DECISION SUPPORT FOR URBAN WIND ENERGY EXTRACTION

DECISION SUPPORT FOR URBAN WIND ENERGY EXTRACTION

DECISION SUPPORT FOR

URBAN WIND ENERGY EXTRACTION

By

RUTH C. COOPER, B.ENG

A Thesis

Submitted to the School of Graduate Studies

in Partial Fulfilment of the Requirements

for the Degree

Master of Applied Science (M.A.Sc.)

McMaster University

Department of Civil Engineering

© Copyright by Ruth C. Cooper, July 2007

ii

MASTER OF APPLIED SCIENCE (2007) McMaster

University

(Civil Engineering) Hamilton, Ontario

TITLE: Decision Support for Urban Wind Energy Extraction

AUTHOR: Ruth C. Cooper, B.Eng. (Carleton University)

SUPERVISOR: Dr. Brian W. Baetz

NUMBER OF PAGES: xxii, 247

ii

ABSTRACT

In the wake of the environmental movement theburgeoning wind energy industry, equipped with arudimentary understanding of the wind resource andlimited socio-economic policies, created today's windenergy paradigm characterised by farms of tower-mounted, 3-bladed, horizontal-axis wind turbines(HAWTs). This research undertaking proposes that aparadigm shift is in order and that commissioning ofdecentralised, in-situ, urban-scale wind energyconversion devices could assist the industry.

Project WEB, which was commissioned by the EuropeanUnion in the framework of Joule III (2000) to explorethe aerodynamic properties of built structures andsuitable wind energy conversion system configurations,coined the terms Urban Wind Energy Conversion System(UWECS) and Building Augmented Wind Turbine (BAWT).The theory behind BAWTs is primarily founded on largeskyscrapers and building-integrated HAWTs. Thisresearch undertaking focused on alternativelyconfigured UWECS and the potential of smaller buildings(e.g., residential and commercial buildings) to producesufficient building aerodynamics-induced windamplification to make urban wind energy generation aviable option.

The assessment of urban morphology-induced windamplification, specifically in support of sitingbuilding-integrated and/or mounted UWECSs, is a veryrecent undertaking. Only a small number of windturbine manufacturers are even exploring thedevelopment of suitably-scaled devices whoseperformance characteristics are tailored to urban windconditions. As such, this research explored thefeasibility of BAWT-theory in an urban setting through

iii

the development of a prototype Urban Wind EnergyPlanning (UWEP) Decision Support System (DSS).

The prototype UWEP DSS focuses primarily onbuilding aerodynamics-induced wind amplification,including consideration of peak-wind seasons.Microsoft® Excel was selected as the platform for theUWEP DSS, supporting development of user forms andintegral databases. This tool is intended for a broadrange of users, including the average home owner, UWECSdevelopers, and energy planners. With minimal userinput, the UWEP DSS determines the mean wind speedwithin the amplification zones, the location of theamplification zones, and the energy that couldpotentially be generated by an appropriately-sitedUWECS.

Two case study applications of the UWEP DSS wereconducted to validate the estimations and demonstratethe capabilities of the tool. The University ofToronto Robarts Library application and the GreenVenture EcoHouse application yielded credible mean windspeed and potential wind energy estimates on comparisonto the online wind atlases. The EcoHouse case studyapplication included the selection and siting ofvarious UWECSs. It highlighted the potential of ahypothetical wind energy conversion device being ableto generate almost 40% of the 700 kWh per month,average household energy demand. Conversely, itdemonstrated that the traditional tower-mountedhorizontal axis wind turbines, situated outside of thepotential amplification zones in accordance withcurrent siting guidelines, would only be able togenerate 5% of the demand.

By demonstrating the prototype UWEP DSS through aninstitutional and a residential application case study,it is hoped that the scope and capabilities of, and the

iv

amplified wind energy potential identified by, thistool will foster further research in urban wind energyplanning, building aerodynamics-induced amplificationassessment, and development of new UWECSs. Theprototype UWEP DSS appears to be the first to estimatebuilding aerodynamics-induced amplification from peakcomposite pressure-gust coefficients published inbuilding codes. Further research is recommended togain a better understanding of sustained, as opposed topeak, wind amplification. The modular nature of theUWEP DSS lends itself to the modifications that willundoubtedly be required as further knowledge isdeveloped in this field.

v

ACKNOWLEDGEMENTS

As I reflect on the last twenty months, I can nothelp but to feel a pang of regret to see this stage ofmy life come to an end. I will surely miss the highsand lows that accompany the thrill of discovery and thedespair in failed hypotheses. I would like to takethis opportunity to acknowledge a few key individualswho supported me through times during which thechallenges appeared insurmountable.

First and foremost, I appreciate having beenprovided with this opportunity to expound on thevirtues of Dr. Brian Baetz. I initially contacted Dr.Baetz out of interest in pursuing graduate studieswithin the department of Civil Engineering at McMasterin the spring of 2005. Without his support andassistance throughout the whole application process, Iwould not have been afforded the opportunity to conductresearch as a Masters candidate. Especially during thelast few months, his continued support, encouragement,and ability to sort the wheat from the chaff, so tospeak, have been invaluable. I would also like tothank Drs. Sarah Dickson and Michael Tait of theDepartment of Civil Engineering for their contributionsas thesis examination committee members.

The association with the Barry Rawn's team at theUniversity of Toronto was a fortuitous accident,instigated by Dr. Ursula Franklin. The data providedby this team of urban wind energy enthusiasts are whathelp to ground the results produced by application ofthe UWEP DSS in reality.

The concept of building aerodynamics induced windamplification, which started as a mere notion, wouldnot have been substantiated had it not been for theresearch conducted by Mr. Neil Campbell and associates

vi

under the auspices of Project WEB. Further research onBAWT theory conducted at TU Delft by Dr. Gerard vanBussel and Dr. Sander Mertens yielded numerous journalarticles, the collection of which I refer to as 'thebible'.

Assessment of the mean wind speed below the urbancanopy layer would not have been possible were it notfor the work of Dr. Robert MacDonald of the Universityof Waterloo. Unfortunately, in attempting to contacthim, I discovered that he had passed away in April of2004.

Last, but by no means least, this researchundertaking would not have been possible without thesupport of the staff and faculty of the Department ofCivil Engineering.

vii

TABLE OF CONTENTS

ABSTRACT....................................iii

ACKNOWLEDGEMENTS..............................v

LIST OF FIGURES...............................x

LIST OF TABLES.............................xiii

NOMENCLATURE................................xiv

1 INTRODUCTION...............................11.1 Climate Change and Government / Industry Response..........................................31.2 The Wind Energy Industry......................71.3 The Building Industry........................111.4 Building Aerodynamics........................141.5 Urban Wind Energy Conversion Systems.........161.6 Scope and Objectives of the Research Undertaking......................................191.7 Thesis Structure.............................21

2 LITERATURE REVIEW.........................222.1 The Interdisciplinary Nature of this Research 232.2 Wind Power Meteorology Overview..............252.3 The Urban Boundary Layer.....................27

2.3.1..........................The Wind Climate31

2.3.2...................Frequency Distributions34

2.4 Boundary Layer Modelling.....................362.4.1..........................Mesoscale Models

372.4.2.........................Microscale Models

38

viii

2.4.3........................Model Combinations39

2.5 The Wind Atlas...............................412.5.1............The Canadian Wind Energy Atlas

472.5.2...........The Ontario Wind Resource Atlas

492.6 Wind Atlas Methodology.......................52

2.6.1...............The Mean Wind Speed Profile54

2.7 Wind Resource Assessment.....................562.7.1.......................Commercial Software

592.8 The Built Environment........................62

2.8.1........................Urban Canopy Layer62

2.8.2............Urban Parameterisation Schemes64

2.8.3...................Mean Wind Speed Profile65

2.8.4.....................Building Aerodynamics67

2.9 Decision Support Systems.....................762.10......................Literature Review Summary

81

3 CONCEPTUAL MODEL DEVELOPMENT..............833.1 Urban Renewable Energy Source Assessment.....843.2 Scope of Applicability.......................863.3 Configuration of the Conceptual Model........873.4 Wind Energy Assessment Methodology...........90

3.4.1.............Meteorological Considerations90

3.4.2..............Morphological Considerations92

4 DECISION SUPPORT SYSTEM DEVELOPMENT.......96

ix

4.1 Overview.....................................984.2 User Interface..............................1014.3 External Databases..........................1034.4 Internal Database...........................104

4.4.1............Roughness Classification Table105

4.4.2..............Urban Parameterisation Table108

4.4.3...........Wind Amplification Factor Table113

4.5 UWEP module.................................1184.5.1....................Wind Data Extrapolator

1214.5.2....................Wind Data Interpolator

1314.5.3............................Wind Amplifier

1344.5.4.........................Energy Calculator

1464.6 Summary Report..............................1484.7 Architectural Configurator..................148

5 APPLICATION OF THE DECISION SUPPORT SYSTEM1505.1 Institutional Case Study....................151

5.1.1.............Annual Wind Energy Assessment1525.1.1.1..................Mean Wind Climate

1535.1.1.2..........Amplified Mean Wind Speed

1605.1.1.3........................Wind Energy

1705.1.1.4.....................Summary Report

1725.1.2...........Seasonal Wind Energy Assessment

173

x

5.1.2.1..................Mean Wind Climate174

5.1.2.2..........Amplified Mean Wind Speed179

5.1.2.3........................Wind Energy181

5.1.3........UWEP DSS Credibility Investigation1825.1.3.1..................Mean Wind Climate

1845.1.3.2.................Wind Power Density

1935.1.3.3........................Limitations

1955.2 Residential Case Study......................196

5.2.1..............Mean Annual Energy Potential198

5.2.2.............UWECS Selection and Placement201

5.2.3.............Architectural Reconfiguration209

5.2.4. .Potential Limitations of the Application210

6 SUMMARY, CONCLUSIONS & RECOMMENDATIONS...2136.1 Summary.....................................2136.2 Conclusions.................................2156.3 Recommendations.............................220

7 REFERENCES...............................228

xi

APPENDICES

Appendix A Boundary Layer Meteorology PrimerAppendix B Wind Atlas Comparison TableAppendix C Topographical Acceleration PrimerAppendix D - 1 Development of the Roughness Classification TableAppendix D - 1a Roughness Classification SchemesAppendix D - 2 Development of the Urban Parameterisation TableAppendix D - 3 Development of the Wind Amplification Factor TableAppendix E Background and Assumptions: Wind Data ExtrapolationAppendix F Background and Assumptions: Wind Data InterpolationAppendix G Background and Assumptions: Wind AmplificationAppendix H Background and Assumptions: Energy CalculationAppendix I - 1 Wind Statistics WorksheetAppendix I - 2 Wind Amplification WorksheetAppendix I - 3 Wind Energy WorksheetAppendix I - 4a Equation SummaryAppendix I - 4b Internal Database Development WorkbookAppendix J - 1a Robarts Library: Summary ReportAppendix J - 1b Robarts Library: Wake Streamline PlotsAppendix J - 2a Green Venture EcoHouse: Summary ReportAppendix J - 2b Green Venture EcoHouse: Wake StreamlinePlotsAppendix K UWECS Photo MontageAppendix L Economic Assessment Tools matrixAppendix M Anemometer Heights of Canadian Meteorological StationsAppendix N UWEP DSS Application

xii

LIST OF FIGURES

Figure 1.1. Brush Windmill Cleveland, OH (1888).. . . .1Figure 1.2. Wind Farm in Coachella Valley, California

(US)..............................................2Figure 1.3. G8 Countries CO2 emissions per capita for

1990 and 2002.....................................3Figure 1.4. Canadian stationary source combustion GHG

Emissions (2004)..................................4Figure 1.5. Canadian electricity generation source mix

in GW (2004)......................................6Figure 1.6. Ontario electricity generation source mix

in GW (2007)......................................7Figure 1.7. Installed wind energy capacity (GW) by

country as of 12/2006.............................8Figure 1.8. Technical renewable energy potential in

Canada in GW by source............................8Figure 1.9. Building Augmented Wind Turbines.......10Figure 1.10. An artists rendition of the Burj al-Taqa

(Energy Tower)...................................12Figure 1.11. General WECS Configurations...........16Figure 1.12. Rooftop wind energy conversion devices.

.................................................17Figure 1.13. Modular Rooftop modular wind energy

conversion devices...............................17Figure 1.14. Building augmented wind energy conversion

systems..........................................18Figure 2.1. The interdisciplinary blocks of urban wind

energy research..................................24Figure 2.2. The Osaka City 'klimatope'.............26Figure 2.3. Urban boundary layer structural sketch. 28Figure 2.4. The terrain dependent velocity shear

profiles.........................................29Figure 2.5. Variation of the primary northern

hemisphere wind climates.........................32Figure 2.6. The components of the instantaneous wind

speed............................................33

xiii

Figure 2.7. Northern hemisphere wind climate frequencydistributions....................................34

Figure 2.8. The Toronto Island airport wind directionrose of June 2005................................35

Figure 2.9. Flow chart of the Riso wind power prediction model.................................40

Figure 2.10. Wind Energy Resource Atlas of the United States...........................................42

Figure 2.11. Analyzed Winter Season Average Wind Resource Maps....................................43

Figure 2.12. Wind resource estimates in the Northwest region of the US.................................44

Figure 2.13. The wind resource in Europe at 50 m a.g.l..................................................46

Figure 2.14. Worldwide status of the wind atlas methodology......................................47

Figure 2.15. The 65 tiles or quadrangles of the Canadian Wind Energy Atlas.......................48

Figure 2.16. Canadian Wind Atlas Quadrangle 40 - Mean Wind Speed at 30m................................49

Figure 2.17. Ontario Wind Resource Atlas demonstration....................................51

Figure 2.18. Wind Atlas Methodology flow chart.. . . .52Figure 2.19. Wind atlas methodology removal &

application of local effects.....................53Figure 2.20. PBL Day and night mean wind speed

profiles.........................................54Figure 2.21. Sketch of mean wind speed profile.. . . .55Figure 2.22. General form of packaged wind resource

assessment software..............................58Figure 2.23. Digital still video aerial imagery

extraction techniques............................65Figure 2.24. Flow regimes within the UCL...........69Figure 2.25. Vortex structure categorisation.......71Figure 2.26. Effect of varying building height on

canyon vortex structure..........................72

xiv

Figure 2.27. Effect of roof configuration on canyon vortex structure.................................73

Figure 2.28. Cross-canyon horizontal mean wind speed variation........................................74

Figure 2.29. The flow chart of the DAD-based WiLDE application......................................77

Figure 2.30. Solar energy decision support system.. 80Figure 3.1. Urban Renewable Energy Source (URES) DSS.

.................................................84Figure 3.2. Conceptual Model of the UWEP DSS.......87Figure 3.3. The vertical stratification of the urban

boundary layer...................................91Figure 3.4. Wind speed histogram (a) and direction

rose (b).........................................92Figure 3.5. Morphology-induced wind amplification.. 94Figure 4.1. Block model of the UWEP DSS............97Figure 4.2. Illustrative representation of the UWEP

module’s methodology............................100Figure 4.3. The Control Centre of the UWEP DSS.. . .102Figure 4.4. The logarithmic wind speed profile.. . .106Figure 4.5. The sky view factor as a measure of street

canyon geometry.................................109Figure 4.6. The geometry of urban morphology.......111Figure 4.7. Satellite plan image of urban subregion

category # 1....................................112Figure 4.8. Cubic array representation of urban

subregion category # 1..........................112Figure 4.9 Neighbourhood-scale aerodynamic effects.

................................................115Figure 4.10. Idealised representations of building-

scale aerodynamic effects.......................116Figure 4.11. Idealised 3D representation of the UBL

sublayers.......................................120Figure 4.12. The WDE submodule of the UWEP module. 122Figure 4.13. The Project Details tab of the Site

Specification form..............................123

xv

Figure 4.14. The Location tab of the Site Specification form..............................124

Figure 4.15. The Topographical tab of the Site Specification form..............................125

Figure 4.16. The Meteorological tab of the Site Specification form (annual).....................126

Figure 4.17. The Meteorological tab of the Site Specification form (seasonal)...................127

Figure 4.18. The mean wind speed profile generated by the WDI.........................................133

Figure 4.19. The Wind Amplifier submodule of the UWEP module..........................................135

Figure 4.20. The Neighbourhood Morphology form.. . .136Figure 4.21. The relationship of urban road type

classifications.................................137Figure 4.22. Typical cross-sectional elements of an

urban road......................................138Figure 4.23. Reconfigured cubic array of category # 1.

................................................138Figure 4.24. The critical angles..................139Figure 4.25. The Aerodynamic Effects form.........140Figure 4.26. The Building Features form...........142Figure 4.27. Linear and angular building-specific

variables.......................................143Figure 4.28. Plan (left) and elevation (right) wake

zone developments...............................144Figure 4.29. Amplification zones..................145Figure 4.30. The Architectural Configurator module of

the UWEP DSS....................................148Figure 5.1. Robarts Library (left) and the Green

Venture EcoHouse (right)........................151Figure 5.2. The southeast side of the J.P. Robarts

Library.........................................152Figure 5.3. The John P. Robarts Research Library.. 153Figure 5.4. A satellite image of a portion of the St.

George campus...................................154

xvi

Figure 5.5. The wind speed histogram and wind direction rose at 80 m..........................155

Figure 5.6. The wind statistics at 80 m (CWEA).. . .156Figure 5.7. Wind direction frequency histogram of the

Robarts Library site............................157Figure 5.8. The mean wind speed profile at the Robarts

Library.........................................159Figure 5.9. Map image of the campus neighbourhood. 161Figure 5.10. The amplified vs. the original mean wind

speed profile...................................163Figure 5.11. The plan and elevation views of the

Robarts Library.................................164Figure 5.12. The UWEP DSS cubic representation of the

Robarts Library.................................165Figure 5.13. The decision process of the Wind

Amplifier.......................................166Figure 5.14. Isometric representation of the building

surface zones...................................167Figure 5.15. The amplified wind speed profiles around

the Robarts Library.............................168Figure 5.16. Turby® - An H-Darrieus VAWT..........170Figure 5.17. Monthly regional mean wind speed

variation.......................................175Figure 5.18. The windy season wind rose at the Toronto

Island Airport .................................176Figure 5.19. The windy season wind speed distribution

by direction at 21 m............................177Figure 5.20. The ENE & WSW mean wind speed profiles.

................................................178Figure 5.21. The WSW peak seasonal and amplified wind

speed profiles..................................180Figure 5.22. Instrumentation placement on the

southwest penthouse.............................183Figure 5.23. Robarts Library Monthly mean wind speed.

................................................184Figure 5.24. Monthly mean wind speed (>80% complete

months).........................................185

xvii

Figure 5.25. The windy season wind rose vs. TIA data veered 90 degrees...............................186

Figure 5.26. Wind direction frequency comparison.. 187Figure 5.27. The windy season wind speed distribution

comparison......................................188Figure 5.28. The primary & secondary direction wind

speed distributions.............................189Figure 5.29. Computational fluid dynamics models

created by CFX..................................190Figure 5.30. Wind power density at 80 m (CWEA) with

Toronto inset...................................194Figure 5.31. Idealised representation of conical

delta-wing vortices.............................195Figure 5.32. The extensions of the Robarts Library.

................................................196Figure 5.33. The Green Venture EcoHouse - Then & Now.

................................................197Figure 5.34. The UWEP DSS representation of the

EcoHouse........................................198Figure 5.35. EcoHouse site wind direction rose

overlay.........................................200Figure 5.36. Power coefficient as a function of tip

speed...........................................202Figure 5.37. UWECSs proposed for installation at the

EcoHouse site...................................204Figure 5.38. Roof ridge and ground mounted UWECSs at

the EcoHouse Site...............................207Figure 6.1. The Bahrain World Trade Center........217Figure 6.2. Primary and secondary wind speed frequency

distributions (TIA).............................222Figure 6.3. Velocity in a Rankine Vortex..........225Figure 6.4. Sun and wind orientation diagram......226

xviii

LIST OF TABLES

Table 2.1. Causes and time scale of temporal wind variation........................................31

Table 2.2. Broad-scale classification of meteorological models............................36

Table 2.3. Wind Resource Estimation Methods.........57Table 2.4. Commercial Wind Resource Assessment

Software.........................................60Table 4.1. The Roughness Classification table of the

UWEP DSS........................................105Table 4.2. The urban subregions of the Urban

Parameterisation table..........................110Table 4.3. Neighbourhood-scale Wind Amplification

Factor table....................................114Table 4.4. Pressure coefficient-based Wind

Amplification Factor table......................117Table 5.1. The wind direction frequencies for the

Robarts Library site............................156Table 5.2. Wind Statistics Summary................158Table 5.3. Representative neighbourhood street

characteristics.................................161Table 5.4. Neighbourhood inter-building spaces and

angles of flow incidence........................162Table 5.5. Summary of the first stage of the

amplification assessment........................162Table 5.6. Summary of the second stage of the

amplification assessment........................167Table 5.7. Characteristic dimensions and amplification

zone areas......................................169Table 5.8. Wind power by building face and zone at the

Robarts Library.................................171Table 5.9. Mean annual wind energy by building face

and zone........................................172Table 5.10. Summary of the WSW amplification

assessment......................................179

xix

Table 5.11. The peak season contributions from the WSWwind direction..................................181

Table 5.12. Mean wind speed and direction summaries.................................................191

Table 5.13. Mean wind speed and associated wind power density comparison..............................193

Table 5.14. Wind energy by building face and zone at the EcoHouse....................................198

Table 5.15. EcoHouse building face-specific assessmentsummary.........................................199

Table 5.16. Performance characteristics of the proposed UWECSs.................................205

Table 5.17. UWECS power generation and percent of rated power.....................................209

Table 5.18. Roof top wind energy by building face at the EcoHouse....................................210

xx

NOMENCLATURE

Nomenclature Primer

Due to conflicting symbol-definitions both within theboundary layer meteorology literature and between thevarious disciplines, some compromises had to be made.For example, h appears as the height of the boundarylayer and the reference height on a building, urepresents the wind speed and the horizontal componentthereof, and P is used to designate both power andpressure.For most cases these conflicts were left unresolvedsince the multiple definitions did not come to bearwithin the same context. For example, B is usedthroughout to designate building breadth, a geostrophicdrag law constant, and a transitional profile equationparameter. In situations where there appeared to be anindustry standard more specifically applicable to thisundertaking, it was adopted for the body of the textand equated to the variables ascribed to the graphs ortables from the literature presented in the section.For example, the designation of building width as Bused within the building code was adopted as opposed tousing B to represent the street canyon width, as isdone in the wind power meteorology literature. Thistype of inconsistency is locally defined as anexception. In other cases, new symbols were introduced(e.g., M for wind magnitude).In general, subscripts are used to maintain consistencyand convention (e.g., D typically designates buildingdepth, so DS is used to designate the smaller of thetwo building plan dimensions (i.e., B & D)). Whenessentially identical variables already adorned withsubscripts occur within the same section, lower andupper case subscripts are used (i.e., CP (pressurecoefficient) vs. Cp (coefficient of performance)).

xxi

Nomenclature is presented in two sections: Acronyms andNotation. Subscripted variables not listed in theNotation section are defined in the Subscripts sub-section of the Notation section.

Acronyms

ABL Atmospheric Boundary Layeragl above ground levelART Advanced Renewable TariffsASCE American Society of Civil Engineersasl above sea levelBAWT Building Augmented Wind TurbineBEPA Biomass Energy Potential AssessmentBLWT Boundary Layer Wind TunnelBRE Building Research EstablishmentBREDEM Building Research Establishment Domestic

Energy ModelBREEAM Building Research Establishment Environmental

Assessment MethodBUWT Building augmented, integrated, and/or

mounted Wind TurbineCanWEA Canadian Wind Energy AssociationCERI Canadian Energy Research InstituteCETC CANMET Energy Technology CentreCFD Computational Fluid DynamicsCHC Canadian Hydraulics Centre (NRC)COM COMpressed (referring to climate data sets)CORINE CoORdination of INformation on the

EnvironmentCPZ Control Potential ZonesCT Canadian TireCWE Computational Wind EngineeringCWEA Canadian Wind Energy AtlasDAD Database-Assisted DesignDD Decimal-DegreesDEM Digital Elevation Model or MapDM Degrees:Minutes

xxii

DMS Degrees:Minutes:SecondsDOE Department of Energy (US)DSM Digital Surface ModelDSS Decision Support SystemDTM Digital Terrain MapEC Environment CanadaECMRF European Centre for Medium Range (weather)

ForecastingEEP Energy and Environmental PredictionEnSim Environmental SimulationEP Energy PlanningeQuest QUick Energy Simulation ToolEU European UnionEWEA European Wind Energy Associationfad frontal area density [f]far frontal aspect ratio [B or L / H]FIWF Fully Independent Wake FlowFRG Federal Republic of GermanyFS Frontal Stagnation (point).fst Format Standard file format, formerly .rpnGEM Global Environmental MultiscaleGESIMA GEesthacht SImulation Model of the AtmosphereGFS Global Forecast SystemGHG Green House GasGIS Geographic Information SystemGPS Global Positioning SystemGS Ground Separation (point)GUI Graphical User InterfaceGWEC Global Wind Energy CouncilHAWT Horizontal Axis Wind TurbineHIRLAM HIgh Resolution Limited Area ModelHOMER Hybrid Optimization Model for Electric

RenewablesHPDM Hybrid Plume Dispersion ModelIBC International Building CodeIBL Internal Boundary LayerICC International Code Council

xxiii

IRF Isolated Roughness FlowISL Inertial Sub-LayerIWA Industrial Wind Action (Group)KAMM Karlsruhe Atmospheric Mesoscale ModelLEED Leadership in Energy and Environmental DesignLES Large Eddy SimulationLIDAR LIight Direction And RangingLSR Least-Squares RegressionLULUCF Land Use, Land Use-Change, and ForestryM Moments (Method of)MASS Mesoscale Atmospheric Simulation SystemMC2 Mesoscale Compressible CommunityML Maximum Likelihood (Method)MM5 Mesoscale Model - 5th generationMML Modified Maximum Likelihood (Method)mn monthMNR Ministry of Natural ResourcesMO Monin-ObukhovMODIS MODerate resolution Imaging SpectroradiometerMOS Model Output StatisticsMOST Monin-Obukhov Similarity TheoryMPAC Municipal Property Assessment CorporationMPC Measure-Predict-CorrelateMRF Medium Range Forecast (MRF)MSC Meteorological Service of CanadaMSFD Mixed Spectral Finite DifferenceNBC National Building CodeNCAR National Center for Atmospheric ResearchNCD&IA National Climate Data and Information ArchiveNCEP National Centers for Environmental PredictionNIMBY Not In My BackYardNOAA National Oceanic and Atmospheric

AdministrationNRCan Natural Resources CanadaNTS National Topographic SystemNWA Numerical Wind AnalysisNWP Numerical Weather Prediction

xxiv

OWA Observational Wind AnalysisOWRA Ontario Wind Resource Atlaspad plan area density [P]PBL Planetary Boundary LayerPCA Principle Component Analysispdf probability density functionPEI Prince Edward IslandPLEA Passive and Low Energy ArchitecturePNL Pacific Northwest LaboratoryPSS Planning Support SystemRANS Reynolds Averaged Navier-StokesRE Renewable EnergyRES Renewable Energy SourceRET Renewable Energy TariffsRETScreen Renewable Energy Technology ScreenRL Robarts Libraryrpm rotations per minuteRPN Recherche en Prévision NumériqueRSL Roughness SublayerRSME Root-Square Mean ErrorRST Retail Sales TaxRWC Regional Wind Climatesar side aspect ratio [D/H] or [W/H]SATP Standard Atmospheric Temperature & Pressure

(25 ºC & 101 kPa) ~ 1.169 kPa

SDSS Spatial Decision Support SystemSEP Solar Energy PlanningSF Skimming FlowSGS SubGrid Scale (LES CFD model based on

filtering)SISL Semi-Implicit Semi-LagrangianSOC Standard Offer Contract (also SOP)SOP Standard Offer ProgramSL Surface LayerSOI Spatial Openness IndexSP Separation Point

xxv

SRTM Shuttle Radar Topographic MissionSTP Standard Temperature & Pressure (0 ºC &

101.325 kPa)~ 1.293 kg/m3

SVF Sky View FactorTAC Transport Association of CanadaTI Turbulence IntensityTIA Toronto Island AirportTKE Turbulence Kinetic EnergyTMY Test Meteorological YearTRY Test Reference YearTSD Time Series DataTVM Temperature Variance MethodUBL Urban Boundary LayerUCL Urban Canopy LayerUHI Urban Heat IslandUQAM Université du Québec à MontréalURES Urban Renewable Energy SourceUS Unites States of AmericaUSGBC US Green Building CouncilUWECS Urban Wind Energy Conversion SystemUWEP Urban Wind Energy PlanningVAWT Vertical Axis Wind TurbineVBA Visual Basic for ApplicationsWAM Wind Atlas MethodologyWAsP Wind Atlas analysis and application ProgramWDE Wind Data ExtrapolatorWDI Wind Data InterpolatorWEB Wind Energy for the Built environmentWECS Wind Energy Conversion SystemWEP Wind Energy PlanningWEST Wind Energy Simulation ToolkitWFLC Wind Farm Location CriteriaWIF Wake Interference FlowWMO World Meteorological OrganisationWPD Wind Power Density W/m2

WWEA Worldwide Wind Energy Association

xxvi

ZED Zero Emissions Development

Notation

A geostrophic drag law constantA cross-sectional area (m2)A intermediate mean wind speed profile equation

parameterAd lot area (m2)Af frontal area (m2)Ap plan area (m2)AS swept area or characteristic area of a wind

energy conversion device (m2)At total area (m2)AZ amplification zone area (m2)a roughness layer height parameter, 2 < a < 5a attenuation coefficientB cross-wind building breadth (m)B geostrophic drag law constantB street canyon width, more typically

represented by W (m)B intermediate mean wind speed profile equation

parameterB/H frontal aspect ratio (far)b Rayleigh pdf parameterCD drag coefficientCg gust coefficientCP pressure coefficientCP Cg peak composite pressure-gust coefficientc Weibull distribution scale factor (m/s)cd elemental drag coefficientCp coefficient of performanceD along-wind building depth (m)D/B fineness ratioD/H side aspect ratio (sar)DC characteristic dimension (m)Di elemental drag (N)

xxvii

DS smaller of the two building plan dimensions (m)

d zero plane displacement (m)d0 zero plane displacement height [z0 + d] (m)E theoretical maximum extractable wind energy

(Wh per time period)Ea actual energy generated by wind energy

conversion device (Wh per time period)Et theoretical maximum extractable wind energy,

based on wind energy conversion device performance criteria (Wh per time period)

coriolis force (s-1)g the force of gravity ~ 9.8 m/s2

g/T buoyancy parameter (m/s2 C)G geostrophic wind speed (m/s)H weighted average building height (m)H height of roughness element (e.g., building)

(m)H1 upwind building height (m)H2 downwind building height (m)H1/H2 cross-canyon building height aspect or

relative height ratioH/B slenderness ratioH/D height-to-depth aspect ratioHe height of building eaves (m)Hm mid-roof height (m)HS smaller of the cross-wind dimensions (e.g., B

or H)h height of the planetary boundary layer (PBL)

(m)h reference height (e.g., He, Hm, or Hr) (m)I turbulence intensityk Weibull distribution shape factorL Ubukhov length (m)L length, typically horizontal length of a

street canyon (m)L/H frontal aspect ratio (far)

xxviii

LB block length of shortest block(m)Lc drag length scale (m)Lg geometric influence scale (m)lm mixing length profilelc mixing or turbulence length scale (m)P wind power (W)Pa actual power generated by wind energy

conversion device (W)p pressure (kPa)QH sensible heat flux (kg/s)q velocity pressure (Pa)R2 correlation coefficientRD rotor diameter (m)RH rotor height (m)r average great circle radius of the earth ~

6372.795 kmS inter-building space width (m)S/H space width-to-height aspect ratioS change in wind speed (for topographical

amplification)SB cross-wind inter-building space width (m)SD along-wind inter-building space width (m)SC orthogonal inter-building space width (m)Sm minimum inter-building space width (m)SV in-line inter-building space width (m)T temperature (C)t timeu horizontal wind speed (m/s)u* friction velocity (m/s)u*/ surface Rossby length scaleui cut-in wind speed (m/s)umax wind speed yielding the most energy (m/s)ump most probable horizontal wind speed (m/s)u0 WECS cut-out or furling wind speed (m/s)

mean wind speed (m/s)mean turbulent component of horizontal wind speed (m/s)

xxix

gust speed (m/s)W street canyon width (m)W/H side, street, or canyon aspect ratio (sar)w vertical wind speed (m/s)

mean turbulent component of vertical wind speed (m/s)

x horizontal distance (m)x0 horizontal adjustment distance length scale

or fetch (m)zH average height of roughness elements ~ height

of the urban canopy layer (UCL) (m)zi height of the inversion layer or top of the

planetary boundary layer (PBL) (m)z0 roughness length (m)z* height of the top of the roughness sublayer

(RSL) (m)zSL height of the top of the surface layer (SL)

(m)zW wake diffusion height (m)

Greek wind shear exponent2 chi-square error delta or change in the variable, which the

symbol precedes height of the internal and/or pertinent

boundary layer (m) degrees latitude () gamma function amplification factor angle of flow incidence () grid orientation ()B building orientation ()R roof-ridge orientation () von-Karman constant ~ 0.4 degrees longitude () tip speed ratio

xxx

p plan area density (pad) (%)f frontal area density (fad) or roughness density (%)s mean building height to street width aspect ratio cardinal sector from which the wind originates () roof pitch or slope () density (kg/m3) standard deviation Reynolds stress (m2/s2) angular velocity of the earth (2/seconds perday) ~ 7.25E-05 rad/s frequency of rotation (Hz) stability correction function comfort parameter

Subscriptsa amplifiedc criticalH mean building height-levell leewardp pedestrian level ~ 1.75 - 2.6 m or primaryr referenceR rateds secondary or sidet topw windward

xxxi

M. A. Sc. Thesis - R. Cooper McMaster University - Civil Engineering

1 INTRODUCTION

The technology to harness the energy of the wind to

create electricity was

developed in the mid-

1800s (Figure 1.1),

while mechanical

application of wind

energy dates back to

the Egyptians circa

2800 BCE (Park, 1981).

An excellent history,

with a mild European

bias, is provided by

the Danish Wind Industry Association (2006). Along a

parallel timeline, Hein's (2003) series of articles

details the development of the omnipotent,

intrinsically connected electricity industry.

Figure 1.1. Brush Windmill Cleveland, OH (1888).Figure Note: The Brush windmill was considered the world's largestat the time, with a rotor diameter of 17 m (50 ft.) and 144 cedarwood rotor blades. The circled figure to the right of the turbineis a man mowing the lawn (Danish Wind Industry Association, 2003)copyright © the Charles F. Brush Special Collection, Case Western

Reserve University, Cleveland, Ohio.

During the 20th Century, further developments of

this technology were rather sporadic. Once the

1


electrical transmission grid was developed and deemed

reliable (~ 1930s (Hein, 2003)), advancements were

limited to crisis periods (i.e., war, oil shortage, and

the advent of the environmental movement).

Correspondingly, the largest collection of relevant

technical literature was published between the mid

1970s and the early 1980s. In the wake of the

environmental movement and the advent of the PC, the

development of numerical models enabled numerous

disciplines to assess the impact of urban development

on the global ecosystem, introducing such terms as

anthropogenic heating, the urban heat island (UHI), and

bioclimatic urban design. Wind-related studies

included assessment of pollutant dispersion for air

quality, natural ventilation, soil erosion, and

pedestrian comfort. In response to the times, the

burgeoning wind industry equipped with a very

rudimentary understanding of the wind resource and

economy of scale-perspectives, created today's wind

energy paradigm that is best illustrated by Figure 1.2.

2


Figure 1.2. Wind Farm in Coachella Valley, California(US).

(Gilbert, 2005).

At the beginning of the 21st Century, heightened

awareness of non-renewable resource depletion, global

warming, and the concept of security of energy supply

have stimulated new interdisciplinary research. But

concepts such as aerodynamic urbanism (Dunster, 2001)

and wind energy planning are still very much hampered

by the original paradigm. This resulted in the mere

relocation of the existing wind farm-sized turbines

into an urban setting and/or mounting the towers on

existing buildings to, unknowingly, ensure exposure to

3


adequate wind speeds. The advent of R-Urbanism,

advocating sustainable integration of rural and urban

areas (Revi et al., 2006), will possibly further

encourage this inappropriately-scaled integration of

wind turbines into urban developments. As remote

locations are finite, and not without various

drawbacks, it is feared that the soldiers of the wind

industry (Figure 1.2) are planning an invasion on the

urban domain (Wind Stop, 2006).

So, why should one even be focussing on electricity

generation, let alone the consideration of wind as a

viable renewable source?

1.1 Climate Change and Government / Industry Response

As of 2004, Canadian greenhouse gas emissions (GHG)

have increased by 26.6% above the 1990 baseline.

Canada is not only one of the largest producers of CO2

emissions, but the leader in establishing a trend in

the wrong direction between 1990 and 2002 in comparison

to the G8 countries and India, Brazil, and China, as

illustrated below.

4


Figure 1.3. G8 Countries CO2 emissions per capita for1990 and 2002.

(Environment Canada (EC), 2006)

Between 1990 and 2002 Canadian GHG emission

contributions from the energy sector alone increased by

25.3%, accounting for over 70% of Canadian GHG

emissions by 2004 (EC, 2006). Energy Industries (i.e.,

public electricity and heat production, petroleum

refining, and manufacture of solid fuels and other

energy industries (Intergovernmental Panel on Climate

Change (IPCC), 1997)), were responsible for 37.9% of

the aforementioned increase (Adejuwon, Herold, & Hanna,

2005). The Land-use, Land-use Change, and Forestry

(LULUCF) sector contributes 10% as the second highest

source of Canadian GHG emissions.

5


Energy sector data are primarily categorised as

pertaining to either stationary or mobile combustion

sources, with the former accounting for the majority of

this sector's GHG Emissions. Analysis of the

stationary combustion source category data suggests

that targeting Electricity and Heat Generation, which

accounts for approximately 36% of the contribution

within this subsector, could have a substantial impact

on Canadian GHG emissions.

36%

22%

14%

12%

11%5%

Electricity and Heat G enerationFossil Fuel IndustriesM anufacturing IndustriesResidentialC om m ercial & InstitutionalO ther

Figure 1.4. Canadian stationary source combustion GHGEmissions (2004).

The percentage contributed by the subcategories of the StationarySource Combustion category of the Energy Sector (EC, 2006).

Canada was one of the first countries to sign the

Kyoto protocol, formally ratified in 2002. The

intensity based, polluter-pays strategy detailed in the

made-in-Canada solution named “Turning the Corner”

6


unveiled in late April 2007, is under close scrutiny by

global experts (Liberal Party of Canada, 2007). On

February 12th, 2007 the Canadian government announced

that $1.5 billion of an anticipated 2006-07 budgetary

surplus would be used to establish the “Canada ecoTrust

for Clean Air and Climate Change”, which is intended to

finance Provincial initiatives (Office of the Prime

Minister, 2007b). On March 6th, 2007 $586.2 million of

this ecoTrust was granted to the province of Ontario

(Office of the Prime Minister, 2007a). In Ontario,

Standard Offer Contracts (SOC) have been in place for

over a year as a financial incentive to encourage

individuals to generate electricity from renewables. A

SOC pays the producer for feeding electricity back into

the grid (i.e., for wind energy-based generation:

$0.11/kWh + $0.0352 for peak hour supply and for solar

energy-based generation: $0.42/kWh). The Retail Sales

Tax (RST) rebate is yet another incentive program in

Ontario, which was expanded in 2004 to include wind,

micro hydro-electric, and geothermal energy systems for

residential premises (Ministry of Revenue, 2007).

From 1990 to 2004 the total electricity generation

capacity in Canada grew 24%, while the associated GHG

emissions over the same time period increased by 35%

(EC, 2006)! Figure 1.5 summarises Canada's hydro-

7


dominated, electricity generation source mix for 2004.

The data concerning renewable energy (RE) sources are

from 2005 and yet renewables still constitute a mere

1.1% of the mix.

10.8

9.7

2.21.1

38.6

3.4

0.2hydrocoalnuclearnatural gasoilrenewablesother fossil-fuels

Figure 1.5. Canadian electricity generation source mixin GW (2004).

Categories are in order of percent contribution. Renewables dataare from 2005 (EC, 2006).

The current electricity generation source mix for the

Ontario, contrary to popular belief, is dominated by

nuclear, not hydro, as illustrated below. Renewables,

primarily due to 395 MW (0.4 GW) of installed wind

energy capacity, come in at 1.5%. Ontario has

established renewable energy targets to increase this

contribution to 5% by 2007 and 10% by 2010 (EC, 2006).

Of all the Canadian provinces, Ontario should be the

8


most motivated to explore a more diverse generation

source mix in light of the blackout in 2003.

36.6%

24.9%

20.6%

16.4%

1.3%

0.2%

1.5%

nuclearhydrocoalnatural gas & oilwindbiom ass

Figure 1.6. Ontario electricity generation source mixin GW (2007).

Categories are in order of percent contribution. Installed nuclearcapacity is 14 GW (Ontario Power Authority (OPA), 2007).

Wind energy may be one answer to increasing the

contribution of renewable sources to the electricity

generation mix and reducing the Energy Industry GHG

emissions.

1.2 The Wind Energy Industry

Wind power is the fastest-growing electricity

source in Canada. From 1995 to 2006 wind energy

capacity grew by almost 4000% to 776 MW (~0.8 GW),

placing Canada 12th in the ranking of countries with

the highest installed wind energy capacity, as depicted

9


in Figure 1.7 (Brazeau, 2007). Growth is not expected

to subside in the near future (EC, 2006).

20.6

11.6

11.68.7

6.3

3.12.62.12.01.71.61.6

21.7

0.8

G erm anySpainUSO thersIndiaD enm arkC hinaItalyUKPortugalFranceNetherlandsC anada

Figure 1.7. Installed wind energy capacity (GW) bycountry as of 12/2006.

The total global capacity of 74.223 GW is enough to power 22.5million homes (Brazeau, 2007).

The Canadian Energy Research Institute (CERI) (2005)

has estimated the technical renewable energy potential

from a variety of sources, totalling 136 GW as

summarised below. To date, Canada has only realised

less than 2% of the 40 GW potential from wind energy.

10


70

40

13

103

solarwindwavesm all hydrotidal

Figure 1.8. Technical renewable energy potential inCanada in GW by source.

Wave energy is estimated to be between 10 and 16 GW (CanadianEnergy Research Institute (CERI), 2005).

Wind was undoubtedly one of the first renewable

resources harnessed for practical application. Upon

reintroduction into the technological era, it is

currently the youngest child of the renewable energy

family. Modern wind industry project initiatives are

primarily based on an assessment of cost-effective

energy production. The large scale of these

undertakings is due to the exponential growth in energy

consumption since the days of the first windmill and

the greatly undervalued cost of readily available

energy generated from non-renewable resources. This

has created the MW windfarms of the common era;

11


shrouded in socio-environmental controversy, attracting

the notion of “not in my backyard” (commonly

abbreviated as NIMBY), the majority of the proponents

appear to be those in whose backyard these machines do

not sit.

Europe, specifically Germany, Spain, and Denmark,

has long been a leader of the wind energy charge.

Meteorological organisations have mapped out the entire

planet, producing estimates of global renewable energy

potential which are astounding. The challenges to

realising this potential are primarily considered to be

financial, due to the current market-based paradigm and

the fiscally-biased restraints it imposes. Working

within these confines, many have long proposed

introduction of Advanced Renewable Tariffs (ART) and/or

Renewable Energy Tariffs (RET) (Gipe, 2007), which

successfully stimulated the industry in Europe. This

advocacy is what lead to the development of Ontario's

SOCs in March of 2006.

Though much work has been done in Europe, the wind

farm paradigm remained, even there, until late into the

1990's when Project ZED (Zero Emissions Development)

was commissioned. Project ZED was one of the first

serious attempts at seeking to enhance and/or

12


concentrate the wind and integrate wind energy

conversion systems (WECSs) into

building architecture. In early

2000, Joule III - Project WEB,

commissioned by the European Union

(EU) and based on Project ZED,

further explored the aerodynamic

properties of built structures and suitable WECS

configurations. These projects coined the terms Urban

Wind Energy Conversion System (UWECS) and Building

Augmented Wind Turbine (BAWT). Until recently, the

outcome of Projects ZED & WEB was only a plethora of

journal articles detailing siting, building

architecture, and WECS design guidelines. The only

drawback to the concepts proposed through these

projects is that a complete re-design of large

skyscrapers is required to put the conceptualised BAWT-

theory into practice.

Figure 1.9. Building Augmented Wind Turbines.Extracted from BDSP Partnership Ltd., MECAL Applied Mechanics BV,

Imperial College, & University of Stuttgart (2001)

Regardless, this wind enhancement work was a

milestone for the wind energy industry. Prior to this,

wind turbine siting was primarily dictated by the

presence of an adequate free-stream wind speed, while

13


wind enhancement was, by and large, unwanted if not

actually considered an outright nuisance. Large-scale

wind farms, though perhaps a solution in some areas,

are not necessarily the only solution. While advocates

of expansive wind farms would argue that small-scale

WECSs are not as efficient, the opponents concerned

about noise, low-frequency vibrations, safety, bird-

kills, and aesthetic impact, would become wind-energy

converts on discovering that these issues could be

rendered inconsequential through installation of small-

scale, in situ, wind energy generation devices.

Large wind farms in remote areas require a costly,

extensive infrastructure of lines to connect the farm

to the grid; often involving deforestation to allow for

the construction of service access roads. The GHG

emissions resulting from the land-use change

(previously identified as the second largest

contributor to GHG emissions in Canada) associated with

such an undertaking could conceivably offset the

reductions achieved through increasing the percentage

of renewables in the generation source mix. Current

initiatives further complicate matters by placing the

wind turbine towers offshore. There are also losses to

be considered when transmitting electricity over great

distances. If the electricity could be created at the

14


point of use the need for this infrastructure, and its

inherent losses, would be diminished if not removed.

By integrating wind and solar energy extraction, the

impacts of inherent energy fluctuations could be

minimised. Through co-operatives, community networks

could be developed, wherein buildings producing more

energy than they required would provide their excess to

those less favourably situated. Making energy

production a local responsibility would take pressure

off the grid, currently primarily powered by non-

renewable resources, and enhance a community’s level of

self-sufficiency.

So, one may still be asking, “What does the building

industry have to do with wind energy?”

1.3 The Building Industry

The building industry as a whole, including urban

planners and developers, is coming to the realisation

that urban sprawl is unsustainable. New concepts,

including urban infill and the rehabilitation of

brownsites are being explored. Councils, coalitions,

and alliances are collaborating all over the world

advocating green building (e.g., the US Green Building

Council (USGBC) and Natural Resources Canada (NRCan) -

Office of Energy Efficiency), as society has come to

15


realise that a large percentage of our energy is

consumed through heating and cooling the built

environment. Developers, architects, and consulting

firms are developing expertise regarding the provisions

of various energy efficiency rating systems (e.g.,

LEED). More aggressive advocates are proposing a whole

building policy to take the industry beyond green (D. L.

Jones, 1998) through efficiency vigilance and the

creation of buildings that produce all of their own

energy, as is the case for the proposed eco- or energy

tower illustrated below.

16


Figure 1.10. An artists rendition of the Burj al-Taqa(Energy Tower).

Proposed for Riyadh, Dubai and Bahrain by Eckhard Gerber'sarchitectural firm and DS-Plan engineering company, Stuttgart

(Thaduesz, 2007).

Ontario’s Building Code was recently under review

to ensure that the necessary provisions are in place

for the green building movement and that there are no

stipulations that may restrict improvements to building

energy efficiency, including provisions for alternative

17


energy extraction devices. There are a myriad of

stand-alone renewable energy (RE) extraction systems

currently on the market, focussing on generating

electricity, heating, and cooling. Solar panels, given

their relatively low complexity, are the leading device

being used, while geothermal properties are being

exploited in the form of heat pumps. In the field of

wind energy, the focus has been on large wind farms and

stand-alone single towers, typically situated in rural

settings. As such, a need has been identified to

explore the feasibility of urban-scale wind energy

extraction.

Given the proliferation of new codes, guidelines,

and alternative energy extraction device options there

is a dependence on multiple data sources, including:

meteorological data (e.g., solar radiation and wind speed),

geological data (e.g., ground water flows, soil composition, and land-use),

morphological (e.g., neighbourhood configuration, building orientation, and architectural features),

RE device performance data, financial data (to determine pay-back period and

feasibility), and socio-economic data.

18


Therefore the task of creating a new building or

refurbishing an old one suddenly becomes even more

complex. What is required is an architectural design

and simulation tool that supports assessment of the

implications of morphological and meteorological site

characteristics on the performance of various RE

extraction devices/systems. Ideally, it would take

seasonal variation into account, balance generation

potential against need, and determine storage and/or

supplementation requirements.

Though there are numerous Decision Support System

(DSS) tools in existence, including guidelines and

checklists, they are for the most part isolated

applications concerned with the assessment of

individual subsystems without consideration of

morphological implications. Since several tools (e.g.,

RETScreen) have already been developed regarding solar

energy assessment, the main focus of this research

undertaking is on small-scale wind energy extraction

devices in an urban setting.

Based on the scale of the modern day wind turbine

and the conventional siting tools in place, it is not

surprising that real life examples of urban wind energy

conversion systems (UWECSs) are difficult to find.

19


That said, wind turbines in an urban setting are by no

means new. In North America, the 1981 installation of

a roof-mounted wind turbine on a community dwelling at

Eleventh St. and Bronx is considered as the pioneer in

urban wind energy (Hurwood, 1981), while numerous other

stand-alone wind turbines in urban settings are

identified in the reports of Campbell & Stankovic

(2001a and 2001b). Unfortunately, the current concept

of a small wind energy industry, advocated by the

Canadian Wind Energy Association (CanWEA), is a bit of

a misnomer as it deals primarily with rural

installations of 30 m tower-mounted three-bladed wind

turbines. This research undertaking correspondingly

advocates, what must now be called, the microscale wind

energy industry in hopes of enacting a paradigm shift.

This microscale perspective will need to account for

neighbourhood morphology and building design details to

reasonably assess the implications of building

aerodynamics on wind energy potential.

1.4 Building Aerodynamics

There is much that the wind energy industry could

learn by exploring the aerodynamic nature of buildings.

Even though ground-level wind speeds are considered to

generally be lower, with turbulence becoming an

20


important factor in urban areas, the presence of

current studies looking to reduce wind levels in

pedestrian corridors suggests that there are building

affected zones that could potentially produce favourable

wind conditions for urban wind energy extraction. Wind

amplification, resulting in a cubed increase of the

wind energy available for extraction and an associated

potential reduction in turbulence, can be achieved

through careful consideration of:

neighbourhood morphology, building feature configuration and orientation, WECS mounting-tower architecture, and WECS design features.

Studies suggest that the combined amplification

produced by designs which incorporate the

aforementioned factors could conceivably increase the

mean wind speed into the WECS by a factor of four or

five, (i.e., 2X due to neighbourhood morphology

(Gandemer, 1977; Canadian Commission on Building and

Fire Codes, 2006) times 1.5X due to building

augmentation (Campbell & Stankovic, 2001a) times 1.4X -

1.8X due to mounting tower architecture (ENECO, 1999)).

These levels of mean wind speed amplification

correspond to a 64 and a 125 fold increase in wind

energy, respectively. This research undertaking

21


primarily focuses on consideration of the neighbourhood

morphology- and building feature-induced wind

amplification, best explained by the building augmented

wind turbine (BAWT) concept, while advocating design

and development of truly urban-scale WECSs.

BAWT theory proposes exploitation of building

aerodynamics-induced wind amplification through the

following three WECS installation configurations:

Roof-top, Near-building ground-level, and Building-integrated.

Building-integrated installations use the geometry of

the building (Figure 1.9) and/or ducting (Figure 1.14)

to amplify the mean wind speed and direct it into the

extraction device.

Project WEB coined the acronym BUWT to encompass

all three configurations. The assessment of urban

morphology-induced wind amplification, specifically in

support of siting building-integrated and/or mounted

wind energy conversion systems (WECSs), appears to be a

very recent undertaking. A growing number of wind

turbine manufacturers are exploring the development of

suitable urban-scale devices (Dutton, Halliday, &

Blanch, 2005; Wineur, 2007). To assist the reader in

22


envisioning how a wind turbine could possibly be

mounted anywhere near, let alone on, a building, urban

wind energy conversion system (UWECS) concepts will now

be discussed.

1.5 Urban Wind Energy Conversion Systems

In general, wind energy conversion systems (WECSs)

are typically either horizontal axis wind turbines

(HAWT) or vertical axis wind turbines (VAWT). Within

the VAWT family, the main configurations are the

Savonius and the Darrieus, as depicted in Figure 1.11

(a) and (b), respectively.(a) Savonius VAWT (b) Darrieus VAWT (c) HAWT

Figure 1.11. General WECS Configurations.(Iowa Energy Center, 2006)

Urban Wind Energy Conversion Systems (UWECSs)

primarily differ from conventional wind turbines in

terms of scale. UWECS designs are generally based on

23


VAWT configurations, with some intended to be

horizontally mounted, which are better suited to

harness energy from omni-directional, skewed, and

turbulent urban winds. There has been a virtual

explosion of UWECS-development. UWECSs are being

developed specifically for roof-top mounting,

graphically depicted in Figure 1.12, including modular

concepts as portrayed in Figure 1.13.

(a) (b) (c)

(d) (e) (f)

Figure 1.12. Rooftop wind energy conversion devices.(a) (b) (c)

24


(d) (e)

Figure 1.13. Modular Rooftop modular wind energyconversion devices.

The reader is directed to the photomontage matrix in

Appendix K for details pertaining to WECS type,

manufacturer, and image credits related to Figures 1.12

- 1.14, inclusive.

Configurations exploiting BAWT theory are roof-

mounting (Figure 1.14 (a)), ground-mounted near

buildings (b), or integrated into a building, as is the

case in the recently constructed Bahrain World Trade

Center in Figure 1.14 (c) and the University of

Strathclyde's ducted wind turbine (d).

(a) (b)

25


(c) (d)

Figure 1.14. Building augmented wind energy conversionsystems.

Excellent reviews of the BUWT industry, including

concepts, research, and current manufacturers are

provided by Blanch (2002) and Wineur (2007). A

comprehensive tabulation of more than 60 WECSs

developed throughout Europe and North America,

including BUWT-potential and performance

26


characteristics, has been compiled by Dutton et al.

(2005). The scope and objectives of this research

undertaking will now be discussed.

1.6 Scope and Objectives of the Research Undertaking

The recent privatisation of various sectors of the

utilities industry, namely natural gas and electricity,

has created supposedly beneficial-to-the-customer

competition amongst the various providers. But the end

result is that the consumer must now deal with the

worry about the future viability of the chosen provider

and continually ‘do the numbers’ to ensure that s/he is

getting the best price. Insecurity, especially in

light of the blackout of 2003, and concern about cost,

prevails. Given the current climate, literally,

politically, and economically, this is an ideal time to

give the security of energy supply and financial

control back to the building owner.

What if one could produce a large portion of their

energy needs right on their own property? The

questions that need to be answered would include:

What is the site's renewable energy (RE) potential(e.g., wind and solar energy generated)?

What efficiency, cost, etc. can be expected of thevarious RE extraction devices?

27


Where is the optimal placement for RE extraction devices?

What are the implications of site layout on RE energy potential?

What is the effect of building configuration on REpotential?

Could individual architectural features enhance REpotential?

What is the efficiency of the dwelling? What is the energy demand incurred by the

building's occupants? What is the estimated energy surplus / deficit? What are the various energy storage options? What are the current and projected utility costs? When will payback be realised?

The scope of this research undertaking proposes to

address assessment of the potential wind energy,

optimal placement and scale of UWECSs, and the

implications of building aerodynamics-induced wind

amplification. Evaluation of other RE sources,

dwelling efficiency, energy demand, and various socio-

economic aspects can be conducted using existing

applications. The reader is directed to Appendix L to

view a matrix of tools supporting such evaluations.

The main objectives of this research are:

Development of a prototype urban wind energy planning (UWEP) decision support system (DSS) for

28


homeowners, urban designers or planners, architects, and UWECS developers;

Assessment of building aerodynamics-induced wind amplification;

Assessment of the seasonal variation of the wind resource; and

Determination of optimal UWECS placement.

In accordance with the objectives and underlying

motivation behind this research undertaking, this

Thesis primarily documents the development of a

prototype urban wind energy planning (UWEP) decision

support system (DSS). It includes a first attempt

approach at estimating building-aerodynamics induced

wind amplification dependent on neighbourhood

characteristics, meteorological conditions, and a

subject building. The UWEP DSS will support wind

energy potential assessment of both an existing and a

proposed building, by providing the capability to make

various building feature-related changes (e.g., roof

type and pitch) in an iterative fashion. It will be

capable of conducting a mean annual as well as a

seasonal or monthly assessment. The proposed tool will

link the user to various online applications and

databases, thereby requiring a minimal amount of user

knowledge concerning meteorology, geography, or

29


building aerodynamics. The UWEP DSS will inform the

user on the:

potentially amplified mean wind speeds at the site,

potential energy that could be extracted, approximate size of an appropriate UWECS, placement location most appropriate for the

subject site, and months with the greatest potential.

1.7 Thesis Structure

This Thesis is divided into the following six

chapters:

Chapter 1 - Introduction Chapter 2 - Literature Review Chapter 3 - Conceptual Model Development Chapter 4 - Decision Support System Development Chapter 5 - Application of the Decision Support

System Chapter 6 - Summary, Conclusions, and

Recommendations

Following this introduction, Chapter Two of the Thesis

is devoted to a literature review, which draws from

numerous interrelated disciplines to explore the

feasibility of BAWT theory in an urban setting. This

chapter includes a review of the current state of the

relatively new field of wind power meteorology, related

30


fields of research in support of defining the flow-

field in the building affected zone, and an overview of

the various applications and DSSs currently in

existence. Chapter Three details the development of

the conceptual model of the proposed UWEP DSS. The

fourth chapter details the development of the UWEP DSS

based on the conceptual model. This chapter includes

discussion on the form, function, and theory behind the

mathematical computations of the proposed tool.

Chapter Five contains a discussion on the performance

of, and results produced by, the UWEP DSS when applied

to case studies. Finally, in Chapter Six, a summary of

the research undertaking is provided, including

conclusions and recommendations for future research.

31


2 LITERATURE REVIEW

Recognising the highly interdisciplinary nature of

this research, an extensive manual and electronic

search was conducted. Initially, an internet search

was performed to identify the various national and

international institutions, organisations, and

associations advocating wind energy. The World Wide

Wind Energy Association (WWEA) (2006) created in 2001

and the Global Wind Energy Council (GWEC) (2007)

established in early 2005, were identified as key

organisations providing a forum for this truly

international pursuit. Even though there is an

overwhelming amount of published information and

ongoing research activity, completed and planned

installations are largely based on the traditional

approach to wind energy extraction, namely through wind

farms and stand-alone rural towers. It was through the

European Wind Energy Association (EWEA) that Project

WEB (Wind Energy for the Built Environment), funded in

part by the European Commission (EC) in the framework

of Joule III, was identified.

Having identified the disciplines and key concepts,

a search of Journals and Conference Proceedings was

conducted through McMaster University’s online

32


Information Portal known as Morris. Databases such as

INSPEC, Science Citation Index, Compendex, SAGE Full-

Text Collection, etc., accessed through Scholars

Portal, Web of Knowledge, and Engineering Village 2,

identified the following key journals:

Atmospheric Environment, Boundary-Layer Meteorology, Energy and Buildings, Energy Conversion and Management, Environment & Urbanisation, Journal of Applied Meteorology, Journal of Wind Engineering, Journal of Wind Engineering and Industrial

Aerodynamics, ReFocus, Renewable Energy, Theoretical and Applied Climatology, Wind & Structures, Wind Energy, and Wind Engineering.

To develop the structure of the proposed Decision

Support System (DSS), regulations, codes, standards,

and guides pertaining to wind turbine siting and

building construction were reviewed. Several articles

pertaining to decision support regarding other

renewable energy sources were also assessed.

33


Additional information was found in various trade

magazines (e.g., North American Wind Power, Wind

Directions, HomePower, and Popular Mechanics). Not

surprisingly, comments by organisations countering wind

energy industry initiatives (e.g. Industrial Wind

Action Group (IWA) (2006)) provided the impetus and

encouragement to pursue unconventional means for wind

energy extraction.

Finally, several software applications and online

assessment tools were explored to determine their

applicability and to further develop the methodology of

the proposed DSS.

2.1 The Interdisciplinary Nature of this Research

The wind in urban and suburban areas is a field of

study of, and interest to, many applied sciences (e.g.

climatology and meteorology, bioclimatic design and

building physics, pollutant dispersion studies, urban

and landscape planning, etc). It is only in the last

10+ years that findings from the field of agricultural

and forest meteorology have been explored for their

applicability to the built environment. Starting with

the formulation of the Urban Heat Island (UHI) concept

by Oke (1978), research proceeded to further

exploration of the correlation between meteorological

34


conditions and the morphological and anthropogenic

impact of urban/suburban developments.

The organisation of this literature review is based

on these interdisciplinary correlations and best

summarised using building blocks, as illustrated in

Figure 2.1.

Modelling, including Global Information

Systems (GIS)

Planning

Urban, inc

luding

landscape

Energy Managem

ent & Pow

er Transmission

Decision Support System(DSS)

Energy SciencesSystem s

ConservationAlternatives

Fluid Dynam ics

BuildingAerodynam ics W ind Engineering

Turbom achinery

Energy Conversion

EngineeringCivil

M echanicalAerospaceElectricalChem ical

Atm ospheric SciencesM eteorologyClim atologyW eather

Sustainable

Urban Boundary Layer

Renewable

W indSolarEtc.

Roughness

Resources

Building SciencesEnergy M anagem entResource M anagem ent

ArchitecturalBuilt Environm ent … . Passive & Low Energy

Environm ental SciencesUrban Environm entPollution DispersionG lobal Change

Im pact Assessm ent

Alternative Energy

Geographic

Information

Syste

m (GIS)

Canopy Layer

EcologyForestryAgricultureGeographyTopography

i.e.Orography

G lobal Political Econom yEco-Socialism

Energy & Resource Policies … Building Codes & Legislation … Incentives & Research Funding

Computer Science

Figure 2.1. The interdisciplinary blocks of urban windenergy research.

The foundation is the field of atmospheric

sciences. Looking to the lower right hand quadrant in

Figure 2.1, geography & ecology are seen as the first

fields to explore earth-atmosphere interactions.

Contributions from these fields include development of

35


the geographic information system (GIS), the concept of

roughness, and the definition of a canopy layer, upon

which the concept of the urban canopy layer (UCL) is

based. The interdisciplinary fields of energy,

environmental, and building science combine various

engineering disciplines under the influence of these

concepts and developments. Ideally, the field of

architecture should encompass the whole.

The core group of disciplines, indicated by the

dot-shaded region within the building block model in

Figure 2.1, is then regulated by the global political

economy and organised through various planning

disciplines. Finally, it is the contributions from the

field of computer science that provide the practitioner

with tools and decision support systems to be able to

put the entirety of this body of knowledge into

application.

Due to the interdisciplinary nature of the research

undertaking and the intrinsic connection between urban

morphology and the local climate, new concepts often

need to be introduced in one section while detailed

discussion thereof is withheld until a later section.

That said, the structure of the literature review

attempts to build background as it proceeds, while

36


reserving detailed discussion to concepts that pertain

directly to the proposed DSS. It is divided into the

following main sections:

Wind Power Meteorology Overview; The Urban Boundary Layer (UBL), including

modelling; Wind Atlases; Wind Atlas Methodology (WAM); Wind Resource Assessment; The Built Environment; and Decision Support Systems (DSS).

2.2 Wind Power Meteorology Overview

Wind power meteorology is an applied science,

founded in boundary layer meteorology and based on wind

climatology and geography, specifically focussed on

wind resource assessment to optimise wind energy

conversion system (WECS) placement (Petersen,

Mortensen, Landberg, Højstrup, & Frank, 1998a). The

'klimatope' (landscape climate) in Figure 2.2 provides

an illustrative portrayal of this combination of

geographic, land-use, and climatic elements (Tanaka &

Moriyama, 2005). The climate analysis map includes

arrows indicating the direction of the regional land

and sea breezes, radar plots identifying locally

observed wind direction by frequency and cardinal

37


sector, and colour-coded dots indicating mean

temperatures.

Figure 2.2. The Osaka City 'klimatope'.Extracted from the prototype of overlaid “Climate Analysis Map” of

Osaka City (Tanaka & Moriyama, 2005).

Wind climatology, based on planetary boundary layer

(PBL) meteorology, deals primarily with the

geographical and temporal distribution of the wind

regime, including both speed and direction. The PBL is

the layer of the atmosphere at the bottom of the

troposphere where both mechanical and thermal processes

occur. A short primer on PBL meteorology is provided

in Appendix A, while terminology is defined in the

Glossary, for the reader seeking more background. This

38


review will focus on the urban boundary layer (UBL)

within the PBL.

The geophysical components of interest are the

terrain topography (including vegetation, built-up

areas, and soil and water surfaces, known as roughness

elements) and the orography or large-scale terrain

features (including hills, cliffs, ridges, escarpments,

and valleys). The retardation effect of the size and

distribution of these surface features on the wind near

ground level is mathematically represented by a

roughness length [z0] parameter (Petersen, et al.,

1998b).

The rudimentary fundamentals of the climatological

aspects of this discipline are covered in the following

sections.

2.3 The Urban Boundary Layer

The urban boundary layer (UBL) is modelled as

either an internal boundary layer (IBL) of the PBL

(Fisher et al., 2004; L. Landberg & Watson, 1994;

Verkaik & Smits, 2001) or as a two-layer PBL (Clarke &

Hess, 1974; Verkaik & Smits, 2001; Verkaik, 1999).

Figure 2.3, based on the earlier works of Oke (1978),

appears in numerous articles by fellow researchers,

representing the most widely recognised and accepted

39


conceptual structure of an unstable daytime UBL. The

boundary layer is considered to be unstable during the

day while the surface is transferring heat to the

atmosphere and stable at night when the heat transfer

is reversed. The addition of a height scale attempts

to add clarity to the notes, including the designation

of the PBL height [zi] ~ u*/f (Verkaik & Smits, 2001;

Clarke & Hess, 1974). The average height of the

roughness elements [zH] is assumed to represent the

approximate height of the urban canopy layer (UCL),

with the term canopy pertaining to a group of roughness

elements. The height of the roughness sublayer (RSL)

[z*] is equated to a zH; where 2 < a < 5 (Fisher et

al., 2004). The inertial sublayer (ISL) lies roughly

between z = a zH and 0.25, where is the boundary-

layer height (MacDonald, 2000 after (Raupach et al.,

1980)). The top of the UBL is designated as or h.

In Figure 2.3 c) the sky view factor (SVF) is used to

define street canyon width-to-height ratio and an

idealised canyon vortex is illustrated. As illustrated

in Figure 2.3, the UBL can be broadly stratified into

mesoscale, local scale, and microscale layers, in order

of increasing complexity and resolution.

Defining the extents and properties of these layers

is crucial to wind power meteorology for estimation of

40


mean wind speed as a function of height and terrain-

induced roughness. Boundary layer theory (i.e., the

two-layer and the IBL) defines how to estimate this

mean wind speed profile within various layers of the

UBL.

a) M esoscale

RURAL

b) Local scale

z

SUBURBAN

c) M icroscale

URBAN

U C L

Inertial Sublayer

Roughness Sublayer

zH

z*

UC L

Urban Boundary Layer

W ind Direction

)z(u

PBL

Surface Layer

Outer Layer

SVF

Surface Layer

Figure 2.3. Urban boundary layer structural sketch.Based on Fisher et al. (2004) and Rotach et al. (2004), modified

after Oke (1987).

The IBL theory is based on well-defined upwind and

downwind roughness transitions, such as the rural

41


through suburban to urban in Figure 2.3 a), which

modify the wind regime as it passes through. Figure

2.4 below further serves to illustrate the effect of

terrain roughness on the mean wind speed profile. It

is proposed within the literature that turbulent

kinetic energy (TKE) is what facilitates mean wind

speed adaptation and/or adjustment to substantially

different terrain (e.g., open country vs. a small town)

(Chan, 2001).

34 m /s

14 m /s

W oodland / Sm all Town

150 21 m /s

225 30 m /s

City Centre Open Country

Height [z] (m

)

75 32 m /s20 m /s

29 m /s

36 m /s

38 m /s

300

)z(u

Figure 2.4. The terrain dependent velocity shearprofiles.

Terrain roughness decreases from left to right (i.e., from CityCentre to Open Country). Modified after Dalgliesh & Boyd (1962).

This original approach worked well when applied to

vegetative stands (e.g., crops or forest) but does not

42


have the two-layer model's advantage of being able to

deal with the multiple roughness transitions typical of

an urban surface. (van Kuik & Bierbooms, 2005 p. 47)

The IBL model cannot accommodate multiple transitions

due to the fetch associated with each roughness region,

which is often longer than the distance between

transition zones (e.g., urban-to-suburban in Figure 2.3

a). The fetch is the horizontal distance required for

the wind profile to adjust to the new roughness. To

address this deficiency, some attempts have been made

to imbed individual IBL models within an IBL model

(Belcher, Jerram, & Hunt, J. C. R., 2003). Models

considering only the surface layer are also called

surface layer (SL) models (e.g., COAST (Verkaik,

1999)). In wind power meteorology, IBL theory is

applied to determine orographic terrain-induced

accelerations.

In the two-layer UBL conceptualisation, the

vertical structure is comprised of a surface layer (ISL

+ RSL) and an outer (mixed) layer (Figure 2.3 a)).

Over urban surfaces the upper and lower limits of the

inertial sublayer (between z/z0 >>1 and z/u* <<1) are

thought to coincide (Rooney, 2001). Two-layer theory

does not factor in orographic terrain-induced

43


accelerations. The theory behind this approach will be

covered in Section 2.6.

The past two decades of extensive research

involving the UBL has resulted in numerous advances in

boundary layer climatology, including a better

understanding of the roughness sublayer (RSL)

(Arnfield, 2003). As the research focus has shifted to

the development of numerical models, improved methods

of evaluating the crucial parameters (e.g., roughness

length [z0]) from urban form, methods of linking

microscale to mesoscale urban climate work, and model

validation are required. To decrease uncertainty,

careful consideration of spatial and temporal relations

is recommended. (Şahin, 2004)

Numerical models are based on the two-layer UBL

theory, the IBL theory, or a combination of the two.

UCL modelling, for the most part, will be covered in

Section 2.8 though some overlap is unavoidable. It is

within the UCL that urban-type roughness elements

induce complex turbulence fields and localised speed-

ups, which are of greatest importance to this research

undertaking. This section of the literature review

will focus primarily on defining the wind climate.

44


2.3.1 The Wind Climate

Wind is a complex, three-dimensional phenomenon

generated by large scale pressure differences caused by

the unequal distribution of solar radiation affected by

both thermal (Souch & Grimmond, 2006) and mechanical

processes. In wind power meteorology, interest

primarily lies in an assessment of the implications of

mechanical processes on the horizontal component of the

wind speed vector. The wind speed is also a function

of the local climate (hence year, time of year, and

time of day). Time series data (TSD) for the

horizontal component of wind speed and the associated

direction are available through most national

meteorological organisations. In Canada, national

climate data are available online through the National

Climate and Information Archive (EC, 2005d).

In order to apply the boundary layer theory, an

accurate representation of the wind climate, including

temporal variation (Table 2.1) and frequency

distribution must be developed.

Table 2.1. Causes and time scale of temporal wind

variation.

45


Cause Tim e ScaleGusts (e.g. turbulence) Sub-second to secondDiurnal cycle DayInversion layers Hour(s)W eather patterns Hours to daysSeasonal cycle (e.g. m onsoon) SeasonalAnnual variation Years

(Ackermann & Söder, 2000) The inversion layer is explained inAppendix A. †Discussion on turbulence is reserved for Section 2.8.

Studies concerning temporal variability (Li & Li,

2005a) have been conducted, investigating diurnal

(Weisser & Foxon, 2003), seasonal (T. Chang, Wu, Hsu,

Chu, & Liao, 2003; Lun & Lam, 2000), and inter-annual

(Voorspools & D’ haeseleer, 2007) time scale

implications. Temporal variability in wind direction

is considered negligible, except in coastal regions

where land-sea breezes are prevalent (Dandou, Tombrou,

Akylas, Soulakellis, & Bossioli, 2005).

To account for inter-annual variation, thereby

minimising the impact of the so-called 'windy years'

(the 90's), a full set of chronological data spanning

at least 30 years is recommended (van Kuik & Bierbooms,

2005). Alternatively, analysis based on one year of

data has been proven to yield predictions accurate to

within 10% (Jamil, Parsa, & Majidi, 1995 p. 623).

Compared to a rural environment, the influence of

seasonal variation within the UBL is reduced due to

46

†


roughness-induced wind field modifications (Radics,

Bartholy, & Pongrácz, 2002).

To reduce the computational effort required to take

into account temporal variation, the development of a

compressed (COM) wind data set is proposed (Celik,

2003). COM is currently available for solar radiation,

representing the important statistics of an entire

month using less than a month's worth of data. Another

useful data summary tool is the test reference or

meteorological year (TRY/TMY). Figure 2.5 illustrates

the (a) monthly and (b) diurnal variation of the mean

wind speed, comprised of instantaneous wind

measurements, of three different wind regimes.

(a) (b)

Figure 2.5. Variation of the primary northern hemisphere wind climates.The magnification portion shows the actual instantaneous wind measurements from which the mean is calculated. Modified after Petersen et al. (1997) p. 12.

47


On closer examination of the instantaneous wind

measurements, highlighted by the magnifying glass in

Figure 2.6 (b) above, the mean and turbulent components

can be seen (as illustrated below). Turbulence, or

gustiness, is generally three-dimensional and defined

as any irregular velocity perturbation superimposed on

the mean. Turbulence can be mechanically and/or

thermally induced, through what are called ground

forcings. The primary mechanical forcings are

frictional drag (i.e., terrain-induced roughness and

form drag, e.g. flow deflection around obstacles,

including orographic terrain features (Roulet, 2004)).

Thermal forcings are primarily due to solar heating of

the ground and buildings, which causes the air in

contact to become warmer than the surrounding air mass,

generating buoyancy-induced thermals.

48


Figure 2.6. The components of the instantaneous windspeed.

(Stull, 1988)

It is due to turbulent momentum flux (i.e.,

Reynolds stress) that the effect of the ever changing

surface forcings is transmitted throughout the boundary

layer (Stull, 1988). In wind power meteorology, a

turbulence intensity factor (TI), based on the standard

deviation of the wind speed, is the most commonly used

indicator of turbulence (Singh, Bhatti, & Kothari,

2006). In boundary layer meteorology, turbulence

kinetic energy (TKE) is used to more fully define the

variance. The reader is directed to the works of

Petersen et al. (1998a) and Stull (1988) for additional

detail.

Wind power meteorology relies on the friction

velocity [u*] parameter for the mathematical

representation of Reynolds stress in the formulation of

the wind profile as a function of terrain-induced

roughness and height, which is discussed in Section

2.6.

2.3.2 Frequency Distributions

While temporal variation is concerned with the

variability of the mean over a specific time period,

49


frequency distributions are used to define the

probability of observing specific wind directions and

speeds about the mean. Referring back to Figure 2.5,

Figure 2.7 below illustrates the wind speed frequency

distributions associated with the same three wind

climates.

Figure 2.7. Northern hemisphere wind climate frequencydistributions.

Extracted from Petersen et al. (1997) p. 12.

The frequency of the wind direction is typically

portrayed in a wind rose, which illustrates frequency

by directional sector. Wind is inherently multi-

directional, but the wind direction at a particular

location can usually be considered as predominantly bi-

50


directional, and occasionally uni-directional as

illustrated in Figure 2.8.

0.0%5.0%10.0%15.0%20.0%25.0%

N

NNE

ENE

E

ESE

SSE

S

SSW

W SW

W

W NW

NNW

Figure 2.8. The Toronto Island airport wind directionrose of June 2005.

The wind speed frequency distribution is portrayed

in histogram form and/or a representative probability

density function (pdf). Histogram bins are often based

on wind speed class, a predefined classification system

representing wind speed ranges (e.g., the Beaufort

scale (University of the Southern Pacific, & UNESCAP,

2004)).

Extensive research has been conducted fitting time-

series wind speed data to pdfs, primarily the log-

normal, Rayleigh (Li & Li, 2005b), and Weibull (Nigim &

Parker, 2007; Jamil et al., 1995; Seguro & Lambert,

2000) distributions. The two-parameter Weibull

51


distribution, below, has been found to yield the most

accurate estimation of the wind speed distribution:

Frequency; (2.1)

where: u is the horizontal wind speed, k is the shape

parameter, and c is the scale parameter. Further

detail regarding the Weibull pdf, including the double-

peaked, bi-Weibull distribution, can be found in

Burton, Sharpe, Jenkins, & Bossanyi (2001).

2.4 Boundary Layer Modelling

The defining equations of boundary layer theory are

the Navier-Stokes equations, which are derived from the

basic principles of conservation of mass, momentum, and

energy. Due to the highly non-linear nature of

atmospheric phenomena within the UBL, numerical models

are used to explicitly or implicitly solve the defining

equations. These models are typically either based on,

or validated by, field observations. Wind tunnel

testing (scale-modelling) appears to be reserved for

the assessment of the complexities associated with the

UCL and terrain orography. Over the last 20 years,

computers have allowed for the development of powerful

numerical models based on the theory developed by

earlier researchers. Much of the current research is

52


focussed on improving the accuracy and increasing the

resolution of these numerical models, while reducing

computational effort.

A simplified classification of meteorological

models, Table 2.2, is provided by Petersen et al.

(1998b). Models are actually comprised of a

combination of these classifications (e.g., MC2 is a

non-hydrostatic, linear, diagnostic, mesoscale,

neutral, turbulent closure model). Complexity

generally increases from left to right in the below

tabulation.

Table 2.2. Broad-scale classification of

meteorological models.Classifica ComplexityDynamics kinematic (mass-consistent), Advection linear, non-linear.Time diagnostic, prognostic.Spatial Global (synoptic), mesoscale > 5 km, Stratifica neutral, non-neutral.Friction frictionless, turbulent closure.Formulatio analytical, spectral, grid point.Type flow model, wind climate model.

Modified after Petersen et al. (1998b) p. 65. † verticalacceleration is considered negligible compared to vertical

pressure gradients and vertical buoyancy forces, which typifies athermally vs. mechanically-driven process. Global scale is

typically in the order of > 100 km.

Models are generally discussed in terms of their

spatial scale, also referred to as resolution.

53


Regional numerical weather prediction (NWP) models are

classified as mesoscale. The most common mesoscale

models used in wind power meteorology are:

Karlsruhe Atmospheric Mesoscale Model (KAMM) - theUniversity of Karlsruhe,

5th generation Mesoscale Model (MM5) - Penn State/NCAR

Mesoscale Compressible Community (MC2) model - MSC,

HIgh Resolution Limited Area Model (HIRLAM) - the European Meteorological Institute, and

Mesoscale Atmospheric Simulation System (MASS) - Meso Inc.

(van Kuik & Bierbooms, 2005)

Unfortunately, a Unix workstation or a Linux-based PC

is required to run these models, in addition to a

thorough understanding of meteorological processes and

the associated parameterisation schemes.

As the aim of this literature review is to provide

a reasonable background for the scope of the associated

research, only a condensed description of the relevant

mesoscale and microscale models on which wind atlases

and associated commercial software development are

based will be provided, highlighting key findings and

limitations.

54


2.4.1 Mesoscale Models

Mesoscale models (A. Martilli, Clappier, & Rotach,

2002) (e.g., Geesthacht Simulation Model of the

Atmosphere (GESIMA) (Mengelkamp, 1999)) are typically

bounded by large-scale climatological parameters (e.g.,

reanalysis data) and use a statistical-dynamical

downscaling procedure (Frey-Buness, Heimann, & Sausen,

1995) to calculate the regional wind climate. The two

main international models are the MM5 and the MC2. The

NCAR series of MMx models, particularly the MM5, and

the MSC's MC2 are both non-hydrostatic models, which

became available at about the same time, in the mid-

1990's (Benoit et al., 1997).

The foundation of the Canadian wind atlases is the

MC2 model. It was developed in the mid-1980s at MSC-

RPN (Recherche en Prévision Numérique) and the

Université du Québec à Montréal (UQAM) (EC, 2005c).

Benoit et al.'s (1997) work documents the developmental

research and early results of the MC2. It is governed

by the fully compressible Euler, or Navier-Stokes,

equations, which include the full time derivatives of

all three components of velocity. “It is probably the

only currently available 3D fully compressible model

55


that uses a semi-implicit semi-Lagrangian (SISL) time

scheme” (Benoit et al., 1997 p. 2382).

Over the last decade these models have been used in

numerous bodies of interdisciplinary research (e.g.,

pollutant dispersion studies by Tong, Walton, Sang, &

Chan (2005) and anthropogenic impact studies by Pino,

Vila-Guerau De Arellano, Comeron, & Rocadenbosch (2004)

and Dandou et al. (2005)). These studies have served

to validate the models and to identify their

limitations. The authors emphasise the importance of

accurate boundary conditions and suggest that some of

the inaccuracies may be attributed to very local

effects unresolved by model resolution. Resolution

improvements can be achieved by imbedding a UCL model,

typically referred to as an urban parameterisation

scheme (to be addressed in Section 2.8.2), into the

mesoscale model (Dandou et al., 2005; Tong et al.,

2005) or by using a microscale model. The reader is

directed to the previously cited references for more

in-depth information pertaining to the theory behind

the models.

2.4.2 Microscale Models

While mesoscale models define the regional wind

regimes, the increased horizontal resolution of

56


microscale models can take the effects of complex

surfaces on the local flow field into consideration.

These models require meteorological data, typically

generated by mesoscale models and provided as

microscale model boundary conditions, as well as

geophysical data. The geophysical data are extracted

from digital elevation models (DEM), Shuttle Radar

Topographic Mission (SRTM) data, and/or land cover and

topographic databases.

In the field of wind power meteorology, the

microscale models, specifically developed to assess the

local wind climate, are most often imbedded in the

associated application software (see Section 2.7).

Examples of microscale models include the Riso

Laboratory of Denmark's Wind Atlas Analysis and

Application Program (WAsP) (Petersen et al., 1998b),

Environment Canada's MS-Micro (Walmsley, Taylor, &

Keith, 1986), and WindMAP (based on MC2).

Computational fluid dynamics (CFD) models, or wind flow

solvers, based on Reynolds Averaged Navier-Stokes

(RANS) equations (e.g., WindSim) are also used for

microscale assessment.

MS-Micro, developed through a partnership between

MSC and Zephyr North, specifically analyses steady-

57


state flow over complex terrain. MS-Micro/3, the

latest version of the linear 3D MS3DJH/3R numerical

flow model (Walmsley et al., 1986), employs the two-

layer UBL model theory to calculate roughness-change

induced flow modifications without requiring “that the

wind data be taken upstream of the site of interest”

(Jacobson & Malte, 2005 p. 11). MS-Micro is only

capable of modelling the steady-state and neutrally

stratified boundary layer over gentle mountainous

terrain without flow separation. For more rugged

terrain and non-neutral atmospheric stratification,

other microscale models must be used (Scholtz & Liu,

2005).

2.4.3 Model Combinations

As a precursor to the modern day wind atlas, one of

the first applications of combined models was in

forecasting the weather. In wind power meteorology,

weather forecasting is a critical component of wind

farm production prediction. The flow chart in Figure

2.9 depicts the components of Zephyr/Predictor (Giebel

& Risø National Laboratory, 2006), a wind farm

production-prediction system, combining mesoscale and

microscale models.

58


Figure 2.9. Flow chart of the Riso wind powerprediction model.

The boxed-in regions represent the mesoscale generation of thesurface or regional wind climate (solid line) and the microscale

generation of the local wind climate (dashed line). Modified afterPetersen et al. (1998b p. 68).

One of the main challenges in numerical modelling

comes from having to weigh computational effort (i.e.,

time to model convergence) against accuracy and/or

resolution. By combining NWP, mesoscale, and

microscale models, optimised for a specific scale or

type of complexity, accuracy and resolution can be

retained without adversely affecting convergence. The

integration of these three types of models has not yet

59


been perfected, but software along these lines has been

developed and proven to yield more accurate results

(Section 2.7.2). (Jacobson & Malte, 2005) The wind

atlas methodology (WAM) used to create wind atlases is

an example of just such an integration.

Wind atlases represent the outcome of the first

stage (i.e., the development of the regional wind

climate (Figure 2.9 - solid-line boxed region)) or the

first two stages (i.e., the development of the local

wind climate (Figure 2.9 - dashed-line boxed region)),

as illustrated above.

2.5 The Wind Atlas

Creation of a wind atlas, a systematic and

comprehensive collection of regional wind climates

(RWC) (Mortensen, 2007), is typically based on the

mesoscale-portion of the WAM, as delineated in Figure

2.9 above. As such, differentiation is made between

wind data, a wind atlas, and wind resource assessment

(see Section 2.7) (Landberg et al., 2003).

In general, the methodology extrapolates detailed

information about the mean wind climate from one

location to another. The Risø National Laboratory

proposes two main methodologies: the observational wind

atlas (OWA) methodology, based on data collected at

60


numerous meteorological stations (Mortensen, 2006) and

the numerical wind atlas (NWA) methodology, based on

mesoscale modelling and a reanalysis data set

(Mortensen, 2006a). Though wind mapping began in the

early 1970s, this review will focus on the Canadian

Wind Energy Atlas and the Ontario Wind Resource Atlas,

both based on NWA methodology. A summary table of the

primary features of these and other atlases is provided

in Appendix B.

The Wind Energy Resource Atlas of the United States

(US) was developed in the late 1970s by Pacific

Northwest Laboratory (PNL) in support of the Federal

Wind Energy Program's aim to assess the nation’s wind

resource. Figure 2.10 indicates the twelve regions for

which wind energy atlases were produced using a wide

variety of data and analysis techniques. Each region

is divided into its individual states (Figure 2.10),

while each state is further subdivided into grid cells

as follows: 1/4° latitude by 1/3° longitude (in the

contiguous US), 1/2° latitude by 1° longitude (in

Alaska), and 1/8° latitude by 1/8° longitude (in

Hawaii, Puerto Rico, and the Virgin Islands).

61


Figure 2.10. Wind Energy Resource Atlas of the UnitedStates.

The 12 regions are outlined in bold and the individual states indashed lines (National Renewable Energy Laboratory (NREL), 1986).

The digitised maps provide mean wind speeds and

power densities at heights of 10 m and 50 m (Figure

2.11). Though seasonal mapping was conducted, no

consideration was given to orographic terrain-induced

variability.

62


Figure 2.11. Analyzed Winter Season Average WindResource Maps.

(NREL, 1986)

Detailed maps provide mean wind speed and wind

power density by class at a height of 50 m (Figure 2.12

(a)). Low resolution topographical maps are provided

at both the region and individual state level (Figures

2.12 (b) & (c)). At the individual state level, only

wind power class contour intervals are depicted (Figure

2.12 (c)).

63


(a) Winter wind power classifications

Figure 2.12. Wind resource estimates in the Northwestregion of the US.

In a) shaded regions identify power classifications @ 50 a g.l.while the legend provides the associated wind power density and

speed (NREL, 1986). A dynamic map is available in the form of theNew York Wind Resource Explorer, powered by Wind Resource

Explorer (AWS Truewind, 2004a).

64

(b) Geographic map of Northwest region

(c) New York state average annual wind power density by class


This atlas is one of the few to include wind

resource certainty ratings, to account for the

complexity of the topography and the reliability of the

available data. It also provides data on met station

locations, including seasonal summary data and

anemometer height. Text summaries of average annual

and seasonal resources are provided at the regional

level. This atlas is freely accessible via the

internet (NREL, 1986). In 1983, the maps were revised

as new site data were gathered. Dynamic interaction

with these maps over the internet is provided through

the United States Annual Wind Resource Potential map

(NREL, 1997). More detailed wind maps have been

created for New Jersey, New York, and Ohio, which are

accessible through AWS Truewind's Wind Resource

Explorer (WRE), an on-line interactive map viewer

(AWS Truewind, 2004b).

The European Wind Atlas was first published in 1989

by Risø National Laboratory, based on the Danish Wind

Atlas. It uses data from 220 meteorological stations

to cover the following 12 countries within the European

Union (EU): Belgium, Denmark, France, Germany (FRG),

Greece, Ireland, Italy, Luxembourg, Netherlands,

Portugal, Spain, and United Kingdom. Weibull

parameters of the wind regime are given at five heights

65


for twelve 30-degree sectors over four different

surface roughness values (Petersen et al., 1998b).

Digitised, low resolution maps are available online

(Mortensen, 2005) (Figure 2.13), but the actual Wind

Atlas (text + CD) must be purchased from the Wind

Energy Department at the Risø National Laboratory. A

rather comprehensive review of the development of the

European Wind Atlas is presented in the two-part

article by Petersen et al. (1998b and 1998a).

66


Figure 2.13. The wind resource in Europe at 50 m a.g.l.Wind resources at 50 meters above ground level for five different

topographic conditions (Troen & Petersen, 1989).

The Atlas includes:

European wind climate data from over 200 stations,including comprehensive statistics and a CD of observed and modelled wind direction roses and wind speed distributions for each station;

67


Colour maps of wind resources of each EU country; and

A handbook for regional and site-specific wind resource assessment, including computational procedures.

On purchase of the Wind Atlas Analysis and Application

Program (WAsP) (see Section 2.7.2), the Danish Wind

Atlas is provided free of charge.

The associated website: 'The World of Wind Atlases'

(Mortensen, 2006b), provides a comprehensive global

listing of other wind atlases, as illustrated below,

and related databases, in addition to providing the

coordinates of all the European & Russian

meteorological stations through Google Earth.

Figure 2.14. Worldwide status of the wind atlasmethodology.

The worldwide status is summarised by country using a grey scalekey as follows: (dark grey) national wind atlases exist, (black)WAsP has been applied for regional and local studies; or (lightgrey) no information was available at the time (Landberg et al.,

68

http://www.windatlas.dk/World/Other.html


2003). Since this map was published, Columbia, Pakistan, and SaudiArabia have developed national wind atlases (Mortensen, 2006b).

2.5.1 The Canadian Wind Energy Atlas

The internet-based Canadian Wind Energy Atlas

(CWEA), a contribution from the EOLE research project

hosted by Environment Canada (EC), was inaugurated in

late 2004. It is based on NCAR/NCEP reanalysis data,

the mesoscale portion of the Wind Energy Simulation

Toolkit (WEST) (based on MC2), and the US Geographical

Survey database for orography & land-use data. The

wind atlas map is subdivided into 65 partially

overlapping tiles, based on the primary quadrangles of

the National Topographic System (NTS) of Canada (Figure

2.15). Statistical post-processed data are available

in RPN standard file format (i.e., *.fst) while mean

wind speed and power are also available in MID/MIF file

format, for use by microscale models (e.g., MS-Micro or

WAsP).

69


Figure 2.15. The 65 tiles or quadrangles of theCanadian Wind Energy Atlas.

(EC, 2005a)

This interactive atlas allows users to superimpose

limited topographic content (i.e., cities, roads, power

lines, and rivers and lakes) onto contour region maps

of elevation, roughness length, mean wind speed, and

power density by tile. Higher resolution maps are made

available through print preview functionality. Figure

2.16 illustrates cities and roads superimposed onto the

mean wind speed map of quadrangle 40 within

southwestern Ontario (EC, 2005b; CMC/RPN - Comm Groupe,

2002).

70


Figure 2.16. Canadian Wind Atlas Quadrangle 40 - MeanWind Speed at 30m.

(EC, 2005a)

The atlas also provides a link through which

mesoscale results can be compared to measurements taken

by meteorological stations in the region of interest.

2.5.2 The Ontario Wind Resource Atlas

The Ontario Wind Resource Atlas (OWRA), released by

Ontario's Ministry of Natural Resources (MNR) in

partnership with Helimax Energy Inc. and AWS Truewind

in early 2005, enhances knowledge of the wind resource

within the province of Ontario. It is based on

71


NCAR/NCEP reanalysis data, MC2, and WindMAP. The

orographic data are obtained from Ontario Geospatial

Database DEMs and SRTM data while topographical land-

cover data are obtained from the MNR's Ontario

Provincial Landcover database and the North American

Percentage Tree Cover database (based on the Moderate

Resolution Imaging Spectroradiometer (MODIS) database).

The combination of a mesoscale and microscale model

(i.e., MASS and WindMAP) are used to produce diurnal

and monthly variation, extreme monthly variation (for

July and January), and a 20-year inter-annual variation

wind speed plot.

This interactive atlas allows users to superimpose

towns, roads, contour elevations, power lines, and

political boundaries, onto contour region maps of mean

wind speed and power density. Land-use is presented in

high resolution thematic detail, as illustrated by the

legend associated with Figure 2.17, instead of mapping

representative roughness length, as is done in the

CWEA. The print preview functionality combines maps,

legends, and the overlay filter menu into a

comprehensive image. Figure 2.17 portrays some of the

atlas' features by zooming in on the greater-Hamilton

area.

72

http://www2.jpl.nasa.gov/srtm/


In general, the OWRA offers increased resolution

over the CWEA for the province of Ontario, through

topographical map detail and the temporal data, and is

freely accessible through the Ontario Wind Resource

Atlas website (Ministry of Natural Resources (MNR),

2006b).

(a) Ontario wind speed map @ 50 m agl

(b) The greater Hamilton area

73


(c) Land-use data legend

Figure 2.17. Ontario Wind Resource Atlasdemonstration.

(a) includes contour line and town overlay; (b) includes roads,towns, contour elevations, and land-use data; and (c) is the land-

use legend. All images were created by the OWRA (MNR, 2006a and2006b).

The OWRA was created through full integration of

NWP, mesoscale, and microscale models, unlike the CWEA,

which only utilises the mesoscale-portion of the WAM.

Figure 2.18 serves to illustrate the general flow chart

of the full WAM. Provincial Wind Atlases have also

been created for Quebec (Ressource Naturelles et Faune,

Québec, 2004), Prince Edward Island (PEI), and New

Brunswick (Gasset, 2005).

Surface W ind Statistics

M icroscale m odel e.g. W AsP

Local wind clim ate

orography

land-use

M esoscale m odel e.g. KAM Mroughness

generator

local orography

local roughness

Clim atology

74


Figure 2.18. Wind Atlas Methodology flow chart.Modified after the KAMM (Petersen et al., 1998b).

Since neither the CWEA nor the OWRA produce monthly

wind statistics, a more detailed understanding of the

WAM is required to estimate monthly, local wind

climates from available data.

2.6 Wind Atlas Methodology

In the previous sections, an overview of the modern

methods used to define the wind climate from synoptic

network station and/or large scale reanalysis data sets

was provided, culminating in the development of the

modern day wind atlas. The primary focus has been on

the climatological aspects of wind power meteorology,

with broad reference to key geophysical components.

The limitations of both the CWEA and the OWRA,

concerning spatial and temporal resolution and the need

for powerful workstations to run the underlying

meteorological models, suggest that a more detailed

understanding of the theory behind the WAM is required.

The WAM is the theory behind wind atlas creation

(Troen & Petersen, 1989; Landberg et al., 2003). It

includes fitting time series wind data to a

representative pdf and wind direction rose. At the

heart of the WAM are the techniques involved in

75


transferring a known wind climate from one location to

another. These techniques estimate the regional wind

climate from a known local climate, through the removal

of local-effect influences (Figure 2.19 upward arrow)

and conversely, estimate a local climate from a known

regional, through application of the site-specific

local-effect influences (Figure 2.19 down arrow).

76


Figure 2.19. Wind atlas methodology removal &application of local effects.

Artistic illustrations credited to Søren Rasmussen (Landberg etal., 2003, modified after (Troen & Petersen, 1989)).

This is done by using boundary layer theory to

estimate the mean wind speed profile in response to

77


energetic and dynamic influences of the local

geophysical reality (Figure 2.20).

W ind Speed

Geostrophic

Figure 2.20. PBL Day and night mean wind speedprofiles.

The free atmosphere is considered to be in a state of geostrophicequilibrium. Due to model resolution and complexity, the profiles

do not represent the wind speed profiles in the UCL (i.e.,consider the bottom of the Surface Layer (SL) to be the top of the

UCL). Based on a combination of figures from Stull (1988).

2.6.1 The Mean Wind Speed Profile

The response of the mean wind speed to terrain

roughness-induced friction (mechanical) as a function

of height [z] and dependent on boundary layer stability

(thermal) is estimated by mean wind speed profiles. As

illustrated in Figure 2.20 (above), diurnal variations

have a profound effect on the nature of this profile.

Figure 2.21 stratifies the profiles within a UBL in

relation to a day-time PBL stratification. The use of

78


these profiles, to estimate mean wind speeds at various

heights within the UBL, is the heart of the WAM.

Figure 2.21. Sketch of mean wind speed profile.

is the mean wind speed; , where is the meanhorizontal and the mean vertical wind speed, respectively.

Mean vertical wind speeds are quite small (i.e., mm/s or cm/s). Gis the geostophic wind speed. The surface layer includes the ISL &

the RSL. Modified after Stull (1988).

Extensive research has been conducted regarding the

formulation and validation of these profiles. Within

the surface layer, local free-convection or the Monin-

Obukhov (MO) similarity theory (MOST) is applied to

develop the mean wind speed profile, also called the

velocity shear profile (Verkaik, 2006; Rooney, 2001; J.

Liu & Kotoda, 1997; Petersen et al., 1998a; de Wit,

Stathopoulos, & Wisse, 2002; MacDonald, 2000). The

79


logarithmic or semi-logarithmic profile (also referred

to as the log law) (Weber, 1999) and the power law

(Akpinar & Akpinar, 2005; de Wit et al., 2002; Petersen

et al., 1998a) are used to represent the velocity shear

profile within this region of the UBL. Above the

surface layer, the general approach is to apply the

geostrophic drag law (Landberg & Watson, 1994; Verkaik

& Smits, 2001), as detailed in Figure 2.21.

Within the wind speed profile equations, the

displacement height [d] and roughness length [zo]

represent the primary length scales parameterising

roughness, while friction velocity [u*] is used to

scale the mean wind speed. When the number of unknowns

exceeds the number of available defining equations,

parameterisation is applied through which unknown terms

are approximated by functions of known quantities (for

which equations exist) and/or empirically determined

constants. Scaling is the practice of using

parameterised variables to non-dimensionalise the

variables of interest, facilitating inter-study

comparison. The majority of more recent research has

been focused on defining and estimating these scaling

parameters. Studies using meteorological data from

network stations (de Wit et al., 2002) and land-use

80


maps (Verkaik, 1999) highlight the need for better

means to estimate these parameters.

Having reviewed the existing mesoscale and

microscale models, wind atlases, and the theory behind

the WAM, we now proceed to the heart of wind power

meteorology, namely wind resource assessment.

2.7 Wind Resource Assessment

The heart of wind power meteorology is wind

resource assessment. Wind resource assessment involves

measuring or estimating the wind statistics at a

particular site of interest and determining the

associated energy that could potentially be extracted

by a wind energy conversion system (WECS). In wind

resource assessment various combinations of techniques

and models, as summarised in Table 2.3 (Landberg et

al., 2003), are applied. The choice of technique

and/or model depends on the availability of a wind

atlas, time series wind climate data, land-use and/or

topographical database, hardware/software, etc., the

nature of the study, and time available to conduct the

study.

81


Table 2.3. Wind Resource Estimation Methods.M ETHODS

1 2 3 4 5 6 7 8CO M PO NENTS Folklore 1 Data 2 M CP 3 Global 4 W AM 5 Site 4 M eso 6 Com bo 7m easurem ents

geostrophic winds database

land-use database

orographic database

local roughness & orography

M esoscale m odel

CFD

m icroscale m odel

geostrophic drag law

Statistical m odels

Observation

‘Measurements’ refers to traditional on-site measurements taken atmeteorological stations often located at airports. An example of

a database of geostrophic winds is the one compiled by theNCEP/NCAR and of a land-use database, the Coordination of

Information on the Environment (CORINE) database. Modified after(Landberg et al., 2003 p. 262).

(1) Folklore include hearsay, flagging vegetation, snowaccumulation, smoke stack plume patterns, etc. (2) measurements

only (3) measure-predict-correlate (MCP) (4) data on scaleindicated (5) namely transference of a known wind climate to asite of interest (6) mesoscale model (7) mesoscale + microscale

models.

Landberg et al. (2003) and Petersen et al. (1998a)

provide a comprehensive overview of the different

techniques that can be used to estimate the wind

resource at a site, including the use of the WAM,

geostrophic wind databases, mesoscale and microscale

models, measured data, and statistical models. Wind

82


resource assessment studies have been conducted at

numerous locations throughout the world (e.g., Basel by

Rotach et al. (2005); Waterloo, Canada by Li & Li

(2005a); Turkey by Sahin, Bilgili, & Akilli (2005);

Taiwan by T. Chang et al. (2003); Vashon Island, USA by

Jacobson & Malte (2005); and Grenada by Weisser & Foxon

(2003)).

Packaging mesoscale wind atlases and microscale

wind flow models together, through a graphical user

interface (GUI) for PC-based systems, produced an

assortment of customised software for wind resource

assessment (e.g., WAsP and WindPro). The majority of

the software that has been developed is in support of

'conventional citing' (i.e., wind farm development).

Additional simplified tools exist for sizing and

selecting stand-alone conventional wind turbines (i.e.,

3 blades, horizontal-axis, and tower-mounted). The

general form of resource assessment software can be

depicted as shown in Figure 2.22.

83


M eteorological Database

M icroscale W ind Clim ate Database

M esoscale W ind Database

Terrain Database

DEM & Land-use

Global Clim ate Database e.g.

NCEP Reanalysis

M esoscale W ind Atlas

M icroscale M odel

M esoscale M odel

Surface Property Generator

Statistics M odule

Statistics M odule

Setup

Initialisation

Classification Schem es

Setup

Figure 2.22. General form of packaged wind resourceassessment software.

Modified after the WEST Flowchart (Pinard, Benoit, & Yu, 2005).

Wind resource assessment ultimately comes down to

determining how much wind energy is available to be

extracted at the site of interest. Wind atlases

typically present this information in the form of wind

power density (WPD) maps.

84


2.7.1 Commercial Software

Table 2.4 provides a summary of some of the most

popular software currently in use for both micro-siting

(i.e., placement of individual WECSs and wind farm

development). Details pertaining to climatological and

geographical (land-use, orography, and roughness)

inputs required by the incorporated mesoscale and

microscale models are not included. The full

capability of either model is not necessarily utilised

within packaged offerings. Accuracy levels are not

presented as there is too much variation, or no

information, pertaining to the various models and

packages in the literature. Ancillary wind data

analysis software (e.g., Windographer and WAsP -

Observable Wind Climate (OWC) Wizard) is also available

for creating wind speed histograms and wind direction

roses.

85


Table 2.4. Commercial Wind Resource Assessment

Software.

Software(Developer) Cost 1 Platform

Model OutputApplic. 6Meso Micro

WindMaps Assessment

U WPD E $HOMER 10 (Mistaya Engineering

FREEWIN 95/98/ME/2000/ XP

EconomicAssessment

Turbine

Renewables DSS

Renewable Energy Calculator 11 (CT)

FREEWIN 95/98/ME/2000/ XP

EconomicAssessment

Turbine

RETScreen (CANMET) FREE

WIN98/NT/2000/ XP

EconomicAssessment

Turbine

Anemoscope 8 (EC&NRC)

$CDN10,000

2

WIN XP ++

WindScope + EnSim3 Wind

Wind Resource Assessment

- WindScope(WEST) ŧ

UNIX /Linux

MC2(CWEA)

MS-Micro/3

Wind

Wind Resource Explorer™ ŧ

Online Viewer Mesomap SiteWind™ Wind

- AWSTruewind

Mesomap Proprietary MASS WindMap Wind

- SiteWind™7 Proprietary MASS MSFD 9

wind flow Wind

WAsP (Reso) 2950 €WIN

98/ME/2000/XP

WindAtlas

WAsPWind flow

Turbine

Turbine(s)

Micrositi

WindSim (WindSim AS)

8250 €WIN

2000/XP++

PHOENICSwind flow

5

Wind Resource

Turbine &Farm

WindPro (EMD)

1500 €12

WIN 98/ME/NT/2000/XP

Modules Turbine &Farm

Wind FarmDesignGH

WindFarmer (GH&P)

7000 €4

WIN2000/XP

+

Modules Farm

86


WindFarm (ReSoft) 4500 €

WIN 98/NT/2000/

XP MS-

Micro/3 Farm ~

Table Notes:

(1) Annual license fee unless indicated otherwise (2) One timecost (3) Environmental Simulation (EnSim) environment - user

interface for the post-processing model results (4) Base moduleonly (5) Phoenics is a CFD Solver (6) Primary application ofsoftware (7) Improves accuracy of Mesomap by incorporating

available site measurements (8) AnemoScope (formerly WindScope(formerly WEST) - first version to be windows-based PC

compatible)) (9) Mixed Spectral Finite Difference (MSFD) (10)Hybrid Optimization Model for Electric Renewables (HOMER) - also

optimises stand-alone configurations of various alternative energytechnologies, grid connection, load assessment, etc ..very

comprehensive (11) Canadian Tire (CT) system building and costingtool - limited to solar arrays and wind turbines sold by CT (ŧ)GIS compatible viewer (+) Pentium IV, 1 GHz (++) Pentium IV, 2

GHz, 512 MB RAM (12) Minimum, only includes Basis + Atlas; 2750 €for farm analysis (incl. PARK).

The software shown in Table 2.4 was, for the most

part, developed within the last decade and founded on

the WAM. The limitation of this software is that it is

primarily for conventional wind turbine siting. At

best, it provides average monthly wind power densities

at the top of the RSL, including the effects of

aerodynamic amplification induced by orographic terrain

elements. Buildings, for the most part, are viewed as

groupings of roughness elements and/or individual

obstacles. This is not surprising since published

siting guidelines for individual turbines and wind farm

location criteria (WFLC) (Baban & Parry, 2001)

recommend such things as unobstructed open areas, in

87


the path of prevailing winds, and preferably near the

crest of a hill or ridge (Canadian Wind Energy

Association, 2005). As models led to model

combinations, the development of commercial software is

prompting the evolution of integrated modelling

systems, which combine commercial software, specialised

microscale models, and GIS-functionality (e.g., WINDA

(Blennow & Sallnäs, 2004)).

In wind power meteorology, no attempt appears to

have been made to estimate the potential aerodynamic

amplification such obstacles may have on the flow;

amplification assessment seems to be limited to

orographic terrain and yet terrain terminology is being

carried over to the urban canopy (e.g., street canyon).

This research undertaking proposes that the

building-wind interaction produces aerodynamic

amplification that increases the wind speed in areas

within the UCL to at least levels experienced aloft.

Providing a way to estimate these winds would be of

great benefit to urban-scale WECS developers, urban

planners, architects, etc. To develop the theory

behind this concept, further understanding of the UCL

and the associated parameterisation schemes is

required.

88


2.8 The Built Environment

In boundary layer modelling, the built or man-made

environment is represented by the UCL. Up to this

point, representation of the influence of an urban area

has generally been limited to its retardation effect on

the wind flow, as defined by the mean wind speed

profile and restricted in applicability to the region

above the UCL.

2.8.1 Urban Canopy Layer

The primary components within the urban canopy are

the buildings and the street canyons. Combined, these

components produce the urban form or morphology (i.e.,

the layout, density, shape, size, and orientation of

buildings and streets within the city). UCL modelling

simplifies the urban morphology by equating it to an

array of obstacles. Plate & Kiefer (2001) after

(Theurer, 1993) proposed that urban complexes could be

categorised into three basic patterns (i.e.,

homogeneous row, heterogeneous rectangular grid, and

real city block).

Wind resource assessment, for the most part, treats

urban terrain as a collection of roughness elements

(buildings) parameterised by roughness length [zo] and

an average height [zH]. These parameters are then used

89


to estimate the wind speed velocity profile above the

UCL using wind speed measurements typically available

from meteorological stations within. At the

microscale, buildings are considered as obstacles,

which impede or significantly reduce the wind speed

(e.g., WAsP uses the dimensions, position, and porosity

of each obstacle to perform shelter modelling (Petersen

et al., 1998b)). This traditional approach fails to

reproduce the vertical structure of the flow

perturbations observed within an urban canopy (Cai,

2000), which is primarily governed by street canyon

energetics and dynamics (Rotach et al., 2005) (i.e.,

sensible heat flux [QH] (thermal) and turbulent kinetic

energy (TKE) (dynamic) (Martilli et al., 2002)). In

modelling the UCL, both horizontal and vertical

resolution must be increased. The use of higher-order

closure scheme CFD models is recommended (Dandou et

al., 2005) to account for mechanical turbulence

generated by airflow over and around buildings (Hanna &

Chang, 1992) and shadowing and street canyon radiation-

trapping (Hildebrand & Ackerman, 1984; Roulet, 2004).

The higher resolution and complexity in the UCL is

modelled using boundary layer wind tunnel (BLWT)

modelling (Plate, 1999; C. A. Miller & Davenport, 1998;

Maruyama & Ishizaki, 1988; Grant, Heisler, Gordon, &

90


Herrington, 1988; Isyumov, 1978), computational fluid

dynamics (CFD) (Dayan, 2006), and mesoscale model-

imbedded urban parameterisation schemes (MacDonald,

2000; Martilli et al., 2002; Grosso, 1998). The

literature also includes research on enhancing existing

IBL (Landberg & Watson, 1994; Verkaik, 1999; Belcher et

al., 2003) and two-layer models (Clarke & Hess, 1974;

Verkaik, 1999; Verkaik, 2006) to more accurately

portray the urban RSL (MacDonald, 2000).

Research pertaining to UCL modelling and urban

parameterisation is being conducted in support of

erosion studies (Grant et al., 1988), air quality /

pollutant dispersion investigations (Fisher et al.,

2004; Rotach et al., 2005; Roulet, 2004; Christen,

Vogt, & Rotach, 2003; Berg, De Wekker, Shaw, Coulter, &

Allwine, 2004), pedestrian comfort and safety (Isyumov,

1978; Plate, 1999) and urban planning (Eliasson, 2000;

de Schiller & Evans, 1998; Bitan, 1992). This

literature can be broadly categorised as either

concerned with flow within and above the street

canyons, or with exchange processes between. A

comprehensive review of the last two years of

multidisciplinary and international urban climate

research has been compiled by Souch & Grimmond (2006).

91


The challenge in modelling the UCL comes from the

need to average the mean and turbulent components of

velocity over both space and time, using RANS

equations, to evaluate the spatial and temporal

complexities of the flow field (Coceal & Belcher,

2005). The resulting momentum equations introduce new

terms that need to be parameterised. Of primary

significance is the parameterisation of elemental

canopy drag [Di] (Coceal & Belcher, 2005). Recent

urban-climate, scale-modelling developments have

focussed on turbulence, dispersion, and radiation

balance, and the need to compliment these findings with

both numerical and field observations. Findings from

both indoor and outdoor experiments indicate that

further research is required regarding the analysis of

thermally-stratified conditions and additional data

need to be provided for basic evaluation of the energy

balance (Kanda, 2006).

2.8.2 Urban Parameterisation Schemes

The theory behind UCL parameterisation (Souch &

Grimmond, 2006; Rotach et al., 2005) is founded on the

parameterisation of plant canopies (MacDonald, 2000;

Belcher et al., 2003; Coceal & Belcher, 2005; Cionco &

Ellefsen, 1998). Early studies on pesticide dispersion

92


in crop fields laid the foundation for pollutant

dispersion studies within the UCL (Belcher et al.,

2003). Urban parameterisation schemes (e.g., the

porosity-drag approach (Dandou et al., 2005 after

(Piringer, 2002)) attempt to capture the significant

features of building and street canyon geometry using

appropriate scaling factors, or parameters, which

account for the thermal (Martilli, 2002) and mechanical

(Martilli et al., 2002) effects of the urban morphology

on the flow, within the fundamental conservation

equations. The primary scaling parameters applicable

within the UCL are the turbulence length [lc], the drag

length [Lc] or friction coefficient [] (MacDonald,

2000), and the horizontal adjustment distance length

scale or fetch [xo] (Coceal & Belcher, 2005).

Extensive research has been conducted defining and

developing these scaling parameters (Coceal & Belcher,

2005; Belcher et al., 2003) and mapping and/or

categorising urban morphology (Souch & Grimmond, 2006;

Fisher-Gewirtzman, Pinsly, Wagner, & Burt, 2005; C. A.

Miller & Davenport, 1998; Ellefsen, 1991; Grosso,

1998).

Related research is being conducted to define urban

terrain zones (Ellefsen, 1991) based on correlating

morphological values (e.g., plan area density (pad)

93


[p] and frontal area density (fad) [f]) to land-use

classification. Application of LIDAR (Souch &

Grimmond, 2006) and GIS-assisted data-extraction from

DEMs (e.g., WINDA (Blennow & Sallnäs, 2004); Tanaka &

Moriyama, 2005; Bishop, 1998) have also been

investigated, as illustrated below.

Figure 2.23. Digital still video aerial imageryextraction techniques.

(a) A section of original image; (b) enhanced orthoimage; (c)digital surface model (DSM); (d) DSM minus digital terrain map(DTM); (e) classified shadows; (f) shack hypotheses; (g) ground

truth data (Bishop, 1998).

2.8.3 Mean Wind Speed Profile

The most important urban effects on the airflow are

summarised as follows:

94


Pressure (form) drag, due to building walls; Friction (viscous) drag, due to street canyon

floors and roofs; Wake diffusion, due to the eddy structure within

turbulent wakes; Mean kinetic energy conversion into TKE in the

intense shear layer at the top of the UCL; and Heat flux modifications, due to shadowing,

radiation trapping, and building heat storage.(Roulet, 2004; Martilli et al., 2002; Roth, 2000)

The first four phenomena are mechanical processes,

while the last, which can lead to the generation of the

UHI effect, is due to thermal processes. Wake

diffusion entails both the mixing and diffusion of

momentum, heat, moisture, and pollutants.

Given the complex effects of urban morphology on

the flow within the UCL and the difficulties in

developing a unified parameterisation scheme, simple

estimations of the wind speed profile are obviously

problematic. In the literature, a simplified

exponential profile has been developed (MacDonald,

2000), which seems to be reserved for qualitative

comparisons vs. quantitative assessment of the actual

wind profile within the UCL (Coceal & Belcher, 2005).

Three-dimensional numerical modelling results

suggest that the urban canopy layer impacts both upwind

95


and downwind flows (Belcher et al., 2003). The

disturbance, or IBL, is roughly the same as the mean

building height in a homogeneous canopy but can extend

to twice that height in heterogeneous canopies and/or

rough terrain (Grosso, 1998).

Ground level wind accelerations are caused by the

vertical pressure gradients that can be created by

taller buildings within the canopy (Coceal & Belcher,

2005) and building aerodynamics. Research attempting

to characterise street canyon flows suggests that three

flow regimes can develop (i.e., the isolated roughness

(f<0.13), wake interference ((0.13<f<0.25), and

skimming flow regimes (f>0.25) (Plate & Kiefer, 2001;

Xie, Huang, & Wang, 2006; MacDonald, 2000 after (Oke,

1992)). These are typified by turbulent, and/or

recirculating flow fields.

2.8.4 Building Aerodynamics

In the previous section urban parameterisation

schemes were discussed, which develop an estimate of

the mean wind speed profile within the UCL, using

scaling parameters based on the underlying urban

morphology. Homogenous zones within the heterogeneous

UCL were defined using key morphological values and

parameters, and the effect of transitioning between

96


zones was explained as primarily governed by the fetch

of each zone.

Building aerodynamics research augments urban

parameterisation schemes by increasing the complexity

of the simplified cubic-array type representation of

the UCL, to include the effects of individual buildings

and their architectural details (e.g., roof shape).

The literature can be broadly categorised by angle of

flow incidence []. Orthogonal flow is seen as

producing Rankine vortex structures within street

canyons and horseshoe vortex structures in the wake of

buildings (Cook, 1985). Parallel flow can produce what

is generally referred to as channelling, (i.e., street

canyon and building geometry-induced flow amplification

similar to the orographic terrain induced speed-ups

previously discussed), while oblique or skewed flow is

characterised by complex vortex structures, including

delta-wind vortices (Zhang, Jiang, & Hu, 2006; Cook,

1990). The majority of the literature pertains to

orthogonal flow studies, including assessment of 2D vs.

3D modelling (Choi & Britter, 2004) and the influence

of roof configuration (Xie, Huang, & Wang, 2005).

Parallel and oblique flow studies range from simple

single canyon channelling (Skote, Sandberg, Westerberg,

Claesson, & Johansson, 2005) to complex networks of

97


buildings and streets (P. J. Jones, Alexander, &

Burnett, 2004), including the effects of passages

through buildings (Blocken & Carmeliet, 2004),

interpreted as the venturi-effect (Bottema, 1999).

Field data collection to map out the complexities

of the flow field involving both temporal and spatial

variation is an onerous undertaking. The need to place

numerous instruments in various locations and heights

and difficulties in measuring direction result in

research in this field having to rely on a combination

of numerical methods and boundary layer wind tunnel

(BLWT) studies. Numerous studies have been conducted

comparing the results obtained using CFD models against

those produced in a BLWT (P. J. Jones et al., 2004;

Dalgliesh & Surry, 2003; Blocken & Carmeliet, 2004).

In BLWT studies point methods (e.g., hot-wire

anemometer or pressure sensor measurements), area

methods (Blocken & Carmeliet, 2004) (e.g. erosion or

scouring of a seeding material (P. J. Jones et al.,

2004)) and/or smoke visualisation (Blocken & Carmeliet,

2004; Guirguis, Hanna, Kotkata, & Gad, 1998) are used

to assess a scaled version of the generated flow field.

To apply BLWT results all of the significant processes

have to be scaled in a consistent fashion so that their

interaction reflects what would be observed in the real

98


atmosphere (Xie et al., 2005). One of the main

challenges in using BLWT methods to study pollutant

dispersion and pedestrian comfort is that low wind

speeds are difficult to maintain (Choi & Britter, 2004;

Xie et al., 2005). The earliest BLWT building

aerodynamic studies were primarily interested in wind

loading, defined by pressure distributions, and dynamic

effects on the building structure. Pedestrian-level

wind studies did not begin until the late 1960s

(Blocken & Carmeliet, 2004).

Numerical methods primarily rely on CFD turbulence

models. These models can be classified as classical or

large eddy simulation (LES). The classical model is

based on RANS equations (e.g., the k– closure model),

which is by far the most used and validated (Zhang et

al., 2006). LES models (Cai, 2000) are based on the

subgrid scale (SGS) filtering model (Murakami, Ooka,

Mochida, Yoshida, & Sangjin, 1999), which is

computationally very demanding and mainly used in

research applications (Xie et al., 2005). At present,

there is some concern regarding the extent and location

of the wind amplification areas (P. J. Jones et al.,

2004) and turbulence structures, in general, generated

by CFD models (Zhang et al., 2006; Gratia, Reiter, & De

Herde, 2005), but results have been found to be

99


sufficiently accurate in the presence of steady winds

(Riley, Gadgil, & Bazaroff, 1996). CFD, validated

through wind tunnel testing, is being applied in

dispersion or air quality studies (Zhang et al., 2006;

Xie et al., 2006; Xie et al., 2005), for pedestrian

comfort (C. Chang, 2006) and natural ventilation

(Gratia et al., 2005) assessments and urban planning

(Jun, 2005). These fields are actually at odds with

each other, since ideal conditions for pollutant

dispersion and natural ventilation often result in

pedestrian discomfort (Chan, So, & Samad, 2001).

Computational wind engineering (CWE), the application

of CFD in the field of wind engineering, is seen as a

relatively new undertaking (Murakami et al., 1999).

At the mesoscale, or perhaps even local level, the

orthogonal flow has been categorised into the following

three regimes: skimming flow (SF), isolated roughness

flow (IRF), and wake interference flow (WIF), as

illustrated below (Xie et al., 2006).

100


Figure 2.24. Flow regimes within the UCL.The dashed-lines represent the wake development. As the streetcanyon width [W] is reduced, the flow regime (characterised by

wake development) transitions from isolated roughness flow (IRF)to wake interference flow (WIF) at W 3H, to finally produce theskimming flow (SF) regime at W 1.5H (Gratia et al., 2005 after

(Hussain & Lee, 1988)), where H is the representative buildingheight.

Given truly isolated buildings in an otherwise

homogeneous terrain (i.e., W > 50H (Sini, Anquetin, &

Mestayer, 1996)), the resulting flow regime is known as

fully independent wake flow (FIWF), wherein classic

vortex shedding occurs. This region does not typically

develop within an UCL due to plan area density. At the

other end of the scale where W <<1.5H, it becomes

evident that the effect of increased areal density has

a limit beyond which further increases have little

effect (Mfula, Kukadia, Griffiths, & Hall, 2005).

101


A wake is the region of flow produced by a

perturbation (e.g., drag) characterised by unstructured

and/or structured turbulence commonly referred to as

eddies and vortices, respectively. van Bussel &

Mertens (2005) used CFD to validate the empirical

equation for the semi-elliptical wake zone proposed by

Wilson (1979). Bottema (1999) used the geometric

influence scale parameter [Lg] (Cook, 1990) to

establish frontal (1Lg) and leeward (4Lg) wake region

lengths (Bottema, 1993) to develop a 2D representation

of the wake region. It is within a relatively small

region of the wake, seen in 3D as having a similar

horizontal profile at the vertical edges of buildings

to that produced vertically at horizontal edges, that

flow amplification is observed (van Bussel & Mertens,

2005). Defining the thresh-hold aspect ratios of each

flow regime is crucial to the development of a

comprehensive description of the flow above, and the

recirculation structures within, street canyons. The

parameters of interest are the geometric aspect ratios,

primarily the street, canyon, or side aspect ratio

(sar) [W/H] and the frontal aspect ratio (far) [L/H],

where L is the length of the canyon normal to the wind

direction. Variation in how these aspect ratios are

defined makes it difficult to compare study findings.

102


Wind tunnel model aspect ratios based on street canyon

geometry (Sini et al., 1996) produce opposite far and

sar definitions from those based on building geometry

(Hussain & Lee, 1980). Other studies create additional

confusion by defining the reciprocals of these ratios

(e.g., H/W (Xie et al., 2006)).

At the microscale within each of the flow regimes,

further stratification has been defined based on the

street canyon vortex structure, as illustrated in

Figure 2.25. Vortex structure categorisation pertains

to orthogonal flow and is primarily based on direction

of rotation and number of vortices within a given

street canyon.

(a) Vortex Structure I(a)

(b) Vortex Structure I (b)

103


(c) Vortex Structure II

(d) Vortex Structure III

Figure 2.25. Vortex structure categorisation.The associated canyon and building aspect ratios are (a) W/H1=0.29& H1/H2=1 (b) W/H1 = 0.5 & H1/H2 = 1 (c) W/H1 = 1.11 & H1/H2 = 0.9

(d) W/H1 = 5.88 & H1/H2 = 1 (Xie et al., 2006)

Additional research involving buildings of varying

heights has explored the implications of cross-canyon

building height aspect ratio (Xie et al., 2006) or

relative height ratio [H1/H2] (Chan et al., 2001) on

vortex structure development as illustrated, where H1

and H2 are the upwind and downwind building heights,

respectively.

104


Figure 2.26. Effect of varying building height oncanyon vortex structure.

The proportion of four buildings is figured to the left of eachconfiguration (Xie et al., 2005 p. 4527)

Detailed investigation of canyon vortex structures

has been conducted through LES, highlighting the

influence of flow direction (Zhang, Weimei Jiang, &

Miao, 2006), secondary flows (Macfarlane & Joubert,

1998), and roof configuration (Souch & Grimmond, 2006)

(Figure 2.27). Louka, Belcher, & Harrison (2000)

suggest that structured vortices are actually very

short lived and weaker than unsteady turbulent

105


fluctuations due to the unsteady position of the shear

layer. Worthy of further exploration are the

inherently higher wind speeds at the radius of the

forced-vortex core of these vortices. It is suggested

that most of the vortices experienced in the field of

wind engineering can be represented by the Rankine

vortex, within which the core is surrounded by a free-

vortex outer region (Cook, 1985). There is very little

in the literature regarding delta-wing vortex

development, as it pertains to skewed flow angles whose

investigation is under-represented.

106


Figure 2.27. Effect of roof configuration on canyonvortex structure.

(Xie et al., 2005 p. 4523)

Figure 2.28 below, summarises the cross-canyon wind

speed variation as observed in each of the numbered

configurations illustrated above.(a) Configuration 1 through 4 (b) Configuration 1, and 5

through 7

107


Figure 2.28. Cross-canyon horizontal mean wind speedvariation.

Situations relate to the configurations illustrated in Figure2.27. Dimensionless cross-canyon distance [X/B], where B is the

canyon width and X is the cross-canyon horizontal distance.Negative velocities correspond to clockwise rotating vortices at

street-level. Extracted from Xie et al. (2005) p. 4525.

The literature pertaining to oblique flow studies

primarily focuses on architectural details (Jamieson,

Carpenter, & Cenek, 1992) (e.g., pilotis or support

columns, passages (Isyumov. & Davenport, 1977), and

setbacks (Yamamura & Kondo, 1993)) and street canyon-

induced wind amplification (P. J. Jones et al., 2004;

Bottema, 1999). The majority pertains to the effect of

the presence of a tall building on pedestrian-level

winds (Yamamura & Kondo, 1993). Explicit formulation

of wind amplification in terms of aspect ratios is

difficult due to complexities and non-linearities. For

the most part, oblique flow research has produced more

qualitative than quantitative findings. These findings

108


categorise various effects by architectural detail

(e.g., venturi, gap, channel, tower, and shelter

effects (Gandemer, 1977)), provide illustrative plots

based on aspect ratios (Chan et al., 2001), and/or

suggest general rules of thumb (Bottema, 1999).

In the field of wind engineering, concerned with

wind loading on structures, decades worth of BLWT study

findings have been summarised in the form of non-

dimensional pressure coefficient tables (Canadian

Commission on Building and Fire Codes, 2005). These

pressure coefficients are dependent on the flow regime

(i.e., mean wind speed and angle of incidence), UCL

parameters, and the geometrical aspect ratios of the

built environment (Gratia et al., 2005; Canadian

Commission on Building and Fire Codes, 2006). In

Canada, these data are published in the National

Building Code (NBC), along with recommended correction

factors intended to assist designers in calculating

mean pressures and forces on built structures and their

components for safety and durability considerations.

The wealth of data available from this field has

only recently been tapped by other fields (e.g.,

natural ventilation (Hussain & Lee, 1980) and

contaminant infiltration (Riley et al., 1996)).

109


Literature on the use of pressure coefficients to

determine local wind velocities, includes correlating

surface pressure fields to the flow regime (Hussain &

Lee, 1980) and Grosso's (1992) parametric pressure

distribution model. The field of wind power

meteorology should further explore the applicability of

these data in the determination of wind amplification

zones in support of truly urban-scale BAWTs.

Under closer scrutiny, there is some concern

regarding the transferability of these pressure

coefficients. One of the primary challenges is that

the intended reference (or design) wind speed used in

calculating the pressure and/or force on the structure

and/or its elements, is a statistically derived maximum

gust. Other challenges are presented through the

various correction factors. Hanzlik, Diniz, Grazini,

Grigoriu, & Simiu (2005) question the neglect of the

influence of building orientation within ASCE 7

'Minimum Design Loads for Buildings and Other

Structures', published by the American Society of Civil

Engineers (ASCE) and referenced by the International

Building Code (IBC), published by the International

Code Council (ICC). Correction factors, in general,

are applied only to a design wind speed while the

pressure coefficients used to calculate the wind-

110


induced pressures and forces remain based on the

results of tests on isolated bodies under specific BLWT

flow conditions (Hussain & Lee, 1980). Based on this

neglect of the influence of neighbouring structures,

Plate & Kiefer (2001) state that standard wind tunnel

results cannot be used in the assessment of urban

areas. The reader is referred to Plate & Kiefer’s

(2001) article for an excellently illustrated and

explained summary of building code correction factors

and BLWT conditions in boundary layer meteorology

terms.

Similar to the notion of codifying diffusion

processes presented by Plate & Kiefer (2001) and the

already codified wind forces/pressures, this research

undertaking proposes the need for codification of

building-induced flow amplification. Most studies are

concerned with forces and pressures, the velocity

field, or the pollutant concentration field, and not

with the relation of the three to each other within the

UCL. More specifically, there appears to be very

little literature pertaining to a correlation between

forces or pressures and mean wind velocities.

In wind power meteorology input, analysis, and

storage of wind, land-use, and urban parameterisation

111


data suggest that a structured data management and

decision support system would be beneficial. As such,

a review of literature pertaining to various systems

currently used in support of renewable energy

assessment was conducted.

2.9 Decision Support Systems

In the literature, a computer-based structured

presentation of a combination of analytical tools

accessing pertinent data in support of making a

decision is generally called a decision support system

(DSS). At the heart of every DSS is the ability to

collect, analyse, and report in a fashion conducive to

making a decision regarding the matter at hand, as

illustrated below for the Wind Load Design Environment

(WiLDE) application.

112


Figure 2.29. The flow chart of the DAD-based WiLDEapplication.

(Whalen, Sadek, & Simiu, 2002)

Numerous DSS approaches have been developed and

explored, which can be broadly categorised based on

data input methodology, the required knowledge-level of

the end-user, and whether the reporting is qualitative

or quantitative. Of primary interest to this research

undertaking are database-assisted design (DAD) models

(Whalen et al., 2002), which obtain data required in

the analysis phase from an integral database. Given

113


the multi-criteria analysis required in this

undertaking, expert systems, based on rules of thumb

and guidelines (Kaminaris, Tsoutsos, Agoris, & Machias,

2006), were also explored. These are often coupled

with qualitative modelling using fuzzy-logic (Rylatt,

Gadsden, & Lomas, 2001), which is seen as justified in

analysis involving high levels of uncertainty (Germano

& Roulet, 2006, after (Paelinck, 1978).

The majority of the DSSs involving assessment of

the interaction between the built environment and

climatological processes have been developed in support

of sustainable urban planning and development (Tweed &

Jones, 2000). Primarily qualitative in nature,

planning support systems (PSSs) (Bishop, 1998), spatial

decision support systems (SDSSs) (Ramachandra, Krishna,

& Shruthi, 2004), and urban planning simulations and

decision support technologies (O’Looney, 2001) produce

guidelines and rules of thumb (de Schiller & Evans,

1998). In the field of passive and low energy

architecture (PLEA), encompassing bioclimatic design

principles and green building design optimization

algorithms (Wang, Rivard, & Zmeureanu, 2005), pre-

design guidelines for passive solar architecture (Rabah

& Mito, 2003) and the concept of an energy and

environmental prediction (EEP) model (Tweed & Jones,

114


2000) have been developed. In the field of wind

engineering the Wind Load Design Environment for Low-

Rise Structures (WiLDE-LRS) has been developed to

provide design guidance by improving on the empirical

nature of the data presented in the building codes

(Whalen et al., 2002).

Energy planning (EP) DSSs focus on the impact of

urban planning concepts on renewable energy potential

(e.g., solar energy planning (SEP) (Rylatt et al.,

2001; Ramachandra, 2006) and biomass energy potential

assessment (BEPA) (Ramachandra et al., 2004)).

Renewable energy source (RES) DSSs (Bishop, 1998) in

support of EP are generally concerned with defining the

technological potential (i.e., the performance of the

renewable energy extraction device) based on the

theoretical climatic potential subject to socio-

economic restraints (Voivontas, Assimacopoulos,

Mourelatos, & Corominas, 1998). With the advent of

commercially available software, existing DSSs

specifically in support of wind energy assessment have,

for the most part, focused on high-level socio-economic

aspects. These assessments involve determination of

the financial and legal aspects of connecting wind

farms to the electrical grid (Kaminaris et al., 2006)

115


and using GIS to define exclusion zones, primarily

typified by urban development (Voivontas et al., 1998).

The key components of a general RES DSS are the:

user interface; data storage, analysis, and mapping capabilities

(e.g., GIS-based (Voivontas et al., 1998)); purpose-built modules or procedures (e.g.,

domestic energy model, BREDEM-8 (Rylatt et al., 2001));

realistic, real-time, interactive visualization (e.g., GIS, GPS, and virtual reality (Bishop, 1998)); and

summary reports.

Figure 2.30 illustrates the interaction between these

components in the form of a solar energy DSS.

The database is typically comprised of land-use,

topographical, and climatic data. Visualisation

includes bioclimatic and orientation diagrams, which

combine architectural and climatic data to identify

what have been called control potential zones (CPZs).

Mahoney tables have also been used to guide design in

relation to climate (Rabah & Mito, 2003).

The following key considerations come to bear in

the development of an effective RES DSS:

source of data (e.g., topographical, land-use, andclimatic);

116


data collection methodology (e.g., database-assisted and/or user-input);

level of required technical expertise; integration of analysis modules; adaptability and re-use of algorithms (Wang et

al., 2005); accessibility & effort-level vs. accuracy; and ease of use.

117


Figure 2.30. Solar energy decision support system.(Ramachandra, 2006)

Task-specific DSSs are being advocated as opposed

to the creation of an all-encompassing, monolithic

system (Bishop, 1998), which may be impossible to

implement and/or yield difficult to interpret results

(Tweed & Jones, 2000). In design, which is primarily

an iterative undertaking, simplicity of input,

processing, and output are vital (Whalen et al., 2002).

2.10Literature Review Summary

This literature review has attempted to organise

the wealth of transferable, interdisciplinary knowledge

in support of the development of an urban-scale RES

DSS. Given that wind power meteorology is a relatively

new field, solutions to perceived limitations

conceivably reside in related disciplines. Though

plagued by inconsistencies, areas of outright dispute,

and serious limitations regarding a quantitative

parameterised assessment of building-induced wind

amplification, there is sufficient material within the

published literature to conceptualise a stratified

approach to the assessment of urban-scale wind energy

potential.

118


An astounding amount of transferable research has

already been conducted in other fields. As an

interesting example, acid rain studies are producing

the urban terrain characterisation-zone maps required

by UCL modellers (Ellefsen, 1991). In the wake of a

growing awareness regarding the impact on human

comfort, air quality, and energy consumption (Eliasson,

2000) that urban architecture can have, there is a

movement underfoot to bridge the gap between scientific

understanding of climatic processes and the associated

impact of design practices (Wisse, 1988). A myriad of

available software, the relatively recent refocus on

bioclimatic design, and the green building movement all

support the need for DSSs linking urban planning,

architectural detail, and their associated climatic

implications. The prevalence of such systems in

support of urban-scale solar energy planning (SEP)

calls for augmentation by an associated urban-scale

wind energy planning (WEP) DSS, given the intrinsic

connection between solar and wind energy potential.

The primary challenge in this research undertaking

will be to assimilate what has been accomplished to

date in the field of wind power meteorology and augment

it with a more detailed appraisal of building-induced

acceleration zones. The uncertainties involving key

119


parameters (e.g., roughness length), wind speed

profiles, and pressure coefficients (potentially

providing a summarisation of building-induced wind

amplification) will need to be resolved, in order to

develop a comprehensive assessment of these wind energy

potential micro-zones. Land-use data, especially

details pertaining to building height, will need to be

gathered and/or generated. It is primarily the

stratified, multi-disciplinary approach to wind field

assessment, within the existing literature, that is

providing the greatest challenge. For example, at the

microscale, drag-based effective roughness length is

not correlated to mesoscale roughness maps. A

generalised approach will need to be developed to

generate the pertinent variables based on available

land-use data to calculate the parameters necessary to

estimate the wind speed profiles. For the most part,

wind speed profile equations and probability

distributions, require an iterative approach to solve

and fit, respectively. Careful consideration of

accuracy vs. simplicity and ease of use will be

required in the development of the proposed DSS.

Given the focus of recent research on validating

numerical models based on empirical and theoretical

equations, one must delve back over a decade to gain a

120


better understanding of the underlying transferable

scientific principles. In general, the literature is

severely lacking with regard to structure-induced wind

amplification, primarily focussing on orthogonal flow-

induced canyon vortex structures. Finally, there is

some concern regarding the compounding of inaccuracies

(i.e., using a probability distribution function

estimation of the mean wind speed to develop the mean

wind speed profile using a logarithmic profile, which

is then used as the reference velocity for building-

induced amplification assessment). Careful

consideration of unavoidable inaccuracies will be

required to calculate appropriate tolerance bands for

the amplified mean wind speed and wind power density

estimates.

The development of a conceptual model for urban

wind energy planning (UWEP) will be discussed in the

following chapter. This conceptual model develops the

methodology upon which the proposed prototype UWEP DSS

is based.

121


3 CONCEPTUAL MODEL DEVELOPMENT

The building industry has made considerable

progress in evaluating the energy and environmental

implications of resource deployment in building

construction using assessment tools such as the

Building Research Establishment's Environmental

Assessment Method (BREEAM) and ECD Energy & Environment

Canada Ltd.'s Green Globes. Numerous guidelines (e.g.,

Leadership in Energy and Environmental Design (LEED)

and R2000/EnerGuide) have been developed to direct

practitioners in the investigation of life-cycle costs

associated with building material and construction

methodology selection. Simulation tools such as the

Sustainable Buildings Industry Council's Energy-10, the

Californian Energy Design Resources program's QUick

Energy Simulation Tool (eQuest), and the Department of

Energy's DOE-2 have been developed, through which the

energy efficiency of the latest building technologies

can be comparatively assessed.

Industry must now move beyond the current

guidelines, which focus primarily on energy efficiency,

to a consideration of how to make a structure self-

sufficient through its ability to produce the energy

its occupants will inevitably consume. An urban

122


renewable energy source (URES) assessment tool would

help achieve this objective. The photovoltaics

algorithm, recently developed for DOE-2.2 suggests that

industry is heading in this direction (Hirsch, 2006).

Unfortunately, the only consideration that wind is

given within existing building energy simulation tools

is with regard to its role in natural ventilation and

the impact of infiltration on building heating and

cooling costs. This research undertaking takes steps

towards future assessment of wind as an energy source

within building energy simulation tools.

3.1 Urban Renewable Energy Source Assessment

The inherent opportunities to maximise renewable

energy (RE) from various sources (e.g., sun, wind,

ground heat, etc.) depend on the urban morphology

(e.g., plan orientation, placement of structures within

the plan, and the architectural detail of the

individual structures) and the meteorological-

morphological interactions (Grosso, 1998). URES

assessment would primarily differ from classic RES

assessment by including an analysis of the interaction

between pertinent meteorological processes and the

underlying morphology. Such an assessment could

support iterative evaluation of building configuration

123


implications on potentially extractable renewable

energy. One possible configuration of a modular URES

DSS is portrayed in Figure 3.1.

DECISION M AKER URES-DSSDATABASESANALYSIS M O DULES

SOCIO -ECO NO M IC M O DULELife-cycle analysis

Energy Policy M akersLegal fram ework & financial incentives

RES ASSESSM ENT M O DULE(e.g. hydro, biom ass, geotherm al, solar, wind, etc.)

Theoretical PotentialBased on m eteorological, hydrological, geological, etc. data.

RES DataG eographicalM eteorological ArchiveW ind Atlasetc.

Available Potential

Influence of local terrain, land-use, built environm ent m orphology, etc.

G eophysical DataTopographicalLand-useRoad ClassificationBldg Classification

Technological Potential

Portion of available energy that can be extracted by the energy conversion device.

Technical DataPerform ance characteristicsSystem efficiencies

ENERG Y CO M PUTATIO N

M O DULE

Energy Dem and Data(including conservation m easures)Forecasting Data

Econom ic DataConstruction costs M aintenance costs DepreciationInflationProject life span

Select a SIte

Can sufficient energy be produced?

Are there negative socio-

econom ic im plications?

Can they be

overcom e or

m itigated?

Yes

LO CAL AUTHO RITIES

UTILITIESon vs. off grid

INVESTO RS

No

Yes

Yes

Viable Project

No

1

2

ENVIRO NM ENTAL IM PACT M O DULE

SO CIO -ECO NO M IC CO ST

No

Figure 3.1. Urban Renewable Energy Source (URES) DSS.Based on Ramachandra (2006) and Voivontas et al. (1998). The

shaded areas highlight areas of focus of the proposed WEP module.

The URES module algorithms would be structured in such

a way as to accommodate the assessment of a variety of

RES for a given location. Modularisation of the

analysis algorithms and integral databases would

produce an extremely versatile framework. It is

124


through the inclusion of detailed morphological data,

parameterised using road and building classification

schemes, that calculations of available potential energy can

truly reflect urban-scale influences (e.g., street-

canyon radiation trapping, shadowing, and building

aerodynamics-induced wind-amplification).

Since there are several solar energy planning (SEP)

DSSs already in existence (Ramachandra, 2006; Rylatt et

al., 2001) and the other RES currently have limited

applicability in an urban setting, the body of this

research will focus on the development of an urban wind

energy planning (UWEP) module. The UWEP module would

fit into a complete URES DSS as illustrated by the

shaded regions within Figure 3.1. Studies have shown

that traditional wind turbine siting guidelines may

result in the selection of suboptimal heights and

locations (Campbell & Stankovic, 2001b). An assessment

of the potentially amplified wind speed in the

neighbourhood of urban morphological features could

have a profound effect on both the design of wind

energy extraction devices and their siting. The body

of this research introduces the concept of urban wind

as a manipulable resource, drawing on theory from

several different fields. Concepts long understood in

the field of meteorology and geography, such as tunnel

125


and hill effects, will be brought to bear in the

assessment of the wind resource within the urban canopy

layer (UCL). Research findings and theory from the

field of building aerodynamics will be used to quantify

building-induced wind amplification phenomena.

The concept of urban-scale building integrated

and/or mounted wind turbines (BUWT) is still in its

infancy. There are only a small number of

manufacturers currently exploring the development of

urban wind energy conversion systems (Campbell &

Stankovic, 2001b). The underlying premise is that if

the configuration of structures on a site, and/or

individual features of a structure, can accelerate the

air flow through the energy extraction device, smaller

urban-scale machines could conceivably become

economically viable. Since these designs are still at

the patent-pending stage, performance characteristics

required to determine how much energy could be

extracted given particular wind conditions are not yet

available. As such, the proposed DSS will focus on

quantifying the energy potentially available for

extraction, leaving the assessment of the socio-

economic feasibility of urban-scale electrical energy

production (Celik, 2003) to other software applications

(e.g., HOMER and RETScreen). A preliminary study of

126


the socio-economic considerations pertaining to wind

turbines in a built environment can be found in Cace et

al.’s (2007) report.

3.2 Scope of Applicability

The planning tool is intended for use by a broad

spectrum of individuals, including, but not limited to

homeowners, engineers, architects, and developers. The

tool would be applicable to both new and existing urban

developments. To facilitate UWEP at an existing site,

and/or in the early stages of urban development, the

proposed tool would need to be flexible, interactive,

and able to extract data from, and provide data to, a

suitable design platform (e.g., AutoCad). For existing

sites, ideally, the site plan would be uploaded into

the assessment tool, to allow for topographical,

geographical, and meteorological detail to be

superimposed. For new developments the superimposed

detail, in conjunction with a predefined optimal plan

and architectural feature configuration, would form the

initial layout. The proposed planning tool, as

conceptualised in Figure 3.2, would incorporate a

database-assisted design (DAD)-type approach and

encompass the shaded portions of the RES assessment

module in Figure 3.1.

127


3.3 Configuration of the Conceptual Model

The platform of the proposed UWEP DSS, illustrated

below, will be based on Microsoft Excel, due to the

various calculations that will be required throughout

the assessment, the need to build databases and store

data, and the desire to accommodate as many future

users as possible. Visual Basic for Applications

(VBA)-based forms will provide the user-interface for

data input, internal database-data extraction, and the

capability to populate fields within the module's

worksheets, wherein the calculations are performed.

This will simplify the exchange of data and protect the

integrity of both the integral database and the

analysis worksheets.

128


DATA INPUT O UTPUT

USER INTERFACE

UW EP M ODULE

EXTERNAL DATABASE

Sum m ary Reporte.g. site and building configuration, am plification zones, estim ated wind energy, peak energy m onths, and recom m endations.

INTERNAL DATABASE

Site Characteristicse.g. location, orientation, land-use category, building and road classification.

Building Characteristicse.g. orientation, height, width, depth, roof type, and pitch.

M eteorological Datae.g. hourly wind speed and direction.

Figure 3.2. Conceptual Model of the UWEP DSS.

Data will primarily be obtained from external internet-

based databases and applications, while parameters

required for mathematical computations will be

extracted from an internal database. The required data

can be categorised as:

Meteorological (e.g., wind speed and direction), Geographical (i.e., latitude and longitude co-

ordinates), Topographical (e.g., land-use and urban

subregion), and Morphological, (e.g., street widths and building

configurations).

129


Meteorological data would be obtained from an internet-

based wind atlas (e.g., the Ontario Wind Resource Atlas

and the Canadian Wind Energy Atlas) or generated from

the time series data (TSD) of a nearby meteorological

station, provided by Environment Canada's internet-

based National Climate Data and Information Archive.

Co-ordinate data would be obtained from an online

database (e.g., Geo-coder). Topographical and

morphological data would be input by the user through

selection of predefined options within an associated

VBA form. The user would be directed to a national-

level, geographical society and/or government regulated

website (e.g., Natural Resources Canada (NRCan) (2006)

Atlas of Canada) to identify land-use categories and

satellite image applications (e.g., MapQuest and

Google Maps) to identify the representative urban

subregion. To limit the scope of this research

undertaking, it is proposed that orographic terrain

(e.g., mountains and hills) feature-induced effects be

assessed by an add-on submodule through future

development of the proposed UWEP DSS. The reader is

directed to the literature of Jacobson & Malte (2005),

Troen & Petersen (1989), and Verkaik (1999) for details

130


regarding the underlying methodology of such an

undertaking.

The UWEP module will be comprised of various

submodules to manipulate the wind speed and direction

data, assess wind amplification, identify the location

and size of the amplification zones, and calculate the

wind energy available for extraction. The UWEP

submodules would extract the appropriate parameters

required in the calculations from specialised tables of

an internal database based on user input. This

internal database will minimise the level of expertise

and user-effort required to parameterise the urban

morphology and building aerodynamics-induced wind

amplification. It would be comprised of the following

tables:

latitude and longitude co-ordinates (for select cities),

roughness parameters (based on land-use categorisation),

urban parameters (based on road and building type classification), and

wind speed amplification factors.

The results of the assessment will be presented in

a summary report, generated from the data, calculated

values, and associated graphs stored in the analysis

131


worksheets. The primary outputs of the proposed tool

will be:

Project details summary (e.g., name, address, comments, etc.)

Morphological (e.g., urban subregion) Subject building dimensions and features (e.g.,

roof type and pitch) Amplification zone areas by building face Wind power density by zone Wind energy by building face Windy season months

Graphical representations will be comprised of user-

saved images and graphs generated by the proposed tool.

The summary report would include recommendations

concerning building orientation and architectural

feature characteristic (i.e., roof type and pitch)

changes. These recommendations could potentially

enhance wind amplification, which in turn would yield

higher levels of wind energy available for extraction.

The proposed methodology of wind energy assessment,

including the meteorological-morphological interactions

resulting in wind amplification, will now be discussed.

3.4 Wind Energy Assessment Methodology

Wind energy assessment requires estimation of

temporally and spatially variable wind statistics. The

132


proposed tool would accommodate both vertical and

horizontal spatial variation by defining the following

scales:

Region, City (mesoscale), Neighbourhood (local scale), and Site (microscale).

The horizontal spatial variation is only considered at

the microscale. Correspondingly, the tool would

perform the following tasks in developing the wind

statistics:

Estimate the regional wind statistics, Develop the mesoscale wind speed profile, Determine the local scale wind amplification, and Calculate the building aerodynamics-induced

microscale wind amplification.

Performance of these tasks involves meteorological and

morphological considerations, which will now be

discussed in turn.

3.4.1 Meteorological Considerations

The first task performed by the UWEP module would

be to assess the spatial and temporal variation of the

wind at the site. Spatial variation is represented by

the mean wind speed profile [ ], as illustrated in

133


Figure 3.3. The UWEP module will calculate the mean

wind speed profile based on the regional wind

statistics. It will consider the mean annual, regional

wind statistics provided by the wind atlases as the

mesoscale wind. The mesoscale wind is considered to

prevail above the surface layer (SL) and remain

relatively constant over several kilometres.z

Urban Boundary LayerSurface Layer

G eostrophic W ind

M esoscale W ind

W ind Direction

)z(u

Figure 3.3. The vertical stratification of the urbanboundary layer.

In regions where there are distinct seasons, a

seasonal assessment would be conducted using time

series data (TSD). The TSD will be used to calculate

the mesoscale wind statistics unless the meteorological

station is greater than five kilometres from the site

in which case the geostrophic wind statistics will be

calculated. The geostrophic wind prevails above the

urban boundary layer (UBL) and remains relatively

constant over distances greater than five kilometres.

In these regions a hybrid solar-wind energy extraction

134


system should be considered (Eke, Kara, & Ulgen, 2005),

given that the wind resource is typically plentiful

during periods when solar energy is less plentiful and

vice versa.

Temporal variation is defined by the regional wind

statistics in the form of a wind speed histogram

(Figure 3.4 (a)) and/or probability distribution and an

associated wind direction rose (Figure 3.4 (b)). Wind

direction is of primary importance when considering

fixed-direction, wind energy extraction devices. Given

that very few sites are truly subjected to uni-

directional winds, with most actually experiencing

omni-directional winds, the data are best presented in

the form of a wind direction rose. By grouping

cardinal sectors by quadrant, two to three dominant

wind directions can often be identified for most sites.

For example, in Figure 3.4 (b) the dominant directions

could be summarised as SSW (combining WSW, SSW, and S),

SE (SSE and ESE), and N. The assessment tool would

primarily rely on the two most dominant directions, to

be referred to as the primary and secondary wind

directions, to determine the presence and type of

aerodynamic effect likely to result in wind

amplification.

135


(a) Wind Speed Histogram (b) Wind Direction Rose

Figure 3.4. Wind speed histogram (a) and directionrose (b).

Having fully defined the wind climate, the UWEP module

would proceed on to the assessment of wind

amplification involving morphological considerations at

both the local and microscale.

3.4.2 Morphological Considerations

The literature concerning the implications of

building aerodynamics on the flow field indicates that

wind amplification is governed by the morphological

parameters of the underlying terrain on several scales

and dependent on the associated flow regime. Though

there is a general awareness of wind amplification,

including folklore pertaining to windy street corners

and/or intersections, the underlying sentiment is best

136


summarised by “wind patterns in complex terrain … are

so complicated that they defy analysis or accurate

description” (H. Liu, 1991 p. 62). The underlying

premise of this body of research is that local scale

and microscale amplification can be estimated using

published comfort parameters and pressure coefficients,

respectively. Wind amplification exceeding a factor of

2 has been recorded (H. Liu, 1991) resulting in a

potential 8X wind power increase. Amplification up to

a factor of 3 has been predicted by CFD analysis of an

urban area (Gandemer, 1977), corresponding to a

potential 27X wind power increase. Determination of

the magnitude and location of amplified wind speed

zones within an urban setting may prove that urban wind

energy extraction is a viable undertaking.

Studies involving the influence of building

configurations on ground-level wind speeds have

developed the concept of a comfort parameter as a

function of turbulence intensity and wind amplification

(Gandemer, 1977), dependent on the morphological

configuration at a local scale. The proposed tool

would correlate the urban subregion morphology to

comfort parameter-based amplification factors for

various aerodynamic effects, within an internal

137


database. Urban subregion morphology would be

characterised by:

plan area density, mean building height, building classification, and road classification, for an assortment of distinct

urban subregions.

In the field of Civil Engineering, specifically

structural design and the assessment of wind loading,

the primary focus is on maximum wind speeds of limited

duration. The empirical relations, correlating

structural geometry to dynamic pressure, are obtained

from this field of study and the associated building

codes. As a first attempt at quantifying the magnitude

of building-aerodynamics induced wind amplification,

the pressure and pressure-gust coefficients provided in

the National Building Code would be used and correlated

to architectural features and building dimensions

within an internal database. The gust coefficient

component will prove to be problematic, requiring

assumptions and simplifications. Further research will

be requiring involving data collection to gain a better

understanding of sustained, as opposed to sporadic,

wind amplification.

138


Wind amplification would be assessed in two stages.

The first stage would consider the wind amplification

potential of the urban subregion's morphology at the

local scale. The channel effect, as illustrated in

Figure 3.5 (a), is one of several aerodynamic effects

that could result in wind amplification at this scale.

The second stage would entail a more detailed

structure-specific assessment, taking into

consideration building dimensions and architectural

features such as roof type and pitch. Figure 3.5 (b)

illustrates the presence of microscale amplification

through depiction of compressed streamlines overtop of

the windward slope of the roof. The urban subregion

morphology and the architectural features of the

subject building have the potential to accelerate the

local wind speed (Dayan, 2006) and affect its

direction. This concept is at the core of this

research undertaking.

(a) Local scaleamplification (b) Microscale amplification

139


Figure 3.5. Morphology-induced wind amplification.Modified after (a) Gandemer (1977) and (b) Dalgliesh & Schriever

(1968).

The final task of the UWEP module will be to

calculate the energy available for extraction, based on

the amplified wind speed, the areas within which these

amplified wind speeds are considered to prevail, and

the wind speed and direction frequency distributions.

The conceptual model developed in this chapter,

pertaining to the assessment of a potentially amplified

wind resource within an urban surface layer with

consideration of plan geometry and architectural

features, is the foundation upon which the assessment

tool will be based. The detailed development of the

prototype UWEP DSS is presented in Chapter Four.

140


4 DECISION SUPPORT SYSTEM DEVELOPMENT

The Urban Wind Energy Planning (UWEP) DSS is

intended to provide decision makers with quantitative

information regarding wind energy potential zones in

support of development, appropriate type-selection,

and/or placement of wind energy conversion systems

(WECSs) within an urban environment. The UWEP DSS does

not address the ever-changing and complex nature of the

associated socio-economic implications of such an

undertaking. This is primarily due to the fact that

these implications are not yet known, since suitable

urban wind energy conversion systems (UWECSs) are still

at the developmental stage. It is further surmised

that the well documented implications associated with

traditional wind turbines will not necessarily be

applicable to these UWECSs. An additional supposition

is that a quantitative assessment of the urban wind

environment will be of tremendous assistance to the

development of suitably sized and configured urban wind

energy conversion systems.

This chapter elaborates on the development of the

UWEP DSS through the following sections, as illustrated

in the block model in Figure 4.1 below:

user interface,

141


external databases, internal databases, UWEP module, Summary Report, and Architectural Configurator module.

In Figure 4.1 the hollow arrows indicate data

exchanges. Module and submodule data are stored within

the analysis worksheets. More involved calculations

are performed in associated calculation worksheets,

which provide summary results to the analysis

worksheets. The related background theory and

assumptions are provided as Appendices, with the

associated computer algorithms and mathematical

computations provided in an attached disk. An overview

of the UWEP DSS will now be provided followed by a

discussion on each of its sections in turn.

142


Figure 4.1. Block model of the UWEP DSS.

143


4.1 Overview

The UWEP DSS is primarily comprised of algorithm-

driven submodules, which perform mathematical

computations. Groupings of task-specific data-entry

forms and worksheets are designated as submodules,

which combine to form the UWEP module. On project

launch, the UWEP module guides the user through the

data-input process using a series of forms. Based on

user-input, the forms extract values from an internal

database to populate appropriate fields of associated

analysis worksheets. The forms and analysis worksheets

contain links to pertinent web-based external databases

(e.g., maps, climate data, and wind atlas applications)

to provide the user with the means to obtain the

required data. To reduce dependence on user knowledge,

the forms primarily contain selectable options. The

analysis worksheets store key data for, and extract

summaries from, the calculation worksheets. The

calculation worksheets perform all the required

calculations and produce the associated graphs. More

detail pertaining to the individual forms, worksheets,

modules, and submodules will be provided in the UWEP

module section. At the end of the data-input sequence,

the Summary Report is presented. The Architectural

Configurator is a sub-path through the UWEP DSS, which

144


provides access to the forms and analysis worksheets

pertaining specifically to building features,

dimensions, and orientation. This sub-path streamlines

the iteration process for the user who can change these

aspects of the subject site to affect wind energy

potential.

The UWEP module is the main module of the UWEP DSS.

It calculates the mean annual or mean seasonal wind

energy by zone through interaction with the user and

iteration between the four submodules, as depicted in

Figure 4.1. The methodology of the UWEP module is

based on the two-layer urban boundary layer (UBL)

model, comprised of a surface layer (SL) and an outer

mixed layer, as previously discussed. The SL is

further stratified into its sublayers, namely the

inertial sublayer (ISL), the roughness sublayer (RSL),

and the urban canopy layer (UCL), as illustrated in

Figure 4.2 (c).

Figures 4.2 (a) through (c) illustrate how the

submodules of the UWEP module delve into the UBL,

sequentially increasing the resolution of the analysis.

Figure 4.2 (a) represents a typical city block in

elevation view, including the associated idealised wind

speed profile [ ]. At this resolution the effect

145


of the street canyons and individual building

dimensions is represented by the parameters tabulated

within the internal database. The initial estimate of

the mean wind speed profile, based on traditional

theory, occurs at this resolution. Figure 4.2 (b) is

the same city block in plan view, with the hashed

blocks pertaining to the buildings shown in (a). At

this resolution user-input assists in calculating the

required street widths and the resulting inter-building

spaces to re-estimate the mean wind speed profile

within the UCL. Figure 4.2 (c) is the elevation view

of the last street of the city block, as indicated by

the section line in (b). This view includes the

designation of the top of the UCL [zH] and the top of

the RSL [z*]. At this resolution, the user is required

to provide the dimensions of the subject building to

assess potential wind amplification.

146


Figure 4.2. Illustrative representation of the UWEPmodule’s methodology.

In this figure z0 represents the roughness length, d thedisplacement height, zH the top of the UCL, z* the top of the

147


roughness sublayer (RSL), zSL is the top of the surface layer(SL), and u(z) the mean wind speed profile. The dashed line

representation of z* in (c) is intended to indicate that the topof the RSL is not necessarily the bottom of the inertial sublayer

(ISL) (based in part on (Fisher et al., 2004) and (Dalgliesh &Boyd, 1962)).

More detail regarding the correlation between the

submodules, the UBL stratification, and user-input will

be provided in the UWEP module section. The main

sections of the UWEP DSS will now be discussed in turn.

4.2 User Interface

The platform of the UWEP DSS is based on

Microsoft Excel, using Visual Basic for Applications

(VBA) macros for navigation, data entry form

interaction, and to automate computational analysis,

including the generation of graphs. The main elements

of the application are the:

Control Centre, data-entry forms, analysis and calculation worksheets, databases, and reports.

The graphical design layout of the individual

worksheets, including the Control Centre, is modelled

after the Wind Energy Project Model of NRCan's

148


RETScreen® International Clean Energy Project Analysis

Software.

The Control Centre, illustrated below, is the

primary user interface and the launch point for all

assessment projects. It is divided into the following

four sections:

Help, Worksheets Reports, and Tools.

Figure 4.3. The Control Centre of the UWEP DSS.

The Help section contains links to files explaining the

methodology behind the UWEP DSS and detailed

instructions on how to conduct an assessment. It

149


includes a glossary of terms and a field legend

explaining the source of the data for each field within

the analysis worksheets. The Worksheets and Reports

sections provide alternate means through which to view

the analysis worksheets and reports. The Tools section

contains links to external database applications, which

assist the user in obtaining the data required to

conduct an assessment.

The data entry forms are the secondary user

interface. The data entry forms are opened by clicking

Project Launch in the Control Centre. The submodule

specific data-entry forms are the primary means through

which to access the databases, analysis worksheets, and

the reports. The user is provided with additional help

through task-specific instructions on the individual

forms. The user has read-only access to the analysis

worksheets and no access to the calculation worksheets.

The UWEP DSS contains an internal and an external

database. The external database is comprised of

internet-based applications and accessible to the user

as previously discussed. The internal database is only

accessible to the analysis worksheets and comprised of

various tabulations. The databases are the heart of

150


the UWEP DSS. The composition of the external

databases will now be discussed.

4.3 External Databases

The external databases are a key component of the

UWEP DSS. They provide the user with the necessary

data to fill out the UWEP submodule forms. Since

climatological and land-use data change over time,

providing access to a variety of free online resources

was considered to be more appropriate than attempting

to create a comprehensive internal database. The user

is able to link to the following external resources

and/or databases through the Control Centre and through

the forms and associated analysis worksheets to which

they pertain:

GeoCoder.ca (Ruci, 2005), NRCan's Atlas of Canada (NRCan, 2006), Google Maps (Google, 2007), MapQuest (MapQuest, 2007), National Climate Data and Information Archive

(NCD&IA) (EC, 2005d), CWEA (EC, 2005a), and OWRA (MNR, 2006b).

The GeoCoder.ca provides longitude and latitude co-

ordinates for a given location in North America. The

Atlas provides topographical land-use coverage detail,

151


enhanced through satellite imagery provided by the map

applications. The NCD&IA provides time series data

(TSD) on wind speed and direction dating back to 1971,

for all registered meteorological stations in Canada

equipped with the instrumentation to collect this type

of data. The CWEA and the OWRA provide regional wind

statistics, comprised of the mean wind speed, the

associated frequency distribution parameters, and the

wind direction frequency distribution.

The majority of the data gathered by the user from

the external database applications must be

parameterised for the UWEP DSS to be able to use it in

calculations. The term parameterisation is being

broadly applied in this context to include conversion

of text variables into numeric variables.

Parameterisation is the primary function of the

internal database, which will now be discussed.

4.4 Internal Database

Within the UWEP submodules, the equations used to

calculate the mean wind speed at various heights and

locations use empirically derived parameters to

represent the complexity of the underlying terrain and

its effects. To ensure repeatability of the

calculations and to reduce the amount of detailed

152


information and associated level of user-skill and/or

knowledge required, the empirically derived parameters

are contained in an internal database. The internal

database is comprised of a:

Co-ordinate data table, Roughness classification table, Urban parameterisation table, and Wind amplification factor table.

The internal database is the heart of the UWEP DSS.

This section provides conceptual overviews and brief

descriptions of the composition of these tables, as

they will be referred to within the UWEP module section

of this chapter. The associated background theory,

assumptions, and calculations pertaining to their

development are provided in Appendix D.

The Co-ordinate Data table is self explanatory; it

contains latitude and longitude co-ordinates for a

select group of cities, primarily within Canada. It is

intended to be of use if a specific building site has

not been determined. An overview of the remaining

tables within the internal database will now be

provided in turn.

153


4.4.1 Roughness Classification Table

The Roughness Classification table of the internal

database of the UWEP DSS correlates land-use categories

to their associated local roughness length [z0] as

tabulated below. The local roughness length represents

the influence of the surface layer on the overlying

flow in two-layer boundary layer methodology.

Roughness length parameterisations are required at

three different resolutions, to be referred to as

mesoscale, local, and microscale, to develop the mean

wind speed profile equations within the UWEP module.

Table 4.1. The Roughness Classification table of the

UWEP DSS.

154


Class Roughness Class Sub-categories Length(#) Description z0 (m )

W ater4 10

Snow4 4

Built features, sm ooth (e.g. airport runways) 3

Desert4 4

Soil 5

Sand (e.g. beaches & dunes)4 1

Built features, sem i-rough (e.g. airport, including buildings & trees) 1

G rassland 7

Built features, rough (e.g. roads, railways & parking lots) 1

Bedrock3 1

Agricultural land 8

Vegetation 6

W etlands 4

Rural area, sm all town (e.g. low buildings & trees) 7

Urban area (e.g. suburbs or m edium -sized town) 7

Forest2 8

5 Urban area, m ixed1 1.5 6

6 Urban area, city centre1 >2 6

4

0

1

2

3

0.15

0.30

# Refs

1.00

0.005

0.05

Table Notes:

1 details to be provided in urban parameterisation. 2 furthercategorised into deciduous, coniferous, mixed, and transitional

(e.g., regrowth, burns, and cuts). 3 (e.g., mine tailings,quarries, outcrops, alvar, and stone or pavement). 4 wind

dependent. Individual category values are weighted averages byfrequency of occurrence within the literature. Smooth, semi-rough, rough, etc. categorisation is based on resulting flow

regime classification.

The local roughness length is the primary length

scale parameter of the mean wind speed equation for a

seasonal assessment. The roughness length is most

often derived from plan area density [p] and mean

roughness element height [H] (Mertens, 2003) or frontal

area density [f] (MacDonald, 2000 after (Plate,

1995)). Further elaboration on these variables, which

155


are the main variables used to characterise urban

terrain, is provided in the Urban Parameterisation

table portion of this section. Theoretically, the

roughness length is the height or zero-plane [d0] at

which the mean wind speed equates to zero through

downward extrapolation of the logarithmic wind speed

profile (Verkaik & Smits, 2001). As such, it defines

the lowest height at which the logarithmic profile is

applicable. Figure 4.4 illustrates how the roughness

length, combined with the displacement height [d],

determines the vertical position of the logarithmic

profile.

zH

dz0 + d

z*

z

0

*zdzlnu)z(u

)z(u

Figure 4.4. The logarithmic wind speed profile.

156


The equation shown is the logarithmic mean wind speed profile,including the displacement height [d]. z0 is the roughness length.

z* is the top of the RSL, zH the top of the UCL, and z0+d is thezero-plane [d0]. Based on Dalgliesh & Boyd (1962).

It is recommended that if the roughness length

exceeds 0.1 m (e.g., urban-type roughness) the

displacement height should be included in the equation

for the logarithmic profile (Taylor & Lee, 1984) as

shown in Figure 4.4, to ensure that the resulting mean

wind speed is undefined below the upper region of the

RSL. The displacement height [d], or zero-plane

displacement, serves to elevate the zero plane to

account for the fact that the top of the RSL is not

necessarily the bottom of the ISL (Wieringa, 1993).

The displacement height parameter, primarily required

in locations characterised by urban-type roughness, is

provided to the UWEP through the Urban Parameterisation

table.

Land-use and topographical data, upon which

roughness length is most often based, can be obtained

from a land-use database, topographical maps, aerial

photos, and satellite imagery, accessible through the

external database(s). The roughness length parameter

can be displayed in various formats, including a

roughness map of the area of interest (EC, 2005a), a

roughness rose corresponding to the up-wind terrain by

157


compass sector, and correlation tables. An extensive

assessment of various classification schemes, with

consideration of the land-use categories represented in

topographical maps within the external database (e.g.,

The Atlas of Canada and the OWRA), prompted the

development of a new roughness classification scheme

based on an amalgamation of findings from the

literature. A correlation table was considered to be

the most appropriate configuration for the Roughness

Classification table within the internal database of

the UWEP DSS. Further details pertaining to the

background theory, assumptions, and calculations

substantiating the development of this table are

provided in Appendix D - 1.

The meteorological parameters of displacement

height and roughness length are related to the urban

morphology through definition of the height of the top

of the RSL [z*]. Accurate estimation of this height is

critical to the development of the Urban

Parameterisation table discussed in the next section.

Within the literature there is much controversy

regarding the height and depth of the RSL. The


pertaining to the height of the RSL are provided in

Appendix E.

158


4.4.2 Urban Parameterisation Table

The Urban Parameterisation table of the internal

database of the UWEP DSS correlates readily

identifiable urban subregions to parameters required by

the UWEP module to perform various calculations. Urban

terrain is primarily characterised by its plan area

density (pad) [p], frontal area density (fad) [f], and

mean building height [H]. Plan area density is the

ratio of the building footprint area to the building

lot area: . These areas can, to some extent, be

determined from land-use databases, aerial photos, and

thematic maps. Frontal area density is the ratio of

the building face area to the building lot area: ,

based on building height and pad. Building height, for

the most part, is estimated using the statistical

probability of a certain height of building existing in

a given urban region and summarised through the

calculation of a plan area, weighted-average building

height (Coceal & Belcher, 2005). A simplified

representation of these areas is provided in Figure

4.6. Considering the effort required by the user to

determine the pad, fad, and weighted-average building

height of a subject site, the development of an

159


integral urban parameterisation scheme was deemed

crucial.

In the literature on the characterisation of urban

morphology, concepts such as the sky view factor (SVF)

(Souch & Grimmond, 2006; Eliasson, 2000), as

illustrated, and the spatial openness index (SOI) or

the ISOVIST “the space visible 360 degrees around a

particular point” (Fisher-Gewirtzman et al., 2005 after

(Benedikt,1979)) have been developed.

Figure 4.5. The sky view factor as a measure of streetcanyon geometry.

The sky view factor (SVF) is a measure of the street geometry asillustrated by the photos of three sites located within a distance

of 200 m in central Göteborg. (a) Open square: SVF = 0.93, (b)Street intersection: SVF = 0.47, and (c) Street canyon: SVF = 0.29

(Eliasson, 2000).

These concepts, primarily used by urban planners, could

have been employed in the development of urban

parameterisations such as the urban porosity model

(Grosso, 1998) and related wind permeability roses

(Steemers, 1997). Unfortunately, no generic yet

160


comprehensive, suitable urban characterisation schemes

were found in the literature, prompting the development

of the Urban Parameterisation table. In the following

paragraphs, an overview of the nature and development

of the Urban Parameterisation table of the UWEP DSS is

provided. Details pertaining to the background theory,

assumptions, and calculations substantiating the

development of this table are provided in Appendix D -

2.

The Urban Parameterisation table simplifies the

heterogeneity of the urban morphology by defining

homogeneous urban subregions. These subregions are

primarily based on a combination of published

tabulations of urban subregion configurations

distinguished by their densities (i.e., frontal and

plan area) and mean building height (Plate, 1995 after

(Theurer, 1993); Coceal & Belcher, 2005 after (Burian

et. al, 2002)). Within the table, the urban terrain is

classified by region as follows:

Suburban, Urban, and City Centre.

These regions are further divided into the following

primary-use categories:

161


Residential, Industrial, Commercial, and Mixed.

The urban subregions are defined through further

delineation of the primary-use categories by street

configuration (e.g., row or block) and mean building

height (e.g., 1-2 storey), as tabulated below.

Table 4.2. The urban subregions of the Urban

Parameterisation table.M ean

Building Height

Plan Area Density

Frontal Area

DensityH

pf

Description (#) (m ) - -Suburban, warehousing, isolated, < 3 s 1 11 0.04 0.03Suburban, business/industrial 2 10 0.30 0.07Urban, residential, 1-2 s 3 8 0.26 0.19Suburban, residential, 1-2 s 4 9 0.15 0.12Urban, industrial, < 3 s 5 9 0.33 0.11Urban, residential, 1-3 s 6 10 0.25 0.15Urban, residential, 3-5 s, r 7 16 0.25 0.10Urban, m ixed industrial/com m ercial, 3 - 4 s 8 12 0.27 0.19Urban, industrial rows, > 3 s 9 20 0.35 0.12Urban, m ixed industrial/com m ercial, ~ 2 s 10 7 0.47 0.14Urban, com m ercial district, > 5 s 11 21 0.28 0.24Urban, residential, 3-5 s, b 12 18 0.30 0.15Urban, city centre (e.g. irregular s, b, & r) 13 34 0.31 0.29

CategoryUrban Subregion

In the table s, r, and b are used to represent storey, row, andblock, respectively.

The urban subregion category numbers simplify the

presentation of parameter tabulations by subregion in

subsequent sections and facilitate database navigation.

Columns titled category in subsequent tabulations,

162


pertain to the urban subregion categories as defined in

Table 4.2.

For each subregion the average building areas

(i.e., frontal [Af] and plan [Ap]) and associated lot

[Ad] and total [At] surface areas were calculated,

based upon various parameters including the densities

and mean building height, as illustrated below. The

frontal area is the area of the surface exposed to the

wind (i.e., Af = H x B) and the plan area is the

building foot-print (i.e., Ap = D x B). The lot

surface area [Ad], defined as the total area [At]

divided by the number of structures, is used in

conjunction with the density values to quantify the

average space [S] between buildings. The total area,

for the purpose of these calculations, is set equal to

the square of the fetch, (i.e., At = x02). These areas

are illustrated in Figure 4.6 (modelled after MacDonald

(2000)), with the associated linear dimensions depicted

in Figure 4.8 below.

163


Figure 4.6. The geometry of urban morphology.

Through calculation of these morphometric

dimensions, the relatively

heterogeneous nature of the

urban subregions is translated

into homogeneous building

arrays. For example, urban

164

W IND S

x0

D

S

B


subregion category # 1, representing a suburban

warehouse district as depicted in the satellite image,

is transformed into the cubic array illustrated below.

Figure 4.7. Satellite plan image of urban subregion category # 1.

(Google, 2007)

The array representations of the urban subregions

within the Urban Parameterisation table provide the

UWEP module with the morphometric dimensions required

to redevelop the mean wind speed profile within the

UCL, define the potential wind

amplification zones, and

extract the appropriate wind

amplification factors from the

Wind Amplification Factor

table. The cross-wind building breadth [B], along-wind

building depth [D], and average inter-building space

[S], as depicted, along with the mean building height

[H] define the cubic array for each urban subregion

category.

Figure 4.8. Cubic array representation of urbansubregion category # 1.

165


The average inter-building space includes parks,

yards, roads, and parking lots, thereby increasing the

average linear space dimension beyond what is normally

encountered between urban buildings. As such, minimum

inter-building space widths [Sm] were assigned based on

observed urban subregion geometry. User-selected road

classification parameters determine the street widths

[W], based on the road classification sub-table of the

Urban Parameterisation table. Additional detail

pertaining to inter-building space widths is provided

in Appendix D - 2.

In the literature, the primary linear dimensions

are non-dimensionalised through formation of aspect

ratios, namely the:

frontal aspect ratio (far) [B/H], side aspect ratio (sar) [D/H], and space [S/H] or street canyon aspect ratio [W/H].

In the field of wind engineering, the following wind

direction-dependent ratios have also been defined:

slenderness ratio [H/B], fineness ratio [D/B] (Cook, 1990), and height-to-depth aspect ratio [H/D] (Canadian

Commission on Building and Fire Codes, 2006).

166


Tabulation of urban subregion specific aspect ratios

within the Urban Parameterisation table assists the

UWEP module in navigating the Wind Amplification Factor

table, to be discussed in the following section. The

Urban Parameterisation table attempts to cover the most

common urban subregion configurations, but is by no

means exhaustive. Differing plan area-to-frontal area

pairings may be in existence, with either higher or

lower mean building heights.

4.4.3 Wind Amplification Factor Table

The Wind Amplification Factor table of the internal

database of the UWEP DSS provides the UWEP module with

wind amplification factors at two scales. An overview

of the key concepts and components of the Wind

Amplification Factor table is provided through the

following paragraphs. Details pertaining to the


substantiating the development of this table are


The wind amplification factor [] is defined as:

, where: (4.1)

167


is the amplified gust speed at a reference

height, and

is the reference gust speed at the same height,

calculated within the UWEP module.

Since wind must pass through the immediate

neighbourhood prior to impinging on an individual

building, the Wind Amplification Factor table is

correspondingly divided into neighbourhood- and building

feature-scale portions.

In the neighbourhood scale portion, the aerodynamic

effect-dependent wind amplification factor [] is

tabulated by effect for urban subregions within which

these effects could possibly occur based on comfort

parameters developed by Gandemer (1977). The

aerodynamic effects, illustrated in Figure 4.9, are

created by the orientation of building groupings in

relation to the primary wind direction. Compiled from

images extracted from Gandemer’s (1977) conference

paper, this figure is merely intended to illustrate the

aerodynamic effects considered during the

neighbourhood-scale portion of the amplification

assessment. As such, the text on the images is

inconsequential. The neighbourhood-scale portion of

the Wind Amplification Factor table is shown below.

168


Table 4.3. Neighbourhood-scale Wind Amplification

Factor table.

Bar Venturi Blocks Passages Tow er Channel Pyram idUrban Sub-region

Category7 1.35 1.65 1.40 1.35 1.80 1.30 1.609 1.34 1.64 1.39 1.34 1.79 1.29 1.5911 1.37 1.67 1.42 1.37 1.82 1.32 1.6212 1.34 1.64 1.39 1.34 1.79 1.29 1.59

W ind Am plification Factor

Aerodynam ic Effect

(a) Bar (b) Venturi

(c) Blocks (d) Passages

(e) Tower (f) Channel

169


(g) Pyramid

Figure 4.9 Neighbourhood-scale aerodynamic effects.

In the building scale portion, wind amplification

factors are primarily tabulated by building surface

zone, based on the composite pressure-gust and pressure

coefficients provided by the National Building Code

(NBC). Idealised representations of the responsible

aerodynamic effects are provided below (Gandemer,

1977). The edge effect can occur along side and/or

overtop of a building. The wake effect primarily

produces turbulence on the leeward side of a building,

while the vortex effect is seen as producing flow

reversal and turbulence on the windward side of a

170


building at street level. All three effects produce

flow acceleration in discrete regions. (a) Edge (b) Wake

(c) Vortex

Figure 4.10. Idealised representations of building-scale aerodynamic effects.

As a first attempt to use published coefficients to

estimate potential wind amplification factors,

simplifications and assumptions were required. For

low-rise buildings pressure coefficients had to be

estimated from composite pressure-gust coefficients.

For the prototype DSS, to simplify wind amplification

factor calculations, a constant gust coefficient equal

to two was assumed. Under actual conditions, given the

171


temporal nature of gusts, this value may vary

considerably.

The building surface zone portion of the Wind

Amplification Factor table is primarily categorised by

building classification (i.e., low-rise and high-rise)

as tabulated below. The low-rise portion of the

database is further classified by roof type and

critical pitch [C]. Angle of flow incidence []

dependency is only considered in the whole building

surface assessments, due to NBC limitations.

Table 4.4. Pressure coefficient-based Wind

Amplification Factor table.

172


BuildingType Building Zone

CriticalRoofSlope[C]

Angle ofFlow

Incidence[ Data

ReferenceHeight[h]

RefNBC

Figure #{{551}}

Side - - Hm I - 8Top

General 0 - 90 0 - 90 Hm I - 7G abled & Hipped

(single-ridge) 7 - H e I - 9

Low-rise

> 7 - Hm I - 11H/DS < 1

G abled(M ultiple-ridge) 10 - H e I - 9

& > 10 - Hm I - 12H 20 m M onoslope 3 - H e I - 9

10 - H e I - 9Sawtoothed> 10 -

CPCg

Hm I - 14High-riseH/DS 1

orH > 20 m

Side & Top 0 90CP&CP*

H I - 15

Stepped flat and monoslope 3 < C 30 roof configurations arealso tabulated within the building code but not included within

the internal database. Modified after Canadian Commission onBuilding and Fire Codes (2006). He designates eaves height and Hm

mid-roof height.

The Wind Amplification Factor table provides

amplification factors for all available combinations of

critical roof pitch and angle of flow incidence by

building type and building element (e.g., roof or

wall), based on the NBC tabulations. These

amplification factors are further tabulated by surface

zone. Details pertaining to the background theory,

assumptions, and calculations supporting the use of

173


pressure coefficients to estimate amplification are


In summary, the inclusion of the Roughness

Classification, Urban Parameterisation, and Wind

Amplification Factor tables within the internal

database of the UWEP DSS reduces the complexity of the

assessment and the level of knowledge required by the

user. Means by which the UWEP DSS interacts with these

crucial integral databases is covered in the following

sections. The main analytical module of the UWEP DSS

will now be discussed.

4.5 UWEP module

The UWEP module is the main module of the UWEP DSS.

It calculates the mean annual or mean seasonal wind

energy through the following submodules:

Wind data extrapolator (WDE), Wind data interpolator (WDI), Wind amplifier, and Energy Calculator.

Extrapolation primarily pertains to upward or

horizontal translations of the wind statistics, while

interpolation pertains to downward translations. The

wind statistics are comprised of the mean wind speed,

the wind speed frequency distribution parameters, and

174


wind direction frequency data. The WDE and WDI

submodules are responsible for calculating the wind

statistics at the regional level and throughout the

entire wind speed profile (Figure 4.11 (a)). The Wind

Amplifier submodule calculates the amplified mean wind

speeds within the amplification zones and the

amplification zone areas for the Energy Calculator

submodule.

Figure 4.11 is a graphical overview of the UWEP

module. The function of the WDE is represented by

Figure 4.11 (a). The WDE is primarily responsible for

calculating the friction velocity parameter [u*] from

the regional wind statistics. The friction velocity is

the velocity scaling parameter of the logarithmic mean

wind speed profile. This empirically derived parameter

quantifies the effect of the turbulence created by the

underlying terrain on the mean wind speed. The

logarithmic profile has been published in various

forms, including forms that attempt to correct for

thermally-induced buoyancy or instability. It is

suggested that at higher wind speeds (i.e., > 3 m/s

(Taylor & Lee, 1984)), the buoyancy effect and the

associated instability is negligible. Averaging

diurnal values also reduces stability effects (Verkaik

& Smits, 2001). The WDE and WDI primarily use the

175


logarithmic profile, without stability correction, for

extrapolation and interpolation of the mean wind speed

above the top of the RSL. The WDI develops the multi-

segment mean wind speed profile (Figure 4.11 (a)) from

above the RSL through the urban canopy layer (UCL) down

to the ground (Figure 4.11 (c)) for the subject

location, based on the friction velocity calculated by

the WDE. The Wind Amplifier calculates the amplified

mean wind speed based on amplification factors and the

mean wind speed calculated by the WDI. It also

calculates the amplification zone areas, which

potentially exist between (Figure 4.11 (b)) and above

(Figure 4.11 (c)) the buildings.

176


Figure 4.11. Idealised 3D representation of the UBLsublayers.

In (a) d is the displacement height or zero-plane displacement, d+ z0 is the zero plane height [d0], and zH is the average height of

177


the buildings. In (c) z* is the height of the roughness layer ~ azH, where 2 < a < 5. Amplification zones potentially exist between(b) and above (c) buildings. Based in part on Fisher et al. (2004)

and Dalgliesh & Boyd (1962).

Each of the submodules will now be discussed, including

the associated data entry forms, analysis worksheets,

and database interfaces.

4.5.1 Wind Data Extrapolator

The wind data extrapolator (WDE) is the base module

of the UWEP DSS. On project launch from the Control

Centre, the WDE offers the user the option of

conducting an annual or a seasonal assessment and,

accordingly, guides the user through the following

steps:

locate the site (geographically), select the representative urban subregion, determine the mesoscale roughness length, determine regional wind statistics, and calculate the local friction velocity.

The seasonal option entails the following additional

steps:

identify a suitable meteorological station, collect hourly time series data (TSD) from the

station, fit the data to a Rayleigh pdf,

178


determine the roughness class of the station's terrain, and

calculate the regional wind statistics.

The data gathered through the above steps are inserted

into the appropriate fields of the Wind Statistics

worksheet, which calculates the mean wind speed, the

associated wind statistics, and the local friction

velocity for the site. Ideally, the average of the TSD

collected from three different meteorological stations,

triangulating the site of interest, should be used to

determine the regional wind statistics. The proposed

prototype UWEP DSS will only be capable of analysing

one year's worth of data from one meteorological

station. A more advanced user may choose to pre-

process data from three different stations and provide

the resulting averaged dataset to the UWEP DSS.

The user interface is the Site Specification form,

which includes the Project Details, Location,

Topographical, and Meteorological subforms or tabs

using the tabbed form-functionality within Excel.

Figure 4.12 depicts the interaction between the Site

Specification form, internal and external databases,

and the Wind Statistics worksheet within the WDE.

179


W ind Data Extrapolator (W DE)

M aps

OW RA

Atlas

NCD&IA

CW EA

W ind StatisticsW orksheet

Urban Param eterisation

Roughness Classification

Co-ordinates

Site SpecificationProject DetailsLocationTopographicalM eteorological

GeoCoder

Figure 4.12. The WDE submodule of the UWEP module.OWRA is the Ontario Wind Resource Atlas. CWEA is the Canadian WindEnergy Atlas. NCD&IA is the National Climate Data and Information

Archive.

The primary tab of the Site Specification form shown

below has fields for entering the name of the

assessment project and additional details and/or

comments.

180


Figure 4.13. The Project Details tab of the SiteSpecification form.

It is through this form that the user indicates whether

an annual or seasonal assessment is being conducted.

The Location tab depicted below is comprised of two

sections. If an actual site has not been selected,

selection of a city within the City section inserts its

co-ordinates, from the Co-ordinate Data table, into the

appropriate fields on the Wind Statistics worksheet.

If a subject site has been selected, the user fills out

the Site section and uses the external GeoCoder to

obtain the latitude and longitude co-ordinates of the

181


site. These co-ordinates are required to locate the

site in the atlases and for distance calculations.

Figure 4.14. The Location tab of the SiteSpecification form.

The Topographical tab (Figure 4.15) provides access

to the external mapping applications (i.e., Google

Maps and MapQuest) and contains the selectable urban

subregion options. Additional information on the urban

subregion categories is accessed through the Category

Details button. The user selects a representative

urban subregion by examining the satellite image with

consideration of the urban subregion characteristics.

The user is instructed to save the map images gathered

into the Projects folder for future reference.

182


Selection of a subregion inserts the subregion's

parameters, extracted from the Urban Parameterisation

table, into the appropriate fields of the analysis

worksheets.

Figure 4.15. The Topographical tab of the SiteSpecification form.

The Meteorological subform functionality depends on

whether an annual or seasonal assessment is being

conducted. There are actually two Meteorological tabs

with the appropriate one activated based on user-

selection of assessment type.

For an annual assessment, the Meteorological tab

links the user to the external wind atlases to gather

183


the wind statistics, as detailed below, at a height of

80 m based on the latitude and longitude co-ordinates

of the site. The user is instructed to save the wind

direction rose to the Projects folder for future

reference.

Figure 4.16. The Meteorological tab of the SiteSpecification form (annual).

If the seasonal option is selected, the seasonal

version of the Meteorological tab depicted below

assists the user in identifying a suitable

meteorological station and the mesoscale roughness

length through the Canadian Wind Energy Atlas (CWEA).

It then links the user to the National Climate Data and

Information Archive (NCD&IA) to download the time

184


series wind data into monthly statistics calculation

worksheets. For locations in Ontario, the OWRA can be

referenced to identify peak wind speed months, to be

referred to as the windy season. Time Series data

analysis within the UWEP DSS is based on either the

windy season or a year's worth of data. Given that a

twelve month sample size has been proven to yield

results accurate to within 5 - 7 % (AWS Truewind,

2007), this sample size is considered acceptable in

light of other uncertainties and to reduce data

handling. Using the Atlas and the co-ordinates of the

meteorological station provided by the CWEA or NCD&IA,

the user selects the land-use category most

representative of the meteorological station's terrain.

185


Figure 4.17. The Meteorological tab of the SiteSpecification form (seasonal).

Selection of a land-use category extracts the

corresponding local roughness length from the Roughness

Classification table of the internal database.

Once all the fields of the Site Specification form

have been filled, the user instructs the WDE to

populate the fields of the Project, Location,

Topography, and Meteorology sections of the Wind

Statistics worksheet and launches the WDI by clicking

the OK or Continue button. The OK button brings the

user back to the Control Centre from where the Wind

Statistics worksheet (Appendix I-1), including the

186


associated mean wind speed profile plot(s) and wind

direction rose(s), can be reviewed. Clicking the

Continue button launches the Wind Amplifier.

The calculations performed by the WDE within the

Wind Statistics worksheet primarily depend on whether

an annual or seasonal assessment is being conducted.

An overview of the main calculations is provided below,

with background theory, assumptions, and detailed

calculations made available through Appendix E.

For an annual assessment, the WDE extrapolates the

regional mean wind speed to the site by calculating the

maximum local friction velocity [u*] using the

logarithmic profile equation as follows: ,

where: (4.2)

is the regional mean wind speed obtained from

the wind atlas at height z,

z is the height of the top of the surface layer

(SL) [zSL] ~ 80 m,

d is the urban subregion displacement height,

z0 is the urban subregion microscale roughness

length, and

187


is von-Karman's constant ~ 0.4.

The WDE obtains z0 and d from the Urban

Parameterisation table based on the urban subregion

selected by the user.

If the seasonal option is selected, the WDE

develops the local wind statistics from a Rayleigh

distribution fit of the TSD obtained by the user from

the NCD&IA, for each of the twelve compass sectors.

These calculations are performed in the monthly

statistics calculation worksheets. Summaries are

posted to the Mean Seasonal and the Mean Monthly

worksheets, which are accessible through the Wind

Statistics worksheet. The mean wind speed is related

to the Rayleigh pdf parameter [b] through .

The regional mean wind speed and maximum local friction

velocity are then calculated using two different

methods depending on the proximity of the

meteorological station to the site.

The WDE considers the regional wind to be either

the mesoscale or the geostrophic wind. The mesoscale

and geostrophic winds are distinguished by the altitude

at, and the fetch over, which they are considered to

prevail. The mesoscale wind speed is defined as the

mean wind speed at the top of the SL [zSL] and is

188


relatively constant over a horizontal distance of

several kilometres. The geostrophic wind speed is the

mean wind speed at a height above the UBL (i.e.,

) and considered constant over horizontal

distances greater than 5 km (Rooney, 2001; Clarke &

Hess, 1974). For an annual assessment, the regional

wind speed is considered to be the mesoscale wind speed

provided by the wind atlases at an elevation of 80 m.

For a seasonal assessment, determination of the

appropriate regional wind speed depends on how far the

meteorological station is from the site.

If the meteorological station is less than 5 km

from the site, the regional wind speed is considered to

be the mesoscale mean wind speed, calculated by

extrapolating the local wind speed to the top of the SL

using the logarithmic profile: , where:

(4.3)

z is the height of the top of the SL;

z0 is the local roughness length provided by the

Roughness Classification table; and

u* is the friction velocity associated with the

mean wind speed and terrain at the meteorological

189


station, calculated using Equation 4.3 with z = 10

m and z0 = the roughness length associated with the

terrain at the meteorological station.

The displacement height [d] is not included in this

calculation based on the assumption that most

meteorological stations are in areas characterised by a

roughness length of less than 0.1 m. The site-

specific, local friction velocity is then calculated

from the regional, mesoscale mean wind speed in the

same fashion as in the annual option.

Use of the logarithmic profile to calculate the

friction velocity for the meteorological station and to

extrapolate values upward from within the RSL is not

entirely appropriate. To compensate for the inaccuracy

of the logarithmic profile below the RSL, attempts have

been made to calculate the maximum friction velocity

through development of a friction velocity profile

(Christen & Rotach, 2004). Unfortunately, extrapolated

wind speeds calculated using the maximum friction

velocity appeared unreasonably high. As such, to ere

on the conservative side, the local friction velocity

associated with the meteorological station is used

instead of the maximum. This practice was deemed

acceptable due to lack of alternative options given the

190


measurement height of the data provided by the climate

data archive and the fact that precedence to do so has

been well established (Verkaik, 2006; Petersen et al.,

1998a; de Wit et al., 2002).

If the meteorological station is more than 5 km

from the site the maximum local friction velocity does

not change, but the regional wind is considered to be

the geostrophic wind as opposed to the mesoscale wind

(Landberg & Watson, 1994) and calculated using the

geostrophic drag law as follows:

, where: (4.4)

A and B are empirically derived constants equated

to 1.8 and 4.5 for a neutral boundary layer,

respectively;

u* is the local friction velocity;

is the coriolis parameter of the meteorological

site; and

z0 is the mesoscale roughness length from the wind

atlas for the region.

Once the geostrophic wind speed is calculated, Equation

4.4 is again applied in an iterative fashion to

determine the local friction velocity at the site, with

191


& z0 pertaining to the location of the site vs. that

of the meteorological station.

There is very little in the literature regarding

the influence of the terrain on wind direction. What

is proposed, based on PBL theory, is that in

extrapolating wind direction from the outer layer to

the top of the surface layer (SL), backing (a counter

clockwise change in direction) reducing the veering

effect of the coriolis force, proportional to the

roughness length is observed (in the northern

hemisphere). The literature does not yield a more

precise definition of this correlation, but proposes

that wind direction remains constant with height within

the logarithmic profile layer (Verkaik, 1999). Since

the wind direction data are gathered from within the

SL, the WDE considers the wind direction to be constant

with height except when the geostrophic wind speed is

calculated for a seasonal assessment. In that case,

the geostrophic drag law constants (i.e., A and B),

mesoscale roughness length, and the calculated maximum

local friction velocity, are used to estimate the

direction of the wind at the surface for each of the

twelve compass sectors through: (Landberg

& Watson, 1994) (4.5)

192


In summary, the WDE:

collects (annual option) or calculates (seasonal option) the regional wind statistics;

develops the local wind statistics at the top of the SL, through calculation of the local friction velocity; and

launches the wind data interpolator (WDI).

The function of the WDI will now be discussed.

4.5.2 Wind Data Interpolator

The Wind Data Interpolator (WDI) is the second

submodule of the UWEP module. Its function is to

estimate the mean wind speed profile from the top of

the SL down to the ground. Provided with data from the

Wind Statistics worksheet generated by the WDE and the

Wind Amplifier, the WDI does not require additional

user-interaction. The function of the WDI is

essentially instantaneously completed within the Mean

Wind Speed Profile calculation worksheet when the user

clicks either the OK or Continue button on the Site

Specification form. The profile plot is posted to the

Wind Statistics worksheet. The WDI is established as a

submodule primarily to allow for future modifications

pertaining specifically to equations involving the

estimation of the mean wind speed profile.

193


The WDI is activated three times within the UWEP

module. The first time, the WDI calculates the mean

wind speed profile from the top of the SL down to the

ground, based on the local friction velocity calculated

by the WDE. The second time, the profile is calculated

from the top of the UCL both upward and downward based

on the amplified mean wind speed at the top of the UCL

calculated by the Wind Amplifier in the first stage of

its amplification assessment. Finally, the WDI

generates the profiles associated with the second stage

of the amplification assessment. This section provides

an overview of the key concepts pertaining to the WDI.

Background theory, assumptions, and detailed

calculations are made available through Appendix F.

The mean wind speed profile developed by the WDI is

comprised of the following three SL sublayer-specific

profile segments:

Mesoscale Transitional, and Microscale.

The mesoscale portion pertains to the ISL where z* < z

< zSL, while the microscale portion defines the profile

below the UCL where z zH. The transitional portion

defines the transition between the mesoscale and

194


microscale portions where zH < z < z*. Figure 4.18

depicts the mean wind speed profile as developed

through this multi-layer approach, compared to the

traditional logarithmic profile extrapolation.

195


Figure 4.18. The mean wind speed profile generated bythe WDI.

The profile segments are as follows: solid - logarithmic, diamond- mesoscale, long dash - transitional, and short dash -

196

zSL

zH

z*


exponential. zSL is the height of the top of the SL, z* is theheight of the RSL, and zH is the height of the UCL.

By establishing the complete profile, as opposed to

merely calculating individual mean wind speeds at

specific heights, the WDI provides the Wind Amplifier

submodule with the flexibility to determine the

appropriate reference wind speed for calculating the

amplified mean wind speed at various heights. The wind

speed profile plot is appended to the Wind Statistics

worksheet accessible through the Control Centre.

Activation of the Wind Amplifier submodule and

generation of the wind speed profile essentially occur

simultaneously. The function of, and the required user

inputs for, the Wind Amplifier submodule will now be

discussed.

4.5.3 Wind Amplifier

The Wind Amplifier submodule of the UWEP module

calculates the amplified mean wind speeds and the

characteristic dimensions and areas of the associated

amplification zones. As previously discussed in the

development of the Wind Amplification database, wind

amplification is assessed in two stages. In the first

stage, aerodynamic effect-induced amplification depends

on the configuration of the neighbourhood and the

197


primary and secondary wind directions. In the second

stage, it depends on the subject building and its

individual features.

The user interface within the Wind Amplifier is the

Building Aerodynamics form, including the Neighbourhood

Morphology, Aerodynamic Effects, and Building Feature

subforms. The interaction between the user form,

internal and external databases, and the analysis

worksheet within the Wind Amplifier submodule is

depicted below. The two stages of wind amplification

assessment, including the associated user-input

subforms and worksheets, will now be discussed.

Background theory, assumptions, and detailed

calculations are provided in Appendix G.

W ind Am plifier

W ind Am plificationW orksheet

Urban Param eterisation

Roughness Classification

W ind Am plification

M apQ uest

G oogle M apsBuilding Aerodynam icsAerodynam ic EffectsNeighbourhood M orphologyBuilding Features

Figure 4.19. The Wind Amplifier submodule of the UWEPmodule.

198


For the first stage of the amplification

assessment, the required neighbourhood configuration

and meteorological data are provided to the Wind

Amplifier by the Urban Parameterisation table, the Wind

Statistics worksheet, and the first two tabs of the

Building Aerodynamics form. The Neighbourhood

Morphology tab (Figure 4.20) assists the user in

determining

road classification, and neighbourhood configuration variables.

These characteristics are used by the Wind Amplifier to

reconfigure the inter-building spaces of the urban

subregion selected by the user in the first part of the

assessment.

199


Figure 4.20. The Neighbourhood Morphology form.

The road types, graphically represented below, are

based on the road classification scheme of the

Transportation Association of Canada (TAC) (1999).

200


Figure 4.21. The relationship of urban road typeclassifications.

(TAC, 1999)

Road classification is correlated to street canyon

width [W] within a sub-table of the internal Urban

Parameterisation table, based on the summation of road

type cross-sectional element widths (e.g., the parking

lane, travel lane, sidewalk, etc.). Illustrated below

are the typical cross-sectional elements of an urban,

local, single lane street with two-side parking.

201

LB

SC

W

Sm

SV


Figure 4.22. Typical cross-sectional elements of anurban road.

(Perry & Associates, 2001)

The neighbourhood configuration variables are grid

orientation [] and block

length [Lb], as illustrated.

The block length is the

length of the shortest

block. The longer block

length is calculated by the

Wind Amplifier using the

user-specified block ratio. The block ratio is the

202


ratio of the short block length to the long block

length, expressed as a percentage.

Figure 4.23. Reconfigured cubic array of category # 1.

The user estimates these

variables by referencing

the images of the site's

neighbourhood. These

variables, along with

primary wind direction []

from the Wind Statistics

worksheet and subregion-specific building dimensions

from the Urban Parameterisation table, are used to

estimate the number of streets within the grid and the

angle of flow incidence []. Estimation of the number

of streets within the grid enables the Wind Amplifier

to distinguish inter-building spaces (i.e., the inline

[SV] and orthogonal [SC] space widths) from street

canyon widths [W], as illustrated in Figure 4.23. The

angle of flow incidence [] is defined as the included

angle between the wind direction vector and the grid

orientation vector [], as depicted in Figure 4.24.

Figure 4.24. The critical angles.

The widths and orientations of the spaces dictate the

amplitude and nature of the potential wind

203


amplification. These calculations are performed within

the Amplification Calculations worksheet, which posts

summaries to the Wind Amplification worksheet.

The Aerodynamic Effects tab illustrated below

provides graphical representations of the aerodynamic

effects to assist the user in selecting those that may

be in effect within their site's neighbourhood.

Figure 4.25. The Aerodynamic Effects form.

204


To determine which effects may be applicable, the user

compares satellite images of the site's neighbourhood

obtained from Google Maps and/or MapQuest, with

consideration of the primary and secondary wind

directions, to the portrayed effect images.

Neighbourhood images stored in the Project folder

earlier in the assessment can be accessed through the

Project Folder button. Wind direction data are

obtained from the Wind Statistics worksheet accessed

through the button of the same name or retrieved from

the Canadian Wind Energy Atlas (CWEA).

For each user-selected aerodynamic effect, the Wind

Amplifier compares the calculated inter-building

spaces, angle of flow incidence, and the applicable

urban subregion morphological dimensions against the

critical morphological dimension- and angle of flow

incidence-limits tabulated within the neighbourhood-scale

portion of the Wind Amplification Factor table of the

internal database. If all criteria are met, the

maximum applicable amplification factor is used to

calculate the maximum local gust speed [ ]:

, where: (4.6)

is the maximum amplification factor, and

205


is the reference gust speed based on the

microscale mean wind speed [ ] at the mean

building height of the urban subregion, provided by

the WDI.

The mean wind speed and gust speed can be related to

one other through the wind speed frequency, pdf

parameter(s) (Nigim & Parker, 2007). The Wind

Amplifier provides the amplified local mean wind speed

to the WDI to recalculate the mean wind speed profile.

This step ensures that the appropriate reference wind

speed is available for the second stage of the

amplification assessment. The first stage essentially

recovers a portion of the wind speed lost through the

theoretical extrapolation of the regional wind speed

into the UCL (Christen & Rotach, 2004).

During the second stage, additional potential wind

amplification is assessed based on the dimensions of

the subject building. The wind amplification factors

pertaining to this stage are tabulated by building

surface zone within the Wind Amplification Factor table

of the internal database. For this stage of the

amplification assessment the Building Features tab

illustrated below assists the user in defining the

key dimensions of the subject building, and

206


roof type and orientation.

Figure 4.26. The Building Features form.

207


The key building

dimensions, as

illustrated, are:

eave height [He]or mid-roofheight [Hm],

along-wind depth[D],

cross-windbreadth [B],

buildingorientation [B] or roof ridge orientation [R],

roof type (i.e., flat, hipped, gabled, monoslope, and saw-toothed), and

roof pitch [].

Figure 4.27. Linear and angular building-specificvariables.

If building orientation is not equal to neighbourhood

grid orientation, the angle of flow incidence [] is

redefined as the included angle between the wind

direction vector and the building orientation vector,

as illustrated. The building dimensions are used by

the Wind Amplifier to extract the appropriate values

from the building-scale portion of the Wind

Amplification Factor table and to calculate the

208


corresponding reference height [h], defined within the

National Building Code (NBC). The reference height

determines the height of the reference wind speed,

provided by the WDI, used in the amplification

calculation.

As this appears to be the first attempt at using

published pressure and composite pressure-gust

coefficients to estimate microscale wind amplification,

the reader is directed to Appendix D – 3 for details

pertaining to the assumptions and simplifications

employed by the prototype DSS. The primary limitation

is that peak gusts, to which produce the composite

pressure-gust coefficients pertain, are temporarily

discrete and not necessarily related to the mean wind

speed.

The extracted amplification factors are used by the

Wind Amplifier to calculate the amplified gust speed [

] by building surface zone, defined in the Wind

Amplification Factor table, through , where:

(4.7)

is the reference gust speed, provided by the

WDI based on the neighbourhood-scale amplification

assessment, at the calculated reference height; and

209


is the applicable amplification factor from the

Wind Amplification Factor table.

The corresponding mean amplified wind speed [ ] is

calculated from the amplified gust speed through the

pdf parameter as detailed in the first stage of the

amplification assessment.

After defining the amplified mean wind speeds, the

Wind Amplifier defines the

limits and areas of the

amplification zones. The

calculation of the limits of an individual

amplification zone is primarily based on wake

development theory.

Figure 4.28. Plan (left) and elevation (right) wakezone developments.

The dashed lines represent the wake streamlines. Modified aftervan Bussel & Mertens (2005).

The idealised development of both plan (left) and

elevation (right) semi-elliptical wake zones (van

Bussel & Mertens, 2005) is illustrated above. The

210


following paragraphs outline the amplification zone

concept. Background theory, assumptions, and detailed

calculations are provided in Appendix G.

Wind amplification is thought to be limited to a

relatively thin band immediately outside of the wake

streamlines illustrated above. The thickness of this

band is approximately 20% of the associated

characteristic dimension (van Bussel & Mertens, 2005).

The characteristic dimension is the building height [H]

for zones above the building corresponding to the

linear dimension of DCt and half the cross-wind

building breadth [B] for

zones on either side of

the building corresponding

to DCs, as illustrated.

The appropriate cross-wind

areas are then calculated

as and

. These areas represent the constant cross-

wind areas of the amplification zones in the along-wind

direction immediately outside of the wake stream line.

The Wind Amplifier calculates the wake streamline

equations, characteristic dimensions, and amplification

zone areas (i.e. side and top) for the four building

211


walls. Detailed calculations are performed in the

Amplification Calculations worksheet, which posts

summaries to the Wind Amplification worksheet.

Figure 4.29. Amplification zones.

To determine whether to calculate a top area [AZt],

side area [AZs], or both, the Wind Amplifier uses the

slenderness ratio [H/B] of the subject building. Cook

(1990) proposes that when 2H/B > 1, the majority of the

wind flows around the sides of the building, while at

2H/B < 1 the majority flows over the building. At 2H/B

1 the wind flows both over and around the building

in a relatively evenly distributed fashion.

In summary, the Wind Amplifier calculates the

amplified mean wind speeds based on the gust speeds and

associated revised pdf parameters, the wake streamline

equations, and the amplification zone areas. Once all

the fields of the Building Aerodynamics form have been

filled, the user instructs the Wind Amplifier to

populate the fields of the Wind Amplification analysis

worksheet and launches the Energy Calculator by

clicking the OK or Continue button. Clicking on the OK

button brings the user back to the Control Centre from

where the worksheets can be reviewed. Clicking the

Continue button generates and opens the Summary Report

212


directly. The function of the Energy Calculator will

now be discussed.

4.5.4 Energy Calculator

The Energy Calculator submodule of the UWEP module

estimates the site's wind energy potential, based on

the amplified mean wind speeds and the amplification

zone areas calculated by the Wind Amplifier. This

submodule does not require a user interface since it

obtains the required data from the worksheets of the

Wind Amplifier and the WDE. All calculations are

performed within the Energy Calculations worksheet with

summaries posted to the Wind Energy worksheet. The

main concepts and equations of the Energy Calculator

will now be discussed. Background theory, assumptions,

and detailed calculations are available through

Appendix H.

The Energy Calculator calculates the theoretical

maximum extractable energy [E] for each building face

by zone (i.e. side and top) through:

, where:

(4.8)

Ei is the wind energy in kWh per time period;

213


i indicates an individual building face zone [AZi],

with each building face having two sides and one

top amplification zone (as illustrated in the last

section of the Wind Amplifier section above);

CP is the coefficient of performance limit (Betz

limit) ~ 0.593;

WPDi is the wind power density in W/m2;

AZi is the amplification zone area in m2;

f()i is the sum of the associated wind direction

frequencies; and

Time Period is one year or one windy season,

depending on whether an annual or seasonal

assessment is being conducted.

The WPD is calculated through: ,

where: (4.9)

is the Rayleigh pdf approximation of the wind

speed frequency distribution: ,

and (4.10)

u is the wind speed class associated with the

frequency.

214


For Equation 4.9, the Rayleigh parameter [b] in

Equation 4.10 is the revised parameter calculated from

the amplified gust speed [ ] by the Wind Amplifier.

The area associated with the building face-specific

amplification zone [AZ] is provided by the Wind

Amplifier. For side amplification zones, as the mean

wind speed varies with height, the pdf parameter

associated with the two-thirds-height mean wind speed

is used in the WPD calculation. Within the literature,

derivation of the Betz limit suggests that the maximum

energy actually available for extraction by a WECS is

approximately 59.3% of the maximum kinetic energy

(Johnson, 2004).

The UWEP module, which contains the WDE, WDI, Wind

Amplifier, and Energy Calculator submodules, stores all

of the data and calculations in the Wind Statistics,

Wind Amplification, and Wind Energy worksheets, from

which the Summary Report is created.

4.6 Summary Report

The Summary Report of the UWEP DSS is generated by

extracting key data and graphs from the Wind

Statistics, Wind Amplification, and Wind Energy

worksheets. The images for the report are obtained

from the Project folder. This report serves two

215


purposes. Primarily, it summarises the site assessment

by identifying the wind energy potential by zone for

the subject site. Secondly, it suggests building

orientation and roof configuration and pitch changes,

which could potentially increase wind amplification

resulting in an increase in the wind energy available

for extraction. If the user is in a position to make

such site adjustments, the Architectural Configurator

connects the user to the appropriate submodule forms

within the UWEP module to do so and rerun the

assessment.

4.7 Architectural Configurator

The Architectural Configurator provides access to

the forms specifically pertaining to building

orientation, roof configuration, and roof pitch data.

It is a structured sub-path within the overall

structure of the UWEP DSS, as illustrated.

Architectural Configurator

Building Aerodynam ics

W ind Am plificationW orksheet

W ind Am plification

W ind EnergyW orksheet

216


Figure 4.30. The Architectural Configurator module ofthe UWEP DSS

The Architectural Configurator can be launched from the

Summary Report or the Control Centre. Once the changes

have been made by the user within the Building

Aerodynamics form, the building-scale stage wind

amplification is reassessed by the Wind Amplifier. The

revised amplified mean wind speeds and amplification

zone areas and characteristics are provided to the

Energy Calculator to recalculate the building zone-

specific, potential wind energy. The Summary Report is

automatically updated.

Within this Chapter the development of the proposed

UWEP DSS was discussed. Lack of similar undertakings

within the literature suggests that this is the first

attempt to develop a prototype DSS in support of

quantifying building aerodynamics-induced wind

amplification for the purpose of estimating wind energy

potential within an urban setting. Unfortunately,

limited suitable data exist in support of validating

the proposed UWEP DSS. Fortunately, time series data

(TSD) collected from a rooftop location on the

University of Toronto's J.P. Robarts Library were made

available. In Chapter 5 the application of the UWEP

DSS to the Robarts Library site, including a comparison

217


of the collected TSD to the wind speeds and directions

calculated by the UWEP module, is discussed. As the

first of two case studies, the Robarts Library site

assessment is intended to lend credibility to the

results generated by the UWEP DSS. The second case

study involves application of the UWEP DSS to determine

the wind energy potential at the site of the Green

Venture Ecohouse.

218


5 APPLICATION OF THE DECISION SUPPORT SYSTEM

The prototype UWEP DSS outlined in Chapter 4 is

believed to be theoretically sound, including the

associated assumptions required to simplify this rather

complex undertaking. To determine whether the

developed tool can yield credible results requires

application of the UWEP DSS in the assessment of

various sites and subsequent comparison of the

calculated values against data collected at these same

sites. Unfortunately, very little suitable urban wind

speed and direction data appear to have been published.

Meteorological stations, for which several decades

worth of data are made readily available, are primarily

located at airports whose terrain characteristics are

not comparable to those of typical urban subregions

(e.g., the categories within the Urban Parameterisation

table of the UWEP DSS). Other urban meteorological

stations collect data at heights above the established

10 m reference height (e.g., Hamilton Royal Botanical

Gardens ID 6153300 ~ 18 m), don't collect wind speed

and direction data (e.g., Toronto City ID 6158355) (EC,

2005d), don't meet the United Nations World

Meteorological Organisation (WMO) Standards (e.g.,

Hamilton Psychiatric Hospital ID 6153298), or are

219


characterised by a combination of the aforementioned.

A tabulation of the historical data collection heights

by station ID for meteorological stations registered

with Environment Canada is provided in Appendix M.

Fortunately, a team of students lead by Barry Rawn

within the Energy Sustainability Community (ESC) of the

University of Toronto (U of T) has been conducting an

assessment of the available wind energy over top of the

Robarts Library on the St. George campus (Rawn, 2006).

In support of this assessment, two years worth of wind

speed and direction data have been collected. This

data, the dimensional details of the Robarts Library,

and preliminary computational fluid dynamics (CFD)

study findings were generously offered in support of

this research undertaking. Although much more data are

required to truly validate the UWEP DSS, this case

study was conducted as a first step in validating the

generated assessment results.

This chapter documents the capabilities of the UWEP

DSS through discussion of two case study applications.

220


Figure 5.1. Robarts Library (left) and the GreenVenture EcoHouse (right).

Photographs are credited to Bell (1978) for the Robarts Libraryand Peterson (2005) for the EcoHouse.

The first case study involves application of the UWEP

DSS in the assessment of the wind energy potential at

the site of the Robarts Library, including a peak

seasonal and a mean annual assessment. The discussion

includes comparison of the amplified mean wind speeds

to the data collected by the U of T's ESC team and a

comparison between the calculated wind power density

(WPD) and the estimate provided by the Canadian Wind

Energy Atlas (CWEA). The second case study pertains to

an assessment of the mean annual wind energy potential

at the site of the Green Venture EcoHouse in Hamilton,

including the selection, siting, and energy generation

potential of appropriate UWECSs.

The chapter is accordingly divided into two

sections pertaining to the institutional and

residential case study applications, respectively.

5.1 Institutional Case Study

The Robarts Library site was selected for the case

study primarily based on the availability of the wind

speed and direction data collected on site. The

221


suitability of an academic campus as a case study site

is further supported by the campus sustainability

movement, which appears to be underway. The Claudette

MacKay-Lassonde Pavilion of the University of Western

Ontario (University of Western Ontario, 2007) and the

Environmental Studies (ES) Green Building of the

University of Waterloo (Office of

Development, University of

Waterloo, 2007) are just a few

examples of green building

projects in support of campus sustainability. The

creation of a Sustainability Office within the Centre

of the Environment at the U of T is yet another example

of campus-based sustainable development. These

initiatives and the multi-disciplinary intellectual

capacity within academia, make a university campus an

ideal site for a case study. This case study will

hopefully act as a catalyst for similar undertakings.

Figure 5.2. The southeast side of the J.P. Robarts Library.In the foreground is the Thomas Fisher Rare Book Library Building;

the head of the peacock (Noyes, 2004).

This case study is documented through two main sections

corresponding to the mean annual and the peak seasonal

energy assessment, respectively. Application of the

222


UWEP DSS to assess the mean annual wind energy

potential at the Robarts Library site will now be

discussed.

5.1.1 Annual Wind Energy Assessment

The mean annual, wind energy assessment portion of

the case study is presented through the following

sections:

Mean wind climate, Amplified mean wind speed, Wind energy, and Summary Report.

These sections will summarise the site-specific user

form data that are provided to the UWEP DSS by the user

and will present tables and graphs summarising the

results at various stages throughout the assessment

process. The detailed calculations, performed within

the calculation worksheets of the UWEP DSS, are

provided as Appendix I - 4 on the compact disc (CD).

5.1.1.1 Mean Wind Climate

The first step of the assessment process involves

the WDE submodule in the determination of the local

wind statistics. During this step the user is required

to fill out the four subforms or tabs of the Site

223


Specification form. The following information is

entered into the fields of the Project Details tab:

Type: Annual Assessment Scope: Site Assessment Project Name: Case Study # 1 - JP Robarts Library Details/Comments: This case study involves the

first step towards validating the UWEP DSS, through the comparison of thecalculated results againstcollected data and CWEA windpower density (WPD) estimates.

In the Site section on the Location

tab, the user enters:

Name: Robarts Library building City: Toronto Province: Ontario Street Address: 130 St. George St. Postal Code: M5S 1A5

Figure 5.3. The John P. Robarts Research Library.(University of Toronto, 2006)

Providing the Geocoder with the address of the Robarts

Library yields the following latitude and longitude co-

ordinates, which the user enters into the corresponding

form fields:

Latitude: 43.664797 ° Longitude: -79.398693 °

224


On the Topographical tab the user is required to

compare the morphology of the campus to the urban

subregion descriptions and characteristics, to

determine which of the urban subregions is most

representative of the campus. To assist the user in

making this comparison, the satellite image of the

campus (Figure 5.4) is obtained by the user using the

satellite imagery application (e.g., Google Maps) and

the address of the Robarts Library. The user is

instructed to save the image to the Project folder of

the UWEP DSS. This comparison suggests that urban

subregion category # 11 is the most representative.

Category # 11 is an urban commercial district

characterised by an average building height greater

than five storeys, a plan area density of 28%, and a

frontal area density of 24%. Category # 13 and # 8,

city centre and mixed commercial / industrial,

respectively could also have been considered but were

dismissed due to their mean building height.

225


Figure 5.4. A satellite image of a portion of the St.George campus.

(Google, 2007)

On the Meteorological tab, the Canadian Wind Energy

Atlas (CWEA) and the Ontario Wind Resource Atlas (OWRA)

buttons link the user to the selected internet-based

wind atlas. For this case study the CWEA is selected

and provided with the site's co-ordinates to obtain the

wind statistics and associated plots at an elevation of

80 m, as illustrated. The user is instructed to save

the wind rose image (Figure 5.5 (b)) to the Project

folder.(a) Mean Wind Speed Histogram (b) Wind Direction Rose

226


Figure 5.5. The wind speed histogram and winddirection rose at 80 m

(EC, 2005a)

With reference to the Turbine formula tab of the Wind rose,

wind speed histogram, turbine formula at a point window, screen-

captured in Figure 5.6 below, the user selects CWEA

from the Data source field options and enters the

following information into the form fields:

Mean: 5.45 m/s Weibull scale parameter: 6.15 m/s Weibull shape parameter: 2.00

227


Figure 5.6. The wind statistics at 80 m (CWEA).

With reference to the data file associated with the

wind rose, available for downloading from the Rose and

histogram tab of the CWEA screen (Figure 5.6), the user

enters the wind direction frequencies for the site into

the compass direction fields, as tabulated below.

Table 5.1. The wind direction frequencies for the

Robarts Library site.

228


N 4.84% S 5.67%NNE 1.05% SSW 13.28%ENE 4.20% W SW 21.08%E 9.04% W 16.99%ESE 4.06% W NW 8.85%SSE 2.61% NNW 8.32%

In selecting the primary and secondary wind directions,

the user should reference the wind rose as opposed to

merely selecting the two directions corresponding to

the highest frequencies. Through grouping cardinal

sectors by quadrant, two to three dominant wind

directions can often be identified for most sites,

which otherwise appear to have relatively omni-

directional frequency distributions. The histogram

plot of the directional frequencies for the Robarts

Library, provided below, perhaps emphasises the tri-

modal nature of the distribution more distinctly than

the wind rose (Figure 5.5 (b)).

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

W NW NNW N NNE ENE E ESE SSE S SSW W SW WD irection

Frequency %

229


Figure 5.7. Wind direction frequency histogram of theRobarts Library site.

The gridded column represents the primary wind direction,including SSW & W, while the secondary wind directions, diagonal

line columns, are identified as NNW (including WNW & N) and E(including ENE & ESE).

With reference to the wind rose and consideration of

sector combinations as illustrated in Figure 5.7, the

user selects the following directions:

Primary wind direction: WSW Secondary wind direction: E or NNW

A comparison between wind statistics provided by the

OWRA and the CWEA for the Robarts Library site is

summarised in Table 5.2.

Table 5.2. Wind Statistics Summary.

Roughness Length

M ean W ind Speed

z0 (m ) u (m /s) k c (m /s) Prim ary SecondaryOW RA 2.50 4.98 1.97 5.62 SSW NCW EA 2.00 5.45 2.00 6.15 W SW E

Data Source

Elevation (m ) 80

W ind DirectionW eibull Param eters

Once the user has entered the mean wind speed pdf

parameters and the wind direction data, the WDE

calculates the friction velocity within the Wind

230


Statistics worksheet. The friction velocity and urban

subregion category # 11 parameters, stored in the Wind

Statistics worksheet, are used by the WDI to generate

the mean wind speed profile. A copy of the Wind

Statistics worksheet is provided in Appendix I - 1.

The plot of the mean wind speed profile, illustrated

below, is displayed on the Wind Statistics worksheet.

231


232


Figure 5.8. The mean wind speed profile at the RobartsLibrary

At this point the user could choose to continue on

to the wind amplification portion of the assessment or

return to the Control Centre to review the Wind

Statistics worksheet prior to continuing. The wind

amplification portion of the Robarts Library site

assessment will now be discussed.

5.1.1.2 Amplified Mean Wind Speed

During this stage the Wind Amplifier assesses the

potential wind amplification in two stages, requiring

the user to fill out the three subforms of the Building

Aerodynamics form. The first stage of the

amplification assessment re-establishes the true mean

wind speed based on the configuration of the

neighbourhood around the Robarts Library and the

primary and secondary wind directions.

With reference to the map image of the campus

(Figure 5.9), obtained through MapQuest and stored in

the Project folder, the user selects the compass

orientation that coincides with the direction of the

longest blocks in the street grid. If the grid

structure does not yield readily identifiable long and

short blocks, the most discernible orientation is

233


selected and 100% is entered into the block ratio

field. The block length is the average length of the

shorter blocks and the block ratio is the ratio of

short block length-to-long block length, both of which

are estimated using the scale on the map. These

variables define the morphology of the campus. In

summary, the following details are entered into the

appropriate fields on the Neighbourhood Morphology tab:

grid orientation [] NNW block length [Lb] 100 m block ratio 50%

Figure 5.9. Map image of the campus neighbourhood.

This form also contains the Road Classification section

from which the user selects both long block-end and

short block-end, representative street characteristics,

234


with reference to neighbourhood images, the road type

diagram on the form, and/or personal knowledge, as

tabulated below.

Table 5.3. Representative neighbourhood street

characteristics.Long Block Short Block

Region Type - Urban UrbanRoad Type - m inor arterial m ajor arterialLanes # 2 2Parking Lanes # one-side two-sideRoad W idth [W ] m 18.4 24.6

Sussex Ave. Huron St.Harbord St. St. G eorge St.

Characteristics

-Representative block-end Street Nam es

The road width is extracted from a sub-table within the

Urban Parameterisation table, based on the user

selected road classification characteristics. The Wind

Amplifier's redefinition of the inter-building spaces

and the angles of flow incidence for the Robarts

Library neighbourhood are tabulated below. These

values assist in determining whether a particular

aerodynamic effect could be producing neighbourhood-

scale amplification on campus.

Table 5.4. Neighbourhood inter-building spaces and

angles of flow incidence.

235


O rthogonal In-line M inim um Angle of Flow Incidence

SC SV Sm

(degrees)17.5 13.2 0.6 90

(m )

Inter-building Space W idth

The Aerodynamic Effects tab provides selectable

illustrations of the aerodynamic effects tabulated

within the Wind Amplification Factor table to assist

the user in identifying the potentially applicable

effects. Selection is based on a comparison between

the presented images and the images of the campus,

further enhanced by user-knowledge of the area. The

Wind Amplifier compares the critical values of the

user-selected aerodynamic effects, tabulated within the

Wind Amplification Factor table, against the campus'

inter-building spaces and angles of flow incidence

(Table 5.4) to determine the maximum applicable

amplification factor. Based on this comparison, the

tower effect is selected, resulting in the calculation

of the maximum local gust speed [ ] and amplified mean

wind speed [ ] as tabulated below. The amplified

mean wind speed is provided to the WDI to regenerate

the amplified mean wind speed profile, as illustrated

in Figure 5.10.

236


Table 5.5. Summary of the first stage of the

amplification assessment.MeanWindSpeed

Gustspeed

Amplification Factor

GustSpeed

Mean WindSpeed

(H) (H) (H) (H)(referen

ce)(referen

ce)(towereffect)

(localmaximum)

(amplified)

m/s m/s - m/s m/s1.39 2.22 1.82 4.05 2.54

All wind speeds are at the mean building height [H] of 21 massociated with urban subregion category # 11.

237


238


Figure 5.10. The amplified vs. the original mean windspeed profile.

The mean building height [H] of the representative

urban subregion is 21 m, while the eaves height [He] of

the Robarts Library is 80 m. As such, neighbourhood

morphology-induced wind amplification is not considered

to be present at the roof height of the Robarts

Library, but is considered to exist at heights below

twice the mean building height of the representative

urban subregion category (i.e. 2H = 42 m).

In the second stage of the amplification assessment

the user selects the appropriate roof type and provides

the following key building dimensions through the

Building Features subform:

building eaves height [He] = 80 m along-wind depth [D] = 102 m cross-wind breadth [B] = 91 m building orientation [B] = NNW roof ridge orientation [R] = N/A roof type ~ flat roof slope [] = 0

239


Figure 5.11. The plan and elevation views of theRobarts Library.

On user-selection of the flat roof type, the roof ridge

orientation field and roof pitch fields automatically

equate to N/A and 0, respectively. The appropriate

cross-wind breadth is identified by recalling that the

primary wind direction is out of the WSW, with

reference to the plan view of the Robarts Library. The

depth is then defined as the dimension perpendicular to

the breadth. The UWEP DSS transforms this triangular

building into its rectangular cross-wind equivalent, as

illustrated below.

240


Figure 5.12. The UWEP DSS cubic representation of theRobarts Library.

The block arrows indicate the building face designations assignedby the Wind Amplifier.

Using the information provided through the Building

Features tab, the Wind Amplifier selects the

appropriate amplification factor(s) from the Wind

Amplification Factor table of the internal database.

The pertinent portion of the flow chart of this

selection process is illustrated below, where I -15

corresponds to the pressure coefficient plot (Figure I-

241


15) from the National Building Code (NBC) (Canadian

Commission on Building and Fire Codes, 2006).

W hole Building

Building Structure

I - 15

~ 90

building surface

N/A

He 20 m No

No

Figure 5.13. The decision process of the WindAmplifier.

I - 15 corresponds to Figure I - 15 from Commentary I of the NBC.

The building surfaces, from this portion of the Wind

Amplification Factor table, are the edge-zones of the

three surfaces perpendicular to an associated windward

building face (i.e., the roof and both sides) as

illustrated below.

242


Figure 5.14. Isometric representation of the buildingsurface zones.

FS is the frontal stagnation point, above which the wind primarilyflows over the top of the building.

Table 5.6 summarises the amplification factors and

building-scale, amplified mean wind speeds for the side

and top surfaces of the four faces of the cuboidal

representation of the Robarts Library.

Table 5.6. Summary of the second stage of the

amplification assessment.MeanWindSpeed

Amplification Factor AmplifiedMean Wind Speed

243


(He) (h)(referen

ce)(side) (top) (side) (top)

m/s - - m/s m/s5.45 1.43 1.58 7.35 8.62

h is the reference height = He (eave height of the RobartsLibrary) for the top and the mid-point of the upper third of the

eaves height (i.e. 67 m) for the side.

Figure 5.15 provides a comparison between the amplified

reference profile and the additional amplification

along side and overtop of the building faces of the

Robarts Library.

244


Figure 5.15. The amplified wind speed profiles aroundthe Robarts Library.

In the final step of the amplification assessment

the Wind Amplifier submodule calculates the critical

245


slenderness ratio of the Robarts Library, the

characteristic dimensions, amplification zone areas,

and the equations for the semi-elliptical wake

streamlines. These calculations are made for each

building face, as tabulated below, based on the

building dimensions.

Table 5.7. Characteristic dimensions and amplification

zone areas.Slenderness Ratio [2H/B] : 1.76

Side Characteristic Dim ension

Top Characteristic Dim ension

Side Am plification Zone Area

Top Am plification Zone Area

D Cs D Ct AZs AZt

- m m m 2 m 2

NW 10.2 16.0 816 1632SW 9.1 16.0 728 1456SE 10.2 16.0 816 1632NE 9.1 16.0 728 1456

Building Face

The wake streamline equations position the

amplification zones in relation to the sides and the

top of the building (Cook, 1990). The wake streamline

plots are provided in Appendix J - 1b. The slenderness

ratio suggests that the wind will primarily flow around

the sides except near the top of the building above the

frontal stagnation (FS) point (Figure 5.14). The FS

point, typically at 2/3H, is the theoretical point at

which the free stream wind velocity is brought to rest,

corresponding to u = 0 m/s and CP = 1 (Cook, 1985).

246


The amplification zone areas are quite large due to the

size of the Robarts Library, which is directly

proportional to the characteristic dimensions used to

calculate these areas.

Ideally, appropriately dimensioned, modular UWECSs

would be mounted both on top and along side of the

Robarts library. The appropriate dimensions are

defined as the characteristic dimensions of the UWECS,

which would produce an area taking up as much of the

amplification zone area as possible. For a horizontal

axis wind turbine (HAWT) the characteristic dimension

is the rotor diameter [RD] and the area, or swept area

[AS], is equal to .

Vertical axis wind turbines (VAWTs) are typically

based on either the Savonius (Figure 1.11a) or the

Darrieus (Figure 1.11b) configuration. The

characteristic dimensions are

the rotor diameter and the

rotor height [RH]. The area of

a Savonius VAWT is equated to

. For a Darrieus-based

VAWT the area is estimated as

247


(D. L. Miller, 1981). Variations on the Darrieus

configuration have been developed, which straighten out

the sides of the classic egg beater shape.

Consequently, this configuration of VAWT is often

called an H-Darrieus (Mertens, van Kuik, & van Bussel,

2003a; Mertens, van Kuik, & van Bussel, 2003b) an

example of which is Turby®, illustrated in Figure 5.16.

Figure 5.16. Turby® - An H-Darrieus VAWT

The area of an H-Darrieus VAWT is (van Bussel,

Mertens, Polinder, & Sidler, 2004).

At the end of the amplification assessment the Wind

Amplifier returns the user to the Control Centre. No

further user input is required for the Energy

Calculator submodule of the UWEP module.

5.1.1.3 Wind Energy

The Energy Calculator calculates the theoretical

maximum extractable energy [Ei] for each building face

by zone [AZi] (i.e., side and top) using Equation

(5.1).

, where: (5.1)

Ei is the wind energy in Wh per time period;

248


i indicates an individual building face zone [AZi],

with each building face having two sides and

one top amplification zone (as illustrated in

Figure 5.14);

, where: (5.2)

CP is the coefficient of performance ~ 0.593

(Betz limit);

is the wind power density in

W/m2, with f(u) representing the wind speed

frequency distribution;

AZi is the amplification zone area in m2;

f()i is the sum of the wind direction

frequencies associated with the wind

directions considered orthogonal to a

particular building face; and

Time Period is one year or one windy season,

depending on whether an annual or seasonal

assessment is being conducted.

The theoretical maximum wind power density (WPD) and

wind power available for extraction at the Robarts

Library by building face and zone are summarised below.

249


Table 5.8. Wind power by building face and zone at the

Robarts Library.

W PD NW SW SE NEZone W /m 2

Side 442.8 47.2 98.2 26.4 27.3Top 707.2 150.6 313.5 84.5 87.3

Building Face

Pow er [P] (kW )

Values corresponding to the side zone are per-side values. The topis at 80 m and the side is at 67 m.

The values are higher within the top zones, since

wind speed increases with height and the top wind

amplification factor is larger than the one applied at

the side of the building. The variation between faces

is due to differing amplification zone areas and the

corresponding wind direction frequencies. As

previously mentioned, the power is based on the

idealised assumption that the UWECS can occupy the

entire amplification zone area. Based on the

aforementioned assumption, the theoretical maximum

extractable energy per building face and zone per year

is summarised in Table 5.9. An appropriately

configured UWECS could conceivably extract wind energy

from both sides of a building face, all faces at the

top edges, etc., in which case it would generate a

summation of the corresponding tabulated values for the

entire building. A comparison of the potentially

250


extractable energy vs. the energy needs of this

building is beyond the scope of this research

undertaking.

Table 5.9. Mean annual wind energy by building face

and zone.

NW SW SE NEZoneSide 413.1 859.9 231.6 239.3Top 1319.6 2746.6 739.8 764.3

Energy [E] (M W h per year)

Building Face

At this point the UWEP module has completed its

tasks within the assessment process and the user is

presented the Summary Report.

5.1.1.4 Summary Report

The Summary Report summarises the assessment

findings, including details pertaining to building

features and/or orientation changes that could

potentially increase the wind energy available for

extraction. The Summary Report in its entirety is

provided in Appendix J - 1.

Unfortunately, the changes suggested by the UWEP

DSS primarily involve roof type and pitch

modifications, which are limited to low-rise buildings

(i.e., He < 20 m) due to NBC data limitations. If low-

251


rise, roof-related NBC data can be considered

transferable to high-rise buildings, retrofitting the

Robarts Library with a sawtoothed roof could

conceivably increase the roof top amplification from

the maximum of 1.58 to 1.71. This would result in a

greater than 25% increase in available wind energy for

extraction. Based on the cubic representation of the

subject building, the building is sub-optimally

oriented. If its largest face, as opposed to its

smallest, were perpendicular to the primary wind

direction, the size of the amplification zone area

orthogonal to the primary wind direction would be

increased. An increase in amplification zone area

allows for a larger UWECS to be installed and more

energy to be extracted. Considering the true

triangular shape of the building, the prevailing winds

out of the SSW and E impinge on building corners. The

nature of the flow field above and around building

corners will be explored in the Discussion section of

this chapter.

One of the shortcomings of an annual assessment

using the wind atlases is that the atlases do not

correlate wind speed frequency to wind direction.

Traditional, tower mounted, horizontal-axis wind

turbines (HAWT) are equipped to align themselves with

252


the wind direction, but due to the size of modern day

turbines, additional external energy is most often

required to do so. Building mounted or augmented wind

turbines (BUWTs), if primarily based on the vertical

axis wind turbine (VAWT) configuration, can utilise

wind from all directions without needing to change

their position, but building-integrated systems are

usually bi-directional or unidirectional. In the

latter case, determination of which sectors yield the

most energy is of utmost importance. This is achieved

by using the UWEP DSS to conduct a seasonal assessment,

as will now be discussed.

5.1.2 Seasonal Wind Energy Assessment

The seasonal mean wind energy assessment of the

Robarts Library site primarily differs from the annual

assessment within the WDE submodule of the UWEP DSS,

affecting wind amplification and the resulting wind

energy potential. This portion of the case study is

accordingly presented through the following sections:

Mean wind climate, Wind Amplification, and Energy calculation.

Unlike an annual assessment, a seasonal assessment

allows the user to identify which months are

253


characterised by higher mean wind speeds and the

corresponding peak energy available for extraction

during these months. Throughout this discussion these

months will be collectively referred to as the windy

season. The time series data (TSD) available for a

seasonal assessment also allow for the correlation

between wind speed and wind direction to be determined.

Establishing this correlation yields more accurate

energy estimations since direction-specific mean wind

speed distributions are used in the calculations.

During an annual assessment, due to wind atlas

limitations, the UWEP DSS assigns the same wind speed

frequency distribution to each wind direction sector

when calculating the potential energy.

A seasonal assessment is based on mean hourly data

over a twelve month period. If the site is located in

Ontario the OWRA can be used to identify the windy

season, reducing the amount of data that need to be

downloaded from the climate data archive. The steps

specific to the seasonal assessment of the Robarts

Library site will now be discussed.


When a seasonal assessment is being conducted, the

initial differences are observed in the Meteorological

254


tab of the Site Specification form. This subform is

different from the one presented during an annual

assessment. For the Robarts Library site, the seasonal

assessment is simplified by determining which months

constitute the windy season using the graph obtained

through the OWRA, illustrated below.

With reference to the monthly mean wind speed

graph, the user selects: November, December, and

January from the drop down menus in the Windiest Months

section of the form. Accessing the CWEA, the following

additional information is entered:

Roughness Length: 2 m Station Name: Toronto Island Airport Latitude: 43.630 Longitude: -79.400

Figure 5.17. Monthly regional mean wind speedvariation.

255


Using the Atlas the land-use in the vicinity of the

meteorological station is approximated to 'Built

features, semi-rough (e.g., airport, including

buildings & trees)', which the Roughness Classification

table equates to a roughness length of 0.03 m. Using

the co-ordinates of the Robarts Library and those of

the Toronto Island airport, the WDE calculates that the

meteorological station is approximately 4 km from the

Robarts Library. As such, the wind statistics from the

meteorological station are used to calculate the

mesoscale wind statistics at 80 m.

The user obtains the mean hourly TSD recorded at

the meteorological station of the Toronto Island

Airport (TIA) from the National Climate Data &

Information Archive (NCD&IA) website. Pasting a

concatenation of the three months of wind speed and

direction data into the appropriate columns of the

Windy Season Statistics Calculations worksheet,

calculates the mesoscale regional wind speed for each

of the twelve cardinal sectors of the season and

produces the associated wind direction rose. The wind

direction rose for the windy season is illustrated

below. The mean wind speed during the windy season was

found to be 1.3X the mean annual wind speed, based on

256


the hourly TSD collected at 10 m at the TIA for the

year 2005.

0.0%5.0%10.0%15.0%20.0%25.0%

N

NNE

ENE

E

ESE

SSE

S

SSW

W SW

W

W NW

NNW

Figure 5.18. The windy season wind rose at the TorontoIsland Airport .

During a seasonal assessment the wind speed

frequency distribution is determined for each of the

twelve cardinal sectors for each month or the windy

season. This case study only involves assessment of

the windy season. Figure 5.19 illustrates the

variation in the sector-specific, wind speed frequency

distributions, by comparing ESE, the direction from

where the lowest mean wind speed is expected, against

ENE and WSW from where the highest mean wind speeds

originate. The wind speed distribution produced by

combining all directions, represented by All in Figure

257


5.19, yields a mean wind speed of ~ 21% less than the

highest mean wind speed associated with the WSW

direction. In other words, the mean wind speed out of

the WSW is ~ 1.26X the windy season mean wind speed.

Since wind energy is proportional to the wind speed

cubed, depending on the direction-specific frequency

distributions and the associated mean wind speeds,

using the windy season mean wind speed to calculate the

wind energy potential could result in an

underestimation by a factor of 2. The potential wind

energy calculated using individual direction-specific

wind speed distributions and the associated directional

frequencies was estimated to be ~ 6% greater than that

calculated based on the windy season mean wind speed

occurring 100% of the time.

258


0%

10%

20%

30%

40%

50%

60%

70%

0 1 2 3 4 5 6 7 8 9 10W ind Speed Class (m /s)

Frequency %

W SW ESE ENE All

Figure 5.19. The windy season wind speed distributionby direction at 21 m.

The WDI generates the mean wind speed profiles at

the site for all twelve of the cardinal sectors. The

mean wind speed profiles corresponding to winds out of

the ENE and WSW at the Robarts Library site are

depicted below. These cardinal sector-specific

profiles are used by the Wind Amplifier to determine

the appropriate reference wind speed during the wind

amplification assessment.

259


Figure 5.20. The ENE & WSW mean wind speed profiles.

260


The wind amplification portion of the seasonal energy

assessment at the site of the Robarts Library will now

be discussed.

5.1.2.2 Amplified Mean Wind Speed

In a seasonal assessment the WDI provides the Wind

Amplifier with monthly or seasonal cardinal sector-

specific, mean wind speed profiles. Neighbourhood-

scale amplification assessment is now month- or season-

and wind direction-specific. Building-scale

amplification assessment uses the average of the mean

wind speeds for the three wind directions considered to

impinge on each of the four faces of the cubic

representation of the Robarts Library. Since this

assessment is primarily interested in peak values, only

winds out of the WSW impinging on the SW building face

will be considered. Table 5.10 summarises the

reference amplified mean wind speed, building-scale

amplification factors, and resulting amplified top and

side mean wind speeds.

Table 5.10. Summary of the WSW amplification

assessment.MeanWindSpeed

Amplification Factor AmplifiedMean Wind Speed

261


(H) (h)(referen

ce)(side) (top) (side) (top)

m/s - - m/s m/s4.90 1.43 1.58 14.17 16.61

The reference height [h] is 80 m for the building-top zone and 67m for building-side zone.

The amplified reference profile and the along side and

overtop profiles resulting from building-scale

amplification associated with the SW building face of

the Robarts Library are depicted in Figure 5.21 below.

262


Figure 5.21. The WSW peak seasonal and amplified windspeed profiles.

The amplification zone areas and wake streamline

equations are the same as in an annual assessment. The

263


peak seasonal potential wind energy at the Robarts

Library site will now be discussed.

5.1.2.3 Wind Energy

In a seasonal assessment, wind energy is calculated

based on a unique wind speed distribution for each of

the twelve cardinal sectors. Since the case study is

interested in peak seasonal values, the wind power

density (WPD), power, and energy per zone associated

with winds out of the WSW are tabulated below. The

power and energy calculations take into account the

frequency of winds originating out of the WSW.

Table 5.11. The peak season contributions from the WSW

wind direction.

W ind Power Density

Power Energy

W PD [P] [E](W /m 2) (kW ) (M W h / yr)

Side 2037.5 186.9 1637.3Top 2303.8 422.7 3702.7

Zone

Building FaceSW

The peak mean seasonal energy potential associated

with winds out of the WSW overtop of the SW face of the

Robarts Library is approximately 3X greater than the

mean annual associated with the same wind direction.

264


Given that wind energy is proportional to the wind

speed cubed, this finding suggests that the windy

season mean amplified wind speed is ~ 1.4X the mean

annual wind speed. Considering that the windy season,

mean wind speed without amplification was previously

determined to be 1.3X the mean annual wind speed and

that 2 - 3X amplification has been found to exist, the

potential peak mean seasonal energy estimation is quite

reasonable.

The Summary Report for a seasonal assessment is

primarily the same as the one generated for the annual

assessment, except that it provides seasonal and/or

monthly estimates vs. mean annual. This added level of

detail is considered crucial in the development of

hybrid solar-wind energy conversion systems, since wind

speeds are typically higher during seasons when solar

radiation levels are lower. It also assists in

determining the most promising cardinal sectors for

uni- or bi-directional, building-integrated UWECS

installations.

This concludes the application of the UWEP DSS to

the case study of the wind energy potential at the site

of the Robarts Library. Prior to delving into the

second case study, a discussion regarding the validity

265


of the results generated by the UWEP DSS application is

provided. In the following section, a comparison

between the amplified mean wind speed as calculated by

the UWEP DSS and the data provided by the U of T team

will be discussed, including directional persistence

and seasonal influences. The section also includes a

comparison between the calculated wind power density

(WPD) and the estimated WPD proposed by the wind

atlases and the limitations of the UWEP DSS with

respect to the case study application.

5.1.3 UWEP DSS Credibility Investigation

This discussion section is primarily intended to

lend credibility to the results produced by the UWEP

DSS in assessing the wind energy potential at the site

of the Robarts Library. To that end, the time series

data (TSD) collected by the roof-mounted anemometer and

wind direction vane were analysed to determine the mean

wind speeds and dominant wind directions on the roof of

the Robarts Library.

The NRG#40 anemometer and NRG #200P direction vane

were mounted near the mechanical room of the southwest

penthouse, approximately 9.8 m above the main roofline

of the Robarts Library. The accuracy of the anemometer

is 0.1 m/s for wind speeds ranging from 5 m/s to 25

266


m/s and undefined below 4 m/s (Gee, 2007). The

accuracy of the wind direction vane is 3.6. In

Figure 5.22, the location of the instrumentation is

indicated by the star in the satellite image plan view.

The arrow on top of the Bahen Centre is from where the

photograph was taken, with the cross hairs indicating

the approximate location of the anemometer. The

simplified elevation sketch corresponds to the

photograph, with the mechanical room portion of the

southwest penthouse detailed in the lower left image.

267


Figure 5.22. Instrumentation placement on thesouthwest penthouse.

This discussion is presented through the following

sections:

Mean wind climate, Wind power density, and Limitations.


The analysis of the time series data (TSD)

collected by the roof-mounted anemometer proved

problematic. Unfortunately, even though mean minutely

data were collected for almost a year and a half, no

data were collected through the summer during either

year and some months were too incomplete to be

considered in the comparison. With the summer season

data missing, the mean annual wind speed could not be

determined from the data provided. As such, the

268


comparison is limited to the months identified as

corresponding to the highest mean wind speed. Figure

5.23 depicts the mean wind speed by month as calculated

from the TSD collected during 2005 and 2006.

0

1

2

3

4

5

6

7

Jan Feb M ar Apr M ay Jun Jul Aug Sep O ct Nov D ec

Mean Wind Sp

eed [u] (m/s)

2005 2006-Inc. 2005-Inc. 2006

Figure 5.23. Robarts Library Monthly mean wind speed.Chart Notes: No data were collected in 2005 between Apr and

August, inclusive. Data collection stopped at the end of May in2006. In 2005 the incomplete months (diagonal-hatch) were Feb and

Mar at 2.8% and 0.1% complete, respectively. In 2006 theincomplete months (horizontal-hatch) were Jan and May at 33.3% and

27.0% complete, respectively.

Considering only the months with greater than 80%

complete data sets, November, December, and February

were identified as the months during which the highest

mean wind speeds occurred. The mean wind speed for

February was greater than that for January, identified

by the OWRA in the seasonal assessment, possibly due to

the fact that the data set for January was only 81.2%

269


complete. The data sets for all the other months

plotted in Figure 5.24 were greater than 95% complete.

The mean wind speed during the windy season was

calculated to be 5.62 m/s.

0

1

2

3

4

5

6

7

Jan Feb M ar Apr Sep O ct Nov D ec

Mean Wind Sp

eed [u] (m/s)

~ 85% W indy Season > 95%

Figure 5.24. Monthly mean wind speed (>80% completemonths).

The column fill indicates: less than 85% complete data set (hatch-marks), greater than 95% (grey), and highest mean wind speed

(black).

Wind direction data were limited by the same

findings as previously detailed pertaining to wind

speed data. Defining the windy season as the three

months during which the highest mean wind speeds were

recorded, as identified in Figure 5.24, the wind

direction rose illustrated below suggests that the

270


primary and secondary wind directions during this

season are NNW/N and S, respectively.

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%N

NNE

ENE

E

ESE

SSE

S

SSW

W SW

W

W NW

NNW

W indy Season TIA veered 90 deg

Figure 5.25. The windy season wind rose vs. TIA dataveered 90 degrees.

Though the same three dominant directions were

identified through analysis of the Toronto Island

Airport (TIA) windy season data, the primary direction

identified from the Robarts Library data is the

direction identified as tertiary at the TIA. There are

several possible explanations for this observation.

Firstly, measurement instrument resolution, the

number of cardinal sectors into which data are

categorised, and definition of category limits can have

271


a profound effect on the identification of primary and

secondary wind directions. Figure 5.26 illustrates how

categorising the wind direction frequencies into four

inter-cardinal directions, as opposed to twelve, NW

emerges as the secondary wind direction for the TIA and

CWEA data vs. E or ENE.

0.0%

5.0%

10.0%

15.0%

20.0%25.0%

30.0%

35.0%

40.0%

45.0%

50.0%

SW NW NE SEDirection

Frequency %

TIA W indy Season RL W indy Season CW EA Annual

Figure 5.26. Wind direction frequency comparison.The abbreviations are TIA = Toronto Island Airport, RL = Robarts

Library, and CWEA = Canadian Wind Energy Atlas.

The Robarts Library wind direction data are in degrees

to four decimal places and the Toronto Island Airport

data are in tens of degrees. The UWEP DSS categorises

TSD into twelve cardinal sectors, centering each sector

within the category limits (i.e., 15 N > 345).

With the TIA data provided in tens of degrees, it is

conceivable that directional data may have been skewed

272


by 5 due to the category limit definitions of the UWEP

DSS. The CWEA groups the frequencies into twelve

cardinal sectors, but does not provide details

regarding category limit definitions.

Secondly, if the TIA wind rose for the same three

months is veered 90 (clockwise) or backed 270, a very

similar wind direction frequency distribution can be

created, as depicted in the dashed-line wind rose

overlay in Figure 5.25. In the northern hemisphere an

increase in friction, represented by an increase in

roughness length in wind power meteorology, causes a

counter-clockwise change in the wind direction or

backing. This finding suggests that the net change in

roughness length between the airport and the Robarts

Library and the height differential (i.e., 10 m at the

airport vs. 80 m at the library) may have combined to

result in either veering or backing of the wind

direction at the data collection site.

Lastly, perhaps the data were collected within

penthouse- and/or building corner-induced wakes below

the wake streamline, characterised by turbulence and

vortices. The reduced mean wind speed observed during

the windy season on the roof of the Robarts Library

compared to that calculated by the UWEP DSS from the

273


Toronto Island Airport (TIA) TSD at 80 m, suggests that

the measurement location was perhaps suboptimal within

the aforementioned wake zones.

0%

2%

4%

6%

8%

10%

12%

14%

16%

0 2 4 6 8 10 12 14 16 18 20 22 24 26W ind Class (m /s)

Frequency %

Robarts W indy Season @ 90 m

TIA W indy Season @ 80 m

Figure 5.27. The windy season wind speed distributioncomparison.

TIA data are from the Toronto Island Airport meteorologicalstation at 80 m, while Robarts data are at ~ 90 m.

Considering only the primary and secondary wind

directions, the Robarts Library data once again yield

slightly lower mean wind speeds even though the data

were collected at ~ 89 m, while the TIA data were

measured at 10 m at the Toronto Island Airport (TIA)

and extrapolated to 80 m.

274


0%

2%

4%

6%

8%

10%

12%

14%

16%

0 2 4 6 8 10 12 14 16 18 20 22 24 26W ind Class (m /s)

Frequency %

Robarts Prim ary W ind Direction

Robarts Secondary W ind Direction

TIA Prim ary W ind Direction

TIA Secondary W ind Direction

Figure 5.28. The primary & secondary direction windspeed distributions.

The profiles associated with the primary and secondary winddirections are represented by the solid lines and dashed lines,

respectively, with diamond markers for the TIA-based data. The TIAdata are the 80 m un-amplified, while the Robarts data are at ~ 90

m.

The supposition that the data were collected below

the wake streamline is thus far supported by the wind

direction discrepancies and the 89 m mean wind speeds

being lower than even those observed at an elevation of

10 m at the Toronto Island Airport. Further support of

this assumption is provided by the wake development

theory-based stream line equations. Considering winds

out of the WSW and the position of the anemometer at

47.4 m downstream of the roof's edge, wake development

theory suggests that the height of the wake stream line

275


at the location of the anemometer is ~ 17.5 m above the

eaves height of the Robarts Library. The anemometer

collecting the data was mounted at ~ 10 m above the

eaves height, conceivably well within the wake.

Initial computational fluid dynamics (CFD)

investigation of a portion of the Robarts Library

subjected to 10 m/s winds out of the SSW (Figure 5.29)

also indicates that the anemometer may be situated

within the turbulent region of initial flow separation.

In the side view of the Robarts Library CFD model

(Figure 5.29 b)), the instrumentation is possibly

located in the green area where a 30% reduction in the

free stream wind speed is expected.

(a) Frontal view.

276


(b) Side View

Figure 5.29. Computational fluid dynamics modelscreated by CFX.

The free stream wind speed was set to 10 m/s for the computationalanalysis. Wind amplification up to 1.4X the free stream wind speedwas predicted by the model above and to the sides of the RobartsLibrary when subjected to skewed flow. The cross hairs on the

inset image represent the approximate location of the anemometerand direction vane (Rawn, 2006).

277


Table 5.12 summarises the mean annual and windy

season mean wind speeds and directions, pertaining to

the case study. The OWRA and the CWEA provide mean

regional values, averaged over 1 km and 5 km,

respectively. The mean wind speed estimates provided

by the atlases differ from each other by ~ 10%. The

UWEP DSS values pertain to the calculated, estimated

mean wind speeds within the amplification zones. The

TIA (Toronto Island Airport) and the Robarts mean wind

speeds (bolded) were calculated based on a Rayleigh pdf

fit of provided data. Based on the TIA data, the mean

windy season wind speed was estimated to be 1.3X the

mean annual. This value is provided as reference for

mean windy season vs. mean annual comparison.

Table 5.12. Mean wind speed and direction summaries.

10 80Source Prim ary SecondaryO W RA 2.20 4.98 SSW NCW EA - 5.45 W SW E

UW EP DSS 2.84 8.62 W SW ETIA TSD 6.15 - W SW ENE

Robarts TSD - 5.62 t NNW S1.3X Am plified M ean Annual 3.69 11.21 - -

UW EP DSS 4.34 13.19 W SW ENE

Height [z] (m )W ind Direction

M ean Annual

M ean W indy Season

M ean W ind Speed [u] (m /s)

ŧ This value is based on a Rayleigh pdf fit of data collected at89 m.

278


In summary, the mean annual wind speeds within the

amplification zones of the Robarts Library, at ~ 1.5X

the un-amplified average regional wind speeds provided

by the atlases, appear reasonable. Additionally, the

UWEP DSS windy season mean wind speed within the

amplification zones at ~ 1.9X the OWRA estimate of the

mean wind speed at 10 m is conceivable on consideration

of published findings regarding pedestrian level wind

speed amplifications ranging from 2X to 4X (Coceal &

Belcher, 2005) and the finding that the windy season

mean seasonal wind speed was 1.3X the mean annual. The

mean wind speed at the TIA is higher than the mean wind

speeds calculated by the UWEP DSS and that provided by

the atlas primarily due to the relatively open terrain

at the airport and the proximity of the lake. The

relative agreement between the mean wind speed

calculated from the windy season portion of the Robarts

Library TSD at 89 m and the CWEA at 80 m is considered

coincidental.

Though 13.19 m/s is ~ 18% higher than the 11.21 m/s

estimate based on 1.3X mean annual amplified wind

speed, given the 10% difference between wind speeds

provided by the atlases for the same elevation it is

conceivable that the actual error is closer to 8%.

Additional consideration should be given to the fact

279


that the mean seasonal assessment was based on the TIA

data collected at 10 m, while the mean annual

assessment was based on the CWEA estimate at 80 m.

Noting that the mean wind speed recorded at the TIA at

10 m is 2.8X that estimated by the OWRA further

emphasises the implications of the use of area-averaged

mean wind speeds and logarithmic extrapolation of the

mean wind speed profile. There is additional error

associated with the instrumentation at the TIA,

parameterisation, and significant digits of values used

in all calculations. Case-specific studies of the

over- and under-predictions produced by the wind

atlases, of up to 40%, are provided by Pinard et al.

(2005) and Yu et al. (2006).

Another way to lend credibility to the windy season

amplified mean wind speed estimates produced by the

UWEP DSS is to consider the OWRA-based projections.

Based on the regional mean wind speed at 80 m provided

by the OWRA and the finding that the windy season mean

wind speed was 1.3X the mean annual wind speed, the

mean wind speed during the windy season could have been

estimated to be ~ 6.47 m/s, as opposed to the 5.62 m/s

measured by the roof top mounted anemometer. If the

instrumentation had been placed within the building

aerodynamics-induced amplification zones, it is then

280


conceivable that given an average amplification of ~ 2

a mean wind speed of ~ 13 m/s could have been observed.

The discussion will now proceed to an exploration of

the wind power density as estimated by the UWEP DSS vs.

the applicable Canadian wind atlases.

5.1.3.2 Wind Power Density

Wind power density (WPD) is the ultimate indicator

of wind energy potential. As WPD is primarily a

function of the wind speed distribution, WPD

calculations require determination of the appropriate

statistical distribution and definition of its

parameters. The UWEP DSS fits all data to the Rayleigh

distribution, whose statistical parameters can be

related to the mean wind speed. Table 5.13 summarises

the mean annual wind speeds and WPDs from the wind

atlases as compared to those calculated by the UWEP DSS

during the mean annual assessment.

Table 5.13. Mean wind speed and associated wind power

density comparison.

281


Height(m ) Side Top Side Top10 2.20 2.84 - 15.54 25.6 -30 3.46 3.93 4.61 - 57.65 109.4 -50 4.09 4.58 6.61 - 92 322.4 -80 4.98 5.45 7.80 8.62 158.79 528.3 707.2100 5.34 8.34 9.22 195.44 642.9 854.4

(m /s)Average W ind Pow er

(W /m 2)Average W ind Speed

O W RA UW EP DSS UW EP DSSO W RACW EA

The wind speeds and associated WPDs calculated by the

UWEP DSS specifically pertain to the amplification

zones, while the wind atlases provide mean regional

values at resolutions of 52 km (CWEA) and 12 km (OWRA).

CWEA only provides low resolution contour plots of

regional wind power density, as illustrated in Figure

5.30.

282


Figure 5.30. Wind power density at 80 m (CWEA) withToronto inset.

The star represents the approximate location of the U of T campus.The wind power density in this region at 80 m is 0 W/m2 > WPD >

200 W/m2.

In summary, the CWEA suggests that the WPD could

conceivably be 25% higher than that estimated by the

OWRA, as one would expect given that wind power is

proportional to the wind speed cubed and that the CWEA

mean wind speed estimate is ~ 1.1X that provided by the

OWRA. The OWRA should be used to assess locations in

283

100

W /m 2

500

300


Ontario as opposed to the CWEA, given its increased

resolution (i.e. 12 km vs. 52 km). The UWEP DSS

proposes that ~ 4.5X higher WPDs could exist within the

building-top amplification zones, corresponding to a

mean wind speed amplification of ~ 1.65.

5.1.3.3 Limitations

The building aerodynamics methodology of the UWEP

DSS is based on cubic array representations of typical

urban subregions for local-scale amplification

assessment and a cubic representation of the subject

building for microscale amplification assessment. Both

the triangular shape and the bridged-sections of the

Robarts Library introduce complexity beyond the

capabilities of the UWEP DSS. Recalling Figure 5.12,

the UWEP DSS idealises the northwest point of the

library as a flat projection of its frontal area. The

true nature of the flow field on either side of this

point in the side zones could be less turbulent and

exhibit greater amplification due to reduced flow

separation.

At the roof level, when the prevailing winds are

perpendicular to any of the three building corners, the

development of delta wind vortices could conceivably be

expected along the edges of the roof as idealistically

284


illustrated below. The effect of the size and location

of these vortices on the magnitude of the wind

amplification and location and size of the

amplification zones, is beyond the scope of this

research undertaking. Interested readers are directed

to the works of Banks, Meroney, Sarkar, Zhao, & Wu

(2000).

Figure 5.31. Idealised representation of conicaldelta-wing vortices.

(Banks et al., 2000)

Finally, it is conceivable that greater than calculated

wind speeds could be

experienced under the

bridged extensions on

either side of the

library's east corner,

connecting the library

285


to the Thomas Fisher Rare Book Library and the Claude

T. Bissell Building, respectively.

Figure 5.32. The extensions of the Robarts Library.The Claude T. Bissell Building is on the right and the Thomas

Fisher Rare Book Library Building on the left (Friedland, 2002).

Having established the relative credibility of the

amplified wind speeds and associated potential wind

energy at the Robarts Library as produced by the UWEP

DSS through the institutional case study application,

the residential case study application will now be

discussed.

5.2 Residential Case Study

The Green Venture EcoHouse in Hamilton, which

serves as an environmental technologies and sustainable

living practices demonstrator, was considered to be an

ideal site for the residential case study. The

Ecohouse, known prior to April 2002 as the Veevers

Estate or Glen Manor, is located in east end of

Hamilton at 22 Veevers Drive, in the neighbourhood of

the Quigley Road / Greenhill Avenue intersection. It

is a pre-Confederation, circa 1850, 1½ storey stone

farmhouse on 1.5 acres of land. Extensive renovations,

which began in 1968, transformed the original row of

286


upper storey windows into 5 individual dormers, added a

two-storey wing and garage in the southeast corner, and

a glass conservatory in the back (Green Venture, 2005).

Late 1960s 1997

Figure 5.33. The Green Venture EcoHouse - Then & Now.Late 1960s picture is from the Veevers' personal collection (GreenVenture, 2005) and the 1997 picture is from a Hamilton Spectator

article (1997).

Given the mandate of the EcoHouse and the new Power

Shift program, designed to facilitate economic

development around local Renewable Energy and Energy

Efficiency products, services, processes and

technologies (Green Venture, 2007), the outcome of this

case study could conceivably serve as a first step in

determining how best to showcase wind energy

technologies at the site of the EcoHouse.

287


This section documents the application of the UWEP

DSS for the assessment of the mean annual wind energy

potential at the site of the EcoHouse. Documentation

of the case study will primarily focus on the

estimation of the energy, which could potentially be extracted on site;

selection, siting, and energy generation of suitable UWECSs; and

demonstration of the Architectural Configurator.

The detailed data entry steps, having been thoroughly

documented in the first case study, are not included.

5.2.1 Mean Annual Energy Potential

To assess the potential wind energy available for

extraction, the UWEP DSS developed a rectangular

approximation of the irregular foot print of the

EcoHouse and assigned cardinal sector labels to the

four building faces, as illustrated.

288


Figure 5.34. The UWEP DSS representation of theEcoHouse.

Extracted from Peterson’s (2004) Landscape Master Plan.

The mean annual energy estimated by the UWEP DSS is the

maximum possible kilowatt hours per month that could

potentially be extracted by hypothetical UWECSs mounted

on the roof or beside the EcoHouse in the building

face-specific amplification zones, as summarised in

Table 5.14.

Table 5.14. Wind energy by building face and zone at

the EcoHouse.

289

WNW

SSW

ESE

NNE


W NW SSW ESE NNEZoneSide 2.2 21.9 6.0 7.6 37.7Top 17.4 173.0 47.2 59.8 297.4

Building Face

Energy [E] (kW h per m onth)Total

A hypothetical UWECS is one whose characteristic or

swept area [AS] is equal to the corresponding

amplification zone area and whose coefficient of

performance [Cp] is equal to the Betz limit (i.e.,

59.3%). The building face-specific zone areas are

summarised in Table 5.15.

Each building face is assessed separately due to

the discrete locations of the amplification zones.

That is to say, wind speed amplification occurring on

the left side of the WNW face cannot apply to winds

originating out of the SSW. It is conceivable, though,

that a UWECS mounted outside the wakes to the left of

the WNW face could be subjected to the amplified winds

originating from the NNE and SSW. There is also the

potential of an appropriately configured UWECS mounted

on the roof to harness the amplified wind energy from

all directions. The advantage of building face-

specific energy estimations is that they allow for

identification of the best site to install a building

integrated UWECS, which can only extract energy from

wind originating out of a portion of the cardinal

290


range. Ideally, such a device would be installed so as

to harness the wind originating from the most frequent

direction (i.e., SSW) and/or the wind direction

associated with the highest mean wind speed. The

frequency of the wind orthogonal to each of the four

building faces is summarised in Table 5.15.

Table 5.15. EcoHouse building face-specific assessment

summary.M ean W ind Speed [u] (m /s) 1.69 2.67

Side Am plification Zone Area

Top Am plification Zone Area

Side Characteristic Dim ension

Top Characteristic Dim ension

[AZs] [AZt] [D Cs] [D Ct]- % m 2 m 2 m m

W NW 7 13.5 27.0 1.5 1.8SSW 55 17.1 34.2 1.9 1.8ESE 19 13.5 27.0 1.5 1.8NNE 19 17.1 34.2 1.9 1.8

Building Face Incident W ind Frequency

Based on the estimates summarised in Table 5.14 and

with reference to the wind direction rose site overlay

provided below, the following UWECS installation

sitings are considered to potentially generate the most

energy:

Roof ridge mounted, and/or Ground mounted on either side of the SSW face.

291


Figure 5.35. EcoHouse site wind direction roseoverlay.

UWEP DSS wind rose overlay on Peterson’s (2004) EcoHouse landscapeplan.

The following section explores appropriate UWECSs for

the aforementioned installation configurations,

estimates their energy production capacity given the

projected conditions at the EcoHouse, and compares

energy production vs. average household need.

292


5.2.2 UWECS Selection and Placement

When endeavouring to select the most appropriate

WECS it is recommended that the characteristics of the

device be matched to the wind statistics at the site as

opposed to the perceived energy demand of the site

(Jangamshetti & Guruprasada Rau, 2001). In other

words, rather than selecting a WECS based on its power

rating, emphasis should be placed on rated, cut-in, and

cut-out wind speeds in relation to the mean wind speed

and wind speed distribution at the installation site.

The rated wind speed is typically the design point of

the WECS, corresponding to its highest efficiency and

rated power. The peak efficiency of a WECS, typically

referred to as the power coefficient or coefficient of

performance [Cp], cannot exceed the Betz limit of ~

59.3%.

As illustrated in Figure 5.36, wind energy

extraction devices are designed for optimal performance

over a range of tip speed ratios centred on the design

point ratio. The graph illustrates the relationship

between the power [Cp] and torque [CM} coefficients and

the tip speed ratio [] for several configurations of

wind energy conversion device. The tip speed ratio is

defined as

293


, where (5.3)

is the frequency of rotation in Hz.

The rotational frequency is a function of machine

design and wind speed. For example, the 3.5 kW

Cleanfield Energy VAWT with a rotor diameter of 2.75 m

and a rated rotational frequency of 170 rpm at a rated

wind speed of 12.5 m/s has a tip speed ratio of 0.3.

According to Figure 5.36, this tip speed ratio is

characteristic of a multi-bladed HAWT (curve B) and a

Savonius VAWT (curve A), corresponding to a maximum

power coefficient of ~ 36% and 20%, respectively.

Unfortunately, the performance curve associated

with VAWTs based on the Darrieus configuration (Figure

1.11b) is not provided in Figure 5.36. Additionally,

rotational frequencies [] are not published for most

of the selected UWECSs. As such, the power coefficient

of each of the selected UWECSs had to be estimated as

detailed in Equation 5.4. Using Equation 5.4, based on

the published 3.5 kW power rating at the rated wind

speed of 12.5 m/s, the maximum power coefficient of the

Cleanfield VAWT is estimated to be ~ 36%. Design

characteristics play a key role in optimising the

efficiency of the device. Figure 5.36 is primarily

294


intended to illustrate that high tip speed ratio

devices are typically more efficient. Unfortunately,

high tip speed ratio machines are usually large and

noisy.

Figure 5.36. Power coefficient as a function of tipspeed.

(Ackermann & Söder, 2000)

If the rated wind speed is much higher or lower

than the mean wind speed at the site, the WECS will be

performing outside of its design range and produce

considerably less than its rated power. If the cut-in

speed, the wind speed at which the WECS starts to

295


produce energy, is above the mean wind speed, the

device will not be producing energy much of the time.

Conversely, if the cut-out speed, the wind speed at

which the WECS stops generating energy, is relatively

low, the wind speeds that produce the greatest energy

will go by unharnessed. To assist in the selection of

an appropriate UWECS, based on the wind statistics of a

site of interest, the literature proposes that the

capacity factor (Jangamshetti

& Guruprasada Rau, 2001) of the candidate UWECSs be

calculated; the higher the capacity factor, the more

appropriate the device. Alternatively, this equation

could be used to determine the rated wind speed [uR],

based on the wind statistics [f(u)], which would

maximise the capacity factor.

Additional consideration should be given to the

characteristic dimensions of the UWECS in relation to

the area and characteristic dimensions of the

amplification zones. The characteristic dimensions of

the three main configurations on which current WECSs

are typically based were defined at the end of Section

5.1.1.2. With reference to the characteristic

dimensions summarised in Table 5.15, the rotor diameter

296


of a HAWT, the diameter of a ground-mounted VAWT, and

the rotor height of roof-top mounted VAWTs should not

exceed 2 m for UWECSs intended for installation at the

site of the EcoHouse (Jenkins & Graham, 1995).

As mean wind speed is a function of height, the

appropriate wind speed reference height needs to be

determined, based on the characteristic dimensions of

the UWECS and any associated mounting apparatus. WECS

performance is typically quoted as producing a rated

power at a rated wind speed, with both parameters

varying from device to device. To compare one UWECS to

another, the rated power should be normalised (Nigim &

Parker, 2007; Jangamshetti & Guruprasada Rau, 2001).

Simple normalisation can be achieved by dividing the

rated power by the rated wind speed. Extensive

summaries of known UWECS performance characteristics

have been compiled by Wineur (2007) and Cace et al.

(2007). Table 5.16 summarises the performance

characteristics, potential on-site energy production,

and mounting configuration of the UWECSs illustrated in

Figure 5.37, which were selected for installation at

the EcoHouse based on the aforementioned criteria.

297


MarlecEngineering

Rutland 910 - 3

SouthwestWindpowerWhisper 100

Windwall B.V.WW 2000

Turby B.V.2.5 kW

WindsaveWS 1000

RenewableDevices

Swift Rooftop

Oy WindsideWS 12

Figure 5.37. UWECSs proposed for installation at theEcoHouse site.

All images were obtained from Wineur (2007), except as follows:Rutland 913 shown from Marlec Engineering (2007) and Oy Windside

WS 30B shown from Raigatta Energy Inc.(2007).

Table 5.16. Performance characteristics of the

proposed UWECSs.

298


Rated Pow er

Rated W ind Speed

Characteristic / Swept Area

Theoretical Power

Coefficient of Perform ance

Energy Produced

UW EC [PR] [uR] [AS] [P] [C p] [Ea] M ounting- (W ) (m /s) (m 2) (W ) % (kW h/m n) -

Rutland 910-3 90 10 0.655 393 23% 5.1 TraditionalW indkraft W B 20 600 10 3.14 1884 32% 36.3 TraditionalW indside W S 12 † 8000 20 12 57600 14% 4.1 GroundW indwall W W 2000 † 2900 10.5 10 6946 42% 44.8 Roof RidgeTurby † 2500 14 5.3 8726 29% 20.4 Roof RidgeSwift Rooftop † 1500 12.5 3.46 4055 37% 22.0 Roof RidgeW indsave W S 1000 † 1000 12 2.41 2499 40% 14.2 Roof RidgeW hisper 100 † 900 12.5 3.46 4059 22% 12.5 Roof RidgeHypothetical UW EC ‡ N/A 6 32.3 N/A 42% 274.2 Roof Ridge

The coefficient of performance values were estimated based onrated power divided by the theoretical wind power. †Values are

average-per-individual-UWECS with room on the roof ridge toinstall two of any individual roof ridge-mounted type.

Performance characteristics for all UWECSs were obtained fromWineur (2007), except as follows: Whisper 100 from Southwest

Windpower (2007) and Swift Rooftop from Renewable Devices EnergySolutions Ltd. (2006).

Traditional mounting refers to a tower-mounted HAWT

at the height recommended by the Canadian Wind Energy

Association (CanWEA), where building aerodynamics-

induced amplification is no longer typically assumed to

be present. Ground and roof mounted configurations are

subjected to a combination of amplified and un-

amplified winds as dictated by placement in relation to

the building. All calculations include wind direction

frequency, summed for each applicable building face.

Actual energy production [Ea] was calculated using

Equation 5.4. It is conceivable that less energy than

estimated will be generated by the selected UWECSs,

given that the calculations were based on an estimation

of the peak power coefficient, only applicable at rated

299


wind speed, and that individual rated wind speeds were

higher than the mean wind speed at the site.

, where: (5.4)

Ea is the actual energy produced in Wh per time

period;

is the dimensionless peak performance or

power coefficient, with

PR representing the rated power and P the

theoretical maximum wind power at the rated

wind speed; and

is the theoretical wind energy

(T. Chang et al., 2003), where: (5.5.)

t is the time period of interest in hours per

time period (i.e., hours per month);

is the wind

power as a function of rated, cut-in, and

cut out wind speed and wind speed frequency

distribution;

300


f(u) is the Rayleigh estimation of the wind

speed frequency distribution (i.e.,

);

is the density of air at standard temperature

and pressure (STP) ~ 1.169 kg/m3;

AS is the characteristic or swept rotor area of

the UWECS;

ui is the cut-in wind speed;

uR is the rated wind speed; and

uo is the cut-out wind speed.

For UWECSs where, either by design or placement, there

was a directional dependency, the actual energy

produced was multiplied by the frequency of the

corresponding direction(s). For example, the roof-

ridge-mounted Windwall WW 2000 was considered as only

being subjected to winds originating out of the SSW and

the NNE for the purpose of energy calculation.

Based on Table 5.16, mounting two Windwall WW 2000

units on the roof ridge (2 in Figure 5.38) and two

Windside WS 12 units at mid-depth of the WNW and ESE

building faces on the ground at the site of the

EcoHouse (1 in Figure 5.38), could potentially result

301


in the generation of 98 kWh per month (mn).

Alternatively, a traditional tower-mounted Windkraft WB

20 with one meter long blades could potentially produce

36 kWh per month.

2

1

1

Figure 5.38. Roof ridge and ground mounted UWECSs atthe EcoHouse Site.

Extracted from Peterson’s EcoHouse Landscape Plan (2004).

Installing the multiple UWECSs is projected to result

in the generation of ~ 15% of the average household

energy need approximated as being ~ 700 kWh/mn (Office

of Energy Efficiency, 2005), while the traditional

installation could potentially produce 5% of this same

need. Due to the size of the traditionally mounted

UWECS and the associated wind turbine-generated wakes,

in relation to the size of the property, one

302


traditional device is possibly all that could be

installed. If another traditional device were

installed, wind direction-specific energy yield of the

downwind installation would be reduced.

The primary thrust of this research undertaking is

highlighted by the hypothetical UWECS, which could

potentially generate 40% of the average household need

at the EcoHouse site. Its theoretically design

characteristics are summarised as follows:

Power coefficient of 42% (based on the highest Cp of the selected UWECSs);

Characteristic (swept) area of 32.3 m2, calculatedthrough , with AZi corresponding to the top amplification zone areas of the four building faces, based on the assumption that it could make use of roof-ridge level amplified wind speeds from all directions through appropriate ducting;

Rated wind speed of 6 m/s, calculating through capacity factor optimisation based on the wind statistics of the site;

Rated power of 1.8 kW, based on the Cp, characteristic area, and rated wind speed;

Cut-in wind speed of 2 m/s; and Cut-out wind speed of no less than 26 m/s.

This hypothetical UWECS demonstrates how urban wind

energy extraction could be an even more viable option

through development of devices tailored to the site-

303


specific wind resource. Table 5.17 summarises the

estimated power generation, rated power, and estimated

power generated as a percentage of the rated power for

the selected UWECS. It also serves to highlight, once

again, that the rated power should not be the primary

factor in selecting an appropriate WECS and that an

appropriately designed UWECS could make the best use of

the on-site wind resource.

Table 5.17. UWECS power generation and percent of

rated power.Estim ated Power

GenerationRated Power

% of Rated Power

[Pa] (W ) [PR] (W ) (% )Hypothetical 375.6 1800 20.9%W indwall W W 2000 61.3 2900 2.1%W indkraft W B 20 49.7 600 8.3%Swift Rooftop 30.1 1500 2.0%Turby 28.0 2500 1.1%W indsave W S 1000 19.5 1000 1.9%W hisper 100 17.1 900 1.9%Rutland 910-3 7.0 90 7.8%W indside W S 12 5.6 8000 0.1%

UW EC

304


This assessment, recommended installations, and the

estimations of UWECS-specific energy generation

potential were based on the existing orientation and

configuration of the EcoHouse. Within the Summary

Report, the UWEP DSS provides recommendations regarding

building orientation and/or roof configuration changes

that could potentially increase local wind

amplification. The recalculation of the potentially

extractable energy on implementation of these

recommended changes is controlled by the Architectural

Configurator. The Architectural Configurator's role in

the EcoHouse case study is discussed in the following

section.

5.2.3 Architectural Reconfiguration

In this case study application, the Summary Report

suggested that building orientation was optimal, on

consideration of the primary wind direction. Regarding

roof-configuration, it suggested that having a

multiple-ridge roof could result in an increase in

roof-ridge wind amplification of ~ 6.7%. The maximum

energy that could potentially be generated by a

hypothetical UWECS associated with this reconfiguration

is summarised in Table 5.18. The estimated average

energy yield, based on the root top total, was ~ 15%

305


higher than for the current configuration of the

EcoHouse roof.

Table 5.18. Roof top wind energy by building face at

the EcoHouse.

W NW SSW ESE NNEZoneTop 20.0 199.5 54.4 68.9 342.8

Building Face TotalEnergy [E] (kW h per m onth)

A reduction in roof pitch could increase roof ridge

amplification by ~ 2%, but the reduction in reference

height due to the change in roof pitch and the

associated lower reference wind speed result in a lower

mean wind speed at the original ridge height. Review

of Figure I - 7 (Canadian Commission on Building and

Fire Codes, 2006 p. I - 13) suggests that roof ridge

amplification increases with an increase in roof pitch

up to 20, beyond which increases in pitch actually

reduce roof ridge amplification.

This concludes the application of the UWEP DSS to a

case study of the wind energy potential at the site of

the EcoHouse. In the final section limitations of the

UWEP DSS with respect to the case study are briefly

summarised.

306


5.2.4 Potential Limitations of the Application

The application of the UWEP DSS in the case study

assessment of the wind energy potential at the site of

the Green Venture EcoHouse is a first attempt at such

an undertaking. The following limitations and

assumptions are presented for consideration on

reviewing the results:

Due to the complexities involved in building

aerodynamics-induced wind amplification estimation

in general, the UWEP DSS incorporates several

simplifications. In this case study the primary

simplification of significance involves the

rectangular representation of the EcoHouse's

footprint. Additionally, the actual building

dimensions were not provided in time for the case

study application and had to be estimated from

data obtained from the Municipal Property

Assessment Corporation (MPAC) (2007), the

assumption that the landscape master plan (Figure

5.35) was to scale, and an average storey height

of 4 m. As a result of the aforementioned, it is

conceivable that both the WNW and the ESE ground-

mounted UWECSs are subjected to winds of reduced

amplification, due to the patio located at the WSW

307


corner of the house and the footprint offset in

the ENE quadrant, respectively (Figure 5.34).

Of additional significance are the five dormers on

the NNE side of the roof ridge, whose impact on

the wind speed is beyond the assessment

capabilities of the UWEP DSS. It is conceivable

that they have a channelling effect on winds

originating out of the NNE, resulting in further

local amplification. It may also be appropriate

to consider the roof as multiple-ridged as opposed

to gabled during the amplification assessment,

especially for winds originating out of the WNW

and ESE perpendicular to these dormers.

The building code figures referenced for building-

scale amplification factors for low rise roof

configurations are intended for structural

component and cladding wind-load estimation.

Since local maximums are of primary importance for

individual components, pressure coefficient

variation with wind direction is not provided.

Review of Figure I - 7 (Canadian Commission on

Building and Fire Codes, 2006 p. I - 13),

pertaining to whole building assessment suggests

that wind amplification at the ridge is higher

308


when the wind direction is primarily parallel to

the ridge. Though preliminary steps have been

taken to allow for the UWEP DSS to differentiate

between perpendicular vs. parallel roof-ridge

angle of flow incidence, the tool is currently

limited to perpendicular angles of flow incidence

for roof-top amplification assessment.

This case study application was based on the mean

annual wind speed. It is conceivable that during

the windy season 2.2X the energy originally

estimated could be generated since the windy

season was previously found to yield wind speeds

1.3X the mean annual wind speed.

The case studies summarised in this chapter

pertained to the application of the prototype UWEP DSS

to both an institutional and a residential potential

wind energy assessment. The institutional case study

application included an assessment of both mean annual

and peak seasonal wind energy potential at the site of

the Robarts Library on the U of T's St. George campus.

Reasonable results were obtained on comparison to

pedestrian-level wind speeds documented in the

literature, the data from the Robarts Library urban

wind power project team, and the wind power density

309


estimates provided by the wind atlases. The

residential case study application included selection

and siting of suitable UWECSs, highlighting the need

for new devices specifically designed for urban wind

conditions, including the potential presence and

implications of wind amplification zones. Chapter six

summarises this research undertaking and elaborates on

conclusions and recommendations, some of which were

alluded to in this chapter.

310


6 SUMMARY, CONCLUSIONS & RECOMMENDATIONS

Chapter six summarises this research undertaking,

including a discussion on the key conclusions drawn and

recommendations for further research.

6.1 Summary

This research undertaking was primarily based on

the postulation that building aerodynamics-induced wind

amplification could potentially yield greater wind

energy than traditionally assumed to be available in

urban developments. The premise was based on building

augmented wind turbine (BAWT) theory, building-

aerodynamics induced wind amplification, wake

development theory, and the emergence of urban wind

energy conversion systems (UWECSs). The objectives

were to develop an urban parameterisation scheme and

wind amplification factor table to quantify potential

wind speed amplification magnitudes and identify the

associated zones of potentially amplified wind speed in

support of developing a methodology to assist

practitioners in the selection of appropriate WECSs and

their placement at a particular location of interest.

The extent of the literature review attests to the

truly inter-disciplinary nature of the field of wind

311


power meteorology. The wealth of knowledge particular

to this field is unfortunately hampered by the wind

energy industry paradigm. The theory, guidelines, and

tools, for the most part, recommend siting of wind

farms and/or stand-alone towers well outside of urban

developments. Additionally, existing wind atlases and

assessment tools primarily provide regional, annual

wind power density estimates, neglecting seasonal and

directional variation. The small body of existing

research advocating building-augmented wind turbines

and urban wind energy conversion systems unfortunately

pertains specifically to high-rise development and

large-scale traditional horizontal axis wind turbines

(HAWT). The existing literature further proposes that

complex flow assessment can conceivably be conducted

only through wind tunnel testing and/or use of

numerical computational fluid dynamics (CFD) methods.

There is growing concern over the land-use

implications of large-scale wind farms and traditional

HAWTs in general. In light of the growing interest in

renewable energy sources, security of energy supply,

and energy self-sufficiency, integrated urban wind

energy conversion systems could be a reasonable

solution. To assess the feasibility of such an

undertaking, a thorough knowledge of building

312


aerodynamics, boundary layer meteorology, and wind

power meteorology is required. On a case-by-case

basis, such assessment services are currently provided

by consulting firms (e.g., BSD Partnership and AWS

Truewind) using proprietary in-house software, which

include CFD capabilities. The cost of such services is

typically beyond what an individual homeowner is

willing to afford. As such, the need for an easily

accessible, user friendly, prototype decision support

tool was identified.

With the small wind energy industry in its infancy,

it was considered of utmost importance that the

developed tool cover as wide a user-base as possible,

including homeowners and entrepreneurial UWECS

developers. To that end, it would need to be based on

a common, popular platform and require as little

scientific knowledge on the part of the user as

possible. In combining boundary layer meteorology

theory with comfort parameters published in pedestrian-

level wind studies and amplification factors estimated

from pressure coefficients published in building codes,

it became conceivable that a simplified assessment tool

could be developed. This combination of inter-

disciplinary theories formed the methodological basis

for the developed urban wind energy planning decision

313


support system (UWEP DSS). The simplifications and

assumptions required to estimate amplification factors

from published composite pressure-gust coefficients

suggests that further research, including data collect

and comparison, is required to gain a better

understanding of sustained, as opposed to sporadic,

wind amplification.

The prototype decision support system estimates the

wind energy potentially available for extraction by

zone, based on the morphology of the subject site. It

employs Visual Basics for Applications (VBA) forms to

connect the user to external internet-based

applications and guide the user through the data input

sequence. At the heart of the application is an

internal database that translates user-selected

descriptive phrases into the parameters required to

perform the calculations. This process yields annual

and seasonal/monthly estimates and generates a summary

report, which includes building orientation and roof

geometry change suggestions to increase wind energy

potential. The modular nature of the prototype tool

allows for future enhancements pertaining to urban

parameterisation schemes, mean wind speed profile

estimations, and wind amplification factor

parameterisations.

314


As an apparent first attempt at quantifying

building aerodynamics-induced wind amplification and

the resulting increase in extractable wind energy, this

research undertaking also emphasises that further

amplification can be achieved through both the design

of the WECS and the structure into or onto which it is

mounted. A better understanding of building

aerodynamics-induced wind amplification will also

result in more appropriate siting of urban wind speed

and direction measurement instruments and the

development of WECSs more appropriate for urban

developments, with the potential to enact a wind energy

industry paradigm shift.

6.2 Conclusions

Numerous findings were made throughout this

research endeavour. The principal conclusions will now

be summarised.

Sustainable development within the wind energy

industry requires a paradigm shift toward de-

centralised, in-situ, energy generation. The

implications of large wind farms, both

substantiated and unsubstantiated, involving low

frequency vibration effects on livestock and

people, bird-kills, line losses, cost of

315


infrastructure, signal disruptions, and land-use

impacts, suggest that a paradigm shift away from

large-scale wind farm construction is crucial to a

sustainable future for the wind energy industry.

Wind power meteorology, as a relatively new field

of applied science, could benefit through greater

inter-disciplinary collaboration. Much of the

theory required to develop suitable UWECSs has

already been well established through research in

fluid dynamics and general physics. This

fledgeling industry should draw from the expertise

of the various disciplines involved in developing

an understanding of fluid flow (e.g., aerospace,

naval, hydro-electric, etc.). It is truly

unfortunate that not much has changed when it

comes to wind turbines except for their size.

The year 2005 is considered by many to be the year

of the birth of the Canadian wind energy industry.

Since then, government-funded financial

incentives, public interest, and daily

developments have inundated the news. The

building industry, supported by energy efficiency-

increasing retrofit program incentives, needs to

move towards energy self-sufficiency from both a

316


security of supply and a sustainability/efficiency

stand point.

Building augmented wind turbine (BAWT) theory has

finally been given real life application in the

form of the Bahrain World Trade Centre. It is not

inconceivable that building aerodynamics-induced

amplification in existence at roof-level and

between buildings in urban developments may be

sufficient to yield the necessary energy

requirements of smaller structures (i.e., a

residential dwelling).

Figure 6.1. The Bahrain World Trade Center

317


The Artist's rendition, on the left, was obtained from DTZ BahrainW.L.L. (2006), while the construction photo on the right was

obtained from Atkins (2007).

Within the wind energy industry, there are

numerous initiatives involving vertical axis wind

turbines (VAWTs) in roof-mounted applications,

with little to no focus on identifying zones

within which building-aerodynamics induced flow

amplification could potentially exist. Instead,

several studies have attempted to redesign the

wind energy conversion device to improve its

performance in less than ideal conditions

(Mertens, 2003a & 2003b). A better understanding

of where more ideal flow conditions exist and the

nature of the flow could result in the development

of more suitable devices.

The current commercially available tools used to

conduct an assessment require either extensive

user knowledge and/or rule out wind energy

potential within urban developments. Services

available to conduct such an assessment are too

costly for most individuals to consider without

prior estimates of potential.

In-depth, accurate assessment of the available

wind resource at a site traditionally entails the

318


collection of at least a year's worth of data.

Prior to embarking on such an exercise, one must

have some knowledge on where best to place data

collection instruments. The initial assessment,

as provided by the prototype decision support

system, yields an estimate of the potentially

available wind energy and identifies the zones

within which one could expect greater potential.

By calculating the extent of the amplification

zones and locating their boundaries through

calculation of the wake streamline equations, the

prototype decision support system suggests where

to site data collection devices and provides

flexibility to UWECS type- and placement-

selection.

The urban parameterisation and wind amplification

factor tables of the internal database of the UWEP

DSS minimise the requirement for end-user

knowledge of building aerodynamics and/or boundary

layer meteorology. The structure of the internal

database lends itself well to future enhancements,

in response to new research findings.

The ability of the UWEP DSS to generate the wind

statistics from time series data, including the

319


associated wind direction roses and direction-

specific wind speed frequency distributions,

allows for monthly and/or seasonal wind resource

assessments to be conducted, a capability which at

this point the wind atlases do not have. Its

additional ability to evaluate peak wind months

also appears to be a first in comparison to

assessment tools currently in existence.

The Summary Report, automatically generated by the

UWEP DSS, condenses the assessment findings

through tabulations, graphs, and images for

further use in subsequent economic feasibility

studies.

The Architectural Configurator of the UWEP DSS

facilitates interactive analysis to explore the

implications of building orientation and

architectural feature configuration on wind energy

potential.

The form-guided data entry process of the UWEP

DSS, using descriptive phrases and images, shields

the user from the complexities involved in the

assessment process.

Selection of Microsoft® Excel as the platform for

the UWEP DSS was primarily based on the fact that

320


it is a common, readily available software

application. The user interface that can be

created using VBA-controlled user forms is

visually appealing, very flexible, and lends

itself well to the development of decision support

systems. Unfortunately, Excel does not

efficiently handle large amounts of data and has a

quite limited table look-up functionality.

Additionally, the graphing capabilities of Excel

do not support generating plots with variables on

the abscissa, as is required when plotting mean

wind speed profiles. Finally, though Excel allows

for formulae to be continually active, unlike

solver-based math applications (e.g., Matlab),

additional VBA code is required when an automated

iterative approach is required to solve an

equation. That said, the combination of VBA and

Excel produces a relatively inexpensive, easy-to-

learn application development tool.

Application of the UWEP DSS to the institutional

case study of the potential wind energy at the

site of the Robarts Library on the St. George

campus of the U of T, yielded credible wind speeds

and wind power densities in comparison to the wind

atlases and the collected time series data.

321


The residential case study at the site of the

Green Venture EcoHouse suggested that

approximately 15% of the energy need of an average

household consuming 700 kWh per month could be

generated by strategically located suitably

configured UWECSs. The energy generation estimate

of a traditional tower-mounted HAWT installed at

the site was a mere 5% of this same need, while

the hypothetical UWECS, tailored to the wind

resource at the site and the wind amplification

zone areas, produced an estimated 40% of the

average household need. Wind energy could

generate a larger percentage of the need if the

need were reduced through energy conservation

practices combined with home energy efficiency

improvements. Though characterised by a slightly

more moderate climate, the average household

energy consumption in the Netherlands is 275 kWh

per month (Bird, Wüstenhagen, & Aabakken, 2002).

As a first attempt to assess building aerodynamics-

induced wind amplification in support of decisions

involving the siting of suitable urban wind energy

conversion systems, this research endeavour has yielded

as many credible results and findings as it has

questions and opportunities for future research. The

322


Recommendations section of this chapter summarises a

number of possible next steps in urban wind energy

research.

6.3 Recommendations

The UWEP DSS appears to be the first attempt to

develop an urban wind energy planning tool. The key to

the UWEP DSS is its internal database, including the

Urban Parameterisation and Wind Amplification Factor

tables. The case studies were founded on simplifying

the morphology of the St. George campus and the

neighbourhood of the EcoHouse through representation as

urban subregion category # 6 and # 11, respectively.

These are only two of the thirteen urban subregion

categories defined within the Urban Parameterisation

table of prototype tool. The UWEP DSS needs to be

applied to other case studies for which comparison data

have been collected at appropriate heights and

locations. Data collected during urban boundary layer

studies in Oklahoma City (Brown, Khalsa, Nelson, &

Boswell, 2004) and Basel, Switzerland (Rotach et al.,

2005) may be ideal for further validation of results

generated by application of the UWEP DSS. Throughout

the development and subsequent application of the

prototype DSS, numerous opportunities for future

323


research were identified. These opportunities are

summarised in the following material.

1) The wind speed frequency distribution. Accurate

estimation of the wind speed frequency distribution is

critical to wind energy estimation. Since wind energy

is proportional to the speed cubed, small errors in

wind speed estimation yield large errors in energy

estimation. Exploration of the Weibull and Rayleigh

distributions' ability to accurately fit the frequency

distribution of the time series data (TSD) suggested

similar findings to those in the literature, namely

that the Rayleigh distribution produces a better fit

for lower wind speed data. On thorough examination of

TSD collected by the Toronto Island Airport

meteorological station and used for the Toronto case

study, the aforementioned acceptable fit appeared to be

limited to a small range of intermediate wind speeds,

with lower and higher wind speeds being either grossly

over- or under estimated. Fitting the data

corresponding to the primary and secondary wind

directions separately, appeared to suggest that there

are two distinct distributions (as illustrated below)

perhaps providing an explanation as to why neither the

Rayleigh nor the Weibull distributions are uncontested

within the literature.

324


0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18W ind Class (m /s)

Freq

uency (%)

Prim ary Secondary

Figure 6.2. Primary and secondary wind speed frequencydistributions (TIA).

It is proposed that the primary wind direction

corresponds to the lower wind speeds associated with

the prevailing wind and can be estimated using the

Rayleigh distribution. The secondary wind direction,

from which higher speeds appear to originate,

corresponds to storms and appears to follow the Weibull

distribution. Since these findings are based on a

single year's worth of data collected at the Toronto

Island Airport, further data would need to be analysed

to confirm the validity and transferability of this

finding.

2) The mean wind speed profile. The mean wind

speed profile equations are crucial to the estimation

325


of the mean wind speed as a function of height and the

morphology of the underlying terrain. In its current

form, the portion of the profile equation applicable

below the mean building height does not retain physical

boundary conditions (i.e., u(0) 0), especially when

reapplied after the amplification assessment. It is

proposed that by using the mixing length profile [lm]

as opposed to the mixing length parameter [lc] in

calculating the attenuation coefficient [a ] of the

exponential profile, this boundary condition would be

restored. The equations proposed by the literature to

calculate the mixing length and the drag length [Lc]

(used to calculate the attenuation coefficient) do not

consider these parameters to be functions of height,

while the mixing length profile equation does. Given

that the frontal area density decreases with height in

heterogeneous urban developments, any parameterisations

that attempt to explain the implications of morphology

on the flow field should be functions of height. Since

the exponential profile and the associated transitional

profile were developed through studies of mean wind

speeds within and above plant canopies of relatively

homogeneous height and density, further research is

required to determine the following:

326


whether the wind speed profile actually varies exponentially below the mean building height within an urban canopy;

the vertical extent over which the exponential profile can truly be considered applicable; and

the effect of heterogeneous canopies on the attenuation coefficient and transitional profile parameters.

3) Urban parameterisation schemes. Urban

parameterisation attempts to develop scaling parameters

that can represent the effects of the underlying

terrain on the wind regime. The key parameters are the

much contested roughness length [z0], zero-plane

displacement [d], and associated friction velocity

[u*]. Most of the wind tunnel studies previously

conducted have explored these parameters using uniform

cubic arrays, characterised by equal plan and frontal

area densities. Unfortunately, this is seldom the case

in urban development and little development has

occurred on this front since 1998 (Grosso, 1998).

Additionally, the logarithmic nature of this portion of

the wind speed profile makes it highly sensitive to

errors in the estimation of these parameters. More

wind tunnel studies need to be conducted using models

more representative of urban terrain.

327


4) Building aerodynamics-induced wind

amplification. This research undertaking involved

amplification assessment on two compounding scales.

Considering the neighbourhood morphology, rudimentary

and near-qualitative comfort parameters from pedestrian

level wind studies were used to calculate the first

stage of the amplification. The resulting amplified

wind speed was then considered to impinge on the

subject building on a more localised scale and pressure

coefficients provided by the national building code

were used to calculate the corresponding amplification

factors. This highly simplified, idealised approach,

especially regarding the low-rise composite pressure-

gust coefficients, is a first attempt to assess

building aerodynamics-induced wind amplification using

the coefficients published in the building codes. A

tremendous amount of additional research is required in

this area. Boundary layer wind tunnel testing,

conducted by the building industry, should be used to

develop a correlation between the pressures and forces

currently measured and the associated mean wind

velocities. These studies should be validated through

on-site data collection.

5) Oblique flow studies. During the wind

amplification assessment the UWEP DSS was required to

328


consider flow as orthogonal to a particular building

face through a range of 45 from orthogonal, due to

lack of data pertaining to the implications of oblique

and/or skewed flow. The literature suggests that flow

can only be truly considered orthogonal through a range

of approximately 12. Computational fluid dynamics

and wind tunnel studies to date have primarily focussed

on orthogonal flow, based on the assumption that it

produces the highest wind loading. Further studies are

required to investigate channelled and skewed flow, in

support of quantifying the amplification levels and

correlating them to various morphological

configurations.

6) Wake interference flow. The UWEP DSS applies

wake development theory to calculate the location and

area of the amplification zones. Unfortunately, the

flow regime in most of the pre-defined urban subregions

should be considered as wake interference flow as

opposed to isolated roughness flow. In wake

interference flow regimes there is insufficient space

between the buildings for full wake development.

Additional research is required to determine the effect

of wake interference on the position and the nature of

the flow within the amplification zones.

329


7) Street canyon vortices. Street canyon vortices,

created by flow orthogonal to

the street canyon are assumed

to yield characteristics

similar to the Rankine vortex

structure. This suggests that

flow amplification could exist

at the interface between the

forced-vortex core and the free-vortex outer region

(Cook, 1985). Additional research is required to

determine the validity and applicability of this

finding, in support of urban wind energy extraction.

Figure 6.3. Velocity in a Rankine Vortex(Cook, 1985)

8) Wind direction. The UWEP DSS, for the most

part, transfers unaltered regional or meteorological

station wind direction frequency data to the subject

site. Analysis of the wind direction data collected at

the Robarts Library suggests that this practice does

not yield a truly accurate assessment of the actual

wind direction frequency distribution at the site.

Additional research is required to establish a

correlation between degrees of backing and meters of

roughness length change.

330


9) UWEP DSS development. Future research should

consider adding the ability to import from, and export

to, various platforms (e.g., AutoCAD or AutoDesk, used

in urban planning and architectural design or ArcView,

a popular desktop GIS). Additionally, a user interface

could be developed between the UWEP DSS and existing

economic assessment tools (e.g., RETScreen and Homer)

and a solar energy planning decision support system

(e.g., RETScreen). Considering the Summary Report, a

graphical portrayal similar to the orientation diagram

superimposed on the plan view of the Robarts Library

below would be of great assistance in the siting of a

hybrid wind / solar energy conversion system. The

orientation diagram could be created by combining the

findings of the UWEP DSS assessment with those of a

solar energy assessment.

331


Figure 6.4. Sun and wind orientation diagram.

Modified after de Schiller & Evans (1998).10) Urban wind energy conversion systems. Current

wind energy initiatives primarily involve the placement

of slightly smaller scale, wind energy conversion

devices into less than ideal flow conditions. It is

proposed that with a better understanding of the true

conditions, more suitably configured wind energy

conversion devices could be developed. For the UWEP

DSS, a tabulation of the dimensional and performance

characteristics of the existing small scale wind energy

conversion devices could be developed to provide an

332


estimate of currently feasible wind energy generation

potential.

333


7 REFERENCES

Ackermann, T., & Söder, L. (2000). Wind energytechnology and current status: A review. [Electronicversion]. Renewable & Sustainable Energy Reviews, 4(4), pp.315-374.

Adejuwon, S. A., Herold, A., & Hanna, J. (2005).Canada: Report of the individual review of the greenhouse gasinventory (Review No. FCCC/WEB/IRI/2004/CAN). Bonn,Germany: United Nations Framework Convention onClimate Change (UNFCCC) Secretariat. Retrieved fromhttp://unfccc.int/files/national_reports/annex_i_ghg_inventories/inventory_review_reports/application/pdf/2004_irr_desk_review_canada.pdf.

Akpinar, E. Kavak, & Akpinar, S. (2005). A statisticalanalysis of wind speed data used in installation ofwind energy conversion systems. [Electronicversion]. Energy Conversion and Management, 46(4), pp.515-532.

Arnfield, A. J. (2003). Two decades of urban climateresearch: A review of turbulence, exchanges ofenergy and water, and the urban heat island.International Journal of Climatology, 23(1), pp. 1-26.

Atkins. (2007). The Bahrain World Trade Center. Atkins ME &India. Retrieved fromhttp://www.atkins-me.com/News.aspx?ItemNo=27.

AWS Truewind. (2004a). The New York Wind Resource Explorer(WRE). AWS Truewind, LLC. Retrieved fromhttp://www.windexplorer.com/wre/CORE/MAIN.ASP?MAP_SERVICE=WRE_NEW_YORK.

AWS Truewind. (2004b). Wind Resource Explorer. Retrievedfrom http://www.windexplorer.com/.

334


AWS Truewind. (2007). AWS Truewind. Retrieved fromhttp://www.awstruewind.com/index.html.

Baban, S. M. J., & Parry, T. (2001). Developing andapplying a GIS-assisted approach to locating windfarms in the UK. Renewable Energy, 24(1), pp. 59-71.

Banks, D., Meroney, R. N., Sarkar, P. P., Zhao, Z., &Wu, F. (2000). Flow visualization of conicalvortices on flat roofs with simultaneous surfacepressure measurement. [Electronic version]. Journal ofWind Engineering and Industrial Aerodynamics, 84(1), pp. 65-85.

BDSP Partnership Ltd., MECAL Applied Mechanics BV,Imperial College, & University of Stuttgart. (2001).Wind Energy for the built environment: The building as aconcentrator. Copenhagen, Denmark: The EuropeanCommission. Retrieved October 29, 2006, fromhttp://www.bdsp.com/web/pdfs/European_Wind_Energy_Conference_2001_Copenhagen_Poster.pdf.

Belcher, S. E., Jerram, N., & Hunt, J. C. R. (2003).Adjustment of a turbulent boundary layer to a canopyof roughness elements. Journal of Fluid Mechanics, 488, pp.369-398.

Bell, G. (1978). Robarts Library viewed from St. George St.Department of Fine Art, University of Toronto.Retrieved fromhttp://www.fineart.utoronto.ca/canarch/ontario/toronto/toronto.jpgs/tor-59.jpg.

Benoit, R., Desgagné, M., Pellerin, P., Pellerin, S.,Chartier, Y., & Desjardins, S. (1997). The CanadianMC2: A semi-lagrangian, semi-implicit widebandatmospheric model suited for finescale processstudies and simulation. Monthly Weather Review, 125(10),pp. 2382-2415.

335


Berg, L. K., De Wekker, S. F. J., Shaw, W. J., Coulter,R. L., & Allwine, K. J. (2004). Observations ofboundary-layer winds in an urban environment. 5thSymposium on the Urban Environment, Aug 23-26 2004, pp. 171-176.

Bird, L., Wüstenhagen, R., & Aabakken, J. (2002). Greenpower marketing abroad: Recent experience and trends(Technical Report No. NREL/TP-620-32155). Golden,Colorado: National Renewable Energy Laboratory(NREL). Retrieved fromhttp://www.eere.energy.gov/greenpower/resources/pdfs/32155.pdf.

Bishop, I. D. (1998). Planning support: Hardware andsoftware in search of a system. Computers, Environmentand Urban Systems, 22(3), pp. 189-202.

Bitan, A. (1992). The high climatic quality city of thefuture. Atmospheric Environment. Part B: Urban Atmosphere,26(3), 313-329.

Blanch, M. J. (2002). Wind energy technologies for usein the built environment. [Electronic version]. WindEngineering, 26(3), p. 125.

Blennow, K., & Sallnäs, O. (2004). WINDA: A system ofmodels for assessing the probability of wind damageto forest stands within a landscape. EcologicalModelling, 175(1), pp. 87-99.

Blocken, B., & Carmeliet, J. (2004). Pedestrian windenvironment around buildings: Literature review andpractical examples. [Electronic version]. Journal ofThermal Envelope and Building Science, 28(2), pp. 107-159.

Bottema, M. (1999). Towards rules of thumb for windcomfort and air quality. Atmospheric Environment, 33(24-25), pp. 4009-4017.

336


Brazeau, L. (2007). 2006 A record breaking year for the globalwind energy industry with Canada now ranking 12th in the world fortotal installed wind energy capacity - Wind energy a growingcontributor to reductions in greenhouse gas emissions. (PressRelease. Ottawa: Canadian Wind Energy Association(CanWEA). Retrieved from http://www.canwea.ca/.

Brown, M. J., Khalsa, H., Nelson, M., & Boswell, D.(2004). Street canyon flow patterns in a horizontalplane: Measurements from the Joint Urban 2003 fieldexperiment. Paper presented at the AMS Fifth Conferenceon the Urban Environment, Vancouver, BC., Session 3 pp. 3.1-3.12. Retrieved fromhttp://ams.confex.com/ams/pdfpapers/80299.pdf.

Burton, T., Sharpe, D., Jenkins, N., & Bossanyi, E.(2001). Handbook of Wind Energy. Toronto, Canada: JohnWiley & Sons, Ltd.

Cace, J., ter Horst, E., Syngellakis, K., Niel, M.,Clement, P., Heppener, R., et al. (2007). Urban windturbines: Guidelines for small wind turbines in the built environment.Europe: Wind Energy Integration in the UrbanEnvironment (Wineur), Intelligent Energy - Europe.Retrieved June 15, 2007, from http://www.urban-wind.org/pdf/SMALL_WIND_TURBINES_GUIDE_final.pdf.

Cai, X. -. (2000). Dispersion of a passive plume in anidealised urban convective boundary layer: A large-eddy simulation. Atmospheric Environment, 34(1), pp. 61-72.

Campbell, N. S., & Stankovic, S. (2001a). Wind energy forthe built environment (Project WEB): Assessment of wind energyutilisation potential in moderately windy built-up areas Report No.JOR3-CT98-0270. The European Commission.

337


Campbell, N. S., & Stankovic, S. (Eds.). (2001b). WindEnergy for the Built Environment (Project WEB). London, UK:BDSP Partnership Ltd.

Canadian Commission on Building and Fire Codes. (2005).Wind load. National Building Code of Canada (NBC) (12thed., pp. 4-16). Ottawa, Ontario: National ResearchCouncil (NRC) - Institute for Research inConstruction (IRC).

Canadian Commission on Building and Fire Codes. (2006).Wind load and effects. User's guide - NBC 2005 StructuralCommentaries (part 4 of division B) (2nd ed., pp. I-2-I-39).Ottawa, Ontario: National Research Council (NRC) -Institute for Research in Construction (IRC).

Canadian Energy Research Institute (CERI). (2005).Energy facts & statistics: Energy fast facts: Canadian fact #18.Retrieved May 28, 2007, fromhttp://www.centreforenergy.com/outsideNav.asp?href=/FastFacts/Canada18.pdf

Canadian Wind Energy Association (CanWEA). (2005).CanWEAs Small Wind Energy Site - Where to put the turbine.Retrieved fromhttp://www.smallwindenergy.ca/en/SmallWindAndYou/YourTurbineSite.html?min

Celik, A. N. (2003). Weibull representative compressedwind speed data for energy and performancecalculations of wind energy systems. [Electronicversion]. Energy Conversion and Management, 44(19), pp.3057-3072.

Chan, A. T. (2001). Atmospheric turbulent boundarylayer development due to a change in surfaceroughness. International Journal of Engineering Science, 39(17),pp. 2001-2014.

338


Chan, A. T., So, E. S. P., & Samad, S. C. (2001).Strategic guidelines for street canyon geometry toachieve sustainable street air quality. AtmosphericEnvironment, 35(24), pp. 4089-4098.

Chang, C. (2006). Computational fluid dynamicssimulation of pedestrian wind in urban area with theeffects of tree. [Electronic version]. Wind &Structures, 9(2), pp. 147-158.

Chang, T., Wu, Y., Hsu, H., Chu, C., & Liao, C. (2003).Assessment of wind characteristics and wind turbinecharacteristics in Taiwan. [Electronic version].Renewable Energy, 28(6), pp. 851-871.

Choi, Edmund C. C., & Britter, R. E. (2004). Re-entrainment around a low-rise industrial building:2D versus 3D wind tunnel study. Atmospheric Environment,38(23), pp. 3817-3825.

Christen, A., Vogt, R., & Rotach, M. W. (2003). Profilemeasurements of selected turbulence characteristicsover different urban surfaces. Paper presented atthe 4th International Conference on Urban Air Quality, Prague.pp. 408-411. Retrieved fromhttp://pages.unibas.ch/geo/mcr/Projects/BUBBLE/docs/prague_christen.pdf.

Christen, A., & Rotach, M. W. (2004). Estimating windspeed at an urban reference height. Paper presentedat the Fifth Symposium on the Urban Environment (AMS),Vancouver. Retrieved fromhttp://ams.confex.com/ams/pdfpapers/79918.pdf.

Cionco, R. M., & Ellefsen, R. (1998). High resolutionurban morphology data for urban wind flow modeling.Atmospheric Environment, 32(1), pp. 7-17.

Clarke, R. H., & Hess, G. D. (1974). Geostrophicdeparture and the functions A and B of Rossby-number

339


similarity theory. [Electronic version]. Boundary-Layer Meteorology, 7(3), pp. 267-287.

CMC/RPN - Comm Groupe. (2002). EOLE wind energy project -MSC/RPN project for wind energy resource and prediction.Retrieved fromhttp://collaboration.cmc.ec.gc.ca/science/rpn/modcom/eole/Home.html

Coceal, O., & Belcher, S. E. (2005). Mean winds throughan inhomogeneous urban canopy. Boundary-LayerMeteorology, 115(1), pp. 47-68.

Cook, N. J. (1985). In Building Research Establishment(Ed.), The Designer's Guide to Wind Loading of Building Structures.Part 1: Background, damage survey, wind data and structuralclassification. London: Butterworths.

Cook, N. J. (1990). In Building Research Establishment(Ed.), The Designer's Guide to Wind Loading of Building Structures.Part 2: Static structures. London: Butterworths.

Dalgliesh, W. A., & Boyd, D. W. (1962). Wind onbuildings. Canadian Building Digest, 28. Retrieved fromhttp://irc.nrc-cnrc.gc.ca/pubs/cbd/cbd028_e.html.

Dalgliesh, W. A., & Schriever, W. R. (1968). Windpressures on buildings. Canadian Building Digest, 34.Retrieved fromhttp://irc.nrc-cnrc.gc.ca/pubs/cbd/cbd034_e.html.

Dalgliesh, W. A., & Surry, D. (2003). BLWT, CFD and HAMmodelling vs. the real world: Bridging the gaps withfull-scale measurements. Journal of Wind Engineering &Industrial Aerodynamics, 91(12-15), pp. 1651-1669.

Dandou, A., Tombrou, M., Akylas, E., Soulakellis, N., &Bossioli, E. (2005). Development and evaluation ofan urban parameterization scheme in the PennState/NCAR mesoscale model (MM5). [Electronic

340


version]. Journal of Geophyscial Research, 110(D10 - D10102),pp. 1-14.

Danish Wind Industry Association. (2006). Guided tour onwind energy. Retrieved fromhttp://www.windpower.org/en/tour.htm.

Dayan, E. (2006). Wind energy in buildings. Refocus, 7(2),pp. 33-38.

de Schiller, S., & Evans, J. M. (1998). Sustainableurban development: Design guidelines for warm humidcities. [Electronic version]. Urban Design International,3(4), pp. 165-184.

de Wit, M. H., Stathopoulos, T., & Wisse, J. A. (2002).Airport wind speeds used for the design in urbanenvironments: The Eindhoven case. Journal of WindEngineering & Industrial Aerodynamics, 90(11), pp. 1289-1298.

DTZ Bahrain W.L.L. (2006). Bahrain World Trade Center -Commercial Complex Brochure. Manama, Bahrain: DTZBahrain W.L.L.

Dunster, B. (2001). SkyZED: The flower tower. RetrievedOctober 2007, 2006, from http://www.zedfactory.com/.

Dutton, A. G., Halliday, M. J., & Blanch, J. A. (2005).The feasibility of building mounted / integrated wind turbines(BUWTs): Achieving their potential for carbon emission reductions(Final Report No. 2002-07-028-1-6). London: EnergyResearch Unit, CCLRC.

Eke, R., Kara, O., & Ulgen, K. (2005). Optimization ofa wind/PV hybrid power generation system.[Electronic version]. International Journal of Green Energy,2(1), pp. 57-63.

Eliasson, I. (2000). The use of climate knowledge inurban planning. [Electronic version]. Landscape andUrban Planning, 48(1-2), pp. 31-44.

341


Ellefsen, R. (1991). Mapping and measuring buildings inthe canopy boundary layer in ten U.S. cities. Energyand Buildings, 16(3-4), pp. 1025-1049.

ENECO. (1999). About WARP TM - Wind amplified rotorplatforms: New technology gives new look towindpower generation. [Electronic version]. ElectricalLine, May/June.

Environment Canada (EC). (2005a). Canadian wind energy atlas(http://www.windatlas.ca/en/maps.php ed.). InVisuInc. Accessed throughhttp://www.windatlas.ca/en/maps.php.

Environment Canada (EC). (2005b). Canadian wind energy atlas– methodology. Retrieved fromhttp://www.windatlas.ca/en/maps.php.

Environment Canada (EC). (2005c). MC2: Mesoscalecompressible community NWP model. Retrieved fromhttp://collaboration.cmc.ec.gc.ca/science/rpn.comm/wiki/doku.php?id=mc2.

Environment Canada (EC). (2005d). The National Climate Dataand Information Archive. Accessed throughhttp://climate.weatheroffice.ec.gc.ca/Welcome_e.html

Environment Canada (EC). (2006). Canada's fourth nationalreport on climate change. Ottawa, Canada: Government ofCanada. Retrieved fromhttp://www.ec.gc.ca/climate/4th_Report_on_CC_e.pdf.

Fisher, B., Joffre, S., Kukkonen, J., Piringer, M.,Rotach, M., & Schatzmann, M. (2004). Meteorologyapplied to urban air pollution problems – Final Report COST 715.Retrieved from http://cost.fmi.fi/wg2/COST715-FinalReport-Dec2004.pdf.

Fisher-Gewirtzman, D., Pinsly, D. S., Wagner, I. A., &Burt, M. (2005). View-oriented three-dimensional

342


visual analysis models for the urban environment.Urban Design International, 10(1), p. 23.

Frey-Buness, F., Heimann, D., & Sausen, R. (1995). Astatistical-dynamical downscaling procedure forglobal climate simulations. [Electronic version].Theoretical and Applied Climatology, 50(3), pp. 117-131.

Friedland, M., Dr. (2002). The University of Toronto: A history.Toronto, Canada: University of Toronto Press.Retrieved fromhttp://www.newsandevents.utoronto.ca/bios/02/history51.htm.

Gandemer, J. (1977). Wind environment around buildings:Aerodynamic concepts. Proceedings of the 4th InternationalConference on Wind Effects on Buildings and Structures, London.pp. 423-432.

Gasset, N. (2005). Atlas éolien du nouveau-brunswick. Accessedthrough http://www.umoncton.ca/cge/atlas_eolien/.

Gee, J. (2007). In Cooper R. (Ed.), Accuracy of the NRG#40anemometer and the NRG #200P wind direction vane. Hinesburg,Vermont: NRG Systems, Inc. [Personal Communication]Also retrieved June 11, 2007, fromhttp://www.nrgsystems.com/store/product_detail.php?cd=11&s=1899;http://www.nrgsystems.com/store/product_detail.php?cd=11&s=1904.

Germano, M., & Roulet, C. (2006). Multicriteriaassessment of natural ventilation potential. SolarEnergy, 80(4), pp. 393-401.

Giebel, G., & Risø National Laboratory. (2006).Zephyr/Prediktor. Retrieved from http://www.prediktor.dk

343


Gilbert, R. (2005). Windmills in blowing sand. FlickrWebsite: Ron's Log. Retrieved with permission fromhttp://www.flickr.com/photos/ronslog/5795246/.

Gipe, P. (2007). Forum on renewable energy tariffs - Marketmechanism for rapid development. Toronto: OntarioSustainable Energy Association (OSEA); David SuzukiFoundation. Retrieved from http://www.ontario-sea.org/Forum/Forum_Presentations/Gipe%20Paul%20Overview%20of%20Forum%20on%20Renewable%20Tariffs%20Toronto%20200.pps.

Global Wind Energy Council. Global wind energy council (GWEC).(2007). Retrieved from http://www.gwec.net/.

Google. (2007). Google TM mapsGoogle. Accessed throughhttp://maps.google.com/.

Grant, R. H., Heisler, Gordon M. Dr., & Herrington, L.P. (1988). Full-scale comparison of a wind-tunnelsimulation of windy locations in an urban area.Journal of Wind Engineering & industrial Aerodynamics, 31(2-3),pp. 335-341.

Gratia, E., Reiter, S., & De Herde, A. (2005). Windpressure distribution influence on naturalventilation for different incidences and environmentdensities. Energy and Buildings, 37(8), pp. 878-889.

Green Venture. (2005). EcoHouse history. Retrieved fromhttp://www.greenventure.ca/ecohouse.asp?ID=105.

Green Venture. (2007). Power shift. Retrieved fromhttp://www.greenventure.ca/gv.asp?ID=274.

Grosso, M. (1992). Wind pressure distribution aroundbuildings: A parametrical model. Energy and Buildings,18(2), pp. 101-131.

Grosso, M. (1998). Urban form and renewable energypotential. Renewable Energy, 15(1-4), pp. 331-336.

344


Guirguis, N. M., Hanna, G. B., Kotkata, M. F., & Gad,I. A. (1998). An investigation of building / windinteraction. [Electronic version]. Renewable Energy,15(1-4), pp. 383-386.

Hamilton Spectator. (1997). Veevers estate. Hamilton,Ontario: Hamilton Spectator. Retrieved fromhttp://www.greenventure.ca/ecohouse.asp?ID=105.

Hanna, S. R., & Chang, J. C. (1992). Boundary-layerparameterizations for applied dispersion modelingover urban areas. [Electronic version]. Boundary-LayerMeteorology, 58(3), pp. 229-259.

Hanzlik, P., Diniz, S., Grazini, A., Grigoriu, M., &Simiu, E. (2005). Building orientation and windeffects estimation. [Electronic version]. Journal ofEngineering Mechanics, 131(3), pp. 254-258.

Hein, J. (2003). The history of the electric industry:1879 - 2003. Closed Circuit, 25, November 3, 2006.Retrieved November 3, 2006, fromhttp://www.wapa.gov/newsroom/cct/2003/sept12/25no193.htm.

Helimax Energy Inc. (2007). Helimax. Retrieved fromhttp://www.helimax.com/.

Hildebrand, P. H., & Ackerman, B. (1984). Urban effectson the convective boundary layer. Journal of theAtmospheric Sciences, 41(1), pp. 76-91.

Hirsch, J. J. (2006). In James J. Hirsch & Associates(JJH), Lawrence Berkeley National Laboratory (LBNL)(Eds.), eQuest ... the quick energy simulation tool (3rd ed.)Enerlogic.

Hurwood, D. L. (1981). Eleventh street & Bronxfrontier: Urban pioneering with wind power. The 2nd

345


Miami International Conference on Alternative Energy Sources, MiamiBeach, Florida. , 4 p. 565.

Hussain, M., & Lee, B. E. (1980). A wind tunnel studyof the mean pressure forces acting on large groupsof low-rise buildings. [Electronic version]. Journal ofWind Engineering & Industrial Aerodynamics, 6(3-4), pp. 207-225.

Industrial Wind Action (IWA) Group. (2006). Facts, analysis,exposure of wind energy's real impacts. Retrieved fromhttp://www.windaction.org/about.

Intergovernmental Panel on Climate Change (IPCC).(1997). Revised 1996 IPCC guidelines for national greenhouse gasinventories: Reporting instructions (Guideline No. Volume 1).Mexico City: The Institute for Global EnvironmentalStrategies (IGES). Retrieved from http://www.ipcc-nggip.iges.or.jp/public/gl/invs4.htm.

Iowa Energy Center. (2006). Wind energy manual., Retrievedfromhttp://www.energy.iastate.edu/renewable/wind/wem/wem-08_power.html.

Isyumov, N., Dr., & Davenport, A. G. (1977). The groundlevel wind environment in built up areas. The 4thInternational Conference on Wind Effects on Building Structures,London. pp. 403-422.

Isyumov, N. (1978). Studies of the pedestrian levelwind environment at the boundary layer wind tunnellaboratory of the University of Western Ontario.[Electronic version]. Journal of Industrial Aerodynamics, 3(2-3), pp. 187-200.

Jacobson, E., & Malte, P. C. (2005). Minimizing thecost of wind energy for Vashon Island – a low windspeed site. (Masters in Mechanical Engineering,University of Washington). , pp. 1-177.

346


Jamieson, N. J., Carpenter, P., & Cenek, P. D. (1992).The effect of architectural detailing on pedestrianlevel winds. Journal of Wind Engineering & IndustrialAerodynamics, 41-44, pp. 2301-2312.

Jamil, M., Parsa, S., & Majidi, M. (1995). Wind powerstatistics and an evaluation of wind energy density.[Electronic version]. Renewable Energy, 6(5-6), pp. 623-628.

Jangamshetti, S. H., & Guruprasada Rau, V. (2001).Normalized power curves as a tool for identificationof optimum wind turbine generator parameters. EnergyConversion, IEEE Transaction on, 16(3), pp. 283-288.

Jenkins, N. H. A., & Graham, J. M. R. (1995).Measurement and prediction of wind energy inassociation with buildings. Paper presented at theThe 17th British Wind Energy Association Conference - Wind EnergyConversion, pp. 223-229.

Johnson, G. L. (2004). Wind Energy Systems (2nd ed.).Englewood Cliffs, NJ: Prentice-Hall, Inc.

Jones, D. L. (1998). The solar office in context.Renewable Energy, 15(1-4), pp. 42-47.

Jones, P. J., Alexander, D., & Burnett, J. (2004).Pedestrian wind environment around high-riseresidential buildings in Hong Kong. Indoor and BuiltEnvironment, 13(4), p. 259.

Jun, L. (2005). Application of fluid dynamics in urbanconstruction planning CVFEM. [Electronic version].Journal of Southeast University (Natural Science), 35(Z1), pp.301-304.

Kaminaris, S. D., Tsoutsos, T. D., Agoris, D., &Machias, A. V. (2006). Assessing renewables-to-

347


electricity systems: A fuzzy expert system model.Energy Policy, 34(12), pp. 1357-1366.

Kanda, M. (2006). Progress in the scale modeling ofurban climate: Review. Theoretical and Applied Climatology,84(1-3), pp. 23-33.

Landberg, L., Myllerup, L., Rathmann, O., Petersen, E.L., Jørgensen, B. H., Badger, J., et al. (2003).Wind resource estimation—An overview. Wind Energy,6(3), p. 261.

Landberg, L., & Watson, S. J. (1994). Short-termprediction of local wind conditions. [Electronicversion]. Boundary-Layer Meteorology, 70(1-2), pp. 171-195.

Li, M., & Li, X. (2005a). Investigation of windcharacteristics and assessment of wind energypotential for Waterloo region, Canada. EnergyConversion and Management, 46(18-19), pp. 3014-3033.

Li, M., & Li, X. (2005b). MEP-type distributionfunction: A better alternative to Weibull functionfor wind speed distributions. Renewable Energy, 30(8),pp. 1221-1240.

Liberal Party of Canada. (2007). Without numbers toprove it, Conservative green plan is “EcoFraud”.[Electronic version]. Media Release.

Liu, H. (1991). Wind Engineering. A handbook for structuralengineers. New Jersey: Prentice Hall.

Liu, J., & Kotoda, K. (1997). Evaluation of surface-layer wind profiles with Heife observations.[Electronic version]. Boundary-Layer Meteorology, 83(1),pp. 27-41.

Louka, P., Belcher, S. E., & Harrison, R. G. (2000).Coupling between air flow in streets and the well-

348


developed boundary layer aloft. Atmospheric Environment,34(16), pp. 2613-2621.

Lun, I. Y. F., & Lam, J. C. (2000). A study of weibullparameters using long-term wind observations.Renewable Energy, 20(2), pp. 145-153.

MacDonald, R. W. (2000). Modelling the mean velocityprofile in the urban canopy layer. Boundary-LayerMeteorology, 97(1), pp. 25-45.

Macfarlane, I., & Joubert, P. N. (1998). Effects ofsecondary flows on developing, turbulent, rotatingboundary layers. Experimental Thermal and Fluid Science, 17(1-2), pp. 79-89.

MapQuest, Inc. (2007). MapQuest maps. Accessed throughhttp://www.mapquest.com/maps/.

Marlec Engineering. Rutland 913 technical data. Retrieved 12May, 2007, fromhttp://www.marlec.co.uk/products/prods/rut913.htm.

Martilli, A., Clappier, A., & Rotach, M. W. (2002). Anurban surface exchange parameterisation formesoscale models. Boundary-Layer Meteorology, 104(2), pp.261-304.

Martilli, A. (2002). Numerical study of urban impact onboundary layer structure: Sensitivity to wind speed,urban morphology, and rural soil moisture. Journal ofApplied Meteorology, 41(12), pp. 1247-1266.

Maruyama, T., & Ishizaki, H. (1988). Wind tunnel teston the boundary layer characteristics above an urbanarea. Journal of Wind Engineering & Industrial Aerodynamics, 28 pt1(1-3), p. 139-148.

Mengelkamp, H. T. (1999). Wind climate simulation overcomplex terrain and wind turbine energy output

349


estimation. Theoretical and Applied Climatology, 63(3/4), p.129.

Mertens, S. (2003). The energy yield of roof mountedwind turbines. Wind engineering, 27(6), pp. 507-517.

Mertens, S., van Kuik, G., & van Bussel, G. (2003a).Performance of a high tip speed ratio H-darrieus inthe skewed flow on a roof. Paper presented at the41st Aerospace Sciences Meeting and Exhibit, Reno Nevada. , AIAA-2003-0523. Retreived fromhttp://www.lr.tudelft.nl/live/binaries/3dcbe092-4334-4d47-9f82-dff9ed15ab5e/doc/aiaa2003-sm.pdf;http://pdf.aiaa.org/preview/CDReadyMASM03_582/PV2003_523.pdf.

Mertens, S., van Kuik, G., & van Bussel, G. (2003b).Performance of an H-Darrieus in the skewed flow on aroof. [Electronic version]. Journal of Solar EnergyEngineering, 125(4), pp. 433-440.

Mfula, A. M., Kukadia, V., Griffiths, R. F., & Hall, D.J. (2005). Wind tunnel modelling of urban buildingexposure to outdoor pollution. Atmospheric Environment,39(15), pp. 2737-2745.

Miller, C. A., & Davenport, A. G. (1998). Guidelinesfor the calculation of wind speed-ups in complexterrain. Journal of Wind Engineering & Industrial Aerodynamics,74-76, pp. 189-197.

Miller, D. L. (1981). Estimating the wind's potentialfor small scale energy generation using availablelocal climatological data. The 2nd Miami InternationalConference on Alternative Energy Sources (1979), Miami Beach,Florida. , 4 pp. 1623-1633.

Ministry of Natural Resources (MNR). (2006a). Ontariowind resource atlas. Retrieved fromhttp://www.ontariowindatlas.ca.

350


Ministry of Natural Resources (MNR). (2006b). Ontariowind resource atlas - maps(http://www.ontariowindatlas.ca/en/maps.html ed.).Ministry of Natural Resources. Accessed throughhttp://www.ontariowindatlas.ca/en/maps.html.

Ministry of Revenue. (2007). Wind, micro hydro-electric andgeothermal energy systems rebate. Retrieved fromhttp://www.rev.gov.on.ca/english/refund/windgeo/.

Mortensen, N. G. (2005). The world of wind atlases - Europesection. Retrieved with permission fromhttp://www.windatlas.dk/Europe/Index.htm.

Mortensen, N. G. (2006). The world of wind atlases -Observational wind atlas. Retrieved fromhttp://www.windatlas.dk/Home/Methodology/OWA.html.

Mortensen, N. G. (2006a). The world of wind atlases - Numericalwind atlas. Retrieved fromhttp://www.windatlas.dk/Home/Methodology/NWA.html.

Mortensen, N. G. (2006b). The world of wind atlases - Windatlases of the world. Retrieved fromhttp://www.windatlas.dk/World/About.html.

Mortensen, N. G. (2007). The world of wind atlases - About thewind atlas. Retrieved fromhttp://www.windatlas.dk/Home/About.html.

Municipal Property Assessment Corporation (MPAC).(2007). AboutMyProperty – 22 Veevers Drive. Retrieved May31, 2007, throughhttp://www.mpac.ca/pages_english/products_services/aboutmyproperty.asp.

Murakami, S., Ooka, R., Mochida, A., Yoshida, S., &Sangjin, K. (1999). CFD analysis of wind climatefrom human scale to urban scale. Journal of WindEngineering & Industrial Aerodynamics, 81(1-3), pp. 57-81.

351


National Renewable Energy Laboratory (NREL). (1986).Wind energy resource atlas of the Unites States(http://rredc.nrel.gov/wind/pubs/atlas/ ed.).National Renewable Energy Laboratory (NREL).Accessed throughhttp://rredc.nrel.gov/wind/pubs/atlas/.

National Renewable Energy Laboratory (NREL). (1997).United States annual wind resource potential - Dynamic map(http://mapserve2.nrel.gov/website/wind_resource1/viewer.htm ed.). National Renewable Energy Laboratory(NREL). Accessed throughhttp://mapserve2.nrel.gov/website/wind_resource1/viewer.htm.

Natural Resources Canada (NRCan). (2006). The atlas ofCanada. Accessed throughttp://atlas.nrcan.gc.ca/site/english/maps/topo/map.

Nigim, K. A., & Parker, P. (2007). Heuristic andprobabilistic wind power availability estimationprocedures: Improved tools for technology and siteselection. Renewable Energy, 32(4), pp. 638-648.

Noyes, S. (2004). Robarts front. Wikimedia Commons:wikipedia.org. Retrieved 2007, fromhttp://en.wikipedia.org/wiki/Image:Robarts_front_750px.jpg;http://en.wikipedia.org/wiki/Robarts_Library.

O’Looney, J. (2001). Sprawl decisions: A simulation anddecision support tool for citizens and policymakers. Government Information Quarterly, 18(4), pp. 309-327.

Office of Development, University of Waterloo. (2007).Faculty priorities. Retrieved June 9, 2007, fromhttp://development.uwaterloo.ca/priorities/faculty.html.

352


Office of Energy Efficiency (OEE). (2005). The stock ofhouses in Canada. Retrieved June 5, 2007, fromhttp://www.oee.nrcan.gc.ca/publications/infosource/pub/energy_use/sheu_e/sheu_1.cfm.

Office of the Prime Minister. (2007a, March 6). PrimeMinister Harper announces ecoTrust funding forOntario - Ontario investment will deliver realresults for Canadians. [Electronic version]. NewRelease.

Office of the Prime Minister. (2007b, February 12).Prime Minister unveils new Canada ecoTrust - Quebecinvestment the first of new funding to deliver realresults for Canadians. [Electronic version]. NewRelease.

Oke, T. R. (1978). Boundary Layer Climates. London: Methuen& Co. Ltd.

Ontario Power Authority (OPA). (2007). Ontario's integratedpower system plan: The road map for Ontario's electricity future(Preliminary Plan). Toronto: Ontario Power Authority(OPA). Retrieved fromhttp://www.powerauthority.on.ca/IPSP/Storage/41/3628_REv._IPSP_brochure_Feb._2007_for_Web_site.pdf.

Park, J. (1981). The Wind Power Book. Palo Alto, CA:Chesire Books.

Perry & Associates. (2001). University of British Columbia midcampus neighbourhood plan - Local street two side parking.Vancouver, Canada: University of British Columbia.

Petersen, E. L., Mortensen, N. G., Landberg, L.,Højstrup, J., & Frank, H. P. (1997). Wind powermeteorology No. Riso I-1206(EN)). Roskilde, Denmark:Riso National Laboratory. fromhttp://www.wasp.dk/Download/DownloadFiles/General/Wind%20Power%20Meteorology.pdf

353


Petersen, E. L., Mortensen, N. G., Landberg, L.,Højstrup, J., & Frank, H. P. (1998a). Wind powermeteorology. Part I: Climate and turbulence. WindEnergy, 1(1), p. 2.

Petersen, E. L., Mortensen, N. G., Landberg, L.,Højstrup, J., & Frank, H. P. (1998b). Wind powermeteorology. Part II: Siting and models. Wind Energy,1(2), p. 55.

Peterson, C. B. (2004). Green venture EcoHouse sustainablelandscape master plan. Guelph, Ontario: C. Brad PetersonEnvironmental Management and Landscape Architecture.Retrieved with permission May 31, 2007, fromhttp://www.greenventure.ca/ecohouse.asp?ID=163;http://www.bradpeterson.ca.

Peterson, C. B. (2005). Green venture EcoHouse. Guelph,Ontario: C. Brad Peterson Environmental Managementand Landscape Architecture. Retrieved fromhttp://www.bradpeterson.ca/images/projects/GV_EcoH01.jpg.

Pinard, J., Benoit, R., & Yu, W. (2005). A WEST windclimate simulation of the mountainous Yukon.Atmosphere-Ocean, 43(2), pp. 259-282.

Pino, D., Vila-Guerau De Arellano, J., Comeron, A., &Rocadenbosch, F. (2004). The boundary layer growthin an urban area. Science of the Total Environment, 334-335,pp. 207-213.

Plate, E. J. (1995). Urban climates and urban climatemodelling: An introduction. In J. E. Cermak, A. G.Davenport, E. J. Plate & D. X. Viegas (Eds.), Windclimate in cities (pp. 23-29). Dordrecht, TheNetherlands: Kluwer Academic Publishers.

354


Plate, E. J. (1999). Methods of investigating urbanwind fields--physical models. Atmospheric Environment,33(24-25), pp. 3981-3989.

Plate, E. J., & Kiefer, H. (2001). Wind loads in urbanareas. Journal of Wind Engineering & Industrial Aerodynamics,89(14-15), pp. 1233-1256.

Rabah, K. V. O., & Mito, C. O. (2003). Pre-designguidelines for passive solar architectural buildingsin Kenya. International Journal of Sustainable Energy, 23(3),pp. 83-119.

Radics, K., Bartholy, J., & Pongrácz, R. (2002).Modelling studies of wind field on urbanenvironment. [Electronic version]. Atmospheric, Chemistryand Physics Discussions, 2(6), pp. 1979-2001.

Raigatta Energy Inc. WS - 0,30B specification sheet. Retrieved12 May, 2007, fromhttp://www.raigattaenergy.com/products.php?pid=20.

Ramachandra, T. V. (2006). Solar energy decision support system. International Journal of Sustainable Energy, 25(1), p. 33.

Ramachandra, T. V., Krishna, S. V., & Shruthi, B. V.(2004). Decision support system to assess regionalbiomass energy potential. [Electronic version].International Journal of Green Energy, 1(4), pp. 407-428.

Rawn, B. (2006). Student-driven projects - Progress inU of T’s energy sustainability community.[Electronic version]. EnviroNews, fall 2006.

Renewable Devices Energy Solutions Ltd. (2006). Swiftrooftop wind energy system. Retrieved 12 May, 2007, fromhttp://www.renewabledevices.com/swift/SWIFT%20Rooftop%20Wind%20Energy%20System.pdf.

355


Ressource Naturelles et Faune, Québec. (2004). Cartes desgisements éoliens exploitables. Accessed throughhttp://www.mrnf.gouv.qc.ca/energie/energie/energie-sources-vent-inventaire.jsp.

Revi, A., Prakash, S., Mehrotra, R., Bhat, G. K.,Gupta, K., & Gore, R. (2006). Goa 2100: Thetransition to a sustainable RUrban design. Environment& Urbanization, 18(1), pp. 51-65.

Riley, W. J., Gadgil, A. J., & Bazaroff, W. W. (1996).Wind-induced ground-surface pressures around asingle-family house. Journal of Wind Engineering & IndustrialAerodynamics, 61(2-3), pp. 153-167.

Rooney, G. G. (2001). Comparison of upwind land use androughness length measured in the urban boundarylayer. Boundary-Layer Meteorology, 100(3), pp. 469-486.

Rotach, M. W., Vogt, R., Bernhofer, C., Batchvarova,E., Christen, A., Clappier, A., et al. (2005).BUBBLE – An urban boundary layer meteorologyproject. Theoretical and Applied Climatology, 81(3-4), pp.231-261.

Roulet, Y. (2004). Validation and application of anurban turbulence parameterisation scheme formesoscale atmospheric models. (PhD, EcolePolytechnique Federale de Lausanne).

Ruci, E. (2005). Geocoder.ca. Ottawa: CarletonUniversity. Accessed through http://geocoder.ca/.

Rylatt, M., Gadsden, S., & Lomas, K. (2001). GIS-baseddecision support for solar energy planning in urbanenvironments. Computers, Environment and Urban Systems,25(6), pp. 579-603.

356


Şahin, A. D. (2004). Progress and recent trends in windenergy. Progress in Energy and Combustion Science, 30(5), pp.501-543.

Sahin, B., Bilgili, M., & Akilli, H. (2005). The windpower potential of the eastern mediterranean regionof Turkey. Journal of Wind Engineering & Industrial Aerodynamics,93(2), pp. 171-183.

Scholtz, T., Dr., & Liu, H. (2005). Wind energy simulationand mapping system (EnSimWE) is in beta testing (Ortech PowerPoint. Mississauga, Ontario: Ortech Power. Retrieved20 January 2007, fromhttp://www.ortech.ca/pdf/power/powerPoints/July%2011,%202005%20-%20Wind%20Energy%20Simulation%20and%20Mapping%20System%20(EnSimWE)%20is%20in%20Beta%20Testing.pdf.

Seguro, J. V., & Lambert, T. W. (2000). Modernestimation of the parameters of the Weibull windspeed distribution for wind energy analysis.[Electronic version]. Journal of Wind Engineering & IndustrialAerodynamics, 85(1), pp. 75-84.

Singh, S., Bhatti, T. S., & Kothari, D. P. (2006). Areview of wind-resource-assessment technology. Journalof Energy Engineering, 132(1), pp. 8-14.

Sini, J., Anquetin, S. A., & Mestayer, P. G. (1996).Pollutant dispersion and thermal effects in urbanstreet canyons. [Electronic version]. AtmosphericEnvironment, 30(15), pp. 2659-2677.

Skote, M., Sandberg, M., Westerberg, U., Claesson, L.,& Johansson, A. V. (2005). Numerical andexperimental studies of wind environment in an urbanmorphology. Atmospheric Environment, 39(33), pp. 6147-6158.

357


Souch, C., & Grimmond, S. (2006). Applied climatology:Urban climate. [Electronic version]. Progress in PhysicalGeography, 30(2), pp. 270-279.

Southwest Windpower. (2007). Whisper 100/200 technicalspecifications. Retrieved 12 May, 2007, fromhttp://www.windenergy.com/documents/spec_sheets/0040_whisper_100-200_spec.pdf.

Steemers, K. (1997). Project ZED - Towards zeroemission urban development. In E. Fitzgerald, & J.Owen Lewis (Eds.), Solar energy in architecture and urbanplanning (Research Digest 8 ed., pp. 11-12). Dublin:The Energy Research Group.

Stull, R. B. (1988). An introduction to boundary layermeteorology. Boston: Kluwer Academic Publishers.

Tanaka, T., & Moriyama, M. (2005). Application of GISto make 'urban environmental climate map' for urbanplanning. Paper presented at the Fifth Symposium on theUrban Environment (AMS), Vancouver. Retrieved fromhttp://ams.confex.com/ams/pdfpapers/79052.pdf;http://ams.confex.com/ams/AFAPURBBIO/techprogram/paper_79052.htm.

Taylor, P. A., & Lee, R. J. (1984). Simple guidelinesfor estimating wind speed variations due to smallscale topographical features. Climatological Bulletin,18(2), pp. 3-32.

Thaduesz, F. (2007, May 11). Zero emissions in thedesert: New tower creates all its own energy.[Electronic version]. Spiegel Online International.

Tong, H., Walton, A., Sang, J., & Chan, J. C. L.(2005). Numerical simulation of the urban boundarylayer over the complex terrain of Hong Kong.Atmospheric Environment, 39(19), pp. 3549-3563.

358


Transportation Association of Canada (TAC). (1999).Geometric Design Guide for Canadian Roads. Ottawa, Ontario:TAC.

Troen, I., & Petersen, E. L. (1989). European Wind Atlas.Roskilde, Denmark: Risø National Laboratory.

Tweed, C., & Jones, P. (2000). The role of models inarguments about urban sustainability. EnvironmentalImpact Assessment Review, 20(3), pp. 277-287.

University of the Southern Pacific (USP), & UnitedNations Economic and Social Commission for Asia andthe Pacific (UNESCAP). (2004). Pilot training onwind energy conversion systems (WECS). Suva, FijiIslands: University of the Southern Pacific.

University of Toronto. (2006). Robarts Library. Retrievedfrom http://content.library.utoronto.ca/general-information/libraries/ROBARTS/.

University of Western Ontario. (2007). Western launches5M green building project. [Electronic version].Western News,April 19, 2007.

van Bussel, G. J. W., & Mertens, S. (2005). Smallwindturbines for the built environment. Paperpresented at The Fourth European & African Conference on WindEngineering (EAWCE4), (Paper # 210). Prague. Retrieved fromhttp://www.lr.tudelft.nl/live/binaries/3dcbe092-4334-4d47-9f82-dff9ed15ab5e/doc/2005Small%20wind%20turbines%20for%20the%20built%20environment.pdf.

van Bussel, G. J. W., Mertens, S., Polinder, H., &Sidler, H. F. A. (2004). The development of Turby®:A small VAWT for the built environment. Paperpresented at the Advanced Small Turbine Technology Session,Global Wind Energy Conference, Chicago. Retrieved fromhttp://www.lr.tudelft.nl/live/binaries/3dcbe092-

359


4334-4d47-9f82-dff9ed15ab5e/doc/Global_Wind_paper.pdf.

van Kuik, G. A. M., & Bierbooms, W. A. A. M. (2005).The status of wind energy. Paper presented at TheFourth European & African Conference on Wind Engineering (EAWCE4)(Paper # 260). Prague.

Verkaik, J. W., & Smits, A. (2001). Interpretation andestimation of the local wind climate. Paperpresented at the 3rd European and African Conference on WindEngineering, Eindhoven, The Netherlands, pp. 43-56.Retrieved fromhttp://www.knmi.nl/samenw/hydra/documents/3eacwe/paper.pdf.

Verkaik, J. W. (1999). A method for the geographicalinterpolation of wind speed over heterogeneous terrain (KNMIHYDRA Project Report). The Netherlands: RoyalNetherlands Meteorological Institute (KNMI).Retrieved fromhttp://www.knmi.nl/samenw/hydra/documents/geograph/interpol.pdf.

Verkaik, J. W. (2006). On wind and roughness over land.Dissertation pp.1-123, Wageningen University andResearch Centre. Retrieved fromhttp://library.wur.nl/wda/dissertations/dis3940.pdf.

Voivontas, D., Assimacopoulos, D., Mourelatos, A., &Corominas, J. (1998). Evaluation of renewable energypotential using a GIS decision support system.Renewable Energy, 13(3), pp. 333-344.

Voorspools, K. R., & D’ haeseleer, W. D. (2007).Critical evaluation of methods for wind-powerappraisal. Renewable & Sustainable Energy Reviews, 11(1), pp.78-97.

360


Walmsley, J. L., Taylor, P. A., & Keith, T. (1986). Asimple model of neutrally stratified boundary-layerflow over complex terrain with surface roughnessmodulations (MS3DJH/3R). [Electronic version].Boundary-Layer Meteorology, 36(1-2), pp. 157-186.

Wang, W., Rivard, H., & Zmeureanu, R. (2005). Anobject-oriented framework for simulation-based greenbuilding design optimization with geneticalgorithms. Advanced Engineering Informatics, 19(1), pp. 5-23.

Weber, R. O. (1999). Remarks on the definition andestimation of friction velocity. [Electronicversion]. Boundary-Layer Meteorology, 93(2), pp. 197-209.

Weisser, D., & Foxon, T. J. (2003). Implications ofseasonal and diurnal variations of wind velocity forpower output estimation of a turbine: A case studyof Grenada. International Journal of Energy Research, 27(13),p. 1165.

Whalen, T. M., Sadek, F., & Simiu, E. (2002). Database-assisted design for wind: Basic concepts andsoftware development. [Electronic version]. Journal ofWind Engineering & Industrial Aerodynamics, 90(11), pp. 1349-1368.

Wieringa, J. (1993). Representative roughnessparameters for homogeneous terrain. [Electronicversion]. Boundary-Layer Meteorology, 63(4), pp. 323-363.

Wind Stop. (2006). Windstop.org - the ugly truth about cape wind.Retrieved from http://www.windstop.org/.

Wineur. (2007). Catalogue of European urban wind turbinemanufacturers. Europe: Wind Energy Integration in theUrban Environment (Wineur), Intelligent Energy -Europe. Retrieved Jun 15, 2007, fromhttp://www.urban-wind.org/pdf/CATALOGUE_V2.pdf.

361


Wisse, J. A. (1988). A philosophy for teaching wind inthe built environment. Paper presented at The 3rdInternational Symposium on “Climate - Building – Housing”, Israel.Retrieved from Energy & Buildings,11(1-3), pp. 157-161.

World Wide Wind Energy Association (WWEA). (2006).World wide wind energy association (WWEA). Retrieved fromhttp://www.wwindea.org/home/index.php.

Xie, X., Huang, Z., & Wang, J. (2006). The impact ofurban street layout on local atmosphericenvironment. Building and Environment, 41(10), pp. 1352-1363.

Xie, X., Huang, Z., & Wang, J. (2005). Impact ofbuilding configuration on air quality in streetcanyon. Atmospheric Environment, 39(25), pp. 4519-4530.

Yamamura, S., & Kondo, Y. (1993). Numerical study onrelationship between building shape and ground-levelwind velocity. Journal of Wind Engineering & IndustrialAerodynamics, 46-47, pp. 773-778.

Yu, W., Benoit, R., Girard, C., Glazer, A., Lemarquis,D., Salmon, J. R., et al. (2006). Wind energysimulation toolkit (WEST): A wind mapping system foruse by the Wind Energy industry. Wind Engineering,30(1), pp. 15-33.

Zhang, N., Jiang, W., & Hu, F. H. (2006). Numericalmethod study of how buildings affect the flowcharacteristics of an urban canopy. [Electronicversion]. Wind & Structures, 7(3), pp. 159-172.

Zhang, N., Weimei Jiang, W., & Miao, S. (2006). A largeeddy simulation on the effect of buildings on urbanflows. [Electronic version]. Wind & Structures, 9(1), pp.23-35.

362

DECISION SUPPORT FOR URBAN WIND ENERGY EXTRACTION

Documents