling (VCC) systems.

Efficiency and part-load characteristics As noted in Table 5.1, sorption cooling technologies essentially present lower efficiencies than vapour compression systems. If the solar collectors’ efficiency is also considered, the overall result is rather poor. However, taking into account the efficiency of power plants in the latter systems and the solar fraction in the former, a relative balance is achieved, and in cases where solar fraction exceeds 50 per cent, there are actually savings in terms of primary energy. Nevertheless, this assumption does not consider the parasitic energy consumption (pump energy). Concerning partial load characteristics, SAC systems perform adequately, producing partial load efficiencies that are as good, or even better, than those for full-load conditions

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 141

Solar Air Conditioning: A Review

141

since their effective capacity can be varied by simply changing the rate at which the solutions are pumped through the systems. In addition, since solar radiation is the prime factor in generating cooling loads, the cooling output of SAC systems follows the airconditioning demand of buildings. Another emerging issue concerns the fact that VCC systems consume energy of a much higher quality – namely, electricity – compared to sorption chillers, which use thermal energy, whether it is solar, waste heat or heat converted from chemical, but nonfossil, energy sources.

Lifetime The lifetime of SAC technologies is of the same magnitude as that of conventional systems or other low-energy cooling technologies. In fact, the relative lack of large moving parts and the corresponding absence of vibrations are leading to longer lifetimes compared to a conventional compressor system. The same hypothesis stands for solar collectors as well, since they have already proved to have a lifetime equal to that of sorption chillers. The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE, 1999) suggests 23 years of service life for absorption chillers, the same for centrifugal electric chillers and 20 years for reciprocating electric chillers.

Turnkey costs The initial investment cost for SAC systems is significantly higher than that for vapour compression systems since SAC systems consist of the sorption heat pump, the solar collectors array, the auxiliary heater and the storage system. Moreover, the costs of the chiller units alone are higher because of the larger number of heat exchangers required. In this sense, the cost of cooling towers is also higher because they have to be of a larger capacity in order to dispose of significantly more and higher temperature heat. In addition, it should be emphasized that today, most VCC systems that are installed are air cooled. The installation costs of SAC systems are higher than of those of VCC systems since the former are larger, heavier and more technically sophisticated, although the lack of rotating parts lowers the foundation costs. Figure 5.9 depicts the capital costs of several solar cooling systems (prices refer to the lower-capacity machines). The cheapest system (single-effect absorption chiller with flat-plate collectors) costs almost five times the price of a vapour compression cooling system. It is obvious that the cost of collectors dominates the capital cost of SAC systems; but it should be emphasized that in such systems, solar collectors are usually all-year-round heat sources, providing space heating and, probably, domestic hot water as well. In a thorough economic study by Syed et al (2002), several SAC configurations (including photovoltaic-driven electric chillers) were compared for a 1MW cooling installation. Syed et al (2002) estimated that SAC systems cost between 650 Euros/kW to 730 Euros/kW for single-effect absorption chillers with flat-plate collectors. The latter constitute the larger portion of the total capital cost, amounting to 50 to 60 per cent. The costs of electric chillers were estimated at approximately 215 Euros/kW, 195 Euros/kW

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

142

3:32 PM

Page 142

S. OXIZIDIS AND A. M. PAPADOPOULOS

Source: adjusted from Kim (2007)

FIGURE 5.9 Cost of solar cooling systems v. vapour compression cooling

and 125 Euros/kW for a centrifugal, water-cooled screw and packaged air-cooled screw vapour compression chiller, respectively.

Operation and maintenance costs The operation and maintenance (O&M) costs of SAC sorption systems are lower than those of VCC. Nevertheless, one cannot fail to notice that systems with water as a refrigerant operate at very low pressure levels, close to technical vacuum conditions, which results in some technical complexity that leads to increased maintenance costs. In particular, vacuum maintenance is critical for the smooth function and lifetime of the system, so special care is needed. In addition, SAC systems require an advanced system of monitoring and control. In general, maintenance costs for absorption chillers range from 0.6 to 1.25 times the maintenance costs of vapour compression chillers (Elsafty and Al-Daini, 2002). Syed et al (2002) estimated maintenance costs at 1 and 4 per cent of the combined capital costs of chiller and cooling tower for the absorption and vapour compression chillers, respectively. Concerning fuels costs, SAC systems (electricity for pumps and fans, and natural gas or oil for auxiliary system) are likely to be cheaper than VCC systems (electricity for pumps, fans and compressors). These points, though, are based on the assumption of a solar collector array large enough for primary energy savings to be achieved and of fairly flat rates for electricity. In the case of escalating load tariffs imposed by electric utilities, SAC performs even more economically (Papadopoulos et al, 2006). Where the cost or availability of water is of importance, the issue arises of using water in the cooling towers, which benefits the economic performance of air-cooled VCC systems. Several studies have showed that annual economic performance of SAC systems is poor compared to that of VCC systems. Balaras et al (2007) calculated that for an office building in Madrid, the best SAC system (single-effect absorption chiller with flat-plate collectors) has 15 per cent higher annual costs than the electric compression cooling system.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 143

Solar Air Conditioning: A Review

143

Payback period According to Henning et al (1998) who studied the efficiency of solar sorption cooling systems in three areas in Germany, Denmark and Italy, the lowest payback periods are achieved with desiccant cooling systems. The payback period does not significantly increase for solar fraction values up to some 70 per cent, so SAC can reach high solar fractions without leading to oversized and, hence, disproportionately expensive collector arrays. With respect to SAC, Tsoustos et al (2003) concluded that even in regions with increased air-conditioning demand and solar energy availability, such as Greece, there is a strong need for investment incentives and for conventional-energy taxation, which would help to internalize the full environmental costs of using conventional fuels for cooling.

Technical availability and reliability Once more, the lack of moving parts and the concept of a thermally driven system result in both exceptional technical availability and reliability. Nevertheless, the absence of an extended sales and technical support network is a major disadvantage that has to be tackled.

Footprint, weight and volume Both the footprint and weight of SAC systems constitute a significant drawback. Sorption chillers are heavier and larger than VCC chillers, so the addition of the needed solar collector array really puts a big constraint on the applicability of SAC to buildings. Generally, absorption chillers are almost two times heavier than water-cooled vapour compression chillers. When considering air-cooled chillers, the difference in weight is even larger. Vapour compression chillers also have half the footprint and volume of absorption chillers. Table 5.2 compares the weight, footprint and volume of three 400kW chillers from two major manufactures. In addition, beyond the need for a quite large area/room for the installation of the sorption chiller (in terms of surface and volume), an extended roof area is also needed for the installation of a large collector array if significant energy savings are to be achieved.

Noise Again, the lack of reciprocating and rotating parts results in a vibration-free and noiseless operation that can be important in specific applications. Typical sound levels for sorption

TABLE 5.2 Physical characteristics of several chillers of 400kW cooling capacity CHILLER TYPE SINGLE-EFFECT WATER-COOLED AIR-COOLED VAPOUR ABSORPTION (WATER VAPOUR COMPRESSION FIRED) COMPRESSION Weight 12kg/kW 7kg/kW 5.5kg/kW Footprint 5m2 2.5m2 2.5m2 3 3 Volume 11m 4.5m 4.5m3 Source: Carrier Corporation (2007); Yazaki Energy Systems Inc (2007)

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

144

3:32 PM

Page 144

S. OXIZIDIS AND A. M. PAPADOPOULOS

chillers are 80dB(A), while for vapour compression chillers, levels of 95dB(A) (depending upon the type of compressor: centrifugal, reciprocating or rotary).

Technical support Currently, there is still a lack of practical knowledge of integrated systems covering their whole life cycle of design, engineering, establishment, control, operation, maintenance and, eventually, disposal. This also reflects the absence of skilled and competent technicians.

Safety In principle, SAC systems are considered to be safer than any conventional systems because of the natural working fluids used and the lack of specialized electricity installation.

Environmental performance A clear and sound assessment of the environmental performance of SAC should definitely take into account the use of solar collectors throughout the year, and not just on a seasonal base. Thus, solar thermal energy should be considered to provide hot water, space heating and space cooling. According to this perspective, there are three parameters to be used for assessing environmental performance: 1 primary energy saving, or primary energy ratio; 2 global warming potential; 3 ozone-depletion potential. Considering these three parameters, SAC appears to have a favourable environmental performance when compared to conventional systems. In fact, most SAC systems use water as a refrigerant, in no case are ozone-depleting substances employed and, when combined with an adequate area of solar collectors, the system can result in essential energy savings for both heating and cooling, performing better both from a primary energy ratio and global warming potential point of view. Lamp and Ziegler (1998) determined the solar fraction necessary for a SAC (absorption) system to consume less primary energy compared to a VCC system, given the electricity generation efficiency depicted in Figure 5.10. Balaras et al (2007) estimated that the primary energy saving for several configurations of SAC systems and for a collector array of 3m2/kWcooling varies from 36 to 53 per cent compared to a VCC system for an office building in Madrid. An assessment of environmental performance should include water consumption and thermal pollution caused by the chillers. In terms of this criterion, sorption chillers with a COP of 0.7 reject 2.43 units of heat for every unit of cooling served, while electric chillers with a COP of 3 reject only 1.33 units of heat for every unit of cooling served. The heat that is rejected in sorption chillers, therefore, results in much higher water consumption in these cooling towers.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 145

Solar Air Conditioning: A Review

1.5

145

Single effect

Primary Energy Ratio PER

COP = 3 1

Double effect

0.5 COP = 5

0 0.4 0.6 0.8 0.2 Solar Fraction f / Efficiency of Power Generation η

0

1

Electrical driven compression chillers Solar assisted sorption chillers Source: Lamp and Ziegler (1998)

FIGURE 5.10 Primary energy ratio of solar-assisted absorption and electrically driven vapour compression chillers

TECHNICAL AND NON-TECHNICAL BARRIERS TO COMMERCIALIZATION There are additional barriers that discourage the commercialization of SAC systems.

Technical barriers Beyond the technical issues reported earlier, additionally technical constraints exist in solar-assisted air conditioning. First, SAC technologies have not yet reached an acceptable level of technological sophistication and are still considered to be under development. There is a need, for example, for cost-efficient and operationally friendly heat storage, as well as a backup system. SAC systems demand advanced electronic control systems, a field that is developed enough to allow for the implementation of SAC systems in buildings, though not necessarily to the point of achieving optimized, unobstructed and reliable daily operation over many years in frequently harsh environments. Over the past decade, progress in the field of building management systems and remote control is expected to provide much needed assistance in this sector (Canbay et al, 2004; Hipkiss, 2005).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

146

3:32 PM

Page 146

S. OXIZIDIS AND A. M. PAPADOPOULOS

Non-technical barriers The main non-technical barrier to adopting SAC systems is the fact that solar refrigeration is still not competitive in terms of the energy services provided by the utilities. Specifically, retail electricity prices were fairly low throughout the period of 1995 to 2005, fuelled by competition that followed liberalization in most European countries, where steeply rising time-dependent tariffs where not common. However, this trend has begun to change and prices are not expected to decrease, due to the demand for new investments in power generation capacities and to CO2 emission charges (Richmond and Kaufmann, 2006). Another significant barrier is the relative lack of public awareness of available systems, together with a lack of environmental sensitivity. In addition, the absence of an appropriate sound legislation, encouraging environmentally friendly energy systems (investment incentive, energy tax and cost of externalities), does not allow SAC systems to take advantage of all aspects of their environmental performance when compared to conventional systems. Limited access to funding also presents a barrier to SAC, which is characterized by a high initial investment. The complexity and cost of different national testing and certification procedures are barriers to trade in solar systems. Indeed, the lack of sound certification and of standardization of SAC systems constitutes another barrier. Finally, architects, developers and construction companies are not interested in adopting active or sophisticated systems such as SAC, but rather adopt conventional solutions such as VCC systems.

EXTERNAL FACTORS THAT WILL ACCELERATE COMMERCIALIZATION The most important external factor that can influence the commercialization of SAC is energy costs. An increase in the cost of electricity will accelerate not only technological development, but the acceptability of such systems. More specifically, if more electricity utility companies adopt time-dependent tariffs or demand-charging schemes, this will definitely lead to an increase of interest for SAC systems. An increase in energy prices can also be initiated by environmental policies that include the costs of environmental externalities within conventional fuel prices. In fact, environmental issues are already putting pressure on electric utilities, placing a heavy burden on the construction and operation of new generation plants. Environmental policies are highly influential in determining the success of SAC systems within the air-conditioning market. In the European Union, in particular, a series of actions involving the promotion and propagation of renewable energy sources (RES) not only refer to research and development (R&D), and dissemination and integration of low-energy technologies, but also highlight regulatory measures as well. Of course, the direction of additional funding towards R&D is a crucial element that may prove to be significant since SAC technologies must continue to evolve if they are to be competitive. In addition, regarding private-sector strategies, SAC systems could be the subject of integrated demand-side management programmes by both electricity and gas utilities since, during the summer, the electricity grid is frequently overloaded, in contrast to gas. Covering peak electricity loads appears to be a particularly significant problem in liberalized electricity markets since private utility companies are not willing to invest the huge capital needed to construct power plants that will be working for only days or weeks

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 147

Solar Air Conditioning: A Review

147

per year. Another factor that may contribute to the penetration of SAC within the commercial market is the expansion of energy service companies (ESCOs), as these can more easily meet the high initial capital investment required by SAC systems, compensating for it by the lower annual energy cost.

A SWOT ANALYSIS APPROACH SWOT is a strategic analysis tool, used widely in industrial management since the 1980s, and has two elements: an external analysis of opportunities (O) and threats (T) and an internal analysis of strengths (S) and weaknesses (W). The external analysis focuses on the external environment and the opportunities and threats that this may represent for technology, a product or investment. Opportunities or threats are defined according to individual perceptions, assumptions about what the future holds or which aspect of the environment is considered. The internal analysis examines the strengths and weaknesses of technology in relation to its users and competitive systems. The aim of the SWOT analysis is to produce a TOWS matrix, which contains the strategies that have to be applied in order to face threats and opportunities by utilizing the strengths and improving weaknesses. When applied to a typical industrial product, the SWOT analysis focuses on six main points: 1 2 3 4 5 6

product to be sold; process of selling the product; customer; distribution; financial aspect (prices, costs and investments needed); and administration of the whole scheme (Humphrey, 2004).

In a slightly varying line of approach, SWOT analysis is considered to incorporate Porter’s Five Forces model as an outside-in business unit strategy tool, which is applied by means of identifying the five fundamental competitive forces (Wu and Wu, 2005): 1 entry of competitors (how easy or difficult is it for new entrants to compete; which barriers exist?); 2 threat of substitutes (how easy can our product or service be substituted with cheaper options?); 3 bargaining power of buyers (how strong is the position of buyers; can they work together to order large volumes?); 4 bargaining power of suppliers (how strong is the position of sellers; are there many or only a few potential suppliers; is there a monopoly?); 5 rivalry among existing players (is there strong competition between existing players; is one player dominant or are all equal in strength/size?). Various published sources also consider the government to be an additional sixth competitive force.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

148

3:32 PM

Page 148

S. OXIZIDIS AND A. M. PAPADOPOULOS

TABLE 5.3 Strategic factors of the external environment INCLINES’ LIST

Ozone depletion Climate change (global warming) Climate change (Higher frequency of heatwaves) Stricter legislations on energy performance of buildings Existing legislation on the promotion of renewable energy sources (RES) Expanding markets in developing countries New approach of energy systems in the liberated market Integration of externalities within energy costs Increasing environmental awareness Inability of electric utilities to deliver peak load Environmental and financial constraints for new power plants Possibility of introducing time-dependent tariffs Increasing demand for building air conditioning Development of energy service companies (ESCOs) High penetration of solar thermal collectors in many countries, increasing rates in many others Lack of need for night-time air conditioning in offices and most public buildings

PROBABILITY POSSIBLE EFFECTS OF SUCCESS ON TECHNOLOGY OPPORTUNITIES Large Mean Mean Mean

PRIORITY MATRIX

Mean Mean

Mean

Small

Small

Large

Mean

Large

Large

Large

Large

Large

Mean

Mean

Mean

Small

Small

Mean

Large

Large

Mean

Small

Small

Large

Mean

Large

Large

Mean

Mean

Large

Mean

Large

Large

Mean

Mean

Large

Mean

Mean

Large

Large

Large

Large

Small

Small

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 149

Solar Air Conditioning: A Review

149

TABLE 5.3 Strategic factors of the external environment (Cont’d) INCLINES’ LIST

Liberalization and deregulation of the electricity market can keep electricity costs low Power companies and customers prefer conventional methods (i.e. TES) in order to smooth out peak load demand Domination of the air-conditioning market by vapour compression cooling (VCC) systems Architects and construction companies are not interested in active solar systems

PROBABILITY OF SUCCESS THREATS Mean

POSSIBLE EFFECTS ON TECHNOLOGY

PRIORITY MATRIX

Mean

Mean

Large

Mean

Mean

Large

Mean

Large

Mean

Small

Small

Source: original material for this chapter

TABLE 5.4 Strategic factors of the internal environment INCLINES’ LIST

Lower energy consumption (favours primary energy ratio) Hybrid systems (desiccant – vapour compression cooling (VCC), absorption and adsorption; natural gas cooling) Provide free solar heating during winter as well Modular and easily adapted to conventional heating and cooling systems

PROBABILITY OF SUCCESS STRENGTHS Large

POSSIBLE EFFECTS ON TECHNOLOGY

PRIORITY MATRIX

Large

Large

Large

Mean

Mean

Large

Large

Large

Large

Mean

Mean

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

150

3:32 PM

Page 150

S. OXIZIDIS AND A. M. PAPADOPOULOS

TABLE 5.4 Strategic factors of the internal environment (Cont’d) INCLINES’ LIST Do not use ozone-depletion potential (ODP) refrigerants Do not contribute to global warming impact (GWI) Thermally driven cooling systems already have a relatively essential share in the air-conditioning market Solar thermal collectors are already proven to be commercially sophisticated and sound technology Many applications are not adequately sophisticated Highly sophisticated systems require advanced control and monitoring Low overall efficiency Large footprint; sizeable systems require substantial free space Need for solar energy storage High initial cost Falling prices for VCC systems Very limited commercial representation Lack of sound certification and standardization of systems Not autonomous systems; require auxiliary (backup) system Lack of imaginary strategies and marketing techniques for renewable energy source (RES) technologies Complete absence of skilled and educated technicians

PROBABILITY OF SUCCESS Mean

POSSIBLE EFFECTS ON TECHNOLOGY Mean

PRIORITY MATRIX

Large

Mean

Mean

Mean

Mean

Mean

Mean

Mean

Mean

WEAKNESSES Large

Large

Large

Large

Large

Large

Large Large

Large Mean

Large Large

Mean Large Large Mean

Mean Large Mean Large

Mean Large Mean Mean

Large

Mean

Mean

Large

Mean

Mean

Mean

Mean

Mean

Large

Small

Mean

for maintenance and repair Source: Original material for this chapter

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

Mean

Hybrid systems (desiccant/ vapour compression cooling, absorption and adsorption/NG cooling)

OPPORTUNITIES

Many applications are not adequately sophisticated

Intensify R&D programmes on SAC systems

Tax (or other) penalties on high (primary) energy-consuming buildings

Provide free solar heating during winter

Low overall efficiency

WEAKNESSES AND OPPORTUNITIES (WO) STRATEGIES

Highly sophisticated systems require advanced control and monitoring

WEAKNESSES High initial costs

Subsidize RES systems

Imposition of restrictions on, and promotion of solar-assisted cooling (SAC) systems to professional actors (architects and construction companies)

Large footprint; sizeable systems require substantial free space

3:32 PM

STRENGTHS AND OPPORTUNITIES (SO) STRATEGIES

Lower energy consumption (favourable primary energy ratio)

STRENGTHS

6/12/08

Existing legislation on the promotion of renewable energy sources (RES)

Stricter legislations on energy performance of building

INTERNAL FACTORS EXTERNAL FACTORS

TABLE 5.5 Strategies for development (TOWS matrix)

ABER_CH_05 Page 151

Solar Air Conditioning: A Review 151

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

High penetration of solar thermal collectors in many countries, increasing rates in many others

Adoption of marketing techniques by new services that can be provided by solar thermal collectors

Introduction of environmental taxes (green taxes) on the price of energy

Incentives to replace old collectors with higher efficiency

Financing R&D projects for cheaper collectors; propagation of solar collectors in order to reduce costs

Allowances and incentives for energy service companies (ESCOs) in delivering energy management within buildings

Involvement of private companies (utilities and manufactures of solar collectors and refrigeration equipment) in R&D and promotional programmes

3:32 PM

Possibility of introducing time-dependent tariffs

Introduction of environmental taxes (green taxes) on the price of energy

WEAKNESSES

152

Promotion and provision of incentives for demand-side management (DSM) programmes from both electric and gas utilities

STRENGTHS

6/12/08

Inability of electric utilities to deliver peak load

INTERNAL FACTORS EXTERNAL FACTORS Integration of externalities within energy costs

TABLE 5.5 Strategies for development (TOWS matrix) (Cont’d)

ABER_CH_05 Page 152

S. OXIZIDIS AND A. M. PAPADOPOULOS

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

Tax penalties on the price of low-efficiency conventional air-conditioning systems

Domination of the air-conditioning market by VCC systems

THREATS

3:32 PM

Tax penalties on the price of low-efficiency conventional air-conditioning systems

Determination of high cost for blackouts or inability to provide electricity

WEAKNESSES AND THREATS (WT) STRATEGIES 6/12/08

Source: original material for this chapter

Price penalties on large electricity consumers (due to cooling or other uses)

Liberalization and deregulation of electricity market can keep electricity costs low

STRENGTHS AND THREATS (ST) STRATEGIES

ABER_CH_05 Page 153

Solar Air Conditioning: A Review 153

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

154

3:32 PM

Page 154

S. OXIZIDIS AND A. M. PAPADOPOULOS

Nevertheless, when the SWOT methodological approach is being expanded from a product to a technology and a market branch, one has to consider factors such as the customer’s awareness, the technology’s image, the importance of its attributes to the potential customers, and the importance of its attributes to decision- and policymakers who determine the boundary conditions of the building sector as a market (Terrados et al, 2007). Furthermore, one has to keep in mind that the complexity of the building market’s response to technological changes makes the assessment of impacts difficult (Tsoutsos et al, 2003). In this sense, SWOT analysis can be useful since it can provide a framework for medium- and long-term planning, especially if economies of scope are to be achieved. This framework takes the shape of a TOWS matrix, providing the results of analysis of the solar cooling market (Papadopoulos et al, 2004) Tables 5.3 to 5.5 present the results derived from the SWOT analysis of SAC technologies, which takes into consideration the features described in this section. The strategic factors of the external and internal environment are presented, as well as a resulting TOWS matrix of the proposed development strategies.

CONCLUSIONS Solar cooling, as an exclusive solution or assisted by a conventional cooling system, cannot be considered a competitive financial and technological alternative to conventional airconditioning systems, at least under the prevailing conditions in the construction and energy markets of most countries. As a result, one can understand its difficulties in entering competitive, conservative and demanding building services’ markets. On the other hand, solar air conditioning potentially offers an elegant model for a clean, sustainable technology, which is consistent with the international commitment to sustainable development. The increasing effort in applied research during the last decade, combined with the collateral advantages arising from the use of solar cooling technologies, together attest to the high potential for commercial applicability. The main advantages of solar airconditioning technologies can by summarized in the: ● reduction of summer peak loads in electricity demand; ● close seasonal and hourly coincidence of solar radiation and a building’s cooling

demand profile; ● absence of any ozone-depleting refrigerants; ● decreased primary energy consumption; ● decreased global warming impact; and ● maximization of the exploitation of solar energy for heating, hot-water production and

cooling by the same solar system. According to this line of thought, the option of solar cooling becomes an increasingly attractive one. Nevertheless, there are significant barriers towards its commercialization. Targeted actions should be adopted concerning R&D, dissemination and promotion of technology, providing effective incentives to users. Another set of actions may refer to the dominant technology of electrically driven VCC in terms of introducing environmental

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 155

Solar Air Conditioning: A Review

155

taxes or penalties for low overall energy efficiency in order to offset the hidden costs of these conventional systems. Nevertheless, the real momentum that solar-assisted cooling can take advantage of is the substantial increase in demand for air conditioning in urban areas over the last decade. It is not logical to expect that this increasing demand can be covered by electricity utilities without incurring a significant economic and environmental penalty. It is therefore reasonable to expect that a significant part of this demand will redefine the market, providing access for new technologies and approaches, including SAC systems. This prospect can offset the current financial unattractiveness of SAC since electric utilities will be forced to shift their marginal costs that are due to increased peak loads to consumers. SAC is a favourable tool for demand-side management in relation to both electric and natural gas utilities. Furthermore, the activities of energy service provision schemes, especially in relation to the management of large buildings, is another tool for propagating SAC technologies in the market, especially since these companies can ensure the financing of the high initial capital investment required by SAC systems. Reforms in the energy market may lead to a new framework for the entire sector: investors, energy producers, brokers and service companies become active in an expanding, more competitive market, utilizing new technologies and taking advantage of developing primary resource potential. But even in this liberalized environment, some form of regulation is necessary. The success story of wind power and the growing popularity of photovoltaics are good examples for such a balancing act: the market has to find its way; but mature technologies have to be available and this can only be done with support during their infancy period. Nevertheless, climate change policies cannot be effective when solely based on economic instruments and technological advances. A broader social consensus is needed for any regulatory measure to succeed.

AUTHOR CONTACT DETAILS S. Oxizidis: (corresponding author) Laboratory of Heat Transfer and Environmental Engineering, Department of Mechanical Engineering, Aristotle University Thessaloniki, Box 483, 54124 Thessaloniki, Greece; tel (30) 2310 996048; fax (30) 2310 996012; [email protected] A. M. Papadopoulos: Laboratory of Heat Transfer and Environmental Engineering, Department of Mechanical Engineering, Aristotle University Thessaloniki, Box 483, 54124 Thessaloniki, Greece

ACKNOWLEDGEMENTS The authors wish to thank the European Commission of Training and Mobility of Researchers Programme (DG Research) for funding the ATREUS project under Research Training Networks, HPRN-CT-2002-00207.

REFERENCES Alvarez, S. and J. L. Molina (2003) ‘Cooling by natural sinks’, in Santamouris, M. (ed) Solar Thermal Technologies for Buildings: The State of the Art, James & James, London, pp140–162

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

156

3:32 PM

Page 156

S. OXIZIDIS AND A. M. PAPADOPOULOS

ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) (1999) ‘HVAC applications’, in ASHRAE Handbook, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA Avgelis, A., S. Oxizidis, A. Stylianou and A. M. Papadopoulos (2005) ‘Energy consumption and indoor quality relationship in a naturally ventilated building’, in Proceedings of the Urban Air Quality Conference, Valencia, Spain, 29–31 March Balaras, C.-A., G. Grossman, H.-M. Henning and C.-A.-I. Ferreira (2007) ‘Solar air conditioning in Europe – an overview’, Renewable and Sustainable Energy Reviews, vol 11, pp299–304 Benach, R. (1984) Chemical Heat Pump/Thermochemical Energy Storage Systems – I: A Reconnaissance, TNO, The Hague BSRIA (2000) World Market for Air Conditioning Update 2001, www.bsria.co.uk/, accessed June 2006 Canbay, C. S., A. Hepbasli and G. Gokcen (2004) ‘Evaluating performance indices of a shopping centre and implementing HVAC control principles to minimize energy usage’, Energy and Buildings, vol 36, no 6, pp587–598 Carrier Corporation (2007) Performance data at www.carrier.com, accessed 31 October 2007 Cartalis, C., A. Synodinou, M. Proedrou, A. Tsangrassoulis and M. Santamouris (2001) ‘Modifications of energy demand in urban areas as a result of climate changes: An assessment for the southeast Mediterranean region’, Energy Conversion & Management, vol 42, pp1647–1656 Chatzidimoula, E.–L. (2004) Contribution to the Natural Ventilation of Internal Spaces via Controlled Use of Ventilation Openings, PhD thesis, Aristotle University of Thessaloniki, Department of Civil Engineering, Thessaloniki, Greece Cheung, K., Y. Hwang, J. Judge, K. Kolos, A. Singh and R. Radermacher (1996) ‘Performance assessment of multistage absorption cycles’, International Journal of Refrigeration, vol 19, pp473–481 Cheung, C. K., R. J. Fuller and M. B. Luther (2005) ‘Energy-efficient envelope design for high-rise apartments’, Energy and Buildings, vol 37, pp37–48 Critoph, R. E. (1999) ‘Rapid cycling solar/biomass powered adsorption refrigeration system’, Renewable Energy, vol 16, pp673–678 Dethman, L. (ed) (2004) Energy Star Home Products Program: Market Evaluation Report No 2, Seattle, WA Dorgan, C. B., S. P. Leight and C. E. Dorgan (1995) Application Guide for Absorption Cooling/Refrigeration Using Recovered Heat, ASHRAE, Atlanta, GA Elsafty, A. and A. J. Al-Daini (2002) ‘Economical comparison between a solar-powered vapour absorption air-conditioning system and a vapour compression system in the Middle East’, Renewable Energy, vol 25, pp569–583 Geros, V., M. Santamouris, S. Karatasou, A. Tsangrassoulis and N. Papanikolaou (2005) ‘On the cooling potential of night ventilation techniques in the urban environment’, Energy and Buildings, vol 37, pp243–257 Henning, H. M. (ed) (2004) Solar-Assisted Air-Conditioning in Buildings: A Handbook for Planners, Springer Wien, New York, NY Henning, H.-M. (2007a) ‘Solar cooling’, in Proceedings of ISES Solar World Congress 2007, Beijing, China Henning, H.-M. (2007b) ‘Solar assisted air conditioning of buildings – an overview’, Applied Thermal Engineering, vol 27, pp1734–1749 Henning, H. M., T. Erpenbeck, C. Hindenburg and S. Paulussen (1998) ‘Solar cooling of buildings – possible techniques, potential and international development’, in Proceedings of Eurosun 1998, Ljubljana, vol III, pp2.20-1–2.20-6 Hipkiss, R. (2005) ‘Energy management needs intelligence’, Energy World, vol 328, pp20–21 Humphrey, W. S. (2004) The Personal Software Process – Technical Report, www.sei.cmu.edu/pub/documents/00.reports/pdf/00tr022.pdf/, accessed June 2004 Katsakos, K. (2007) The Effect of Electric and Electronic Equipment on the Cooling Load of Buildings, Diploma thesis, Aristotle University of Thessaloniki, Department of Mechanical Engineering, Laboratory of Heat Transfer and Environmental Engineering, Thessaloniki, Greece Kim, D.-S. (2007) Solar Absorption Cooling, PhD thesis, Technische Universiteit Delft, The Netherlands Lamp, P. and Z. Ziegler (1998) ‘European research on solar assisted air conditioning’, International Journal of Refrigeration, vol 21, no 2, pp89–99

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 157

Solar Air Conditioning: A Review

157

Li, Z. F. and K. Sumathy (2000) ‘Technology development in the solar absorption air conditioning systems’, Renewable and Sustainable Energy Reviews, vol 4, pp267–293 Mugnier, D. (2006) Thermally Driven Water Chillers: Market Status and New Developments, IEA SHC Programme Task 38 Meeting, Aix les Bains, France Papadopoulos, A. M. (2001) ‘The influence of street canyons on the cooling loads of buildings and the performance of air conditioning systems’, Energy and Buildings, vol 33, pp601–607 Papadopoulos, A. M., S. Oxizidis and N. Kyriakis (2003) ‘Perspectives of solar cooling in view of the developments in the air-conditioning sector’, Renewable and Sustainable Energy Reviews, vol 7, pp419–438 Papadopoulos, A. M., D. H. Oxizidis and I. Samlidis (2004) ‘Constrains and potential for the propagation of solar refrigeration in the buildings sector’, Ktirio, Scientific Edition, vol A–B, pp41–50 Papadopoulos, A. M., A. Stylianou and S. Oxizidis (2006) ‘Impact of energy pricing on buildings’ energy design’, Management of Environmental Quality, vol 17, no 6, pp753–762 Richmond, A. K. and R. K. Kaufmann (2006) ‘Energy prices and turning points: The relationship between income and energy use/carbon emissions’, Energy Journal, vol. 27, no 4, pp157–180 Santamouris, M. (2001) ‘Solar and natural resources for a better efficiency in the built environment’, in Gordon, J. M. (ed) Solar Energy: The State of the Art – ISES Position Papers, James & James, London Santamouris, M., N. Papanikolaou, I. Livada, I. Koronakis, C. Georgakis, A. Argiriou and D. N. Assimakopoulos (2001) ‘On the impact of urban climate on the energy consumption of buildings’, Solar Energy, vol 70, pp201–216 Sayigh, A. A. M. and J. C. McVeigh (1992) Solar Air Conditioning and Refrigeration, Pergamon Press, New York Schawe, D. (2001) Theoretical and Experimental Investigations of an Adsorption Heat Pump with Heat Transfer between Two Adsorbers, PhD thesis, Institut für Kernenergetik und Energiesysteme der Universität Stuttgart, Stuttgart, Germany Srikhirim, P., S. Aphornratana and S. Chungpaibulpatana (2001) ‘A review of absorption refrigeration technologies’, Renewable and Sustainable Energy Reviews, vol 5, pp343–372 Syed, A., G. G. Maidment, R. M. Tozer and J. F. Missenden (2002) A Study of the Economic Perspectives of Solar Cooling Schemes, CIBSE National Conference, Part 2, London, October Terrados, J., G. Almonacid and L. Hontoria (2007) ‘Regional energy planning through SWOT analysis and strategic planning tools.: Impact on renewables development’, Renewable and Sustainable Energy Reviews, vol 11, pp1275–1287 Tsoutsos, T., J. Anagnostou, C. Pritchard, M. Karagiorgas and D. Agoris (2003) ‘Solar cooling technologies in Greece: An economic viability analysis’, Applied Thermal Engineering, vol 23, pp1427–1439 Wongsuwan, W., S. Kumar, P. Neveu and F. Meunier (2001) ‘A review of chemical heat pump technology and applications’, Applied Thermal Engineering, vol 21, pp1489–1519 Wu, X. and Z. Wu (2005) ‘Research on strategic environmental analysis of small hydropower enterprises in China: A case study of ZhpngDaSanChuan Hydropower Ltd’, in IEEE International Engineering Management Conference Proceedings, vol I, Article number 1559115, pp184–189 Yazaki Energy Systems Inc (2007) Performance data at www.yazakienergy.com, accessed 31 October 2007 Ziegler, F. (2002) ‘State of the art in sorption heat pumping and cooling technologies’, International Journal of Refrigeration, vol 25, pp450–459

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 123–157

ABER_CH_05

6/12/08

3:32 PM

Page 158

ABER_CH_06

6/12/08

3:33 PM

Page 159

publishing for a sustainable future

6

Experimental Methods in Ventilation M. Sandberg, H. Lundström, H. O. Nilsson and H. Stymne

Abstract This chapter describes different experimental methods related to the function of ventilation. Methods of thermal comfort are also treated. A classical, comprehensive review paper on this subject is found in Hitchin and Wilson (1967). Much information provided by this paper is still valid, although today the standard technology is digital, not analogue. In this chapter we examine a range of developments in the field since Hitchin and Wilson’s paper. With arrays of sensors in digital cameras and heat cameras, etc., we can record velocities, temperatures and concentrations over whole fields in space instead of point-wise. Adding tracers that follow air movement is a qualitative visualization method that has now become a quantitative method for recording velocities over large fields. Optical techniques such as laser Doppler anemometry are now a standard technique for recording the velocity vector at a specific point. Passive tracer gas techniques enable one to take measurements in the field without tying up expensive and bulky equipment. There has been a shift from pure technical issues to a focus on people. Concern about air quality has led to development of manikins that can breathe. Instruments for assessing thermal comfort have been developed that combine the effect of radiation, velocity and temperature in an index.

■ Keywords – ventilation, thermal comfort, measurement techniques, temperature, velocity, tracer gas

VENTILATION People are the customers and consumers of ventilation. Therefore, it is of great importance to plan the indoor environment with people in mind, and not just building surfaces and air. The focus must be on people as the final control point, not just the building itself. This position is necessary in order to develop healthy, realistic and sustainable indoor environments. In many companies today, the costs involving personnel are about 100 times higher than energy costs. It is therefore very uneconomical to save energy at the expense of people’s well-being. Ventilation refers to the process of introducing outdoor air, distributing the air to different rooms and finally distributing the air within the room. The purpose of supplying

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210 doi:10.3763/aber.2008.0206 ■ © 2008 Earthscan ■ ISSN 1751-2549 (Print), 1756-2201 (Online) ■ www.earthscanjournals.com

6/12/08

160

3:33 PM

Page 160

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

air is to dilute contaminants generated within the room that may be harmful or annoying and, finally, to remove them from the room. The first contaminant considered was bioeffluents, followed by emissions from building materials; today the focus is on small particles. A prerequisite for air to maintain good hygiene is that the supply of air must be cleaner than indoor air. It is now acknowledged that heating, ventilation and airconditioning (HVAC) systems may be a pollutant source. The contaminant exposure in a space is a balance between pollutant generation rates and sinks, where the main sink is ventilation. The amount of air required to reduce the concentration of contaminants is the hygienic ventilation flow rate and is related to the issue of indoor air quality (IAQ). Ventilation requirements have been explored in several studies. Figure 6.1 displays the result from a study carried out at the Danish Technical University. Figure 6.1 shows that the majority of people have very modest requirements, although others may be very sensitive. Frequently, ventilation systems have a dual role to play: in addition to being a means of controlling indoor air quality, ventilation systems are used for cooling or heating. When used as a cooling system, ventilation flow rates often exceed hygienic flow rates. Ventilation systems provide a means of transporting outdoor air into and from a building. The two main types are mechanical (fan powered) and natural ventilation. With a fan-powered system, the ventilation flow rate can, at least in theory, to a great extent be controlled. Natural ventilation relies on outdoor air being introduced into a building through purpose-provided openings or cracks in the building’s fabric. The engine for this flow is provided by the kinetic energy in the wind and the difference in potential energy between columns of air with indoor air density and outdoor air density, respectively. The ventilation flow rates in a naturally ventilated house are inherently disparate due to varying driving forces. A combination of mechanical and natural ventilation is known as hybrid ventilation (Heiselberg, 2006). Building ventilation rates are important and need to be measured, including the whole building ventilation flow rate and the flow rate to individual rooms. The flow into a building

Air quality just acceptable (%)

ABER_CH_06

50 40 30 20 10 0 0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 Ventilation rate (I/s person)

Source: Clausen (2007)

FIGURE 6.1 Percentage of people who find air quality just acceptable at a given ventilation rate

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 161

Experimental Methods in Ventilation

161

Annoyance Total annoyance

draught noise dry air particles smell gaseous pollutants

Air flow rate Source: original material for this chapter

FIGURE 6.2 Sketch of the dependence of concentration and velocity upon ventilation flow rate

is through the ventilation system and through leakages. An additional path of inflow to a room is through inflow of ‘unused’ air from other rooms (i.e. air that has not been contaminated). A necessary side condition for a proper supply of outdoor air is that it will not cause thermal discomfort by creating substantial heat loss from a part of the human body. This trade-off between the positive effect of diluting the contaminant and increasing velocity is shown in Figure 6.2. Velocities of the order of 15cm/s can be considered a draught. Recording such low velocities is a challenge. Figure 6.2 points out that the first remedy is source control.

THE SYSTEM VIEW Before ventilation air, taken from outdoors, arrives inside, it passes through a number of processes that will affect its quality (concentration) and temperature. Figure 6.3 shows the different parts in the system that must be dealt with when exploring ventilation and thermal comfort. People are the customers and consumers of ventilation, and in terms of their preferences and requirements, are not alike. The room is the enclosure that is closest to the individual. Ventilation air is supplied to the room, and the shape of the room affects the motion of air generated by the ventilation system. Through radiation exchange, people sense surface temperature. The temperature difference between room surfaces and

Human

Room

Building

Building Envelope

Outdoor Ambient

Source: original material for this chapter

FIGURE 6.3 The sequence outdoor ambient to the human

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

6/12/08

162

3:33 PM

Page 162

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Flow rate [m3/h]

ABER_CH_06

300 250 200 150 100 50 0 Mar May Jul Sep Nov Jan Mar

Source: adapted from Stymne et al (2006)

FIGURE 6.4 Weekly average values of ventilation flow rate as a function of a season in a house ventilated by mechanical extract ventilation

room air generates the movement of air that penetrates the occupied zone. Within the occupied zone, air movement may be caught by upward flowing air, which envelopes people at a higher temperature than is the case with ambient air. A building is subdivided into apartments or offices, which, in turn, are subdivided into rooms. Large airflows are generated through doorways, which act as conduit between rooms. Flow contact between floors is established by the flow through stairwells. Air then enters the building through the building envelope through purpose-provided openings or cracks. In the case of open windows and open doors, large quantities of air flow in and out of a building. Even if a building is ventilated by a mechanical ventilation system, ventilation flow rate varies with time (see Figure 6.4). In a sufficiently tight house, an extract ventilation system creates an under-pressure, and, in theory, ventilation is not influenced by minor disturbances, with a resulting nearly constant total flow rate. Figure 6.4 depicts a peak in flow rate during the summer. The cause is probably the behaviour of the occupant, who leaves windows partly open during shorter or longer time periods. As a result, the influence of occupant behaviour is an important factor. With respect to ventilation, occupancy behaviour transforms technically identical houses into different houses. Figure 6.5 shows the measurements of ventilation airflow rate during

1 0,8 0,6 0,4 0,2 0

0

0,25 0,5 0,75 1 1 Air change rate [ h ]

Source: adapted from Stymne et al (2006)

FIGURE 6.5 Distribution of air change rate in similar naturally ventilated dwellings (unbroken line) and similar mechanically extract ventilated (dashed line)

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 163

Experimental Methods in Ventilation

163

identical weather conditions in terraced houses, of which 29 were naturally ventilated and 22 extract ventilated. The distribution shown is the cumulative distribution depicting the fraction of houses with an air change rate less than the value given on the vertical axis. The fact that technically identical houses are not alike is an important consideration when making a survey of the ventilation conditions in a building stock. A sufficiently large number of technically identical buildings must be included in the sample. For example, the near ambient, located in an urban area with a high building density or a in a rural area, is particularly important when the building is ventilated through natural ventilation. Planning experiments and selecting suitable instrumentation will necessitate that one knows beforehand the ambient that will be recorded. Therefore we will, in order of priority, go through the characteristics of the individual parts in the sequence depicted in Figure 6.3. We only deal with spaces that people normally inhabit and therefore exclude areas such as crawl spaces.

PEOPLE MICROCLIMATE AND TEMPERATURE REGULATION The human body’s temperature regulation system permits physiological adjustment to thermal stress and provides thermal comfort under a variety of conditions. Heat is primarily produced through metabolism, which emanates from digestion and muscle exercise. In normal conditions, this results in an average ‘deep’ body temperature of about 37°C. Within a certain interval, the body’s temperature control system strives to maintain this temperature when internal or external thermal disturbances arise. The effects resulting from a change in climatic conditions produce a reaction within the body concerning physiological autonomic responses, but also activate behavioural regulation. Autonomic regulation is controlled by the hypothalamus, which regulates the different avenues of heat loss. The human body’s temperature-regulating centre works in a similar fashion to a thermostat. Temperature set point may change during different physiological conditions.

Source: Hans Lundström University of Gavle FIGURE 6.6 (Left) A person simulator produces 100W of heat; (right) the photographer Magnus Mattsson

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

164

3:33 PM

Page 164

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Thermo-sensitive receptor nerve endings send signals to the regulation centre. In a cool environment, stimulation of the cold receptors may lower the set point, and heat loss can decrease by means of vasoconstriction and shivering. In a warm environment, on the other hand, the set point becomes elevated and heat loss can increase by means of vasodilatation in the skin and sweating. The temperature regulatory centre is connected to receptors in the skin, as well as in the deep body core and the brain itself. These receptors consist of a net of fine nerve endings that are specifically activated by heat or cold stimuli. These temperature receptors are especially sensitive to rapid changes in temperature and are highly susceptible to adaptation. The number of active receptors determines, to some extent, the sensation of temperature (Åstrand and Rodahl, 1986). A person at rest produces approximately 100W of heat. If clothing and ambient conditions are adequate, the same amount of heat is lost to the environment. The ‘whole body’ heat balance is therefore in equilibrium, and the individual feels thermally neutral. With increasing ambient temperature, convection and radiation diminish. Sweat evaporation has to compensate for this in order to balance the heat production of 100W. Sweating is associated with a sensation of warmth and, eventually, discomfort. In a cooler environment, convection and radiation increase, producing total heat losses greater than 100W. The physiological response is to reduce skin and extremity blood flow in order to lower the external temperature gradient. In this case, the individual feels cool or cold and uncomfortable. A normal response is to add or remove clothing. In other words, for comfortable climatic conditions, dry heat loss can only vary within a certain narrow range. In a similar way, the heat balance of a skin segment can be analysed. Optimal local heat balance and a ‘comfortable’ skin temperature can only be maintained within a specific span of convective, radiative and conductive heat loss. Higher local heat losses will be felt as cool or cold, and lower local heat losses as warm. The human body’s response to the thermal environment has been found to depend primarily upon six factors (Parsons, 1993): 1 2 3 4 5 6

air temperature; mean radiant temperature; air velocity; relative humidity; physical activity; and thermal resistance of clothing.

These factors are critical in steady state conditions. If the exposure is short or intermittent, the length of the exposure can also have an impact. The final decision on whether a combination of these factors represents a comfortable situation or not depends upon the sensory reception of the climate. In order to produce correlations between this human thermo-sensation and different climatic disturbances, it must be assumed that the human body cannot differentiate between the heat loss sensations of draughts, radiation or conduction. Body build, sex, age, geographical differences, food and beverages, as well as several other environmental factors, also influence the perception of climatic comfort. The effect

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 165

Experimental Methods in Ventilation

165

of these components is not taken into consideration in any of the above indices. They may have an impact upon the perception of thermal comfort and eventually contribute to the occasionally significant individual variation in response.

TWO DIFFERENT WAYS OF LOOKING AT THERMAL COMFORT Two different types of models are widely known for predicting steady-state thermal comfort. One model is based on the heat balance of the human body. Another approach assumes, to a certain degree, an adaptation to the thermal environment: 1 Predicted mean vote (PMV) model: a method for calculating a steady-state thermal comfort index derived from heat balance calculations and climate chamber studies. The model assumes a relation between optimal thermal conditions, using the steadystate heat balance equation for the human body and thermal comfort ratings from panels of subjects. 2 Adaptive principle: a relation for steady-state thermal comfort obtained from large field studies. The model assumes that people will adapt to thermal conditions, using only indoor and outdoor temperatures as dependent variables.

THE PREDICTED MEAN VOTE (PMV) PMV (Fanger, 1970) describes the conditions for climatic comfort, and methods and principles for evaluating and analysing different environments from a climate comfort point of view, with the indices of predicted mean vote (PMV) and predicted percentage dissatisfied (PPD). The concept defines conditions that must be fulfilled in order for a person to be in ‘whole body’ climate comfort. An additional condition is the absence of any local climate discomfort. The intention was to develop a comfort equation where just the above-mentioned six factors are needed in order to calculate in which thermal state a ‘normal’ person is found. These calculation and estimation methods now comprise an international standard (ISO 7730, 2006) and are used all over the world. The model is, in essence, a regression equation that relates PMV on the seven-point thermal sensation scale of a group of people exposed to a certain environment to the calculated result of the basic heat balance equation. This equation uses the heat balance for the human body and assumes a connection between the deviation from optimal thermal balance and thermal comfort vote. The greater the deviation, the more the comfort vote departs from zero. The PMV equation is primarily a steady-state model. It is a semi-empirical equation for predicting the mean rating on an ordinal rating scale of thermal comfort for a group of people. PMV is currently the most widely used thermal comfort index. ISO Standard 7730 uses limits of PMV and PPD, as well as local recommendations, as definitions of a comfort zone.

THE ADAPTIVE PRINCIPLE The premise of the adaptive approach is expressed by the adaptive principle: If a change occurs that produces discomfort, people react in ways that tend to restore their comfort.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

166

3:33 PM

Page 166

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

The adaptive principle consequently suggests that people will adapt to certain climatic conditions. For instance, in warmer climate, when environmental thermal comfort parameters point at a higher PMV, people will become adapted to the higher temperatures and still feel comfortable. As an analogy, a corresponding ‘PMV principle’ could be written as: No change should occur that produces discomfort. In a properly designed environment, will people maintain their comfort? The adaptive principle (Nicol and Humphreys, 2001) is supported by field comfort studies in many different environments. From these global field studies of thermal comfort ratings and temperatures, Humphreys (1976) found that comfort temperature differed between groups of people. This means that comfort conditions calculated with the heat balance models did not fully agree with the comfort conditions found in the field. By linking the comfort rating to human behaviour, the adaptive principal links the comfort temperature to the climatic situations in which individuals find them. Comfort temperature is a result of the interaction between the subjects and the thermal environment. Nicol and Humphreys (2001) also conclude that people with more opportunities to adapt themselves to the environment will be less likely to suffer discomfort. The adaptive model is essentially a regression equation that relates the desired temperature indoors to the monthly average temperature outdoors. The only input variable used is the average outdoor temperature, which has an indirect impact upon the human heat balance. Consequently, the adaptive model does not include the six classical thermal parameters that have an impact upon the human heat balance and, therefore, upon thermal sensation.

CLIMATE INDICES WITH SIMILAR PRINCIPLES A number of indices have been developed, all according to similar methods, by transforming the ‘real’ environment to a ‘standard’ environment, which gives the same thermal climate experience: ● ● ● ●

effective temperature (ET); resultant temperature (RT); corrected effective temperature (CET); new effective temperature (ET*, or ET star); ● standard effective temperature (SET); ● PMV* (PMV star).

EQUIVALENT ‘EXPERIENCED’ TEMPERATURE Equivalent temperature (ET) is a recognized measure of the effects of non-evaporative heat loss from the human body (Dufton, 1936, Madsen et al, 1984, Nilsson, 1999). It is particularly useful whenever complex interactions of various heat fluxes are present. Equivalent temperature is derived from the operative temperature by including the effect of air velocity on a heated body.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 167

Experimental Methods in Ventilation

167

TABLE 6.1 Description of the connection between the measured quantities and equivalent temperature ACTION Increased air speed ↑ Decreased air temperature ↓ Decreased mean radiant temperature ↓ Decreased air speed ↓ Increased air temperature ↑ Increased mean radiant temperature ↑

INFLUENCE ↓ Lower teq ↓ Lower teq ↓ Lower teq ↑ Higher teq ↑ Higher teq ↑ Higher teq

Source: original material for this chapter

The well-known operative temperature (top) only considers air temperature and mean radiant temperature, and is defined for the actual air velocity, whereas equivalent temperature (teq) is defined for a standard low air velocity. One advantage of teq is that it expresses the effects of combined thermal influences in a single figure, and is easy to interpret and explain. It is particularly useful for differential assessment of climatic conditions. However, the underlying hypothesis is that the teq value always represents the same ‘subjective’ response irrespective of the kind of combinations of heat losses. As mentioned earlier, it is often difficult to explain the combined effects of different heat losses to people. It is therefore very useful to convert these values into something easier to understand, such as ‘experienced’ temperature, which is a more straightforward concept. This equivalent ‘experienced’ temperature (teq) is then calculated with Equation 1 according to (Nilsson, 2004a): teq = t s – RT ⋅ q "T

[1]

where: ● q”T = measured manikin heat loss during the actual conditions (W/m2); ● RT = total insulation, seated, no clothing (m2K/W); ● ts = manikin surface temperature (°C); ● teq = equivalent temperature of the uniform, homogenous environment (ºC).

The equivalent temperature in Table 6.1 shows low values, as could be expected. The results should be compared to the climate of an unclothed person, sitting in the air stream during the same conditions as the manikin. The definition today (Nilsson and Holmér, 2003) reads: ‘the equivalent temperature (teq) is the temperature of an imaginary enclosure with the mean radiant temperature equal to air temperature and still air in which a person has the same heat exchange by convection and radiation as in the actual conditions’.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

168

3:33 PM

Page 168

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

actual non-uniform environment

tr = 26°C

uniform enclosure teq = 24 °C

ta = 22°C

_ tr = ta = teq= 24°C va 0 m/s

va = 0.2 m/s R+C

=

R+C

Note: It is assumed that the posture, the activity level and the clothing are the same in both environments. The dry heat exchange, as well as the equivalent temperature, becomes the same in both environments. The heat exchange can, of course, be negative in warm environments and positive in cold environments. Source: Nilsson (2004a)

FIGURE 6.7 A thermal manikin is exposed to two different environments, one actual with non-uniform and one ‘imaginary’ with uniform climatic conditions

THE ROOM The flow in a room can be subdivided into two regimes: primary airflow generated by the air from the supply device and secondary flow generated by entrainment of air into the primary airflow. The primary air stream is a jet when isothermal air is supplied, and a buoyant jet when ventilation air is heated or cooled. If supplied along a horizontal surface with a low velocity, it becomes a gravity current if warm air is supplied at the ceiling or cool air is supplied at floor level. The primary air stream can further be subdivided into a near field region and a far field region. Within the near field region, the supply device leaves a footprint on the flow, whereas in the far field region, details of the effect of the supply devices have disappeared. A room is an enclosure that constrains airflow, and a characteristic feature of room airflow is the occurrence of recirculation zones. Outdoor air is limited to the supply of outdoor air through the ventilation systems; but internal airflow rates can be much larger. A characteristic feature of room airflow is the strong interaction between airflows generated by different sources. Heat transfer occurs at surfaces with a temperature that varies from room air temperature; there is also an exchange of heat by radiation between surfaces of different temperatures. Walls are lined by boundary layer flows, which are seldom fully developed, but exist in a transition between laminar and turbulent flow. An upward flowing boundary layer flow surrounds individual people, starting at the body’s ankles. In general, airflow in a room is not fully developed, which makes it difficult to predict flow according to computational fluid dynamics. Most of the frequency content of turbulent kinetic energy is below 2Hz.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 169

Experimental Methods in Ventilation

169

SUPPLY AIR TERMINALS Supply of air to a room can occur through a ducted system with supply air terminals at the end of the duct. The other alternative is ventilation through vents mounted in the walls, which is used in connection with mechanical extract ventilation and natural ventilation. Figure 6.8 shows supply air terminals for mechanical ventilation. With respect to supply velocity Binlet one makes a distinction between high-velocity supply devices, medium-velocity devices and low-velocity devices. A high-velocity supply can amount to 10m/s. A mediumvelocity supply device or low-velocity supply device is used when supplying air directly to the occupied zone. Local ventilation occurs where air is supplied directly towards people’s faces from a ventilation supply device located, for example, in a computer (Melikov, 2004). A new type of local ventilation is where air is blown from a perforated cushion around the neck of a person and upwards (Nielsen et al, 2007; see Figure 6.8(d)).

PARAMETERS CHARACTERIZING SUPPLY DEVICES AND AIR STREAMS FROM SUPPLY DEVICES The parameter characterizing an isothermal supply is the flow force (momentum flux) (Newton). The specific momentum flux is the flow force per unit mass: ⎡ 4⎤ 2 m m = q UBInlet = AInlet (UBInlet ) ⎢⎢ 2 ⎥⎥ ⎢⎣ s ⎥⎦

[2]

where q is the volumetric flow rate and Ainlet is a characteristic opening area of the supply terminal. The momentum flux of supply air terminals with a flat front can be determined by blowing against a balance and recording the reaction force (Nordtest, 2000). The Reynolds number is defined as: Re =

U A Inlet ν

=

m ν

[3]

where ν[m2/s] is the kinematic viscosity. The parameter characterizing a non-isothermal supply with temperature difference ∆T (K) between supply air and room air is the specific buoyancy flux B: B = qg

∆T T

⎡m4 ⎤ ⎥ ⎢ ⎢ 3⎥ ⎢⎣ s ⎥⎦

[4]

where g (m/s2) is the acceleration of gravity. Alternatively the specific buoyancy flux can be express in the load EC (W), removed by the ventilation system: B=g

EC ρC pT

[5]

where CP is the specific heat at constant temperature and ρ is the density. When the ventilation air is used for cooling or heating and supplied with a low velocity, we obtain a buoyant jet. In the beginning, the momentum flux dominates; but the buoyancy flux gradually becomes more important. When supplied vertically, the transition

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

170

3:33 PM

Page 170

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

(a)

(b)

(c)

(d)

Source: original material for this chapter

FIGURE 6.8 Supply air terminals: (a) High-velocity supply; (b) medium-velocity supply; (c) low-velocity supply intended for displacement ventilation; (d) local ventilation

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 171

Experimental Methods in Ventilation

g

171

1,85 lT

q

g

1,85 lT

q Source: original material for this chapter

FIGURE 6.9 Interpretation of thermal length

from momentum driven to buoyancy driven is given by the thermal length, sometimes called jet length, lT: Round buoyant jet lT =

m3 / 4 B1/ 2

. Plane buoyant jet lT =

m B2 / 3

(in metres).

[6]

The thermal length is only dependent upon the flow force of the supply air and the heat to be removed by the ventilation system. With no buoyancy, B = 0 and the thermal length is infinite. The physical meaning of the thermal length is illustrated in Figure 6.9. Cold air is supplied upwards. Initially, the momentum dominates; but the buoyancy directed downwards gradually becomes relatively stronger, and when it takes over, the flow turns downwards. Pure buoyancy sources are heat sources and sinks. Pure buoyancy sources give rise to plumes. A human body is a heat source; an example of a heat sink is a passive chilled beam.

VELOCITY IN THE OCCUPIED ZONE The highest velocity occurs in the primary air stream, where predicting velocity is of the utmost importance with respect to the risk of draught. In a room, airflow is constrained by room surfaces. Therefore, the room length L in the direction of the flow and the room width W are significant parameters. In a long room, the jet from a ceiling-mounted supply device may separate from the ceiling before it reaches the opposite wall. If it reaches the opposite wall, it follows the wall downwards and continues along the floor and a backflow is generated. Between air streams, there is an exchange of momentum (for an example, see Carrilho da Graca and Linden, 2003). The ratio between the velocity in the room and

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

172

3:33 PM

Page 172

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

the inlet velocity can be written as UR ( x ) UInlet

⎛ b (x ) lT ⎞⎟⎟ l L ⎜ = f ⎜⎜Re , T , , ⎟. ⎜⎜⎝ A Inlet A Inlet W L ⎟⎟⎠

[7a]

For isothermal supply the number of parameters is reduced: UR ( x ) U Inlet

⎛ b (x )⎞⎟⎟ L ⎜ = f ⎜⎜Re, , ⎟. ⎜⎜⎝ A Inlet W ⎟⎟⎠

[7b]

For small Reynolds numbers, there is a dependence on the Reynolds number (T. Valentino, personal communication). In particular, the length of the potential core (region of the jet with supply velocity) of the jet is a function of the Reynolds number. With increasing Reynolds number, the length decreases to finally attain the asymptotic value of about six times the characteristic length AInlet of the supply terminal. This may be important in relation to personal ventilation. The second parameter reflects the constraint by the room length L on the development of the flow. As mentioned earlier, the length of the room must be larger than six times the characteristic length of the supply air terminal in order for a jet to be developed. Some modern supply devices are so large (Elvsén and Sandberg, 2008) that the flow discharged from the supply device will not reach an asymptotic state. The third parameter provides information about the forces dominating from the beginning, the momentum of the supply air or the buoyancy of the supply air. It can be expressed in terms of supply velocity, characteristic length of the supply device and the temperature difference: lT A Inlet

=

U Inlet ∆T g T

=

A Inlet

1 Ars

[8]

where Ars is what is called the densimetric Archimedes number of the supply within ventilation engineering. It is used in the design of ventilation (Nielsen, 2007). In fluid mechanics, this parameter is called the densimetric Froude number. If this parameter is of the order of 1, the flow is dominated by buoyancy from the very beginning. One example is displacement, where cold air is supplied with a low-velocity diffuser. The fourth parameter is the ratio of the width b of the supply to the width of the room. When the width of the flow begins to approach the width of the room W, the further development of the flow will be constrained by the room. For isothermal flow, the increase in width is proportional to the entrainment coefficient α. For an axisymmetric jet with an entrainment coefficient of 0.0535, the growth in width is: b (x ) = 2αx = 0.107x .

[9]

For a flow dominated by buoyancy, the spread is controlled by the density difference. Figure 6.10 shows how a buoyant flow (cold air) discharged from a low-velocity air

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 173

Experimental Methods in Ventilation

173

Source: Sandberg and Mattsson (1993)

FIGURE 6.10 A flow dominated by buoyancy (Ars = 0.97): Overhead view of position of the front on the floor

terminal is spread. The tests were conducted in a model with water as the operating fluid and salt was added to the supply to obtain a density difference. The point XD indicates the position where the air from the supply air terminal hits the floor. From the point of impact the air is spread as a gravity current on the floor. It is reasonable to assume that the front is spread with a velocity proportional to ∆T g h where h is the local thickness of the air stream: T t

b (x (t )) = k ∫ 0

g

∆T (x (t )) T

h (x (t ))dt

[10]

where k is a parameter to be determined experimentally. The last parameter is the ratio between the thermal length and the length of the room in the direction of the supply. There is always a risk that in the case of a supply of cold air at ceiling level, the jet may separate from the ceiling and descend into the occupied zone and cause a draught. It is difficult to predict where it will separate. However, in order to avoid separation from the ceiling, a rule of thumb is that the thermal length should be much larger than the room length. Figure 6.11 shows the velocity at a point in the occupied zone as a function of the supply flow rate. The supply air terminal is a high-velocity supply air terminal located at ceiling level. With a supply of air that features the same temperature as room air, velocity increases linearly with the flow rate (supply velocity). If we need to heat the room with warm air, the supply air remains under the ceiling at low flow rates, which gives rise to

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

6/12/08

174

3:33 PM

Page 174

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Velocity in the occupied zone

ABER_CH_06

Cooling Isothermal Heating

Supply velocity Source: original material for this chapter

FIGURE 6.11 Velocity at a point in the occupied zone where there is a supply of isothermal heated air and cold air

low velocities. With the supply of isothermal air or warm air, the velocity in the occupied zone is less than the supply velocity. If we supply cold air with a low velocity, the velocities in the occupied zone will be high due to the tendency for the colder and heavier air to descend into the occupied zone. In displacement ventilation, buoyancy dominates and the maximum velocity in the occupied zone close to the supply device may be greater than the supply velocity.

A CHARACTERISTIC ROOM VELOCITY SCALE FOR BUOYANCY-DRIVEN FLOW Horizontal flow driven by buoyancy is comprised of gravity currents (see Figure 6.12). A characteristic room velocity scale based on the specific buoyancy flux and the room width is: ⎛ B ⎞1/3 U R = ⎜⎜ ⎟⎟⎟ (in m/s). ⎜⎝W ⎠ [11] This velocity scale can be used to assess the velocity in the room generated by displacement ventilation (Sandberg and Etheridge, 1996, p449; Cehlin and Sandberg, 2007).

UR [m/s]

W

Radial

Two-dimensional

Source: original material for this chapter

FIGURE 6.12 Assessment of velocity in a room ventilated by displacement ventilation

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 175

Experimental Methods in Ventilation

175

Source: Sandberg (1994)

FIGURE 6.13 Time constant

This velocity scale is only dependent upon the width of the room and the load to be removed by the ventilation system. It can therefore be calculated in the early design stage. If the velocity is larger than the comfort criteria required, a displacement ventilation system cannot be used.

TIME CONSTANT OF THE VELOCITY FIELD IN THE ROOM Knowledge of the velocity field’s time constant is necessary for determining sampling rate and integration time (number of samples) required to record the mean velocity and standard deviation of the velocity fluctuations within a certain accuracy. The time constant is a measure of the memory of the velocity field. Samples should be independent to provide new information. If the time separation between two samples is twice the time constant, they are regarded as independent. Sampling more frequently does not yield additional benefits. The time constant can be determined from the autocorrelation function of the velocity fluctuations. Figure 6.13 displays the recorded time constant in different parts of a room. Within the occupied zone, the time constant can be as low as 1s. The inset in Figure 6.13 shows an example of two energy spectra derived from velocity fluctuations. The mean velocity is the same, but the frequency content is different. To be able to retrieve information about the frequency content of velocity fluctuations without distortion, the velocity signal must first pass a low-pass filter. Through this process, the frequency content is limited to a maximum frequency fmax. To avoid aliasing the sampling interval, the filtered signal should, according to the Whittaker-Shannon theorem, be less than 1/(2 fmax).

THE VENTILATION PROCESS Under ideal conditions in an airtight room, outdoor air is supplied to the room from a welldefined location and extracted from the room at a well-defined point. Furthermore, the

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

176

3:33 PM

Page 176

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

assumption is made that the contaminants released within the room do not react with each other. The resulting concentration due to a release of contaminant in the room depends upon both how the air and the contaminant are distributed within the room (see Figure 6.14). The distribution of room air is quantified by its age at each point in the room. The local mean age of air τp at an arbitrary point P describes the length of time that the flow of air ‘molecules’ passing by P has, on average, spent within the building. The local mean age of air within and leaving the room is independent of the airflow pattern and is equal to: τe =

V q

[12]

where V is the volume of the room. The volume divided by the flow rate is also called the nominal time constant of the ventilation system. Since there is a local mean age of air at each point, it is meaningful to define the mean age (τ) of the whole volume of air within the room. The time τt, that it takes, on average, to replace the air within the room is: τr = 2 τ .

[13]

The flow rate, q (kg/s), of outdoor air provides the dilution capacity in the sense that the * in the room is: lowest concentration attained due to a local release with a rate m *

C=

m . q

[14]

This relation is only based on mass conservation (what goes in, goes out) and holds true irrespective of airflow pattern within the room. This concentration is attained in the extract point where all contaminants released within the room are mixed with ventilation air.

C= m q q

q Air

C o n ta

τe=

V q

mi na

nt

m Source: original material for this chapter

FIGURE 6.14 Concentration depends upon both how the contaminant and the air are distributed

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 177

Experimental Methods in Ventilation

177

An ideal state is uniform mixing where the concentration everywhere is equal to the concentration in the extract. At steady state, the total amount M (kg) of contaminants in the room (according to chemical engineering parlance) is given by the equation: i

M = m / τC

[15]

where τ—C is the hold-up time, which is the same as the mean age of contaminant leaving the room. Equation 15 provides a recipe for air-quality control: to minimize the amount of contaminant in the room, the contaminant should be removed as fast as possible and not be spread throughout the room. This is the basis for using exhaust hoods (see Figure 6.15). However, if a spread of the contaminant cannot be avoided, there are two strategies. One is to mix it with the ambient air in order to attain the minimum concentration (Equation 14): this is the idea behind mixing ventilation. The other principle is displacement ventilation, which is based on the principle of stratified flow caused by the density difference between layers of air with different temperatures. This air distribution principle can be used when there is a need for cooling. There are two layers in the room: one lower layer of air with supply air quality and an upper layer with polluted air. The height of the interface depends upon the heat load and the flow rate. Figure 6.16 displays a room ventilated by displacement ventilation. Two people sitting down were simulated. A tracer

Source: original material for this chapter

FIGURE 6.15 Kitchen exhaust hood used for laboratory test of capture efficiency: Visualization with smoke and a laser light sheet; oil is sprayed into a hot pan by means of a spray-nozzle

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

178

3:33 PM

Page 178

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Source: Stymne et al (1991)

FIGURE 6.16 Concentration recorded in a room ventilated by displacement ventilation

gas source was located on the body of the ‘person’ to the right. The concentration distribution is shown. The reference concentration is the concentration in the extract, which has been put equal to 1. In Figure 6.16, the two layers appear clearly and there is no contaminant in the lower layer. If the room had been ventilated by a mixing ventilation system, the concentration should (apart from a region close to the source with enhanced concentration) have been equal to 1 everywhere.

THE RELATION BETWEEN THE LOCAL MEAN AGE OF AIR AND THE CONTAMINANT CONCENTRATION For one and only one source there is a direct correspondence between the local mean age of air. If a passive contaminant is generated in each point of the room with the same rate everywhere then the concentration at point P is proportional to the local mean age at that point CP =

•

M τp V

[16]

˙ is the total release rate within the room. This relation is the theoretical where M background to the homogeneous emission method for recording the local mean age. THE BUILDING FLOWS BETWEEN ROOMS Rooms stay in contact with each other through flows generated by the density difference of air masses or by pressure differences introduced as a result of mechanical ventilation or wind loading. Flow generated by a density difference can be subdivided into two cases. In bulk density flow, the greatest difference in density (temperature) is between the two

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 179

Experimental Methods in Ventilation

179

rooms. The other case is boundary layer flow, where the main difference in density is between room air and room surface. Boundary layer flow occurs when one has a roomdominating heat source or a heat sink in the room. In the bulk density case, the only source of buoyancy is due to different room air temperatures. The room air surfaces are assumed to be in thermal equilibrium with room air and have attained the same temperature. These conditions can be achieved by carrying out experiments with water as an operating fluid and adding salt to the water to obtain a density difference corresponding to the density difference in air. This experimental method is called the salt bath method and is a simple method to study flows generated by density difference. An advantage of using water is that the flow can easily be visualized by adding various coloured dyes. Different parts can be tinted simultaneously with different colours. Fluorescent dye lit up with a laser light gives a clear picture of the flow. In the experiment depicted in Figure 6.17, salt has been added to the water in the compartment on the righthand side. The door between the adjacent rooms has suddenly been opened. A bidirectional flow with the heavier fluid (the cooler air) moves towards the left as a gravity current, whereas the lighter, heavier fluid moves towards the right. Figure 6.18 shows a boundary layer flow generated by a heat source in a building’s hall. The boundary layer flow passes into an adjacent room as a unidirectional flow where it is spread as a three-dimensional gravity current. Density differences between floors in a house give rise to a bidirectional flow. Figure 6.19 shows the flow in a stairwell studied with the salt bath method. The flow is visualized with the shadowgraph method (Settles, 2001), which makes the density difference visible. On the reverse of the model there is a transparent paper so that a distant light source, such as a projector, can render the density difference visible. Flows between floors are dealt with in Heiselberg and Li (2007).

Source: Claes Blomqvist, University of Gavle

FIGURE 6.17 Bidirectional bulk density flow through a doorway studied in a two-dimensional model: The cooler air moves towards the left as a gravity current along the floor

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

180

3:33 PM

Page 180

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Source: Blomqvist (2008)

FIGURE 6.18 Unidirectional boundary layer flow through an opening

Source: Blomqvist and Sandberg (2004)

FIGURE 6.19 Flow in a stairwell studied with the salt bath method: The density difference is made visible with the shadowgraph method

DIFFERENT ROUTES FOR THE SUPPLY OF OUTDOOR AIR TO A ROOM A building consists of many rooms and, therefore, many individual zones. In terms of airflow within a building, a zone is a space that is sharply marked off from its surroundings with respect to the transport of air. Normally, a zone is physically marked off from its surroundings by partition walls. Zones can also be established by pressure control achieved by supplying air. Figure 6.20 displays the different routes for flow of outdoor air to a room. The top image depicts direct supply of outdoor air through a ventilation system. In the middle is shown air supply through the ventilation system and infiltration through external walls. However, outdoor air can also arrive from other rooms. What matters is whether or not the air has a dilution capacity when arriving inside. It has a dilution capacity if it is ‘unused’ air – that is, it has not been used for dilution of a contaminant released in the actual room. The bottom image shows the purging flow rate U, which is the flow rate of

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 181

Experimental Methods in Ventilation

181

q

q

q Source: original material for this chapter

FIGURE 6.20 A ventilation system, a ventilation system with infiltration, and purging flow rate

outdoor air to the ventilated space that is available for dilution of contaminants in the actual room. According to the multi-zone theory, the purging flow rate can be expressed as follows (Sandberg, 1984): Ui = qi + ∑ Pij q j j ≠i

[17]

where: ● Pij = the transfer probability from zone j into zone i; ● qi = flow rate of outdoor air directly to zone i.

Within a room there is strong recirculation; nevertheless, there is a purging flow of ‘particles’ that never return. The maximum purging flow rate is equal to the total airflow rate to the building. The supply of outdoor air through a ventilation system is recorded in the duct. The supply of air through a ventilation system, as well as infiltration, can be recorded with the tracer gas technique, otherwise known as the constant concentration method. The purging flow rate can be obtained by a constant release of tracer gas and recording the equilibrium concentration.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

182

3:33 PM

Page 182

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

OUTDOOR AMBIENT The ambient surrounding a building is important with respect to wind and pollutants. When a house is provided with a natural ventilation system, wind environment is important. One engine for natural ventilation is the kinetic energy of wind. The outdoor ambient climate affects both wind speed and wind profile. If the building is located in open terrain or in an urban area, this is an important consideration. Pollutants in the ambient (e.g. small particles), pollen and exhaust from vehicles may penetrate the building and affect indoor air quality. Pollutant distribution around buildings should be considered when choosing the location of ventilation intakes.

WIND ENVIRONMENT The logarithmic mean wind profile over terrain is as follows (Stull 1988): ⎛⎜ z − D ⎞⎟ ⎟⎟ ⎜ z 0 ⎟⎠

U −⎜⎜ U( z ) = ∗ e ⎝ κ

U∗ =

τw ρ

and

friction velocity

[18]

[19]

where: ● ● ● ●

z0 = roughness height; D = zero-plane displacement height; τw= surface stress; κ = von Karman constant.

Values of between 0.41 and 0.37 have been reported in published papers. The roughness height z0 is defined as the height where the wind speed becomes zero. Protrusion of objects (e.g. buildings, forests, etc.) above a surface displaces the entire flow upwards. To account for this, a displacement height D is introduced. Historically, the zero-plane displacement height was introduced in order to be able to retain the logarithmic profile over a rough surface (see Figure 6.21).

Logarithmic profile U(z) Transition layer h z0+D D

Mean building height z0

0 Source: adapted from Meystayer in Fenger et al (1998)

FIGURE 6.21 Definition of displacement height (D) and roughness height (z0 )

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 183

Experimental Methods in Ventilation

183

Source: Leif Claesson, University of Gavle

FIGURE 6.22 Wind-tunnel test of a hybrid-ventilated school at a scale of 1:200: In the background is shown the spires and the roughness elements

In a wind tunnel, the atmospheric boundary layer (ABL) flow can be reproduced. As a result, a wind tunnel can serve as an experimental tool for exploring the effect of the ambient on the pressure on buildings or the flow through openings. Models of buildings and the ambient are typically reproduced on a scale of between 100th to 1000th of the original. Wind tunnel technique is an established experimental technique. The ABL over different types of terrain is generated by roughness elements and spires (see Figure 6.22) (Blomsterberg et al, 2007). The properties of the boundary layer are quantified in terms of the shape of the vertical velocity profile and roughness height and friction velocity. The roughness height is determined from the vertical velocity profile, and the wall shear stress can be recorded with a Preston tube (Goldstein, 1996). There are standards for doing wind tunnel tests (VDI, 2000).

FLOW THROUGH THE ENVELOPE A flow (infiltration) occurs through the building envelope as a result of leakages consisting of small cracks or small holes. To the list of parameters in Equations 7a and 7b that deal with mechanical supply, we must now add another parameter, t/√A, where t is the length (wall thickness) and √A is the linear size of the opening area A. Figure 6.23 shows a wind tunnel study of wind-driven flow through openings. The arrow indicates wind direction. For small t/√A, the opening is a sharp-edged opening; but for large t/√A, the opening is a duct and the wind is ducted into the room. The flow depicted in Figure 6.23 has been visualized by spraying semolina powder on the floor. The sand erosion method is thoroughly dealt with in Dezi (2006).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

184

3:33 PM

Page 184

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

t/√A=0.16

t/√A=1.29 Source: Leif Claesson, University of Gavle

FIGURE 6.23 Visualization by the sand erosion method of wind-driven flow through openings

MEASUREMENT OF HEAT FLUX FROM THE HUMAN BODY MEASUREMENTS WITH THERMAL MANIKINS A heated thermal manikin represents the ultimate heat flux transducer. The whole manikin surface is heated to, and controlled at, the same ‘skin’ temperature as the human body

Source: Nilsson (2004a)

FIGURE 6.24 The heated full-scale thermal manikin with 33 individually controlled zones: This manikin was especially constructed for climate evaluation in 1991

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 185

Experimental Methods in Ventilation

185

surface. The early manikins comprised only one zone. Today’s manikins provide 16, 20, 35 or even more individually heated and controlled zones. To obtain a single figure for teq, local heat losses are added and the total value divided by the total body surface area gives the dry heat loss to be used for the teq computation.

Manikin with constant surface temperature In this case, the surface temperature of the manikin is uniform and constant (approximately 34°C) over all zones.

Manikin with distributed surface temperature Here, the manikin has a distributed temperature over all zones. Usually, the temperature declines from the torso to the extremities.

Manikin with constant heat loss This manikin type operates with constant heat loss across the different zones. Small radial as well as inward axial heat losses may influence the measurements by this technique.

Manikin with adaptable surface temperature Here, the surface temperature of the manikin is allowed to change as a function of dry heat loss using an expression derived from the basis of comfort criteria.

Manikin with heated sensors This method measures the surface temperature of the sensors at constant heat loss with 18 small sensors on a manikin body, each containing two heated surfaces with two different power levels. The surface temperature of the two surfaces with known heat loss is measured and a linear model is used to calculate the surface temperature for an unheated surface.

SIMPLIFIED OR ADVANCED MANIKINS Researchers around the world have developed many different geometrical configurations in order to simulate a human body through a manikin. These manikins are different with

Source: Magnus Mattsson, University of Gavle

FIGURE 6.25 Dummies simulating the heat generated by a person

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

186

3:33 PM

Page 186

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

FIGURE 6.26 Manikin with 20 individually controlled segments Source: Nilsson et al (2007)

respect to size, form, heat generation, number of zones, regulation principles, etc. The evolution of these manikins is shown in Figures 6.25 to 6.27.

Simple Figure 6.25 shows a simple but still efficient dummy for simulating the heat generated by human occupancy. In this case, there is constant heating with no regulation or measurement of heat loss (Mattsson, 1999; see also Figure 6.6, where the temperature of a dummy and a real person is compared). These dummies are relatively cheap, and a large number can be manufactured to simulate a group of individuals – for example, as a class of students.

Standard Figure 6.26 shows Comfortina, a thermal manikin of ‘Nille type’ belonging to the University of Aalborg, Denmark. Comfortina has individually controlled segments enabling assessment of local thermal comfort or discomfort. It measures heat loss or equivalent temperature, with up to 20 independently controlled zones (Nilsson et al, 2007).

Advanced Figure 6.27 displays an advanced automotive manikin (ADAM), which measures heat loss at 120 independently controlled zones. Manikin data is transferred to a computer model that simulates human thermoregulatory responses in real time. Still another computer model predicts human thermal comfort (Fan, 2006; see also www.nrel.gov/vehicles andfuels/ancillary_loads/adam.htmlUT).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 187

Experimental Methods in Ventilation

187

Source: NREL (2006)

FIGURE 6.27 Manikin with 120 independently controlled zones

The variations do, of course, reflect different subjects of interest, such as effects on airflow, thermal comfort, as well as pollutant production and exposure.

OTHER HEAT FLUX MEASUREMENT TECHNIQUES Today, heat flow and equivalent temperature are measured with various instruments and devices. The most developed method is a man-sized heated manikin, which realistically simulates the three-dimensional heat exchange of a human body.

MEASUREMENT OF CONVECTIVE AND RADIATIVE HEAT LOSSES WITH HEATED SENSORS Heated sensors can be used to determine body surface heat losses through convection and radiation.

ELLIPSOID SENSORS A heated ellipsoid sensor may serve as a representative physical model of the human body (Brüel and Kjær, 1982; Madsen et al, 1984). The heat exchange of the sensor is assumed to correspond to the total (and uniform) convective and radiative heat exchanges of the human body surface. The instrument directly calculates equivalent temperature.

FLAT SURFACE SENSOR Heated flat surface sensors give reliable and relevant estimates of dry heat losses. The three-dimensional nature of human heat transfer requires special arrangements of several flat sensors or measurements on several locations – for example, on the surface of a body-shaped dummy. In order to provide a single figure for the overall teq values,

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

188

3:33 PM

Page 188

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

individual values are added and area is weighted in a representative way (Mayer and Schwab, 1993).

LOCAL DISCOMFORT METER Here, a double-sided heated skin element is used (Madsen et al, 1992). The difference in mean heat flux (MHF) from the two opposite elements can directly be transformed to a thermal asymmetry. The electrical signals received from the skin element are transformed by a microprocessor into a value called perceived heat flux (PHF), which is equivalent to the sensation of local thermal discomfort. Calibration of the instrument is done in the same way as for the manikins.

MEASUREMENT OF BASIC CLIMATIC PARAMETERS With measurements of air temperature, mean radiant temperature and air velocity, radiant and convective heat losses can be calculated and teq determined using appropriate equations.

POINT MEASUREMENT TECHNIQUES FOR TEMPERATURE AND VELOCITY Classical techniques are point-measuring techniques. Measurements of temperature, air velocity, etc. are performed at discrete positions. Measuring at fixed positions corresponds to the Eulerian approach to fluid mechanics. Observations are made at fixed positions and the flow passes by. During the last two decades, there has been substantial advances in whole field measuring techniques, which makes it possible to take continuous measurements over parts of a room. Here, measurements of velocities are based on the Lagrangian approach to fluid mechanics, where the displacement of tracer particles ideally follows air motion. Many of the measurement techniques are indirect in that the methods rely on the physical interpretation of the quantity measured. Hot-wire techniques, for instance, are based on relating heat transfer to velocity. A Pitot tube senses pressure, and velocity is inferred from the measured pressure by using the Bernoulli equation. Flow rate is recorded by relating the rate of change of tracer gas concentration to flow rate.

TEMPERATURE MEASUREMENTS CONTACT MEASUREMENTS Years ago, the mercury thermometer was the ultimate choice for measuring the indoor climate; today, however, electrical thermometers such as resistance thermometers and thermocouples are usually used. Electrical thermometers can be made small enough to have a fast time response and may also be connected to data loggers and computers. But the accuracy is not necessarily higher; a mercury thermometer can be calibrated to read 0.001K. Electrical thermometers for laboratory use are usually resistive temperature sensors or thermocouples. Resistive temperature sensors rely on the fact that the electrical resistance of a material changes as its temperature alters. Two key types are the metallic devices and thermistors. Metallic sensors are usually made of a platinum coil or film, while thermistors consist of a piece of semiconductor material made of metal oxides

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 189

Experimental Methods in Ventilation

189

pressed into a small bead and covered with glass or epoxy. The thermocouple, which is the most commonly used temperature sensor, has a quite different working principle. A thermocouple is created when two dissimilar metal conductors are joined. The contact point produces a small voltage as a function of temperature. This thermoelectric voltage is known as the Seebeck effect and can be used to determine temperature. Significant features of platinum resistive sensors are high reliability and excellent longterm stability. They are well standardized and are very common in industrial applications and whenever high accuracy is important, such as with reference sensors for calibration. However, they are not the first choice for air temperature measurements in indoor climate research. Thermistors have low cost and high resolution, and they can be made very small. The smallest bead thermistors available are approximately 0.1mm in diameter. Thermistors are used in many low-cost temperature instruments and are the standard sensor employed in the great variety of digital household thermometers available on the market. This does not mean that thermistors are inaccurate. They can be calibrated to read one thousandth of a degree, and standardized types now available guarantee an accuracy of ± 0.2K in the room-temperature range. Thermocouples are easy to use. In many applications, the thermocouple wire itself can be employed as a sensor simply by twisting the two wires together. Thermocouple wire is available in different sizes and with different insulations. For indoor climate use, 0.2mm to 0.5mm diameter wire is common; but wires as small as a few micrometres in diameter are available. There are, in fact, two junctions in a thermocouple circuit. One is at the measurement point and is called the hot junction; the second one is known as the cold junction and appears at the connection inside the instrument. Since a thermocouple only measures the temperature difference between two junctions, it is necessary to know the temperature at the cold junction in order to calculate the temperature at the measuring point. The cold junction temperature is measured by means of a temperature sensor inside the instrument. Consequently, any errors in the measurement of the cold junction temperature will be added to the total measurement error. In a low-cost instrument, this sensor may have poor accuracy and is sometimes not even specified. This source of errors is often overlooked when using thermocouples. A typical thermocouple holds an uncalibrated accuracy of approximately ± 0.3K in the roomtemperature range; but the total error is highly dependent upon the instrumentation used. The maximum resolution attained is in the order of 0.05K without using specialized instruments. Measuring temperatures in still air invites some pitfalls due to the poor heat conduction of air. In order to measure the resistance of a resistive temperature sensor, a small current must be passed through the sensor, causing self-heating of the sensor. It is thus important to design the measurement electronic circuit so that self-heating is low enough to allow for temperature measurements of the desired accuracy. For this reason, in more expensive data loggers, the excitation current may be controlled. The thermocouple does not generate self-heating or suffer from this problem. With thermocouples, however, errors may occur if the thermocouple is mounted in a temperature gradient. Since the metal wires in the thermocouple are good thermal conductors, heat may be conducted through the

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

190

3:33 PM

Page 190

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

wires out to, or from, the measuring junction, thus causing a different temperature in the junction than in the surrounding air. Errors may also be introduced due to radiation exchange between the sensor and the surrounding surfaces. This is usually not a problem in the occupied zone under normal conditions where the temperature difference between the air and surrounding surfaces are low. But when measuring temperatures in an environment with surfaces at a temperature that is quite different from air temperature, such as large windows, chilled ceilings or heating radiators, this effect should always be considered. Here, the thermocouple may have an advantage. If the insulation on the thermocouple wire is stripped off on a piece close to the measuring junction, we will have a sensor with small emissivity that is less sensitive to radiation effects. Bare metal wires have a much smaller emissivity than insulated wires.

NON-CONTACT MEASUREMENTS The above-mentioned temperature sensors are contact sensors, which means that they must be in thermal contact with (and ideally in thermal equilibrium with) the media in which the temperature is measured. Strictly speaking, a contact temperature sensor is only able to measure its own temperature and it is important to mount the sensor so that it attains the temperature of the media that we want to assess. There are also non-contact temperature meters, which make use of the fact that all objects emit radiation. The amount of radiation can be measured and related to temperature using the Planck law of radiation. Using this technique, surface temperature can be measured remotely. Infrared radiation thermometers have long been employed in industry for measuring high temperatures, where contact sensors are difficult to apply. Today, the availability of lowcost infrared semiconductor sensors has opened a new broadened market for infrared temperature measurements. Handheld infrared thermometers are now available at reasonable prices and are excellent for making a quick check of surface temperatures in rooms. However, it cannot be used for measuring air temperature. The Planck law only holds for a perfectly black surface. A real surface, usually called a grey surface, emits less radiation than a black surface. The radiation from a grey surface is related to the black surface radiation by a correction factor called emissivity. The emissivity has a value in the range of 0 to 1, where 1 applies to a black surface. Thus, in order to use the thermal radiation as a measure of temperature, the emissivity must be known. High emissivity surfaces, such as building materials, painted surfaces or human skin, hold an emissivity close to 0.9. Low-cost instruments have the emissivity fixed at that value, and with this assumption we can measure temperatures at this kind of surface with an accuracy in the order of ± 1K, well enough for many indoor climate applications. More expensive instruments have provision for the emissivity to be set manually, which allows for measurements with higher accuracy, provided that the emissivity is known. The more sophisticated instruments also are provided with optics that allow for a small measuring spot. Typically, a measuring spot of approximately 20mm diameter can be achieved at a distance of a couple of metres from the instrument. A fundamental characteristic of infrared thermometers is, however, that measuring temperature on low emissivity surfaces, such as metals, is very difficult and prone to large errors.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 191

Experimental Methods in Ventilation

191

AIR VELOCITY MEASUREMENTS In the occupied zone, velocity measurements are performed in combination with air temperature measurements in order to investigate the risk of draught. But air movements in other zones, such as close to air terminals, may also be of interest. In the occupied zone, velocities are low, typically in the range of 0.1m/s to 0.5m/s, while the velocity from an air inlet terminal may be up to a few metres per second. Air velocity measurements may also be used in ducts as a means of estimating the volume flow. There are basically two types of instruments in use for measuring air velocities in indoor climate research: thermal anemometers and laser Doppler anemometers. Thermal anemometry has been around for almost 100 years and is a well-established technique. Laser Doppler techniques have evolved rapidly during the last decades; but laser Doppler anemometers (LDAs) are still very expensive. The two techniques are really very different, each one having its pros and cons. A thermal anemometer relies on the fact that convective heat loss from a heated body varies with velocity, while the LDA measures the movement by small seeding particles following the air stream.

THERMAL ANEMOMETERS There are essentially two types of thermal anemometers that are used for indoor climate measurements: the hot-wire anemometer and the hot-sphere anemometer. The hot-wire anemometer has a sensor consisting of a thin metal wire of platinum or tungsten, typically 5µm in diameter and a few millimetres long. The sensor is heated by means of electrical joule heating. It is usually operated in constant temperature mode, which means that electrical current through the sensor is controlled by means of electronic controlling circuitry, so that the temperature of the sensor wire is kept at a constant value. The electrical power delivered to the sensor to maintain its temperature constant is a measure of the heat transfer from the sensor to the air and can be related to velocity. This relationship must be achieved by means of calibration to known velocities. The main characteristics of the constant temperature hot-wire anemometer are a fast response and high spatial resolution, making it useful for unsteady and turbulent flows. Sampling rates of up to 100kHz are possible. Hot wires are prone to drift with time and need to be frequently recalibrated. Accuracy depends upon many parameters (and, of course, upon the quality of the calibration equipment used). A typical value is ± 5 per cent and up to ± 1 per cent in extreme cases. The original application area for hot-wire anemometers is high velocity measurements in wind tunnels. Here, the wire is operated at a high temperature. However, for low velocity flows, as in indoor climate applications, the wire temperature must be kept low, typically less than 100°C, in order to minimize errors due to natural convection from the heated sensor wire. A decrease in wire temperature results in an increase in sensitivity to air temperature. As the sensor wire temperature is constant, any change in air temperature results in a change in temperature difference between the wire and the air and will thus affect the output from the anemometer and appear as an error in the velocity measurements. This means that hotwire anemometers usually must be temperature compensated when measuring low velocities. Temperature compensation is accomplished by measuring the air temperature with a separate temperature sensor and using a compensation scheme in the software.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

192

3:33 PM

Page 192

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

The hot-wire anemometer is most sensitive to the velocity component perpendicular to the wire and must be oriented properly in order to get correct readings. Room air movements may not have a well-defined direction and in these situations the hot sphere anemometer is often preferred. The hot-sphere anemometer has a spherical sensor, usually consisting of a glass sphere typically 2mm to 3mm in diameter, covered with a nickel layer. This type of anemometer is essentially omni-directional and may be regarded as an air speed meter or draught meter, mostly used in a free air stream such as in the occupied zone. It is not suitable for measurements close to surfaces, such as in boundary layer measurements. The time response is limited to about 20Hz. In order to minimize disturbances due to natural convection from the sensor itself, the sensor temperature is low, typically 20K to 30K above ambient air temperature, and this is accomplished by operating the anemometer in the constant temperature difference mode. This means that the hot sphere temperature is controlled by electronic controlling circuitry such that the temperature difference between the sensor and the ambient air is always kept at a fixed value. For this reason, the hot sphere anemometer is always equipped with a temperature sensor close to the velocity sensor. It is important to bear in mind that in order to give correct measurement results, the two sensors must be immersed in air with equal temperature. This must always be considered when temperature gradients are expected in the flow, as may be the case, for instance, in natural convection flows. In addition to the above-mentioned types of thermal flow sensors, a variety of sensors of different configurations exist for use in different applications. Hot-wire anemometers may be equipped with more than one sensor wire in order to make it possible to sense more than one component of the flow field. The most common is the X-probe with two perpendicular wires, which can detect the velocity vector within a cone of ± 10º. Sensors made of a glass fibre rod covered with a nickel layer (fibre film sensors) are sometimes used instead of the solid wire sensors. Fibre film sensors are not as fragile as the solid wire sensors and are often used in high-velocity measurements in wind tunnels or for measurements in water. Hot-sphere anemometers are sometimes equipped with bead thermistors as sensors instead of the nickel-covered glass sphere. Thermistors can make reasonably good omni-directional anemometers at low cost, and it is not uncommon that indoor climate laboratories assembly their own thermistor anemometers for measuring velocities at many points simultaneously (McNair, 1972; Lundström et al, 1990). A fundamental drawback with thermal sensors is that they cannot sense flow reversals, so measurements in circulating and highly turbulent flows should be handled with caution. Strictly speaking, we may say that when measuring velocities by means of thermal anemometers, we must have some a priori information about the character of the flow field. Visualization with smoke may accomplish this. By applying a puff of smoke into the flow and by watching where it is going, we can estimate the velocity and direction, but also get rough information about turbulence intensity. This is a simple but very useful technique that should not be overlooked when investigating low velocity flows in air. For more detailed information about thermal anemometry and the practical problems involved, the reader is referred to the available literature in the field (e.g. Christman and Podzimek, 1981; Fingerson, 1994; Bruun, 1995; Melikov et al, 1998a). An extensive investigation

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 193

Experimental Methods in Ventilation

193

of uncertainties in low-velocity measurements when using hot-sphere anemometers is presented by Popiolek et al (1998) and Melikov et al (1998b). Lundström et al (2007) have investigated the use of hot-wire anemometers at low velocities and varying air temperatures.

LASER DOPPLER TECHNIQUES In the laser Doppler anemometer, the scattered light from small seeding particles passing through the intersection of two laser beams is measured. The laser beams interfere in the intersection (the measuring volume of the LDA), and when a particle passes, the scattered light has a frequency that is proportional to the velocity of the particle. The frequency depends upon particle velocity, as well as laser wavelength and the angle between the two laser beams. The laser wavelength is very stable, and once set up the laser Doppler anemometer needs no recalibration. Light from the measuring volume is scattered in all directions, although the light scattered in the forward direction from the laser beams has the highest intensity. Today, many laser Doppler anemometers use backscatter and have the feature that both laser optics and receiving optics, as well as the detector, may be assembled in one single unit. The measuring volume is slightly ellipsoid shaped and the size depends upon the diameter of the laser beams, the angle between the beams and the focal length of the optics used. Typical size of the measuring volume is a few millimetres, down to about 0.05mm. However, it is important to remember that the sampling rate cannot be exactly controlled because a measurement occurs when a seeding particle passes the measuring volume. Measurements are intrusion free and the accuracy is generally higher than for thermal anemometers – figures as high as 0.1 per cent have been reported. In contrast to thermal anemometers, the laser Doppler anemometer measures the true velocity component. For measurements of more than one velocity component, more laser beams of different wavelengths are added, and laser Doppler anemometers that simultaneously measures one, two or all three velocity components are available. A drawback, however, is the need for seeding. In air, seeding is usually accomplished by adding smoke. Supplying smoke to the measuring volume without disturbing low-velocity flow is sometimes tricky and deserves some exercise. In addition, a confined space will soon be filled with smoke, and measurements must be interrupted for ventilation. The comparably large measuring volume and the fact that use of a laser Doppler anemometer close to a surface may introduce errors from reflections make the laser Doppler anemometer less useful for boundary layer measurements. Here, the hot-wire anemometer usually performs better. The hot-wire anemometer is also superior when making time-series analysis due to the difficulty in controlling the sampling rate in the laser Doppler anemometer.

ULTRASONIC ANEMOMETERS Although uncommon for measuring indoor velocities, the ultrasonic anemometer is another type of anemometer worth mentioning here. Ultrasonic anemometers are based on the measurement of the transit time of transmitted sound waves between two points. This can be accomplished in different ways; but commercial anemometers available today usually utilize the pulse method. Here, the anemometer measures the time taken

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

194

3:33 PM

Page 194

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

for an ultrasonic pulse to travel from one transducer to the other, and then compares it with the time taken for another pulse to travel in the opposite direction. By using three pairs of transducers, the true velocity vector can be calculated. The ultrasonic anemometer was presented for the first time during the 1940s; but commercial anemometers have been on the market only for a couple of decades. The main application is in meteorology for outdoor wind measurements. These types of ultrasonic anemometers are rather large, but are sometimes used for indoor climate measurements. However, the future trend seems to be smaller-sized ultrasonic sensors, and we will certainly see more of these anemometers in the future. The sampling rate is typically 20Hz for ultrasonic anemometers and accuracy in the order of a few per cent may be achieved.

MEASURING VELOCITY AND TEMPERATURE WITH WHOLE-FIELD MEASURING TECHNIQUES Whole-field techniques are relatively new and under development. An overview of wholefield measuring techniques, including absorption tomography for recording concentrations, is given in Sandberg (2007). All techniques are optical in the sense that pictures are taken by an array of sensors sensible to different wavelengths of light depending upon the application. The velocity components can be recorded with particle image velocimetry (PIV) or particle streak velocimetry (PSV). Both methods are based on adding tracer particles, which ideally follow air motions; a digital camera records their displacement during a specified time interval. This corresponds to the Lagrangian formulation of fluid mechanics. With PIV, the displacement of groups of particles is recorded in contrast to the PSV method, where the displacement of a single particle is recorded. PIV can provide high-resolution information over small regions, whereas the PSV method can cope with large areas. Temperature distribution can be recorded and visualized with an infrared camera and a measuring screen that ideally attains room air temperature.

PARTICLE DISPLACEMENT In both PIV and PSV, velocity is determined by measuring the displacement ∆X in room coordinates: ∆X = X(t + ∆t) – X(t) = (X(t + ∆t) – X(t), Y(t + ∆t) – Y(t), Z(t + ∆t) – Z(t)) of seeded particles in a time interval ∆t. For a known time step ∆t, the components of the velocity vector U(X, t) are derived from the fundamental definition of velocity: U (X, t ) =

⎛ X (t + ∆t ) − X (t ) Y (t + ∆t ) −Y (t ) Z (t + ∆t ) − Z (t ) ⎞⎟ ∆X = ⎜⎜ , , ⎟⎟. ⎝⎜ ⎠ ∆t ∆t ∆t ∆t

[20]

Equation 20 is the definition of velocity, and both PSV and PIV are therefore direct methods. A planar light sheet is used and the velocity components lying in the light sheet are the in-plane components. Traditionally, the notation assumes that the X and Y components are the in-plane components, while the Z component is the out-of-plane component. A well-defined time step is generated by a pulsed light source or by interrupting the light beam by a chopper or a shutter.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 195

Experimental Methods in Ventilation

195

Light source Image coordinate system

Light sheet

y

Image

x

Lens

CCD Array

d0 di

Y Interrogation volume

X

Tracers

Z Z

Room coordinate system

Source: Sandberg (2007)

FIGURE 6.28 Illumination system, flow markers and a digital camera

BASIC EQUIPMENT FOR MEASURING VELOCITY Here, the objects are markers following air motion. The markers are illuminated by light of a suitable wavelength (see Figure 6.28). Through an optical system, the light is turned into a light sheet, usually with a cylindrical lens. The light sheet constitutes the measuring volume. With a system measuring two velocity components, proper measurements require that the flow field is essentially two dimensional and the light sheet is oriented accordingly. To determine the out-of-plane component, stereo-photogrammetry can be used. At least two imaging systems are then required and the thickness of the light sheet is set to approximately 10cm.

TRACER SEEDING SYSTEM Here, the air must be seeded by particles that follow air motion. As a result, there is a compromise between the need for particles to be small enough to follow air motion, but large enough to reflect sufficient light in a suitable illumination. In the PIV mode, particles with a diameter of approximately 1µm are used. Smoke generators – for example, by boiling paraffin oil – can generate such particles. In the PSV mode, individual particles must be identified; therefore, the particles must be larger because light sources are not strong enough to produce a good exposure of small particles. In the PSV mode, the particles used in air are normally neutral buoyant heliumfilled soap bubbles with a diameter of about 1mm to 2mm. According to simulations by Muller et al (2001), a soap bubble follows variations in air velocity reasonably well, up to a frequency of 100Hz. The requirements on tracer particles for PIV are dealt with in Melling (1997).

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

196

3:33 PM

Page 196

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Interrogation volume Flow field x y z d0 d1

x

Object plane

y z

xc

Image plane Interrogation spot

Interrogation analysis of the images

xc

Resulting vector

Displacement

Source: adapted from Soloff et al (1997)

FIGURE 6.29 Particle image velocimetry (PIV) recording and interrogation process

PARTICLE IMAGE VELOCIMETRY (PIV) An overview of PIV is given in Adrian (1991, 2005). The PIV recording and interrogation process are illustrated in Figure 6.29. As a result of using small-diameter particles (droplets), a pulsed high-power laser is employed as a light source. Seeding density is high and the displacement of whole groups of particles is recorded. Through correlation analysis, the displacement of the group of particles is identified. A drawback with PIV is the small area that the system can manage: approximately 0.2 x 0.2m; nevertheless, detailed information can be obtained. PIV systems are sold commercially.

PARTICLE STREAK VELOCIMETRY (PSV) In PSV, seeding density is low and the movement of individual tracers can be recorded (a synonym for particle streak velocimetry is particle tracking system). As a result, PSV can only record where there are tracers. In order to measure areas between tracers, data must be interpolated. An advantage of PSV is that the system can cope with a large area and standard lighting systems can be used. A drawback is that within the interpolated regions, accuracy is lower at about 20 to 25 per cent.

MEASURING AIR TEMPERATURE WITH INFRARED THERMOGRAPHY Air is transparent to infrared radiation and therefore a measuring screen that adopts air temperature must be used (see Figure 6.30), making this technique intrusive. A modern

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 197

Experimental Methods in Ventilation

197

Supply device

Measurement screen

Source: original material for this chapter

FIGURE 6.30 Orientation of the measuring plane

infrared camera typically has a resolution of 320 x 240 pixels, which is much lower than commercial charge-coupled device (CCD) cameras and is sensitive to infrared radiation in the range 7.5 to 13µm. The absolute accuracy is about ± 2°C, while the relative accuracy is about ± 0.1°C. A measuring screen is affected by radiation from ambient room surfaces, which gives rise to a systematic error. A remedy for minimizing error is to cover the two sides of the screen with different materials (Cehlin, 2006). One side can be of paper, while the other may be aluminium foil. Figure 6.30 shows how the measuring screen can be placed when recording the radial supply temperature from a low-velocity diffuser for displacement ventilation placed at floor level. Disturbance is minimized when the screen is placed in the symmetry plane.

Source: Sandberg (2007)

FIGURE 6.31 Supply of warm air from a ceiling diffuser: Temperature distribution recorded with an infrared camera

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

198

3:33 PM

Page 198

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

By taking repeated measurements along the other positions indicated in the figure, a three-dimensional picture of the temperature distribution can be obtained. This information may be used in computer-aided design (CAD) programmes showing how a floor area is affected by the direct supply of the chilled air to the occupied zone. A solid screen restricts the applicability of the method. Porous screens (10mm thick with holes of 9mm in diameter) oriented horizontally have been used to record the temperature in a plume above a heat source (Streblow et al, 2007). Figure 6.31 shows an application of this method. Here, the recorded temperature distribution of the supply of warm air from a ceiling diffuser is depicted.

TRACER GAS METHODS Ventilated rooms feature strong recirculation; therefore, the velocity field does not reflect the rate by which a contaminant is removed. Tracer gas methods are the ventilation measuring method because the tracer gas is a substitute for a passive contaminant. Through ventilation, the tracer gas is diluted and the rate of dilution provides information about the airflow rate to a room and parameters related to the ventilation process. A tracer gas can also be seen as an indicator of air ‘molecules’. We can therefore record the time that it takes to replace the markers in the room with non-marked air molecules coming from outside. Tracer gas methods are the most general technique for measuring ventilation airflow rates in rooms, including infiltration. They are the only methods available for recording the total flow rate, including infiltration, to a room, and are the only methods for recording the age of air and the time that it takes to replace air.

TRACER GASES Ideally, a tracer gas should be a gas not normally present in the indoor environment. In the review by Hitchin and Wilson (1967), tests are mentioned that feature radioactive tracers. Today, there is a great concern about health effects and the impact of gas on the environment. Common tracer gases are sulphur hexafluoride (SF6 ) and nitrous oxide (N2O). Occasionally, carbon dioxide (CO2) is used as a tracer gas despite the fact that it is present in people’s ambient and exhalation air. A gas used in connection with passive tracer gas techniques is hexafluorobenzene (C6F6). This gas can be detected at low concentrations.

METHODS OF DETECTING TRACER GASES A gas analyser consists of a detector and a gas sampling unit. A gas chromatograph with an electron capture detector or mass spectrometer can detect sulphur hexafluoride and hexaflurorobenzene at low concentrations. A gas chromatograph is primarily a laboratory instrument. With an infrared gas analyser or a photoacoustic detector, one can detect nitrous oxide, carbon dioxide and sulphur hexafluoride. These two types of analysers can be used in field trials.

TECHNIQUES FOR INJECTING TRACER GAS AND SAMPLING Here, the two main techniques are active or passive. In an active technique, the pressure in the gas cylinder is used for injecting the gas. The gas is mixed with room air using

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 199

Experimental Methods in Ventilation

199

Metal wire

Source: original material for this chapter

FIGURE 6.32 Components in a passive system

mixing fans or other means that promote amalgamation. Air samples are taken by using a pump, and when sampling at many points, electro-mechanical valves are used for switching between sampling points. Passive systems are based on molecular diffusion both for injecting and sampling gas. As a gas source is used, a small container is provided with a capillary or a permeation membrane in order to control the emission rate. Figure 6.32 depicts a source with a capillary. The emission rate can be controlled with a metal wire. Molecular diffusion is temperature dependent and careful calibration is required for the actual gas used. For passive sampling, an adsorption tube is used (see Figure 6.32).

METHODS FOR MEASURING FLOW RATE Various methods for measuring the flow rates to a house or air change rates by tracer gas are dealt with in ISO 12569 (2000). Depending upon injection strategy, we can classify them as the decay method, the constant injection method, the constant concentration method and the arbitrary injection method. The decay rate method is the simplest method and is used to obtain discrete flow rates over short periods of time. A small quantity of tracer gas is mixed into the room air.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

200

3:33 PM

Page 200

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

The source is then removed and the decay in the concentration of tracer gas is measured over a period of time. Mixing fans may be used to ensure that the tracer gas concentration is the same in all points of the room. The method is based on the fact that when there is complete mixing, the local mean age in its point is equal to τp. The coefficient of the exponential decay is equal to the inverse of the local mean age and thus: Q = V λe .

[21]

To obtain the flow rate, an additional parameter, the volume V of the ventilated zone, must be known. The decay method is therefore an indirect method. With the constant injection method, a tracer gas is released with constant and known mass (volume) rate. According to Equation 10, the attained equilibrium concentration gives the flow rate as: •

m q= . C

[22]

This is therefore a direct method that is used, both with active and passive methods, for estimating long-term average flow rates and the airflow through long ventilation ducts. With a time-varying flow rate, bias is associated with the method. In order to achieve the average flow rate, we need to know the average of the inverse concentration; but we only know the inverse of the average concentration: 1/C. These two quantities differ from each other. The constant concentration method is used for continuous airflow rate measurements in one or more zones. Here, the concentration of tracer gas in a zone is measured and a computer controls the tracer gas injection rate so that the concentration in each zone is kept at a constant target value. By keeping the concentration constant, the effect of flow through doorways is neutralized and the inflow of outdoor air to each individual room (zone) can be recorded (see Figure 6.20). By using system identification methods, the flow rate can be retrieved from the time history of the concentration generated by a time-varying injection of tracer gas. This is a so-called inverse problem that is very sensitive to the accuracy of the data. One example of the application of the method is found in Okuyama (1990).

METHODS FOR MEASURING THE LOCAL MEAN AGE OF AIR For field measurements, the most common method is the homogeneous emission method (see Equation 16), based on use of passive tracer gas techniques. The room is . subdivided into a number of sub-volumes: Vi. The injection rate m i into sub-volume Vi is proportional to the volume of the sub-volume: •

mi =

.

Vi • M V

[23]

where M is the total injection rate into the room. The decay technique is also relatively simple. The local mean age at the point P is obtained from the area of the graph of the decay curve and the initial concentration C(0):

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 201

Experimental Methods in Ventilation

201

∞ −

τP

=

∫ CP (t )dt 0

[24]

C ( 0)

Procedures for doing the measurements can be found in ISO 16000-8 (2007).

MEASURING DUCT FLOW Perhaps the most widely used flow-metering principle is to record the pressure difference over some obstruction in the duct. In a laboratory set-up, orifice flow meters are the most common and rely on measuring the pressure difference over a sharpedged orifice placed in the duct. This is a standardized and very accurate method. Orifice flow meters are seldom used for field measurements; but modern ventilation installations may be provided with similar devices that feature connections for a differential pressure meter for monitoring the flow. Sometimes a bend in the ducts is used for this purpose. The volume flow can also be calculated from velocity measurements in the duct. Since the flow profile in the duct is unknown, velocities in a number of points in the duct must be measured, and the flow is calculated from the velocity values according to a standardized scheme. The velocity is usually measured by means of a hot-wire anemometer or a Pitot tube. The instrument is introduced through a hole in the duct wall and traversed across the duct. The Pitot tube (named after Henri Pitot in 1732) measures the fluid velocity by converting the kinetic energy of the flow into pressure using the Bernoulli equation. There are instruments on the market that can directly measure the flow through an air inlet or exhaust terminal. These instruments consist of a hood, which is placed over the air terminal and through which the airflow must pass. The hood is equipped with a flow-measuring device based on either differential pressure measurements or a hot-wire type of instrument. The accuracy is about ± 5 per cent for these instruments. The bag method can be used to measure the true volumetric flow through an air exhaust terminal. The method relies on measuring the inflation time of a polyethylene bag, which is suddenly connected to the air terminal. Flow in ducts may be measured using the constant injection method and monitoring the concentration. Equation 22 gives the flow rate. Here, it is essential that the duct is sufficiently long that injected tracer gas will be fully mixed with the air.

DATA COLLECTION AND DATA HANDLING Today, measuring instruments are often delivered as complete systems that connect directly to a computer and come with a dedicated software package. This applies to, for instance, laser Doppler anemometers, modern hot-wire anemometers and gas monitors. For measuring devices that deliver an analogue output signal, such as temperature sensors, a separate data logger must be used. A data logger can be a multi-channel analogue/digital (A/D) card inserted in the computer or a separate instrument connected

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

202

3:33 PM

Page 202

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

to the computer’s serial bus. During the last few years, many manufacturers have presented a large variety of small low-cost data loggers using USB connection. But there are great differences between different data loggers, and it is important to ensure that the data logger has the resolution and accuracy needed. Thermostats have a high output and are easily connected to most data loggers, while thermocouples are more demanding. Thermocouples, which are the most commonly used temperature sensor for laboratory use, have a low output signal, and a good-quality instrument is recommended. For field measurements, where the measuring points are often spread over an apartment or a whole building, several small data loggers may be beneficial. During the last decade, small battery-operated data loggers have become available from several vendors. They are equipped with typically 1 to 5 inputs to which sensors for temperature, humidity, etc. can be connected. These type of loggers store data in their internal memory for weeks or months and can then be connected to a PC for saving data onto the computer. More sophisticated models feature connection to a host computer via wireless radio communication or via a telephone line. The availability of user-friendly software has considerably reduced frequently timeconsuming computer work. Today, LabVIEW (a trademark of National Instruments) is probably the most commonly used software dedicated to measurement and data handling. Many instrument manufacturers supply software drivers for operating their instruments with LabVIEW.

ASSESSMENT AND EVALUATION METHODS COMFORT ZONE DIAGRAM EVALUATION In order to make comfort evaluation independent of clothing, the construction of new comfort zone diagrams can be made by inserting any seated total insulation available (Nilsson, 2004b). Equation 25 shows how a relationship between the equivalent temperature level and mean thermal vote (MTV) can be established for each manikin body part. The heat loss corresponding to a certain level of comfort, or discomfort, in the diagram is consequently considered to be the same. The shape of the zones is, however, changed according to the clothing used: teq ,zone = t s − RT . (a + b . MTVzone )

[25]

where: ● ● ● ● ●

teq,zone = equivalent temperature in the zone (°C); ts = manikin surface temperature (here 34°C) (°C); RT = total insulation, seated (m2K/W); a, b = linear regression constants (W/m2); MTVzone = mean thermal vote in the zone (here no data).

The equation is valid for an interval of seated whole-body total insulation (RT) between 0.9clo and 1.9clo. By using Equation 25 and the information in Table 6.1, it is, for the first time, possible to make a comfort zone diagram (see Figure 6.33) that applies to a specific

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 203

Equivalent temperature, teq (°C)

Experimental Methods in Ventilation

35

203

too hot 1.5

30 warm but comfortable

25

0.5 20

neutral

15

−0.5 cold but comfortable

10

−1.5

5 too cold L. F oot R. F o ot L. L ow leg R. L ow leg L. T high R. T high Pelv is Hea d Top of h ead L. H and R. H a nd L. F orea rm R. F orea rm L. U ppe r ar R. U m ppe r ar m Che st Bac k All

0

Note: This figure shows the results with no clothing (IT = 0.9clo) for 17 segments and ‘all’ of the manikin body. Abbreviations refer to L = left; R = right; U = upper. A spreadsheet can be downloaded from the Thermal Manikin Network (Nilsson, 1999) as well as the CFD Benchmark sites (see www.cfd-benchmarks.com). Note that the comfort zones of acceptance are much narrower for summer clothing, except for the ‘less sensitive’ head and hands. Source: www.cfd-benchmarks.com

FIGURE 6.33 Comfort zone diagrams adapted for Comfortina-type manikins

clothing combination used in a given situation. One must insert in Equation 25 corresponding local values of the surface temperature ts (usually 34°C), a and b, and local total insulation value RT of the clothing and air layer, together with a MTVzone for the four borders of the three shaded comfort zones in the legend of Figure 6.33. Now teq,zone can be calculated for those four borders, for all zones and the whole body. The plotted result forms the evaluation background in the clothing-independent comfort zone diagram. This is only done once for each clothing combination and reflects the insulation distribution of the clothing used. As could be expected, diagrams for clothing with less insulation show an increased sensitivity in most zones, except the normally unclothed head and hands. The opposite, decreased sensitivity for heat loss variations, can be observed for diagrams with increased clothing insulation. These new methods have also recently become an international standard: ISO 14505: Ergonomics of the Thermal Environment – Thermal Environment in Vehicles. The comfort zone diagram shown in Figure 6.33 corresponds to the figures D1 and D2 in the new standard (ISO 14505-2, 2006). A new benchmark study is providing requirements for the design and development of computer manikins and computational fluid dynamic (CFD) benchmark tests for comfort evaluation. The main idea is to focus on people. That the comfort requirements of occupants should be the decisive factor in thermal climate will prevail. It is therefore important to use comfort simulation methods that originate from people, not just the temperature of surfaces and air. In order to obtain accurate results from the numerical simulation, knowledge or correct simulation of the manikin surface heat transfer is of utmost importance. The near

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

204

3:33 PM

Page 204

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

surface flow field in a room or around a heated body is characterized by a combination of natural (free) and forced convection, developing boundary layers. The restricted validity of the heat-transfer models often used originates from the assumptions that have been made to solve special boundary layer flows. These assumptions are consequently not always valid for flows that can be commonly found in the indoor environment. Nielsen et al (2003) introduced two benchmark tests focusing on the airflow around virtual thermal manikins or computer-simulated persons (CSPs). Now a new benchmark test for a CFD manikin or a CSP is available. One purpose of this test is to create a series of very detailed and accurate full-scale measurements to serve as a comparison with CFD predictions. The ideas behind a CFD manikin benchmark test, which define the boundary conditions around a real, as well as a CFD, manikin, are as follows: ● It is of great importance to be able to verify that simulated heat losses equal

measured heat losses converted into equivalent ‘experienced’ temperature in order to support comparisons with human experiences presented in clothing-independent comfort zone diagrams. ● If different versions of virtual CFD manikins can be tested with the same boundary conditions, it is possible to make comparisons, and perhaps make some new decisions, on a geometrical level of the design, turbulence model used, type of grid, etc. ● It is very useful to have the results presented not only as a whole-body influence, but also with local information on how the thermal climate varies over the human body. The development of these virtual models gives us a more efficient and complete complement to traditional evaluation of the thermal environment. ● Different avenues of heat loss from a building’s occupants are best recorded directly with a thermal manikin or some other heat loss-detecting instrument. Today, it is frequently more time-consuming to measure indirect discrete values of air and radiant temperatures, as well as air velocity. This is, however, foreseen to change in the near future when it becomes more important to have better control of the indoor environment due to increased demands and new legislation and regulations.

TEST ROOMS AND SCALE MODELS A model of a building or a room can be a full-scale replica of a building or room, or a building or room on a reduced scale (scale models). The advantage of a model is that it can easily be altered and sophisticated instrumentation can be used. Topics, for example, that can be studied in a model are airflow pattern, the risk of draught and the risk of short-circuiting of air.

FULL-SCALE REPLICA In the design process, full-scale rooms are used when there are special requirements concerning ventilation and cooling or heating. These special requirements can either be a non-standard room shape or, for example, a high cooling load in terms of W/m2 floor area, which introduces a risk of draught. Special conditions will make it uncertain whether standard solutions will work; as a result, a full-scale room is built to test that the thermal comfort requirements can be met. An advantage of a full-scale room is that people can be exposed to the climate and therefore can ‘sense’ it. Special full-scale test rooms (climate

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 205

Experimental Methods in Ventilation

205

chambers) are built with the purpose of exploring people’s sensation of climate. ‘Test people’ are used in such experiments and two important issues related to the use of such people are the question of internal and external validity. In such experiments, one variable is frequently varied (the independent variable), often in such a way that the test people are unaware of the change. The internal validity of the experiment concerns itself with whether the observed effects were caused only by the independent variable. The external validity tells us to what degree the findings can be generalized to other people and other situations (Liebert and Langenbach-Liebert, 1995). As shown by Oseland (1995), a context effect on thermal assessment may be obtained in the sense that the judgement of the thermal conditions can be influenced by the surrounding environment. Therefore, it is important that the appearance of the test room is similar to a real room. This means that the test room must be decorated and furnished. If an installation does not perform well in a laboratory test, it will in all probability not work in practice due to future disturbances and difficulties.

SCALE MODELS In general, a scale model is used when it is not possible to do a full-scale test in the laboratory or in the field. The aim of using a scale model may be as follows: ● During the design process, predictions can be made by using scaling laws to transform

results from scale models to those at full scale (Etheridge and Sandberg, 1996, Chapter 14). These scaling laws are based on the premise that there will be similarity between the model and the full scale, which implies that the non-dimensional numbers relevant to the actual physical process involved, such as a Reynolds number, a Froude number and a Peclet number, will be equal. The principle problem with a model in reduced scale is that the different physical phenomena are affected to varying degrees by the change in scale. Before the introduction of computational fluid dynamics, scale models were used, but are now less frequent due to the problems with scaling of non-isothermal phenomena. ● To isolate a physical process – for example, the effect of a source of buoyancy can be studied by using a solution of salt and water to generate a density difference. This method is sometimes called the salt bath method. With the use of water as an operating fluid, there is no radiation exchange between room surfaces. As a result, we cannot expect that the recorded density distribution exactly represents the temperature distribution in air. The scaling is based on the fact that the reduced ∆ρ gravity g ′ = g is the same in the model and at full scale. An example of this ρ approach is found in Hunt and Linden (1999), who studied the combined effect of wind and buoyancy in natural ventilation. Examples of the use of the salt bath method can be seen in Hunt and Linden (1999, Figures 10, 17 and 18). ● To calibrate a CFD model with the aid of data from the scale model. By doing so, the scale model becomes the real model. The rationale behind this approach is that despite the fact that the model is not ‘real’, it has all the complexity of reality and all the physical processes present in reality.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

206

3:33 PM

Page 206

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Scale models of buildings are used in wind tunnels in order to study wind effects in the built environment, as well as the dispersion of contaminants.

DISCUSSION AND CONCLUSIONS Buildings are large objects and a room is a large space. This large scale in itself introduces a problem: one cannot measure everywhere. A strong recirculating flow exists inside rooms, and the velocity in the occupied space is very low (in the order of 10 to 20cm/s). It is crucial to perform measurements on several occasions; in real buildings, measurements are carried out to quantify the performance of the system. Mock-up rooms are built in laboratories when designing a building and in climate chambers when carrying out research. The application of computational fluid dynamics has renewed the need for measurements with the purpose of calibrating (validating) CFD models. Development of virtual manikins to be used in computers is an ongoing process. The basic idea behind these virtual manikins is to connect the results from measurements using thermal manikins to human sensation, and then to implement virtual manikins with the same properties as the thermal manikins. The crucial thing is to predict the correct heat transfer rate (see Brohus and Nielsen 1996; Loomans, 1998; and Nilsson, 2004a). This has been shown to be difficult, leading to renewed interest in measuring the heat flux from different parts of the human body with segmented thermal manikins. Among thermal climate variables, velocity is the most difficult variable to record. In order to assess draught, the speed (magnitude) of the velocity field can be recorded pointwise with omni-directional probes. The velocity vector field can then be recorded with whole-field measuring techniques. With digital cameras, pictures of tracers following air motion are recorded. Particle image velocimetry (PIV) is a commercial technique for measuring two-dimensional velocity fields at a high accuracy. However, PIV can only cope with fields in the order of 0.2 x 0.2m. Due to the limited size of the region that can be measured, PIV is not a general-purpose method for studies of room air. Another disadvantage is that it requires a high-powered laser light source. One non-commercial technique is particle streak velocimetry (PSV), which can cope with areas up to 10m2. An advantage here is that standard light sources can be used. With PSV, velocity is recorded only at locations where a tracer happens to be located. Within regions where there is no tracer, data interpolation is required and the interpolation may introduce a relative error of up to 20 per cent. Due to the problem of scaling non-isothermal flow, CFD and multi-zone models have, to a great extent, replaced scale models. However, there is a renewed interest in the use of scale models for calibrating CFD models. In this approach, the scale model is the real model. The rationale behind this is that despite the fact that the model is not, in fact, real, it has all the complexity of reality and all the physical processes present in reality. Its small size, furthermore, makes it easy to instrument. A characteristic feature of internal flows in buildings is the occurrence of recirculating air. As a result, the velocity field does not adequately provide information about the transport of contaminants. The tracer gas technique is the ventilation measuring technique because it is possible to record the dilution of a tracer gas – and the purpose of ventilation systems is to dilute and remove contaminants in air. This is a versatile

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 207

Experimental Methods in Ventilation

207

method: the rate of outdoor air entering a building can be recorded and the distribution of air within a room can be quantified in terms of the air’s age. Malfunctioning of systems through short-circuiting can be easily detected. In a building with many rooms, the rooms are connected to each other in a complex way. Outdoor air, for example, can reach a room through different passages. The purging flow rate is the amount of air available for dilution and can easily be recorded through the tracer gas technique. Full-scale test rooms are an excellent tool where people can sense the indoor climate. As shown by Oseland (1995), a context effect on thermal assessment may be obtained in the sense that people’s perception of thermal conditions can be influenced by the surrounding environment. As a result, it is important that the test room appears similar to real life. In general, laboratory tests can be seen as a ‘screening’ procedure; if a system does not work in the laboratory, it will not work in reality. This chapter is an attempt to briefly explain some of the more well-known measurement techniques used in indoor climate research. Due to continuous development in the field, a review of this kind may, of course, never be complete. However, we have tried, as closely as possible, to cover current state-of-the-art technology. We have tried to apply a practical approach, avoiding unnecessary formulae (which, in any case, may be found in textbooks). The intention is, rather, to inform the reader of the different techniques available, their pros, and cons, and the benefits expected. For more detailed information, the reader is referred to existing literature in the field (e.g. Doeblin, 1990) for a general approach on measurement techniques; for fluid mechanics measurements, see Goldstein (1996) and Tavoluaris (2005).

AUTHOR CONTACT DETAILS M. Sandberg: (corresponding author) Department of Technology and Built Environment, Laboratory of Ventilation and Air Quality, University of Gavle, Sweden; [email protected] H. Lundström: Department of Technology and Built Environment, Laboratory of Ventilation and Air Quality, University of Gavle, Sweden H. O. Nilsson: Department of Technology and Built Environment, Laboratory of Ventilation and Air Quality, University of Gavle, Sweden H. Stymne: Department of Technology and Built Environment, Laboratory of Ventilation and Air Quality, University of Gavle, Sweden

ACKNOWLEDGEMENTS We gratefully acknowledge the help from our colleagues Claes Blomqvist, Leif Claesson, Per-Åke Elvsén, Magnus Mattsson and Hans Wigö.

REFERENCES Adrian, R. J. (1991) ‘Particle imaging techniques for experimental fluid mechanics’, Annual Review of Fluid Mechanics, vol 23, pp261–304 Adrian, R. J. (2005) ‘Twenty years of particle particle image velocimetry’, Experiments in Fluids, vol 39, pp159–169 Åstrand, P. and K. Rodahl (1986) Textbook of Work Physiology, 3rd edition, McGraw-Hill, New York

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

208

3:33 PM

Page 208

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Blomqvist, C. (2008) ‘Conversion of electric heating in buildings: An unconventional alternative’, Energy and Buildings, under review Blomqvist, C. and M. Sandberg (2004) ‘Air movements through horizontal openings in buildings: A model study’, The International Journal of Ventilation, vol 3, pp1–10 Blomsterberg, Å., M. Sandberg and Å. Wahlstrom (2007) ‘Behovsstyrd hybridventilation-mer än förstärkt självdrag’, in Broschyr 2007, Formas, Stockholm Brohus, H. and P. Nielsen (1996) ‘CFD models of persons evaluated by full-scale wind channel experiments’, in Proceedings of the 5th International Conference on Air Distribution in Rooms (Roomvent 96), Technical University of Denmark and Danvac, pp137–144 Brüel and Kjær (1982) Thermal Comfort Meter Type 1212: Instruction Manual, No 033-0699, Brüel and Kjær Sound and Vibration Measurement A/S, Nærum, Denmark Bruun, H. H. (1995) Hot-Wire Anemometry, Principles and Signal Analysis, Oxford University Press, Oxford Carrilho da Graca, G. and P. F. Linden (2003) ‘Simplified modeling of cross-ventilation airflow’, ASHRAE Transactions, vol 109, pp1–14 Cehlin, M. (2006) Visualization of Airflow, Temperature and Concentration Indoors, KTH Research School, University of Gävle, Gävle Cehlin, M. and M. Sandberg (2007) Time Evolution of Gravity Currents Discharged from Low Velocity Diffusers, Roomvent 2007, Helsinki, Finland Clausen, G. (2007) ‘Human interaction with the indoor environment – scientific background for design criteria: Human comfort, health and productivity’, in Seppänen, O. and Säteri, J. (eds) Roomvent 2007, FINVAC ry, Helsinki Christman, P. J. and J. Podzimek (1981) ‘Hot-wire anemometer behaviour in low velocity air flow’, Journal of Physics E: Scientific Instruments, vol 14, pp46–51 Dezi, G. (2006) An Assessment of Wind Comfort by Sand Erosion, University of Eindhoven, Eindhoven Doeblin, E. O. (1990) Measurement and Systems: Application and Design, 4th edition, McGraw-Hill, Eindhoven, Holland Dufton, A. (1936) ‘The equivalent temperature of a warmed room’, JIHVE (now the Journal of CIBSE), vol 4, pp227–229 Elvsén, P.-Å. and M. Sandberg, (2008) ‘Buoyant jet in a ventilated room: Velocity field, temperature field and air-flow patterns analyzed with three different whole-field methods’, Building and Environment, accepted for publication Etheridge, D. and M. Sandberg (1996) Building Ventilation: Theory and Measurement, John Wiley, Chichester Fan, J. (ed) (2006) Proceedings of the 6th International Thermal Manikin and Modeling Meeting, Hong Kong, 16–18 October 2006 Fanger, P. (1970) Thermal Comfort, Danish Technical Press, Copenhagen Fenger, J., O. Hertel and F. Palmgren (eds) (1998) Urban Air Pollution: European Aspects, Kluwer Academic Publisher, Dordrecht, The Netherlands Fingerson, L. M. (1994) ‘Thermal anemometry, current state, and future directions’, Review of Scientific Instruments, vol 65, pp285–300 Goldstein, R. J. (1996) Fluid Mechanics Measurements, Taylor and Francis, Washington, DC Heiselberg, P. (2006) ‘Hybrid ventilation’, in Santamouris, M. and Wouters, P. (eds) Building Ventilation: State of the Art, James & James, London Heiselberg, P. and Z. Li (2007) ‘Experimental study of buoyancy driven natural ventilation through horizontal openings’, in Proceedings of Roomvent, vol 2, Helsinki, Finland, pp141–150 Humphreys, M. (1976) ‘Field studies of thermal comfort compared and applied’, Journal of the Institution of Heating and Ventilating Engineers, no 44, pp 5–27 Hunt, G. R. and P. F. Linden (1999) ‘The fluid mechanics of natural ventilation – displacement ventilation by buoyancy-driven flow assisted by wind’, Building and Environment, vol 34, pp707–720 Hitchin, E. R. and C. B. Wilson (1967) ‘A review of experimental techniques for the investigation of natural ventilation in buildings’, Building Science, vol 2, pp59–82

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

3:33 PM

Page 209

Experimental Methods in Ventilation

209

ISO 12569 (2000) Thermal Performance of Buildings: Determination of Air Change in Buildings – Tracer Gas Dilution Method, International Organization for Standardization, www.iso.org ISO 7730 (2006) Moderate Thermal Environments: Determination of the PMV and PPD Indices and Specification of the Conditions for Thermal Comfort, International Organization for Standardization, www.iso.org ISO 14505-2 (2006) Ergonomics of the Thermal Environment – Thermal Environment in Vehicles, International Organization for Standardization, www.iso.org ISO 16000-8 (2007) Determination of Local Mean Ages of Air in Buildings for Characterizing Ventilation Conditions, International Organization for Standardization, www.iso.org Liebert, R. M. and L. Langenbach-Liebert (1995) Science and Behavior: An Introduction to Methods of Psychological Research, Prentice Hall International, New York Loomans, M. (1998) The Measurement and Simulation of Indoor Airflow, PhD thesis, Eindhoven University of Technology, Eindhoven Lundström, H., C. Blomquist, P. O. Jonsson and I. Pettersson (1990) ‘A microprocessor-based anemometer for low air velocities’, in Proceedings of the 2nd International Conference of Engineering Aero- and Thermodynamics of Ventilated Rooms, Roomvent 90, Oslo, Norway, 13–15 June, ppB1–27 Lundström, H., M. Sandberg and B. Moshfegh (2007) ‘Temperature dependence of convective heat transfer from fine wires in air: A comprehensive experimental investigation with application to temperature compensation in hot-wire anemometry’, Experimental Thermal and Fluid Science, vol 32, pp649–657 Madsen, T., B. Olesen and N. Kristensen (1984) ‘Comparison between operative and equivalent temperature under typical indoor conditions’, ASHRAE Transactions, vol 90, part 1, pp1077–1090 Madsen, T., E. Soehrich and Z. Popiolek (1992) ‘Measurement of draft sensation by a new skin element’, Thermal Insulation Laboratory, Technical University of Denmark, vol 8 Mattsson, M. (1999) ‘On the efficiency of displacement ventilation, with particular reference to the influence of human physical activity’, Centre for Built Environment, Royal Institute of Technology, Gävle, Sweden Mayer, E. and R. Schwab (1993) ‘Evaluation of heat stress by an artificial skin’, Proceedings of Indoor Air 93: The 6th International Conference on Indoor Air Quality and Climate, vol 6, pp73–78 McNair, H. P. (1972) ‘A combined thermistor’s thermometer and anemometer’, in Cockrell, D. J. (ed) Fluid Dynamic Measurements, Leicester University Press, Leicester, pp250–255 Melikov, A. K. (2004) ‘Personalized ventilation’, Indoor Air, vol 14, pp157–167 Melikov, A. K., T. Sefker, Z. Popiolek, I. Care and W. Finkelstein (1998a) ‘Requirements and guidelines for low-velocity measurements’, ASHRAE Transactions B, vol 104, no 1, pp1529–1542 Melikov, A. K., T. Madsen, and G. Langkilde (1998b) ‘Impact of natural convection on the accuracy of low-velocity measurements by thermal anemometers with omnidirectional sensor’, ASHRAE Transactions, vol 104, no 1, part B, pp1519–1528 Melling, A. (1997) ‘Tracer particles and seeding for particle image velocimetry’, Measurement Science and Technology, vol 8, pp406–1416 Muller, D., B. Muller and U. Renz (2001) ‘Three-dimensional particle streak tracking (PST): Velocity measurement of a heat exchanger inlet flow’, Measurement Science and Technology, vol 30, pp645–656 Nicol, J. and M. Humphreys (2001) ‘Adaptive thermal comfort and sustainable thermal standards for buildings’, in Moving Thermal Comfort Standards into the 21st Century, CD ROM Nielsen, P. V. (2007) ‘Analysis and design of room air distribution systems’, HVAC and Research, vol 13, pp987–997 Nielsen, P. V., S. Murakami, S. Kato, C. Topp and J. H. Yang (2003) Benchmark Tests for a Computer Simulated Person, Version of 7 November 2003, Aalborg University, Aalborg, Denmark Nielsen, P. V., N. M. Bartholomaeussen, E. Jakubowska, H. Jiang, O. T. Jonsson, K. Krawiecka, A. Mierzwjewski, S. J. Thomas, K. Trampezynska, M. Polak and M. Soennichsen (2007) ‘Chair with integrated personalized ventilation for minimizing cross infection’, in Roomvent 2007, FINVAC ry, Helsinki

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_06

6/12/08

210

3:33 PM

Page 210

M. SANDBERG, H. LUNDSTRÖM, H. O. NILSSON AND H. STYMNE

Nilsson, H. O. (1999) The International Thermal Manikin Network, http://hem.passagen.se/nilssons/mer_manikin_network.htm Nilsson, H. O. (2004a) Comfort Climate Evaluation with Thermal Manikin Methods and Computer Simulation Models, Royal Institute of Technology, University of Gävle and the Swedish National Institute for Working Life, National Library of Sweden, Sweden, http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-3726UT Nilsson, H. (2004b) ‘Evaluation and visualisation of perceived thermal conditions’, European Journal of Applied Physiology, Special Issue 5I3M, pp714–716 Nilsson, H. and I. Holmér (2003) ‘Comfort climate evaluation with thermal manikin methods and computer simulation models’, International Journal of Indoor Air Quality and Climate, vol 13, pp28–37 Nilsson, H. O., H. Brohus and P. V. Nielsen (2007) ‘CFD modeling of thermal manikin heat loss in a comfort evaluation benchmark test’, in Proceedings of the Roomvent 2007 Conference, Helsinki, Finland, 13–15 June, www.roomvent2007.org Nordtest (2000) NT VVS 122-2000 Jet Momentum Flux: Determination Based on Weighing of Air Issuing from a Supply Device, Nordtest, Espoo, Finland NREL (National Renewable Energy Laboratory) (2006) ‘ADAM: Thermal Manikin’, www.nrel.gov/vehiclesandfuels/ ancillary_loads/adam.html Okuyama, H. (1990) ‘System identification theory of the thermal network model and an application for multi-chamber air flow measurement’, Building and Environment, vol 25, pp349–363 Oseland, N. A. (1995) ‘Predicted and reported thermal sensation in climate chambers, offices and homes’, Energy and Buildings, vol 23, pp105–115 Parson, K. (1993) Human Thermal Environments, Taylor and Francis, London Popiolek, Z., A. K. Melikov, F. E. Jorgensen, W. Finkelstein and T. Sefker (1998) ‘Impact of velocity and temperature fluctuations on the accuracy of low-velocity measurements indoors by thermal anemometers, ASHRAE Transactions, vol 104, no 1, part B, pp1507–1518 Sandberg, M. (1984) ‘The multi-chamber theory reconsidered from the viewpoint of AOR quality studies’, Building and Environment, vol 19, pp221–233 Sandberg, M. (1994) ‘Measurement techniques in room air flow’, in Roomvent ‘94, Krakow, Poland Sandberg, M. (2007) ‘Whole-field measuring methods in ventilated rooms’, HVAC and Research Journal, vol 13, pp951–970 Sandberg, M. and Etheridge, D. (1996) Building Ventilation, John Wiley, Chichester Sandberg, M. and M. Mattsson (1993) Density Currents Created by Supply from Low Velocity Devices, Research report, National Swedish Institute for Building Research, Gävle Settles, G. S. (2001) Schlieren and Shadowgraph Techniques, Springer Verlag, Berlin Soloff, S. M., R. J. Adrian and Z.-C. Liu (1997) ‘Distortion compensation for generalized stereoscopic particle image velocimetry’, Measurements Science Technology, vol 8, pp1441–1454 Streblow, R., R. Rank and D. Muller (2007) ‘An analysis of thermal plume above a cylindrical heat source in a room with a vertical temperature gradient’, in Seppänen, O. and Säteri, J. (eds) Proceedings of Roomvent 2007, FINVAC ry, Helsinki Stull, R. B. (1988) An Introduction to Boundary Layer Meteorology, Kluwer Academic Publishers, Dordrecht, The Netherlands Stymne, H., M. Sandberg and M. Mattsson (1991) ‘Dispersion pattern of contaminant in a displacement ventilated room – implications for demand control’, in Proceedings of the 12th AIVC Conference, Ottawa, Canada Stymne, H., G. Emenius and C. A. Boman (2006) ‘Passive tracer gas measurements of long term variation of ventilation in three Swedish dwellings’, International Journal of Ventilation, vol 5, pp333–343 Tavoluaris, S. (2005) Measurements in Fluid Mechanics, Cambridge University Press, New York VDI (2000) Umweltmeteorologie – Physikalische Modellierung von Strömungs- und Ausbreitungsvorgängen in der atmosphärischen Grenzschicht – Windkanalanwendungen, Verein Deutscher Ingenieure e.V., Düsseldorf Germany

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 159–210

ABER_CH_07

6/12/08

3:34 PM

Page 211

publishing for a sustainable future

7

A Review of Optical Properties of Shading Devices Athanassios Tzempelikos

Abstract The optical properties of shading devices have an impact upon the daylighting and thermal performance of a building’s perimeter spaces. This chapter presents an overview of past and current approaches in computing the optical properties of shading devices. Optical properties are determined by the different components of visible and solar transmittance, reflectance and absorptance. Emphasis is given to the three major types of shading used today: venetian blinds, roller shades/screens and draperies. Comparisons between experimental procedures and analytical and numerical modelling techniques are presented, and limitations in existing models and standards are discussed. The final section describes the latest generalized methods and models for characterizing complex fenestration/shading systems.

■ Keywords – shading; optical properties; venetian blinds; roller shades

NOMENCLATURE

V E Es F1→2 J s h I f nr Po g AS D

standard photopic luminous efficiency function illuminance (lux) normalized spectral solar power distribution view/shape factor from 1 to 2 irradiance (W/m2) length/slat width (m) distance between slats (m) irradiance (W/m2) view factor number of inter-reflections portion for each specular inter-reflection (percentage) total solar energy transmittance (percentage) aspect ratio (percentage) Diameter (m) ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

doi:10.3763/aber.2008.0207 ■ © 2008 Earthscan ■ ISSN 1751-2549 (Print), 1756-2201 (Online) ■ www.earthscanjournals.com

ABER_CH_07

6/12/08

212

S L dw

3:34 PM

Page 212

ATHANASSIOS TZEMPELIKOS

material spacing (shades) (m) luminance (cd/m2) solid angle (sr)

GREEK LETTERS

α αs αp β γ γs θ λ ρ σ τ ϕ Λ Ω

absorptance/solar altitude (percentage/degree) relative surface solar altitude (degree) profile angle (degree) blind tilt angle/surface inclination (degree) surface orientation (degree) relative surface solar azimuth (degree) solar incidence angle/polar coordinate (degree) wavelength (m) reflectance (percentage) shining factor (percentage) transmittance (percentage) solar azimuth/polar coordinate (degree) equivalent conductance for inward-flowing fraction (W/m2K) cut-off angle (degree)

SUBSCRIPTS/SUPERSCRIPTS

f b dir diff sol vis s ex p sky sys h gl bl total in sc sh x y i, j, k

front back direct diffuse solar visible spectral/solar/surface exterior profile sky radiation system property hemispherical glazing venetian blind total solar energy property inward flow/interior screen roller shade horizontal property vertical property surface indices

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 213

A Review of Optical Properties of Shading Devices

213

INTRODUCTION The optical properties of shading devices have a decisive impact upon the thermal and daylighting performance of perimeter spaces. It is therefore surprising that, currently, there is no widely recognized procedure for measuring or calculating shading optical properties (ISO 15099, 2003). Consequently, many shading manufacturers provide limited or incomplete technical specifications of the optical and thermal properties of their products. The data provided tends to be integral values that cannot satisfactorily resolve the often strong angular dependency of shading devices. As a result, the products cannot be reliably modelled in energy simulation tools since the software engines usually import sets of properties from existing databases (WINDOW 5/THERM, 2003). Optical properties of window shading refer to transmittance (τ), absorptance (α) and reflectance (ρ) of window shadings – with or without the window itself (for both the visible and the solar spectrum). This chapter presents an overview of modelling and experimental approaches previously used to predict the properties of shading devices, and reviews the assumptions and limitations of these methodologies. Despite the large number of shading devices available, the chapter focuses on the three most common types of moveable shading devices used today in buildings: venetian blinds, roller shades and draperies. Exterior fixed window obstructions, such as light shelves, awnings, overhangs and fins, are excluded from the review since the properties of these systems are generally known. Moreover, occupants usually have no control over these fixed systems; therefore, they are generally easy to model and use in building energy simulation software. The location of the shading layer (interior, intermediate or exterior) is an important factor; in this review, all three possible locations for the shading layer are considered: interior, exterior and between the glazing panes (integrated). Research on shading devices is divided into three categories: analytical models, numerical models (e.g. ray-tracing, Monte Carlo) and experimental methods. Hybrid models also exist (e.g. combining some experimental results with computer simulation). The chapter summarizes and extends the findings of an earlier review report (Tzempelikos, 2005).

OPTICAL PROPERTIES OF WINDOW SHADING DEVICES The optical properties of windows equipped with shading devices depend upon: 1 the type of glazing(s) used; 2 the type of shading device used; and 3 the location of the shading device with regard to window glazing(s). Venetian blinds, roller shades and draperies can be treated as shading layers located parallel to the glazing plane, with intimate thermal and optical contact. The case of venetian blinds is, of course, more complicated since they are a ‘discrete’ shading device, in contrast with shades and draperies, which are considered ‘continuous’ or ‘homogeneous’ systems. The shading layer is usually modelled as a one-dimensional layer similar to a pane or film, so the optical properties of the shading system are

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

214

3:34 PM

Page 214

ATHANASSIOS TZEMPELIKOS

Source: Original material for this chapter

FIGURE 7.1 Glazing system with N layers

a function of the device geometry and the position in the assembly. In general, the equations remain similar to those for window glazing. The solar optical properties of windows are usually calculated from the optical properties of the individual glass layers. The three basic quantities – transmittance (τ) (assumed the same in front and back), front reflectance (ρf) and back reflectance (ρb) – are used (usually given for normal incidence) and they are wavelength dependent. These properties can be estimated according to ISO 9050 (2003), or using any other standard tool or method with acceptable results. Figure 7.1 shows a glazing system consisting of N glass layers separated by non-absorbing gas layers. Solving the recursion equations between layers (i) and (j), including inter-reflections, yields: τi , j − 1 τ j , j τi , j = [1] 1 − ρfj , j ρbj −1,i

ρ fi , j = ρ fi , j −1 +

ρbj ,i = ρbj , j +

τ i2, j −1 ρfj , j 1− ρfj , j ρbj −1, i τ 2j , j ρbj −1, i

1 − ρbj −1,i ρfj , j

[2]

[3]

A detailed explanation of the above may be found in Rubin et al (1998); Hollands et al (2001); ISO 9050 (2003) or ASHRAE (2005). Because the above properties are a function of wavelength, spectral-averaged values are obtained by integration over wavelength. Moreover, angular properties for coated and uncoated glazings are calculated using optics

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 215

A Review of Optical Properties of Shading Devices

215

theory. A description of glazing and window system properties is available in ASHRAE (2005). The WINDOW software (WINDOW 5/THERM, 2003) uses a large database of thermal and optical properties of different types of glazings and windows, and can perform a detailed optical and thermal analysis of all kinds of window systems (excluding shading devices). Therefore, it is used by most of the major advanced building simulation software to accurately simulate thermal and optical properties of windows (Zmeureanu, 1998). DOE-2 software imports fenestration analysis results from WINDOW 4.1 (Reilly et al, 1995). TRNSYS software has a simplified option where the user inputs the window transmittance (type 19) or window U-value (type 35), but also offers a more detailed option (type 56), where the results of WINDOW software could be used as input (TRNSYS, 2003). In ESP-r software, reflection, absorption and transmission of short-wave radiation are evaluated separately for the visible and solar spectral portions, employing a large optical properties database supported by an automated link to WINDOW software. In ESP-r, advanced glazings and shading device systems can be supported. The thermal/optical properties can vary with time to support the modelling of blinds and shading devices (which can be controlled as a function of time, irradiance or temperature). ESP-r’s special materials facility can be used to impart behaviour to individual glazing layers (i.e. transmittance switching in the case of an electrochromic layer). However, modelling a venetian blind system in ESP-r is quite complex: building the geometry is not simple; the optical properties are given for five incidence angles only; there are no correlations for convective heat transfer between the blind and the air, etc. Therefore, researchers and designers cannot use the software to model blinds. The IMAGE project (Clarke et al, 1998) was a major European project aimed at assessing the overall performance of advanced glazing systems using computer simulation. In contrast to shading devices, a substantial amount of work has been done in optical and thermal characterizations of complex glazings/prisms (Lorenz, 1998; Sullivan et al, 1998; Beck et al, 1999; Karlsson and Roos, 2000, Lorenz, 2001), producing analytical models based on detailed geometry and optics theory for the considered glazing system. A particular characteristic of a shading device compared to ‘normal’ glazings or films is that the incident solar radiation may change direction while being transmitted or reflected at the layer. For a complete evaluation of the spatial distribution of daylight, calculations using the full matrix of transmission, forward and backward reflection and absorption at each component for every angle of incidence are required. Consequently, it is important to know the following solar properties of the solar shading device for each angle of incidence: ● Transmittance: ● direct-to-direct transmittance: τdir,dir ; ● direct-to-diffuse transmittance: τdir,diff ; ● diffuse-to-diffuse transmittance: τdiff,diff .

The same respective parameters have to be calculated for reflectance: ρdir,dir, ρdir,diff and ρdiff,diff .

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

216

3:34 PM

Page 216

ATHANASSIOS TZEMPELIKOS

● Absorption:

Direct: αdir = 1 − τdir ,dir − τdir ,diff − ρdir ,dir − ρdir ,diff ;

[4]

Diffuse: αdiff = 1 − τdiff ,diff − ρdiff ,diff .

[5]

In general, the above properties are a function of wavelength. Usually, a distinction is made between visible – or luminous/daylight – transmittance, τvis, and solar transmittance, τsol. For both cases, the transmittance function has to be evaluated as a function of incidence angle θ. The visible transmittance at incidence angle θ can be defined as: τvis (θ ) = ∫ D65 (λ)V (λ)τ (λ,θ) d λ

[6]

where D65(λ) is the spectral relative power distribution of the Commission Internationale de l’Eclairage (CIE) standard illuminant and V(λ) is the standard photopic luminous efficiency function. The solar transmittance at incidence angle θ is defined as: τsol (θ ) = ∫ E s (λ)τ (λ,θ )d λ

[7]

where Es(λ) is the normalized spectral power distribution of solar radiation. For a window equipped with a layer-type shading device, the equations used for glazings remain the same in concept. However, transmittance and reflectance are split into a backward and a forward value. These two values are not necessarily equal because of redirection of radiation. The sum of (τdir,dir + τdir,diff) is equal to the direct-to-hemispherical transmittance τdir,h (and similarly for the reflectance). Note that according to ISO 15099 (2003), there is no existing international standard for measuring these optical properties. Once a beam transmitting through or reflecting from a solar shading device is split into a direct and a diffuse part, the diffuse part continues its route through the system. This implies that even for normal incidence solar radiation, for all other panes, films and shading layers in the window, the τdiff,diff and ρdiff,diff values are required. Thus, the values for normal incidence provide insufficient information. The following sections describe the methodologies and concepts used to extract information about optical and solar properties of venetian blinds, roller shades and draperies. In most cases, the shading device itself (without the window) is examined. Starting from standards and analytical models, a review is provided of computer-based numerical methods and experimental methods, and/or combinations than can be used for simulating the optical properties of shading devices. Finally, the latest methods and models of complex fenestration/shading systems are presented.

VENETIAN BLINDS Venetian blinds are possibly the most complex type of shading in terms of optical and thermal properties. They consist of separate, equally spaced horizontal louvres; therefore, their properties change with louvre characteristics, tilt angle and angle of

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 217

A Review of Optical Properties of Shading Devices

217

Source: Original material for this chapter

FIGURE 7.2 Between-glazing and exterior venetian blinds

incidence (see Figure 7.2). Most publications refer to interior daylighting conditions and control of blinds without discussing their optical/thermal properties (Lee and Selkowitz, 1995; Lee et al, 1998; Dubois, 2003). Overviews of daylighting systems include several types of venetian blinds and describe their function and characteristics, but provide very limited (Laar and Grimme, 2002) or no information (Kischkoweit-Lopen, 2002) about their optical (or thermal) properties. Because of the complexity of these systems and limitations in existing models, many researchers work experimentally, trying to measure the optical/thermal properties of different types of venetian blinds. The experimental findings can be well used for simulation and perhaps this is the most accurate way of modelling venetian blinds. The disadvantage of experiments is that they have to be repeated for all kinds of blinds, which is very time-consuming. Analytical and numerical models could handle all types of venetian blinds by changing the corresponding parameters. The disadvantage of analytical models for venetian blinds is that they are usually complex. However, there are some basic standards and analytical methods that can be used to model their properties.

ASHRAE METHOD The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE, 2005) uses the following approach: first, it categorizes venetian blind systems (as well as roller shades and draperies, as discussed later) in the exterior, between glass and interior. For all cases, solar attenuation coefficients are used. These quantities serve the purpose of the so-called ‘shading coefficient’ for glazings and they describe the interaction of the incident radiation with the shading system. In the case of exterior venetian blinds, blind systems are separated into four categories, depending upon the colour and the width over spacing ratio. For average values of transmittance and particular values of profile angles, unshaded fractions and exterior solar attenuation coefficients are given (discussed in the following section). Sample light transmittance, reflectance and absorptance data for interior medium-coloured venetian blinds (slat angle 45° and width-to-slat spacing 1.2) are provided for normal incidence only. ASHRAE provides more data on specific venetian blind products – mainly on thermal characteristics. For cases that are not covered by the tables given in ASHRAE (2005), a standard procedure for calculating optical properties based on Equation 1 is described in

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

218

3:34 PM

Page 218

ATHANASSIOS TZEMPELIKOS

detail. In general, the methods and information described in ASHRAE (2005) are not considered a systematic, detailed analysis of venetian blind systems, but provide some data for estimating optical and thermal properties of particular venetian blind products (the data, however, is not generalized). Another disadvantage is that venetian blinds are assumed to be perfect diffusers.

ISO 15099 ISO 15099 (2003) also presents a solution technique for multilayer solar-optical models. For a shading device consisting of parallel slats, optical properties can be determined as a function of slat properties, geometry and position. The calculation is based on three assumptions: 1 The slats are non-specular reflecting. 2 Any effects on window edges may be ignored. 3 Due to the assumption of non-specular reflection, a slight curving of the slats may be ignored. Based on these assumptions, the procedure considers two adjacent slats and subdivides each of the slats into five equal parts, front (f) and back (b) (see Figure 7.3).

Source: ISO 15099 (2003)

FIGURE 7.3 Discretization of venetian blinds

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 219

A Review of Optical Properties of Shading Devices

219

Every slat is divided into five elements (the benefit of considering more elements is supposed to be negligible). Notice that different properties can be assigned to every element, particularly to every side of the slat. The process described below has to be solved for every wavelength band required by the properties of the elements or by the rest of the transparent system where the shading device is installed. The irradiance/illuminance on each subsurface (f,i) and (b,i), with i from 0 to n (n = 6 here, including the vertical side as n= 0) and normally for each spectral interval is calculated from: E f ,i = ∑ [(ρf ,k + τb ,k )E f ,k i Ff ,k →f ,i + (ρb ,k + τf ,k )E b ,k i Fb ,k →f ,i ]

[8]

E b ,i = ∑ [(ρb ,k + τ f ,k )E b ,k i F b ,k →b ,i + (ρ f ,k + τb ,k )E f ,k i F f ,k →b ,i ]

[9]

k

k

where Fp→q is the view (shape) factor from surface p to surface q. The radiosity from the external environment (Jex) is equal to the incident solar radiation, and it is assumed that there is no internal radiation source (reflections from room interior, etc.).

Diffuse-to-diffuse transmission and reflection Because specular reflection is excluded, view factors between sub-surfaces and between the external environment and layers can be calculated using known diffuse radiation exchange theory (Siegel and Howell, 1972). Diffuse-to-diffuse transmission is the ratio of the radiation reaching the internal environment, divided by the incident solar radiation, after solving the following equations (this procedure is not explained in ISO 15099, 2003): E b ,n

τdiff , diff =

Jex

[10]

for each wavelength. Similarly, for diffuse-to-diffuse reflection: ρdiff ,diff =

E b ,ex Jex

[11]

Direct-to-direct transmission and reflection The beam radiation passing through the slats without touching can be calculated from the tilt angle and aspect ratio of the slats for any incidence angle θ (see Figure 7.4). The directto-direct transmittance is equal to: τdir ,dir =

E dir ,dir Jex

[12]

for each wavelength and is a function of the incidence angle. No information is given on how to actually calculate the direct-to-direct transmission. Obviously, the direct-to-direct reflectance (ρdir,dir) is equal to zero since the slats are assumed to be perfect diffusers.

Direct-to-diffuse transmission and reflection First, the parts of the slat (k) directly irradiated are computed for a certain incidence angle, θ (see Figure 7.5). The view factors between the external environment and directly irradiated

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

220

3:34 PM

Page 220

ATHANASSIOS TZEMPELIKOS

parts are equal to unity, whereas view factors between the inside and outside are set to zero to exclude the direct-to-direct component. The obvious equations are given for each incidence angle: ⎪⎧⎪ E f ,n ⎪⎫⎪ ⎪ ⎪ ⎪⎧⎪τdir ,diff ⎪⎫⎪ ⎪⎪ Jex ⎪⎪ ⎪⎬ [13] ⎨ ⎬ = ⎪⎨ ⎪⎪ρdir ,diff ⎪⎪ ⎪⎪E b ,n ⎪⎪ ⎪⎩ ⎪⎭ ⎪ ⎪ ⎪⎪ J ⎪⎪ ⎪⎩ ex ⎪⎭

Jex Edir,dir

Source: Original material for this chapter

FIGURE 7.4 Direct-to-direct transmittance

Source: ISO 15099 (2003)

FIGURE 7.5 Directly irradiated parts of venetian blind

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 221

A Review of Optical Properties of Shading Devices

221

Absorption Radiation that is neither transmitted nor reflected is absorbed by the blind, according to Equations 4 and 5 for the direct and the diffuse components. This is the procedure recommended by ISO 15099 (2003). All of the calculated quantities are essentially a function of wavelength. Therefore, spectral averaging over the visible and the solar spectrum should give the average visible and solar transmittance, reflectance and absorptance of venetian blind systems. Following the above procedure, ISO 15099 (2003) has an appendix containing tables of the optical properties of venetian blinds for four different types of slats, two different slat angles (45° and 80°) and incident solar radiation values, including a translucent white slat system.

ENERGYPLUS METHODOLOGY EnergyPlus (2007) models venetian blinds as a series of equally spaced slats that are oriented horizontally or vertically (Crawley et al, 2002; EnergyPlus, 2007). The procedure is similar to the one described by ISO 15099 (2003); but the geometry and equations are all clearly stated and analysed. All of the slats are assumed to have the same optical properties. The overall optical properties of the blind are determined by the slat geometry (width, separation and angle) and the slat optical properties (front-side and back-side transmittance and reflectance). Blind properties for direct radiation are also sensitive to the ‘profile angle’, which is the angle of incidence in a plane that is perpendicular to the window plane and to the direction of the slats. The following assumptions are made in calculating the blind optical properties: ● The spectral dependence of inter-reflections between slats and glazing is ignored;

spectral-average slat optical properties are used. ● The slats are perfect diffusers. They have a perfectly matte finish so that reflection

from a slat is isotropic (hemispherically uniform) and independent of angle of incidence (i.e. the reflection has no specular component). This also means that absorption by the slats is hemispherically uniform with no incidence angle dependence. If the transmittance of a slat is non-zero, the transmitted radiation is isotropic and the transmittance is independent of the angle of incidence. ● Inter-reflection between the blind and wall elements near the periphery of the blind is ignored. ● If the slats have holes through which support strings pass, the holes and strings are ignored. Any other structures that support or move the slats are ignored. The direct-to-direct and direct-to-diffuse transmittance of a blind is calculated using the slat geometry depicted in Figure 7.6, which shows the side view of one of the cells of the blind. For the case highlighted, each slat is divided into two segments so that the cell is bounded by a total of six segments, denoted by s1 to s6. The lengths of s1 and s2 are equal to the slat separation, h, which is the distance between adjacent slat faces. Lengths s3 and s4 are the segments illuminated by direct radiation. In the case shown in Figure 7.6, the cell receives radiation by reflection of the direct radiation incident on s4 and (if the slats have non-zero transmittance) by transmission through s3, which is

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

222

3:34 PM

Page 222

ATHANASSIOS TZEMPELIKOS

Note: In the figure, the profile angle is shown as ϕs (αp in the text) and slat angle is shown as ϕb (βp in the text) Source: EnergyPlus (2007)

FIGURE 7.6 View of a cell formed by adjacent slats showing how the cell is divided into segments, si, for the calculation of direct solar transmittance; and side view of a cell showing case where some of the direct solar passes between adjacent slats without touching either of them

illuminated from above. The goal of the blind direct transmission calculation is to determine the direct and diffuse radiation leaving the cell through s2 for unit direct radiation entering the cell through s1.

Direct-to-direct blind transmittance As shown in Figure 7.6, some of the direct radiation passes through without hitting the slats. Using the geometry, the equation is as follows: τdir ,dir = 1 −

w h

[14]

where w can be expressed as a function of slat angle and profile angle: w =si

cos(β − αp ) cos(αp )

[15]

where s is the slat width, αp is the profile angle and β is the slat angle.

Direct-to-diffuse transmittance This is performed by using the radiosity method and view factors, similarly to the method described in ISO 15099 (2003). However, all of the equations are explicitly presented and the view factors between the six segments (two on each slat and two on the openings) are computed using the cross-string method.

Diffuse-to-diffuse transmittance Assuming that all surfaces are perfect diffusers, each slat is divided in two new sections of equal length and the radiosity theory is again applied. The diffuse-to-diffuse calculations are performed separately for solar, visible and long-wave slat properties in order to obtain

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 223

A Review of Optical Properties of Shading Devices

223

Note: ϕs denotes the profile angle, αp in the text Source: EnergyPlus (2007)

FIGURE 7.7 Side view of horizontal blinds for calculating optical properties for sky and ground diffuse radiation

the corresponding solar, visible and long-wave blind properties. Moreover, the diffuse-todiffuse solar properties of blinds are separated based on radiation coming from the ground and radiation coming from the sky. The properties are obtained by integrating sky and ground elements, as described in Equation 16 and Figure 7.7, treating each element as a source of direct radiation of intensity I(αp) incident on the blind at profile angle αp: π 2

sky τdiff , diff

=

∫ ⎡⎣⎢τdir, dir (αp ) + τdir, diff (αp )⎤⎦⎥ i Isky (αp ) i cos(αp )d αp 0

π 2

∫ Isky (αp ) i cos(αp )d αp

. [16]

0

If the sky radiation is uniform, then Isky is independent of αp and Equation 16 becomes: sky τdiff , diff

π 2

= ∫ (τdir ,dir + τdir ,diff ) i cos(αp )d αp .

[17]

0

For the ground-diffuse reflectance, αp is integrated from –π/2 to 0. A correction is also made to the blind solar/optical properties in order to account for the amount of radiation

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

224

3:34 PM

Page 224

ATHANASSIOS TZEMPELIKOS

incident on the blind that is reflected or absorbed by slat edges (thickness). This correction is not available in any other method.

Blind-glazing system EnergyPlus (2007) also provides all of the required equations for calculating transmittance and absorptance values for a combination glazing/blind system for both interior and exterior blinds. These analytical expressions, presented below, can be directly used for building simulation models.

Interior blind The direct-to-hemispherical transmittance (sum of τdir,dir+τdir,diff) for a glazing/blind system with an interior blind is given by: bl, f bl , f gl, b ⎞ ⎛ τdiff i ρdir ⎜⎜ bl , f , diff (αp ) i ρdiff ⎟ ⎟⎟ sys gl bl τdir τ ( α ) + τ ( α ) + ⎟⎟ p dir , diff p , h (θ, αp ) = τdir (θ ) i ⎜ dir , dir ⎜⎜ bl, f gl, b ⎟⎠ 1− ρdiff i ρdiff ⎜⎝

[18]

where θ is the incidence angle, αp the profile angle, ‘gl’ stands for glazing and ‘bl’ stands for blind. The diffuse system transmittance is equal to: sys τdiff , diff =

gl bl, f τdiff i τdiff , diff bl, f gl, b 1 − ρdiff i ρdiff

[19]

where the blind diffuse transmittance is described in Equation 17 for sky or ground radiation.

Exterior blind The direct-to-hemispherical transmittance for a glazing/blind system with an exterior blind is given by: bl, b ⎞ gl gl, f ⎛ gl bl τdiff i ρdir i ρdir ⎜⎜ gl , diff ⎟ ⎟⎟ τdir ,diff (αp ) i τdiff sys bl, f τdir ( θ , α ) = τ ( α ) i τ ( θ ) + . + ⎜ ⎟⎟ p dir , dir p ⎜ dir ,h gl, f bl, b gl, f bl, b ⎜⎜⎝ ⎟⎠ 1− ρdiff i ρdiff i ρdiff 1− ρdiff

[20]

The diffuse system transmittance is equal to: sys τdiff , diff =

bl, f gl τdiff , diff i τdiff gl, f bl, b 1− ρdiff i ρdiff

.

[21]

Overall, the ISO 15099 (2003) method is similar, in principle, to the method explained thoroughly in EnergyPlus (2007), except that the slats are divided into two parts instead of five. The ISO 15099 (2003) and EnergyPlus (2007) results agree to within 12 per cent, except for very small transmittance values. A comparison table given in the EnergyPlus guide (2007, Table 28) provides more detailed data for particular products.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 225

A Review of Optical Properties of Shading Devices

225

PREVIOUS RESEARCH ON THE SOLAR OPTICAL PROPERTIES OF VENETIAN BLINDS The first (and pioneer) extensive study on the effect of slat-type shading on heat and light transmission was published by Parmelee and Aubele in 1952. For the first time, a distinction between specular and diffuse slat surfaces was made. Transmittance and absorptance of specular and diffuse slats were then plotted for different spacing (opening) ratios, blind tilt angles and profile angles based on analytical expressions presented in the appendices of the paper. Opening ratios were expressed as a function of blind geometry and profile angle. The main assumptions in this study were: ● ● ● ●

the slats have negligible thickness; slats are flat and opaque; edge effects are negligible; and graphs provided are strictly applicable to monochromatic radiation.

One of the most significant efforts in analytically determining blind solar optical properties was completed by Pfrommer et al (1996). An analytical model derived from the fundamental geometrical and physical characteristics of blind systems was developed for solar radiation transport through slat-type blinds so that it could be directly used in a thermal simulation engine. Light transmission through blinds was separated according to four paths: 1 direct-to-direct transmittance; 2 direct-reflected transmittance from a slat surface, then processed as specular/diffuse/combinations; 3 diffuse-to-diffuse directly transmitted; and 4 diffuse-reflected transmittance on a slat surface and then transmitted. Direct-to-direct transmittance is calculated from geometrical properties of the blind (opening ratio described by Parmelee and Aubele, 1952). For direct-reflected transmittance, a shining factor, σ, is used to separate direct-to-diffuse and specularreflected components. Direct-to-diffuse transmittance is calculated by recursive calculations for two inter-reflections: τdir ,diff = (ρ i f1 + ρ 2 i f2 i f3 ) i (1 − τdiff )

[22]

where ρ is the slat reflectance (assumed constant and independent of wavelength), f1 is a view factor between the illuminated slat area and the inside, f2 is between the illuminated slat area and the upper slat, and f3 is between the upper slat and the inside. These view factors were approximated based on a method for infinitely small illuminated areas (i.e. configuration factors). The term (1 – τdiff) accounts for the absorbed and directly reflected component. Transmittance of the specular-reflected component is calculated using the possible number of inter-reflections from the system geometry, nri, and the portion for each specular inter-reflection, Poi: spec τdir = ∑ρ

nr 1

i Poi .

[23]

i

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

226

3:34 PM

Page 226

ATHANASSIOS TZEMPELIKOS

Diffuse-to-diffuse transmittance is computed by integrating incoming diffuse radiation between the cut-off angles (Ω1, Ω2) (see Figure 7.8) and is given by: Ω2

∫

τdiff , diff =

Isky Ω1 90°

∫

i cos(Ω) i τ (Ω) d Ω [24]

Isky i cos(Ω) i d Ω

−90°

where Isky is the sky radiation and τ(Ω) is calculated from: τ (Ω) = 1 −

s i sin(β ) − s i cos(β ) i tan(Ω) , h

[25]

s being the slat width, β the slat angle and h the distance between slats. An excellent comparison of Parmelee and Aubele (1952) with the Pfrommer et al (1996) model and with experimental measurements identifying limitations and future work was published by Chantrasrisalai and Fisher (2004). The Advanced Window Information System (WIS) is a European software tool for calculating optical and thermal properties of commercial and innovative window systems, a big part of which are venetian blinds. One of the unique elements in this software is the combination of glazings and shading devices. In fact, the way in which WIS treats the solar optical properties of a ‘layer-type’ shading device has been the basis

Cut-off angle Ω1 Cut-off angle Ω2

Source: Pfrommer et al (1996)

FIGURE 7.8 Description of cut-off angles

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 227

A Review of Optical Properties of Shading Devices

227

of ISO 15099 (2003). A comparison between WIS, Monte Carlo simulation and radiosity results is available by Van Dijk and Oversloot (2003). In the WIS optical sub-model for blinds, the spectral direct-hemispherical reflectance of the slats measured at normal incidence is used throughout, although light strikes them at different angles when passing between slats. The main assumptions are, first, that the slats are considered Lambertian surfaces and, second, that the curvature of the slats is neglected. Using the same procedure with ISO 15099 (2003), the slats are separated into five sections, as depicted in Figure 7.3. For diffuse-to-diffuse properties, view factors are calculated based on a three-dimensional model (LAMAS), developed under the European PASCOOL project (Molina et al, 2000). Parasol version 3 (Hellstrom et al, 2007) is another computer programme for calculating the solar optical and thermal properties of windows with shades and blinds. Users can input direct and diffuse transmittance and reflectance. For shades, the transmittance is independent of the angle of incidence. For blinds, the ISO 15099 method is used, in general, and specular and diffuse reflection are taken into account. Effective properties of absorptance, reflectance and transmittance of venetian blinds were calculated by Yahoda and Wright (2004), based on a radiosity theory. Results are presented as a function of slat aspect ratio and blind tilt angle in categorized graphs. In addition, the effect of curved slats was analysed by modifying view factors to account for self-viewing. Kuhn (2006) developed a generalized analytical model for the solar optical properties of interior/exterior blinds, shades and other shading/glazing systems. The total solar energy transmittance for a glazing/shading system with interior venetian blinds is calculated by the following expression as a function of solar incidence angle, θ, profile angle, αp, blind tilt angle, β, and wavelength, λ: in gtotal (θ, β,λ) = τtotal (αp ,θ, β,λ) + qtotal (αp ,θ, β,λ)

[26]

where τtotal (αp , θ, β, λ) is the radiation (at wavelength λ) entering the room, equal to: τtotal (αp ,θ, β,λ) = τ gl (θ,λ) i τ bl (αp , β,λ) +
   Directly transmitted

gl bl (λ) i τdiff (β,λ ) τ gl (θ,λ) i ρbl (αp , β,λ) i ρdiff gl 1 − ρbl (αp , β,λ) i ρdiff (λ)
   [27] Reflected back inside (from glass )

and qintotal(αp, θ, β, λ) is the inward-flowing fraction of radiation (at wavelength λ), equal to: gl τ gl (θ,λ) i ρbl (αp , β,λ) i αdiff (λ) in q total (αp ,θ, β,λ) = g gl (θ,λ) − τ gl (θ,λ) + (1−κ ) i gl bl
 1− ρ (αp , β,λ) i ρdiff (λ) Absorbed in glazing 


Reflected by blind and then absorbed in glazing

⎛ Λinternal ⎞⎟ ⎟⎟ i ⎜⎜⎜1 − ⎜ Λ2 ⎟⎠ (λ) ⎝
  bl

+ τ gl (θ,λ ) i

α (αp , β )

bl 1− ρdiff

[28]

gl ( β ) i ρdiff

Absorbed by the blind

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

228

3:34 PM

Page 228

ATHANASSIOS TZEMPELIKOS

where: ● κ∈[0,1] is an adjustable parameter for the outward-flowing fraction of energy that is

reflected by the interior blind and then absorbed by the glazing; ● α gl is the glazing absorptance; ● Λ2 = 18W/m2K; ●

Λinternal =

1

1 U gl

+

1 , where Ugl is the glazing thermal conductance. Λ2

The model assumes that the solar optical properties are known (previously measured) for each wavelength (i.e. τbl(αp, β)) and identifies the parameters that determine each property. It can be used for exterior blinds (appropriately modified) and for other shading systems with known properties. Wright and Kotey (2006) developed another general method for predicting beam and diffuse radiation through multilayered glazing/shading systems. The method allows for unequal front and back layer properties, and detailed solutions were obtained using matrix manipulation. Kotey and Wright (2006) also published the algorithm for the case of venetian blinds. A disadvantage of all of these methods is that the direction of transmitted light cannot be predicted. Therefore, the models are good for predicting solar radiation transmitted through systems of blinds; but they are not accurate enough for detailed interior work, plane illuminance prediction and daylighting simulation. Another model for the optical properties of venetian blinds was inspired by the fact that, for many cases, the slat reflectance is highly anisotropic, strongly peaked about the specular direction (Breitenbach et al, 2001). The spectral specular and diffuse reflectances at normal incidence are required as inputs to the model. In this study, Breitenbach et al (2001) measured the visible and total solar energy transmittance of an integrated venetian blind system using a gonio-spectrometer for different slat angles and incidence angles. They then developed a model as follows: for the light that is diffusely reflected and that undergoes two or more inter-reflections, a fraction F of it is assumed to proceed as a specular component at each reflection. The fraction 1-F is reflected backwards. It retraces its path, with a fraction absorbed at each reflection on its way back. This model therefore contains one adjustable parameter, F, whose magnitude depends upon the degree to which the diffusely reflected light is peaked about the specular direction. The fraction of incident light undergoing n specular reflections was also calculated and a good agreement was obtained between measured and predicted values. The ALTSET project was part of the European Commission’s Standards Measurement and Testing programme from 1996 to 1999. Its main objective was to develop procedures to measure the angular dependence of luminous and total solar energy transmittance of complex glazings. Fixed and variable venetian blinds were included in the project. An overview of recent developments in modelling optical and thermal properties of complex windows (Rosenfeld et al, 2000) explains and compares available models, such as the Breitenbach et al (2001) model with measurements and WIS results.

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 229

A Review of Optical Properties of Shading Devices

229

Due to the complexity of the problem and inaccuracies in the existing models, many researchers prefer to work experimentally, obtaining the optical properties of venetian blinds and then using them to create simulation models. An example is the work published by Papamichael and Beltrán (1993), named the integrated directional coefficients method. This hybrid method is based on the combination of scale-model photometry and computer simulation. Physical scale models are first used to experimentally determine a set of ‘directional illuminance coefficients’. These are defined as the ratio of the interior illuminance on a point due to light coming from a set of spherical coordinates (ϕ,θ), divided by the exterior illuminance due to light incident from a direction normal to (ϕ,θ). The general state-of-the-art method is described in the section on solar optical properties of complex fenestration and shading systems (page 233). An experimental procedure for measuring the average solar transmittance of nonhomogeneous shading devices, such as venetian blinds, is presented by Aleo et al (1994), to be used together with the LAMAS project. Two pyranometers were used (one exterior and one moving interior), and the average direct and diffuse solar transmittance was measured for three different types of venetian blinds at several tilt angles. In another study, the total and the diffuse-to-diffuse solar transmittance of an integrated venetian blind system was measured and modelled as a function of blind tilt angle and profile angle (Athienitis and Tzempelikos, 2002; Tzempelikos and Athienitis, 2003). In the same study, the ‘optimum’ blind tilt angle, which functions as a means of blocking direct sunlight while maximizing view, was calculated as a function of blind and solar geometry.

ROLLER SHADES, SCREENS AND DRAPERIES ROLLER SHADES Roller shades are probably the most commonly used shading devices due to their easy operation and relatively low price. They are usually manually controlled to satisfy occupants’ needs concerning privacy and glare issues (see Figure 7.9). The optical

Source: Chapter author

FIGURE 7.9 Examples of interior roller shades

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

6/12/08

230

3:34 PM

Page 230

ATHANASSIOS TZEMPELIKOS

properties of roller shades are generally assumed to be constant and independent of incidence angle: they are considered perfect diffusers. EnergyPlus (2007) model ‘shades’ are perfect diffusers; however, a distinction between ‘shades’ and ‘screens’ is made, as explained below. This means that direct radiation incident on the shade is reflected and transmitted as hemispherically uniform diffuse radiation: there is no direct component of transmitted radiation. It is also assumed that the transmittance, reflectance, and absorptance are the same for the front and back of the shade and are independent of angle of incidence. In this case, the glazing/shade system transmittance is equal to: sys τdiff (θ ) =

τ gl (θ ) i τsh

[29]

gl 1 − ρdiff i ρsh

where τsh, ρsh, ρdiffgl are constant diffuse properties of the shade and the glass and τgl(θ) is the glass transmittance, which depends upon the incidence angle. In published work about modelling transmitted daylight through roller shades (Tzempelikos and Athienitis, 2007), this same approach was followed. Nevertheless, recent studies have shown that roller shades do not always perform like perfect diffusers, although they could have a specular component and directional properties, depending upon their construction and material (M. Collins, pers comm, 2007). Moreover, there is a strong dependence upon wavelength – visible and the solar optical properties are not equal (see Figure 7.10).

20 Transmittance (30 deg)

18

Transmittance (45 deg) 16 Shade transmittance (%)

ABER_CH_07

Transmittance (60 deg)

14 12 10 8 6 4 2 0 0

500

1000

1500

2000

2500

Wavelength (nm) Source: Measurements by Dr M Collins, University of Waterloo

FIGURE 7.10 Spectral and directional transmission properties (different incidence angles) of a common roller shade

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 231

A Review of Optical Properties of Shading Devices

231

SCREENS EnergyPlus (2007) differentiates between ‘shades’ and ‘screens’, defining ‘screens’ as exterior insect protection devices made up of metallic or non-metallic materials. The model assumes that the screen is composed of intersecting orthogonally crossed cylinders, with the surface of the cylinders assumed to be reflecting diffusely. The cylinder diameter and spacing are supposed to be known and equal in both dimensions (see Figure 7.11). The exterior screen is modelled as a planar semi-transparent sheet having specular transmittance that is dependent upon the angle of incidence of beam solar radiation. The screen transmittance algorithm includes two components. The first accounts for the blockage of the sun’s rays by the screen material (i.e. the beam solar radiation passing through the screen openings without hitting the screen material). The aspect ratio (AS) of the screen must initially be determined. This is equal to the screen material (cylinder) diameter, D, divided by the material spacing, S: AS =

D S

[30]

Then, the direct beam transmittance is computed as a function of the aspect ratio (AS), the surface solar azimuth, γs (equal to the difference between the surface orientation, γ, from true north, and the solar azimuth, ϕ), and the relative surface solar altitude, σs (equal to the difference between the surface inclination, βs (from vertical), and the solar azimuth, ϕ). For a vertical window/shade, αs would be equal to the solar zenith angle (or 90°). Given the screen diffuse reflectance, ρsc, and the screen aspect ratio, AS, the model takes the direction of solar sc incidence, αs and γs, and calculates the direct beam transmittance, τdir , dir (αs , γs ) , as follows: y sc x τdir , dir (αs , γs ) = τdir (αs , γs ) i τdir (αs , γs )

[31]

where τ xdir and τ ydir are the horizontal and vertical components of direct beam transmittance of the shade, respectively, calculated by the solar and screen geometry: ⎡ cos(αs ) i cos(γs ) ⎤⎥ π x )] ⎥ [32] τdir (αs , γs ) = 1− AS i ⎢⎢ cos(ξ ) + sin(ξ ) i tan(µ) i 1 + cot 2 [ − cos−1( cos(µ) 2 ⎥⎦ ⎢⎣

Source: EnergyPlus (2007)

FIGURE 7.11 Screen model rendering of intersecting orthogonal crossed cylinders

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

232

3:34 PM

Page 232

ATHANASSIOS TZEMPELIKOS

and: y τdir (αs , γs ) = 1 − AS

⎡ × ⎢⎢ cos[tan−1(tan(αs ) i sec(γs ))] + sin[tan−1(tan(αs ) i sec(γs ))] i tan(αs ) i ⎢⎣

1+ cot2 (

⎤ π − γs ) ⎥⎥ [33] 2 ⎥⎦

where µ and ξ are given by: µ = cos−1 cos2 (αs ) i cos2 (γs ) + sin2 (αs )

[34]

⎡ cos(αs ) i cos(γs ) ⎤ ξ = tan−1 ⎢ tan(µ) i sec(cos−1( ))⎥ . ⎢ ⎥ cos(µ) ⎣ ⎦

[35]

The second part, τ scdir,diff (αs, γs), accounts for the additional flux of transmitted beam solar radiation by diffuse reflectance (scattering) from the screen material (direct-to-diffuse transmittance) and is calculated with an empirical algorithm (EnergyPlus, 2007). It should be noted that although this method has some limitations (standard screen geometry, equal spacing, empirical diffuse inward fraction), it is the only analytical method available today for transmission through screens. Moreover, although EnergyPlus (2007) refers to ‘screens’, the method can be well applied to roller shades using the appropriate aspect ratio.

DRAPERIES ASHRAE (2005) categorizes draperies based on appearance, yarn colour and weave according to the chart shown in Figure 7.12. Classes may also be approximated by eye. All of the existing models assume that draperies are perfect diffusers, with constant properties independent of the incidence angle. The only detailed study on the solar optical properties of draperies was completed recently by Kotey et al (2007). The drapery was modelled as a series of uniformly arranged rectangular pleats with optical properties that are dependent upon geometry, the optical properties of the fabric and the profile angle of incident radiation. The fabric was assumed to transmit and diffusely reflect any incident beam radiation. A radiosity-based method was then applied to compute the transmittance and reflectance for different folding ratios and profile angles.

SUMMARY The experimental techniques discussed in this chapter, although giving an estimation of transmittance values for venetian blind systems, can be used in simulation only if the blind/glazing system is assumed to be a perfect Lambertian surface. However, this is not usually the case since most venetian blinds have a specular component and the spatial distribution of transmitted daylight cannot be modelled accurately. Numerical methods provide a solution to this problem. In some cases, ray-tracing techniques using RADIANCE (Reinhart and Walkenhorst, 2001) or Monte Carlo-based techniques (Tsangrassoulis et al, 1996; Campbell and Whittle, 1997) were used to predict daylight in rooms equipped with exterior venetian blinds. The following section presents the latest developments and

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 233

A Review of Optical Properties of Shading Devices

233

Source: ASHRAE (2005)

FIGURE 7.12 Categorization of fabrics for draperies/shades

state-of-the-art methods for measuring and computing optical properties of complex fenestration/shading systems.

SOLAR OPTICAL PROPERTIES OF COMPLEX FENESTRATION AND SHADING SYSTEMS: LATEST DEVELOPMENTS AND STATE OF THE ART An accurate evaluation of daylight distribution through complex fenestration systems (such as venetian blinds) requires knowledge of solar optical effective properties as a function of both incident and emerging directions of light. This is achieved using the concept of bidirectional transmission (or reflection) distribution functions (BT(R)DFs). The idea was first proposed by Klems and Warner (1995) and was then developed by Andersen (2002). BTDFs are defined as follows (see Figure 7.12):

BT (R )DF (θ1, φ 1, θ2 ,φ 2) =

L2 (θ1,φ 1,θ 2 ,φ 2 ) L1(θ1,φ 1) i cos(θ 1) i d ω 1

=

L2 (θ 1,φ 1,θ 2 ,φ 2 ) ⎡ cd ⎤ ⎢ ⎥ ⎢ (m 2lx ) ⎥ E1 (θ 1) ⎢⎣ ⎥⎦

[36]

where: ● (θ1,ϕ1) and (θ2,ϕ2 ) are the polar coordinates of the incoming and transmitted light flux,

respectively;

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

234

3:34 PM

Page 234

ATHANASSIOS TZEMPELIKOS

● L1(θ1, ϕ1) and L2(θ1, ϕ1, θ2, ϕ2) are the luminances of an element of incoming and

transmitted light flux; ● dω1 is the solid angle subtended by incoming light flux; and ● E1(θ1) is the illuminance on the sample plane due to the incident light flux.

BT(R)DFs are measured with bidirectional gonio-photometers. Most of these techniques are based on a scanning process, which is usually very time-consuming. However, there are other methods that are based on a video approach, detecting the light through digital video capture after being collected on a device-specific projection surface. As explained by Andersen and Scartezzini (2005) and Andersen and de Boer (2006): . . . light emerging from the sample is reflected by a diffusing triangular panel towards a charge-coupled device (CCD) camera, used as a multiple-points luminance meter. The diffusing coating of the screen is necessary both because the camera must be able to capture light reflected by any area of the screen independently of the location of this area and to avoid any correlation between the camera’s position and the

Source: Andersen and Scartezzini (2005) FIGURE 7.13 Bidirectional reflection (a) and transmission (b) distribution functions

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 235

A Review of Optical Properties of Shading Devices

235

(b)

(a)

Source: Andersen and Scartezzini (2005) FIGURE 7.14 Detection of the light transmitted through a sample: (a) specular component against diffuse transmission; (b) light transmission and detection with the digital imaging-based photogoniometer

measurements. After six 60° rotations of the screen-camera system, the emerging light distribution is fully determined. For reflection measurements, some additional constraints appear due to the conflict of incident and emerging light flux. The incoming beam needs to penetrate the measurement space and reach exactly the sample surface, therefore requiring a special opening through the measurement space envelope and the removal of screen covers when the latter is obstructive. Andersen et al (2003b) also presented the first extensive comparison of BTDF data with ray-tracing simulation. Particular applications for venetian blinds can be found in Andersen et al (2003a, 2003b, 2005). The inclusion of the specular component in the assessment of BTDF (Andersen and Scartezzini, 2005) set the framework for future work on computing solar optical properties of venetian blind products. Using measured or simulated BTDFs, researchers can model the optical properties of venetian blinds and then use one of the available methods for processing transmitted flux with known directional properties. Reinhart and Andersen (2006) used gonio-photometer and integrated sphere measured data and then used RADIANCE with the Perez et al (1993) sky model and a daylight coefficients approach (Tregenza, 1983) in order to model the performance of a translucent panel. De Boer (2006) presented a method for including BTDF raw data sets into an appropriate format for daylighting simulation. The new version of EnergyPlus incorporates BTDFs in its daylighting simulation engine, according to Crawley et al (2002).

CONCLUSION An overview of past and current approaches for computing the optical properties of shading devices was presented in this chapter; emphasis was placed on venetian blinds, roller shades/screens and draperies, the most commonly used systems. Although the international standard ISO 15099 (2003) describes a general procedure for calculating

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

236

3:34 PM

Page 236

ATHANASSIOS TZEMPELIKOS

these properties, there is still no existing standard for experimental measurement procedures. Several analytical and numerical methods have been developed in order to model the optical behaviour of fenestration systems and shading attachments. In most of the studies, objectives are to calculate the effective solar optical properties: direct-to-direct, direct-to-diffuse and diffuse-to-diffuse transmittance and reflectance. For direct-to-direct properties, the quantities are computed based on the shading geometry (e.g. blind tilt angle, openness ratio and screen construction) and the solar geometry. For direct-to-diffuse properties, empirical calculations are often used for estimation of the specular component, based on material properties and geometry. Otherwise, the shading surface is assumed to be a perfect diffuser and radiosity-based methods are utilized. Finally, for diffuse-to-diffuse properties, a radiosity analysis is performed, based on the system geometry and incoming radiation from the sky and the ground. The main assumptions in most of the models are that: ● The blinds are considered perfect diffusers. ● Proper inter-reflections are ignored. ● Slat curvature/thickness/edge effects are not considered.

Some of the analytical methods – which can be used directly in building simulation software – overcome the above limitations by modifying transmittance modes. However, the models assume that some of the basic properties are already known (from measurements). Moreover, the direction of reflected/transmitted light is usually unknown and therefore the models are not accurate enough for detailed daylighting simulation. Venetian blinds are probably the most complex of all shading systems; therefore the majority of research has focused on this type of shading. Roller shades are modelled as perfect diffusers that are independent of shade transmittance, and their properties are considered constant on both shade surfaces. An advanced model for simulating optical properties of screens was only found in EnergyPlus (2007). More research is needed in this direction since, depending upon shade construction and material properties, the specular component could have an effect and, as a result, the direct-to-diffuse transmittance may not be modelled accurately. Ray-tracing techniques provide another viable solution and include the specular component. Recently, development of bidirectional transmission (or reflection) distribution functions (BT(R)DFs) permitted a more accurate evaluation of radiation transmission/reflection through complex fenestration/shading systems. This method set the framework for modelling the solar optical properties of any fenestration/shading system. Transmitted/reflected components are computed as a function of both incident and emerging directions of radiation. These functions are measured with bidirectional gonio-photometers that scan a hemisphere around the measured sample and can include the specular components. Comparison of BTDF measured data with ray-tracing simulation has shown good agreement. The bidirectional transmittance distribution functions (measured or simulated) can be used to model the optical properties of complex shading systems; researchers may then use one of the available methods for processing

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 237

A Review of Optical Properties of Shading Devices

237

transmitted flux with known directional properties (i.e. ray-tracing) in order to evaluate detailed illuminance distributions inside buildings.

AUTHOR CONTACT DETAILS Athanassios Tzempelikos: Solar Buildings Research Network, Department of Building, Civil and Environmental Engineering, Concordia University, 1455 de Maisonneuve West, EV. 6-139, Montreal, Quebec, Canada H3G 1M8; www.solarbuildings.ca; tel 514-848-2424/7080; [email protected]

ACKNOWLEDGEMENTS The author would like to thank the National Research Council of Canada for financial support, and especially Aziz Laouadi and Christoph Reinhart from the Institute of Research in Construction in Ottawa for providing resources and useful technical comments.

REFERENCES Aleo, F., S. Sciuto and R. Viadana (1994) ‘Solar transmission measurements in outdoor conditions of non-homogeneous shading devices’, in Proceedings of the European Conference on Energy Performance and Indoor Climate in Buildings, Lyon, France, pp1074–1079 Andersen, M. (2002) ‘Light distribution through advanced fenestration systems’, Building Research and Information, vol 30, no 4, pp264–281 Andersen, M. and J. de Boer (2006) ‘Goniophotometry and assessment of bidirectional photometric properties of complex fenestration systems’, Energy and Buildings, vol 38, pp836–848 Andersen, M. and J. L. Scartezzini (2005) ‘Inclusion of the specular component in the assessment of bi-directional distribution functions based on digital imaging’, Solar Energy, vol 79, no 2, pp159–167 Andersen, M., M. Rubin and J.-L. Scartezzini (2003a) ‘Comparison between ray-tracing simulation and bi-directional transmission measurements on prismatic glazing’, Solar Energy, vol 74, no 3, pp157–173 Andersen, M., J. L. Scartezzini, M. D. Rubin and R. C. Powles (2003b) ‘Bi-directional light transmission properties assessment for venetian blinds: Computer simulations compared to photogoniometer measurements’, in Proceedings of ISES Solar World Congress, Gotenborg, Sweden, pp187–198 Andersen, M., M. D. Rubin, R. C. Powles and J. L. Scartezzini (2005) ‘Bi-directional transmission properties of venetian blinds: Experimental assessment compared to ray-tracing calculations’, Solar Energy, vol 78, no 2, pp187–198 ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) (2005) ASHRAE Handbook: Fundamentals, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA, Chapter 30 Athienitis, A. K. and A. Tzempelikos (2002) ‘A methodology for simulation of daylight room illuminance distribution and light dimming for a room with a controlled shading device’, Solar Energy, vol 72, no 4, pp271–281 Beck, A., W. Körner, O. Gross and J. Fricke (1999) ‘Making better use of natural light with a light-redirecting double-glazing system’, Solar Energy, vol 66, no 3, pp215–227 Breitenbach, J., S. Lart, I. Langle and J. L. J. Rosenfeld (2001) ‘Optical and thermal performance of glazing with integral venetian blinds’, Energy and Buildings, vol 33, pp433–442 Campbell, N. S. and J. K. Whittle (1997) ‘Analyzing radiation transport through complex fenestration systems’, in Proceedings of 5th IBPSA Conference, Prague, Czech Republic, pp173–180 Chantrasrisalai, C. and D. E. Fisher (2004) ‘Comparative analysis of one-dimensional slat-type blind models’, in Proceedings of SimBuild 2004, Boulder, CO, pp1–10

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

238

3:34 PM

Page 238

ATHANASSIOS TZEMPELIKOS

Clarke, J. A., M. Janak and P. Ruyssevelt (1998) ‘Assessing the overall performance of advanced glazing systems’, Solar Energy, vol 63, no 4, pp231–241 Crawley, D. B., L. K. Lawrie, C. O. Pedersen, R. K. Strand, F. C. Winkelmann, W. F. Buhl, M. J. Witte, R. J. Henninger, D. E. Fisher and D. Shirey (2002) ‘EnergyPlus: New, capable and linked’, in Proceedings of 2nd eSim Conference, Montreal, Quebec, Canada, pp244–251 De Boer, J. (2006) ‘Modelling indoor illumination by complex fenestration systems based on bidirectional photometric data’, Energy and Buildings, vol 38, pp849–868 Dubois, M.-C. (2003) ‘Shading devices and daylight quality: An evaluation based on simple performance indicators’, Lighting Research and Technology, vol 35, no 1, pp61–76 EnergyPlus (2007) Engineering Reference: The Reference to EnergyPlus Calculations, US Department of Energy, Washington, DC Hellstrom, B., H. Kvist, H. Hakansson and H. Bullow-Hube (2007) ‘Description of Parasol 3.0 and comparison with measurements’, Energy and Buildings, vol 39, pp279–283 Hollands, K. G. T., J. L. Wright and C. G. Granqvist (2001) Solar Energy: The State of the Art, ISES Position Papers, James & James, London, Chapter 2 ISO 15099 (2003) Thermal Performance of Windows, Doors and Shading Devices: Detailed Calculations, 1st edition, International Organization for Standardization, Geneva ISO 9050 (2003) Glass in Building: Determination of Light Transmittance, Solar Direct Transmittance, Total Solar Energy Transmittance, Ultraviolet Transmittance and Related Glazing Factors, International Organization for Standardization, Geneva Karlsson, J. and A. Roos (2000) ‘Modeling the angular behavior of the total solar energy transmittance of windows’, Solar Energy, vol 69, no 4, pp321–329 Kischkoweit-Lopin, M. (2002) ‘An overview of daylighting systems’, Solar Energy, vol 73, no 2, pp77–82 Klems, J. H. and J. L. Warner (1995) ‘Measurement of bi-directional optical properties of complex shading devices’, ASHRAE Transactions, vol 101, no 1, pp791–801 Kotey, N. A. and J. L. Wright (2006) ‘Simplified solar optical calculations for windows with venetian blinds’, in Proceedings of the 31st Conference of the Solar Energy Society of Canada Inc (SESCI) and 1st Solar Buildings Conference (SBRN), Montreal, Quebec, Canada, 20–24 August 2006 Kotey, N. A., J. L. Wright and M. R. Collins (2007) ‘A simplified method for calculating the effective solar optical properties of a drapery’, in Proceedings of the 32nd Conference of the Solar Energy Society of Canada Inc (SESCI) and 2nd Buildings Conference (SBRN), Calgary, Alberta, Canada, 10–14 June 2007 Kuhn, E. (2006) ‘Solar control: A general evaluation method for façades with venetian blinds or other solar control systems’, Energy and Buildings, vol 38, pp648–660 Laar, M. and F. W. Grimme (2002) ‘German developments in daylight guidance systems: An overview’, Building Research and Information, vol 30, no 4, pp282–301 Lee, E. S. and S. E. Selkowitz (1995) ‘The design and evaluation of integrated envelope and lighting control strategies for commercial buildings’, ASHRAE Transactions, vol 101, no 1, pp326–342 Lee, E. S., D. L. Di Bartolomeo and S. E. Selkowitz (1998) ‘Thermal and daylighting performance of an automated venetian blind and lighting system in a full-scale private office’, Energy and Buildings, vol 29, pp47–63 Lorenz, W. (1998) ‘Design guidelines for a glazing with a seasonally dependent solar transmittance’, Solar Energy, vol 63, no 2, pp79–96 Lorenz, W. (2001) ‘A glazing unit for solar control, daylighting and energy conservation’, Solar Energy, vol 70, no 2, pp109–130 Molina, L. J., R. I. Maestre and E. Lindawer (2000) ‘A model for predicting the angular dependence of the optical properties of complex windows including shading devices’, in Proceedings of PLEA, Cambridge, UK, pp765–770

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 239

A Review of Optical Properties of Shading Devices

239

Papamichael, K. and L. Beltrán (1993) ‘Simulating the daylight performance of fenestration systems and spaces of arbitrary complexity: The IDC method’, in Proceedings of 3rd IBPSA Conference, Adelaide, Australia, pp509–515 Parmelee, G. V. and W. W. Aubele (1952) ‘The shading of sunlit glass’, ASHRAE Transactions, vol 58, pp377–398 Perez, R., R. Seals and J. Michalsky (1993) ‘All-weather model for sky luminance distribution – preliminary configuration and validation’, Solar Energy, vol 50, no 3, pp235–245 Pfrommer, P., K. J. Lomas and C. Kupke (1996) ‘Solar radiation transport through slat-type blinds: A new model and its application for thermal simulation of buildings’, Solar Energy, vol 57, no 2, pp77–91 Reilly, M. S., F. C. Winkelmann, D. K. Arasteh and W. L. Caroll (1995) ‘Modeling windows in DOE–2.1E’, Energy and Buildings, vol 22, pp59–66 Reinhart, C. F. and M. Andersen (2006) ‘Development of a radiance model for a translucent panel’, Energy and Buildings, vol 38, pp890–904 Reinhart, C. F. and O. Walkenhorst (2001) ‘Validation of dynamic: Radiance-based simulations for a test office with external blinds’, Energy and Buildings, vol 33, pp683–697 Rosenfeld, J. L. J., W. J. Platzer, H. Van Dijk and A. Maccari (2000) ‘Modeling the optical and thermal properties of complex glazing: Overview of recent developments’, Solar Energy, vol 69, no 6, Supplement, pp1–13 Rubin, M., K. Von Rottkay and R. Powles (1998) ‘Window optics’, Solar Energy, vol 62, no 3, pp149–161 Siegel, R. and J. Howell (1972) Thermal Radiation Heat Transfer, McGraw-Hill Inc, Columbus, OH Sullivan, R., L. O. Beltrán, E. S. Lee, M. D. Rubin and S. E. Selkowitz (1998) ‘Energy and daylight performance of angular selective glazings’, in Proceedings of Thermal Performance of Exterior Envelopes VII, Clear Water, FL, pp319–328 Tregenza, P. R. (1983) ‘The Monte Carlo method in lighting calculations’, Lighting Research and Technology, vol 15, no 4, pp163–170 TRNSYS 15 (2003) Reference Manual, TRANSSOLAR/TESS Tsangrassoulis, A., M. Santamouris and D. Assimakopoulos (1996) ‘Theoretical and experimental analysis of daylight performance for various shading systems’, Energy and Buildings, vol 24, pp223–230 Tzempelikos, A. (2005) Literature Review on Convection Heat Transfer Correlations in Skylight Cavities, and on Thermal and Optical Prediction Models of Window Shadings, Final report submitted to IRC–NRCC, May 2005 Tzempelikos, A. and A. K. Athienitis (2003) ‘Modeling and evaluation of a window with integrated motorized venetian blinds’, in Proceedings of 3rd ISES World Congress, Gotenburg, Sweden Tzempelikos, A. and A. K. Athienitis (2007) ‘The impact of shading design and control on building cooling and lighting demand’, Solar Energy, vol 81, no 3, pp369–382 Van Dijk, D. and H. Oversloot (2003) ‘WIS, the European tool to calculate thermal and solar properties of windows and window components’, Proceedings of the 8th IBPSA Conference, Eindhoven, The Netherlands, pp259–266 WINDOW 5/THERM (2003) NFRC Simulation Manual, Lawrence Berkeley National Laboratory, CA Wright, J. L. and N. A. Kotey (2006) ‘Solar absorption by each element in a glazing/shading layer array’, ASHRAE Transactions, vol 112, no 2, pp3–12 Yahoda, D. S. and J. L. Wright (2004) ‘Methods for calculating the effective long-wave radiative properties of a venetian blind layer’, ASHRAE Transactions, vol 110, no 1, pp463–474 Zmeureanu, R. (1998) Defining the Methodology for the Next Generation HOT2000 Simulator, Internal report submitted to CANMET

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 211–239

ABER_CH_07

6/12/08

3:34 PM

Page 240

ABER_Index

6/12/08

1:52 PM

Page 241

Index ABL see atmospheric boundary layer absorption contaminant interactions 59, 62–3, 66, 81, 88 shading devices 215–16, 221 solar cooling 128–31, 138 see also sorption acceptance angles 36, 45 access, solar 99 active tracer gas methods 198–9 ADAM see advanced automotive manikin adaptive principle 165–6 adsorption contaminant interactions 59–60, 62–3, 66, 79, 88 solar cooling 128–9, 131, 138 see also sorption adsorption isotherm models 66 advanced automotive manikin (ADAM) 186–7 Advanced Window Information System (WIS) 226–7 air contaminant modeling 63–4 mean age 176, 178, 200–1 temperature 102, 196–8 urban design 102, 105–6 velocity 191–3 see also ventilation air conditioning 26, 123–57 air phase diffusion 68–73, 75 albedo 105, 112–13 alternative cooling technologies 126 ALTSET project 228 American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) 6–14, 29, 217–18, 232 anemometers 191–4, 201 anidolic systems 50–1 architectural design 95–121, 125 Arrhenius-type law for temperature dependency 80–1 artificial sky 52 ASHRAE see American Society of Heating,

Refrigerating and Air-Conditioning Engineers aspect ratios 97, 104–6, 114, 231 atmospheric boundary layer (ABL) 183 attics 13–15 axisymmetric jets 172 bag method 201 bead thermistors 189, 192 behaviour 162, 166 benchmark comfort zone tests 203–4 bidirectional flow 179 bidirectional transmission (or reflection) distribution functions (BT(R)DFs) 233–4, 236–7 bidirectional transmittance distribution function (BTDF) 52–3, 233, 235–6 bimolecular reactions 87–8 blind systems 35–9, 213, 216–29, 232, 236 boundary layers 59, 62–5, 179–80 BTDF see bidirectional transmittance distribution function BT(R)DFs see bidirectional transmission (or reflection) distribution functions building materials sorption and diffusion 69–73 urban design 112–13, 115 see also non-porous materials; porous materials bulk density flow 178–9 buoyancy 169–70, 172–5 California 3–4, 6, 14–21, 28–9 canopy systems 49 carbon/methanol adsorption heat pumps 131 Carnot diagram 134, 137 CCDs see charge-coupled devices ceilings 14–15, 50, 197 CFD see computational fluid dynamics models charge-coupled devices (CCDs) 234 chemical reaction heat pumps 131–2, 138

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248

ABER_Index

6/12/08

242

1:52 PM

Page 242

ADVANCES IN BUILDING ENERGY RESEARCH

chemical transformations 86–8 chemisorption cooling technologies 128–9 Chicago, Illinois 22–3 city design 95–121 see also urban areas climate urban 95–121, 125 zones 11, 17 climatic parameter measurement 188 closed systems 129 clothing 202–3 coefficient of performance (COP) 130–1, 135, 138, 144 comfort 125, 164–6, 202–4 Comfortina 186, 203 commercial level cool roofs 22–3 SAC 140–7 competitive aspects contaminant interactions 82–5 SAC systems 123–4, 146–7, 154 compound parabolic concentrators (CPCs) 37–9, 51 computational fluid dynamics (CFD) 63–4, 203–6 computer-simulated persons (CSPs) 204 constant concentration method 181, 200 see also tracer gas technique constant injection method 200 contact temperature measurements 188–9 contaminants interactions 57–3 tracer gas methods 198–9, 206–7 ventilation 160, 176–8, 206–7 continuous sorption processes 136 control systems 35, 43, 145 convective loss measurements 187–8 convective mass-transfer coefficients 63–5 cool coatings 113 cooling solar-assisted 123–57 vegetation 107–10 cool-roof standards 1–32 COP see coefficient of performance core systems 34, 41–51 costs cool roofs 16, 18–19, 21 daylighting systems 54 SAC systems 141–2, 146

CPCs see compound parabolic concentrators credits, cool roof 6, 8–9, 12–15, 23, 25 CSPs see computer-simulated persons cup method 75–6 cut-off angles 226 data loggers 201–2 daylighting systems 33–56, 217 DAYSIM 53 decay rate method 200–1 DEM see digital elevation model densimetric Archimedes number 172 density building ventilation 178–9 urban 97–8, 101–2, 105 dependency, temperature 80–1 deposition velocity approach 86–7 desiccant cooling systems 129, 137–8, 143 design, city 95–121 desorption 59, 63, 88, 128 see also absorption; adsorption; sorption developed countries 140 diffuse reflectance 228 diffuse-to-diffuse transmittance and reflectance 219, 222–6, 236 diffuse transmittance 235 diffusion 60–2, 87–8, 221, 230, 236 diffusion-cell methods 75 diffusion coefficients 68–77, 80–1, 84 diffusivity 68–75, 78, 80–5 digital cameras 195 digital elevation model (DEM) 106–7 digital image-based photogoniometers 235 dilute liquid solutions 63 dilution capacity 176, 180 dimensions of rooms 171–4 direct beam transmittance 231 directional illuminance coefficients 229 directional transmission 230 direct-reflected transmittance 225 direct-to-diffuse transmission and reflectance 219–22, 236 direct-to-direct transmittance and reflectance 219–22, 225, 236 direct-to-hemispherical transmittance 224 discontinuous sorption processes 135 displacement height 182 displacement ventilation 174–5, 177–8 DOE-2 software 215 doorways 179

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248

ABER_Index

6/12/08

1:52 PM

Page 243

INDEX

draperies 232–3 ducts 13, 181, 201 economic level SAC systems 141–2 urban planning 100 effects, number of 130–1 efficiency see energy efficiency electrical thermometers 188 electricity cool roofs 2–5 SAC systems 140, 146 elemental models 58–61 ellipsoid sensors 187 emission profiles 77 emissivity 190 energy conservation 22–3, 98 consumption 110–1 efficiency 1–32, 140–1 energy performance factor (EPF) 26 EnergyPlus 221–4, 230–1 energy savings cool roofs 2–5, 16, 18–21 daylighting systems 33 SAC systems 144–5 vegetation 110 Energy Star Label, EPA 26–7 entropy–temperature (S–T) diagrams 135–6 envelopes 25, 183–4 environmental level parameter modeling 79–81 SAC systems 124, 143–4, 146, 148–50 urban climates 96 Environmental Protection Agency (EPA), US 26–7 EPA see Environmental Protection Agency EPF see energy performance factor equivalent temperature (ET) 166–7 ESP-r 53, 215 ET see equivalent temperature European glazing systems performance 215 exhaust hoods 177 experienced temperature 167 exposure, solar 99, 101–4 exterior fixed window constructions 213, 217, 224 extract ventilation 162

243

fan-powered ventilation see mechanical ventilation FBC see Florida Building Code fenestration systems 34–41, 233–5 see also windows fibre optics 43–7 Fick’s law of diffusion 60–1, 74 field comfort studies 166 flat surface sensors 187–8 Florida Building Code (FBC) 23–5 flow markers 195 flow rate measurement 199–200 flows between rooms 178–80 forced convection 65 Fresnel lenses 45–6 full-scale replicas 204–5, 207 gaseous contamination interactions 57–93 glass fibres 44 glazing systems 113, 214 see also fenestration systems; windows global urban cooling demand 126 global warming 125 gonio-photometers 52, 234–6 gonio-spectrometers 228 Green Building Rating System, LEED 27–8 ground-diffuse reflectance 223–4 ground surface temperature 111 Guam 25–6 Hawaii 25 heat balance, human 164 heat flux 109, 184–7 heating, ventilation and air conditioning (HVAC) systems 125, 140, 160 heat island effect 97, 107, 112, 125 heat pumps 127–38 Heliobus system 48 heliostats 41–3, 45–6 Henry’s law for dilute liquid solutions 63 Henry’s law for ideal absorption 66, 68, 82, 85, 88 heterogeneous processes see chemical transformations Himawari system 45 HOEs see holographic optical elements holographic optical elements (HOEs) 41 homogeneous emission method 200 homogeneous reactions 87 horizontal light pipes 48–9 horizontal venetian blinds 223

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248

ABER_Index

6/12/08

244

1:52 PM

Page 244

ADVANCES IN BUILDING ENERGY RESEARCH

hot-sphere anemometers 192 hot-wire anemometer 191–2, 201 humans behaviour 162, 166 microclimates 98 temperature regulation 163–5 humidity 68, 84–6 HVAC see heating, ventilation and air conditioning systems hybrid solar lighting systems 45–6 hygienic ventilation flow rates 160 IAQ see indoor air quality IAS see ideal adsorption theory ideal adsorption theory (IAS) 83–4 IECC see International Energy Conservation Code illumination systems 195 IMAGE project 215 indoor air quality (IAQ) 57, 64, 75, 160 infiltration 111, 180–1, 183, 198 inflatable heliostats 43 infrared thermography 196–8 infrared thermometers 190 initial concentrations 77–8 injection strategies 198–200 inlet velocity 171–2 insulation 7–8 integrated directional coefficients method 229 internal environment 149–50 International Energy Conservation Code (IECC) 22 ISO 14505 203 ISO 15099 218–21, 224, 235–6 isothermal supply 169, 172, 174 Jeffersonian grids 103–4 jet length 171 Knudsen diffusion 68, 73–4, 78 Kuhn’s model 227–8 LabView 202 Lambertian surfaces 227, 232 landscaping 107–8 Langmuir equation 82–4 Langmuir-Hinshelwood model 88 Langmuir’s theory of adsorption 59, 61–2 large eddy simulation (LES) 106 laser-cut light-deflecting panels (LCPs) 40–1 laser Doppler anemometers (LDAs) 191, 193

LCPs see laser-cut light-deflecting panels LDAs see laser Doppler anemometers Leadership in Energy and Environmental Design (LEED) 27–8 LEED see Leadership in Energy and Environmental Design LES see large eddy simulation light pipes 47–9 linear adsorption isotherm models 60, 68, 74, 77–8, 81, 83, 88 liquid fibres 44 lithium bromide/water absorption heat pumps 130, 138 living standards 125 local discomfort meters 188 local equilibrium sorption models 88 local urban cooling demand 126 low-sloped roofs 15–16, 18–19, 22, 28–9 luminaires 46–7 macroscopic models 59 maintenance costs 142 mandatory energy efficiency standards 5–6 manikins 184–7, 203–4, 206 market level daylighting systems 33–4 SAC systems 137–47, 155 mass roof assemblies 26 material contamination interactions 57–93 mean age of air 176, 178, 200–1 mechanical ventilation 160, 162, 169–70 medium pairs 139 mercury thermometers 188 metallic sensors 188–9 meteorological models 97 microclimates human temperature regulation 163–5 urban 95–121, 125 Microlouvers 39 mixed convection 65 mixing ventilation 177 mixtures of contaminants 82–4 modelling material and gaseous contamination 57–93 shading devices 229, 236 thermal simulation 53 see also simulation molecular diffusion 68, 73–4, 78, 80 momentum flux 169–70 Monte Carlo-based techniques 232 multi-component sorption 82–3

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248

ABER_Index

6/12/08

1:52 PM

Page 245

INDEX

natural convection 65 natural cooling 125 natural ventilation 160, 162–3, 182 neighbourhood typology 106–7, 114 nodal models 63–4 noise 143–4 non-contact temperature measurements 190 non-isothermal supply 169 non-porous materials 68, 88 non-residential buildings 2–4, 8–9, 15–19 non-white roofing materials 5 occupancy trends 125 Okasolar-W system 38 one-phase models 60, 62–3, 69–71 one-sink model 59 open systems 129 operation costs 142 optical properties of shading devices 211–39 optimization 98–9 orientation of streets 103–4, 114 orifice flow meters 201 outdoor environment 2, 180–4 parabolas 37–9 Parasol version 3 227 parks, urban 107–11 particle image velocimetry (PIV) 194–6, 206 particle streak velocimetry (PSV) 194–6, 206 partition coefficients 67–8, 75, 81 partition diffusion 79 part-load characteristics 140–1 passive tracer gas methods 199 paving materials 112–13, 115 payback period 143 peak-power demand 2, 126 perfect diffusers 221, 230, 236 performance buildings and microclimates 98 cool roofs 6, 8, 12, 26 COP 130–1, 135, 138, 144 daylighting systems 51–3 light pipes 47 SAC systems 144 sorption heat pumps 135–6 photogoniometers 52, 234–6 photometric measurement 52 physically based modeling 57–93 Pitot tube 201

245

PIV see particle image velocimetry Planck law 190 planning, urban 96–115 plastic fibres 43–4 platinum resistive sensors 189 PMV see predicted mean vote point measurement techniques 188 policy level 146 pollution 106, 182 polymeric materials 78, 80–2, 85–6 porous materials contaminant interactions 62–3, 78, 80–2, 84, 86–9 diffusion 60–1, 68–77, 80 Porter’s Five Forces model 147 predicted mean vote (PMV) 165 predictive approaches 114 prescriptive aspects 5–6, 8–9, 12–18 prices air conditioners 126 SAC systems 146 primary airflow 168, 171 primary energy 126–7, 145 prismatic systems 39–40 Prismex 46–7 prism light 47, 53 profile angles 221, 229 prototype daylighting systems 50 prototype inflatable heliostats 43 pseudo absorption constant 62, 66 PSV see particle streak velocimetry pulse method 193–4 purging flow rate 181 Ra see Rayleigh number RADIANCE 232, 235 radiant barriers 18–21 radiant exchange 104–5 radiation 190 see also solar radiation radiative loss measurement 187–8 radiative properties 7–9 radiosity theory 227 Rayleigh number (Ra) 64 ray-tracing techniques 53, 232, 236 Re see Reynolds number rebate programmes 28 recursive equations 214, 225 reflectance bidirectional 234 diffuse 228 direct-to-diffuse 219–22, 236

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248

ABER_Index

6/12/08

246

1:52 PM

Page 246

ADVANCES IN BUILDING ENERGY RESEARCH

direct-to-direct 219–22, 225, 236 ground-diffuse 223–4 roller shades 230 screens 231 shading devices 215–16 solar 1–2, 7–9, 12, 15–16, 24–9 venetian blinds 219–20 reflecting blinds 35 refraction 39–41 refrigerants 126, 128–31 see also cooling regional urban cooling demand 126 regulation of human temperature 163–5 regulations 125 relative humidity 68 reliability 143 replicas 204–5 residential buildings 5, 8–9, 13–16, 18–22, 24 resistive temperature sensors 188–9 Retrolux blind 37 Reynolds number (Re) 64, 169, 172 roller shades 229–30, 236 roof gardens 111 roofs 1–32, 112–13 room velocity scale 174–5 room ventilation 168–78 roughness height 182 SAC see solar-assisted cooling safety 144 salt bath method 179, 205 Sc see Schmidt number scale models 204–6 Schmidt number (Sc) 64 screens 197–8, 231–2 secondary airflow 168 seeding 193, 195 sensors 43, 187–92 Sh see Sherwood number shades 233 shading 34–9, 99, 110–11, 211–39 Sherwood number (Sh) 64 silica gel/water adsorption heat pumps 131, 138 simulation cool roofs 8, 10–12 daylighting 51–3 manikins 184–7 shading devices 229, 236 see also modelling

sink effect 58, 61–2, 66, 79–80 site-specific weather data 115 size of cooling system 139 size of room 171–4 skylights 41–2 slat properties 218–29 solar-assisted cooling (SAC) 123–57 solar lighting systems 45–6 solar radiation access 99 cool roofs reflectance 1–2, 7–9, 12, 15–16, 24–9 exposure 99, 101–4 shading devices 215–16 transmittance 216 solar reflectance index (SRI) 12, 27–9 Solux System 46 sorbent/refrigerants 130–1, 138–9 sorption chemical transformations 87 coefficients 69–72 cooling technologies 127–37 correlations 79 diffusion coefficients 76 elemental models 59–60 local equilibrium 88 multi-component 82 parameters 66–8 processes 61–2 see also absorption; adsorption source/sink behaviour 58 Spanish grids 103–4 spectral aspects 230–1 specular aspects 225, 227–8, 235 SRI see solar reflectance index S–T see entropy–temperature diagrams stairwells 179–80 standards comfort zones 203 cool-roof 1–32 steep-sloped roofs 16–21, 28–9 stoichiometric coefficient 87 streets 97, 103–6, 114 strengths, weaknesses, opportunities, threats (SWOT) analysis 147–54 structural shading 99 subsidiarity 99 summer point multipliers 23 supply velocity 169–70, 172–4 surface chemistry models 87–8 surface diffusion 73–5, 80

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248

ABER_Index

6/12/08

1:52 PM

Page 247

INDEX

SWOT see strengths, weaknesses, opportunities, threats analysis system identification methods 200 system view 161–3 TALISYS 53 technology alternative cooling 126 environmental effects 148–50 temperate climates 102 temperature anemometers 191–2 building and paving materials 112–13 contaminant modeling 79–81 diffusivities 81–4 ET 166–7 heat pumps 130–6 measurement 188–90, 194–9 microclimates and human regulation 163–5 street aspect ratio 104–5 thermal manikins 185 vegetation 107–11 test people 205 test rooms 204–6 texture analysis 107 thermal anemometers 191–3 thermal aspects building and paving materials 112–13 comfort 125, 164–6 cool roofs 1–2, 7–9, 15–16, 24–9 thermal-driven active solar cooling technologies 127–37 thermal manikins 184–7, 206 thermal transmittance 9–10, 14 see also transmittance; U factor thermistors 188–9, 192 thermochemical heat pumps 131–2 thermocouples 189–90, 202 thermometers 188 threats, opportunities, weaknesses, strengths (TOWS) matrix 147, 151–4 time constants 175 time varying flow rate 200 Title 24 standards, California 6, 14–21 TOWS see threats, opportunities, weaknesses, strengths matrix tracer gas methods 181, 198–9, 206–7 see also constant concentration method tracer seeding system 195 tracking mechanisms 43–4, 47

247

transmittance bidirectional 234 diffuse 235 direct-to-diffuse 219–22, 236 direct-to-direct 219–22, 225, 236 direct-to-hemispherical 224 roller shades 230 screens 231 shading devices 215–16 venetian blinds 219–29 see also thermal transmittance trees 109–10 TRNSYS software 215 tropical climates 102 turnkey costs 141–2 twin chamber methods 75 two-phase models 60–3, 72 typology of buildings 106–7, 114 U-factor 8–10, 14 see also thermal transmittance ultrasonic anemometers 193–4 United States of America (US) 1–32 urban areas climate research 95–121 cooling demand 124–6 density 97–8, 101–2, 105 SAC systems 155 US see United States of America Van’t Hoff equation 81 vapour compression cooling (VCC) 124, 126, 141–5 VCC see vapour compression cooling vegetation 107–11, 114–15 velocity deposition 86–7 measuring 188, 191–9, 206 ventilation 171–5 venetian blinds 35–6, 213, 216–29, 232, 236 ventilation 103, 159–210 video approaches 234–5 virtual manikins 204, 206 visible transmittance 216 visualization 184, 192 VOCs see volatile organic compounds volatile organic compounds (VOCs) 61, 66, 77, 79–80 volume flow 201 volumetric expansion coefficient 65 voluntary approaches 26–8

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248

ABER_Index

6/12/08

248

1:52 PM

Page 248

ADVANCES IN BUILDING ENERGY RESEARCH

water/ammonia absorption heat pumps 130, 138 weather data 97, 115 whitewashing 112 whole-field measuring techniques 194–9 wind 111, 182–3 windows 214–15 see also fenestration systems

WINDOW software 215 wire sensors 189, 191–2 WIS see Advanced Window Information System zeolite/water adsorption heat pumps 131 zonal models 63–4

ADVANCES IN BUILDING ENERGY RESEARCH ■ 2008 ■ VOLUME 2 ■ PAGES 241–248