Proceedings of theFourth International Conference on

Wind Effects onBuildings and Structures

HEATHROW 1975

Edited byDr Keith J Eatonau I L 01 NG RESEARCH ESTABLISHMENT

UNITED KINGDOM

CAMBRIDGE UNIVERSITY PRESSCAMBRIDGE

LONDON' NEW YORK' MELBOURNE

~

SESSION 6 - MEASUREMENT TECHNIQUES (continued)

7 The measurement of non- steady wind forces by R E Whitbreadon small-scale building models

Page 567

8 Discussion on Session 6 Rapporteur: H van Koten 575

SYMPOSIUM ON 'DESIGNiNG FOR WIND'

SESSION 7 - PRACTICAL APPLICATION

1 Summary paper for Session 7 by C W Newberry 585

PART I

2 Tests to determine wind speeds for.scouring and blow-off of rooftop gravel

3 The effects of wind on people in thevicinity of buildings

4 Design criteria for stability of cylindricalshells subj ected to wind loading

5 Glass and curtain wall: Effects of high winds:Required design criteria

6 A proposed European Code of Practice:Current work of the ECCS towards specifi­cation of the effect of wind on structures

by Dr R J Kind

by T V Lawson and A D Penwarden

by K S Prabhu, Dr S Gopalacharyulu andProfessor D J J ohus

by H S Saffir

by Dr D Sfintesco and Dr T A Wyatt

591

605

623

633

643

7 On the design of glass pane against windloading

8 Discussion on Session 7: Part I

by Professor H Ishizaki, S Myoshi andT Muira

Rapporteur: Professor W H Melbourne

655

663

PART II

Rapporteur: Dr T A Wyatt

by Dr M J O'Rourke and Professor R A Parmelee699

by Dipl-Ing H Panggabean, Dr G I Schuellerand Dr F M Wittmann 707

by Dr E Simiu 721

by J W Vellozzi (presented by Dr E Simiu) 735

by Dr G R Walker 745

755

669

685

by Dr P K Stansby and Dr L R Wootton

by Professor B G Korenev (not presented)

Equivalent static wind loads for tallbuilding design

Tall guyed tower response to wind loading

Investigations of wind design criteria usinga statistical simulation model

Discussion on Session 7: Part II

The value of wind research to civilengineering

On damping wind-induced vibrations offlexible structures

Serviceability analysis for Wind loads onbuildings

Reliability based design of slenderstructures under wind action

9

10

11

12

13

-----14

15

16

'Crown Copyright reproduced with the permission of the Controller of Her Majesty's Stationery Office.'

THE EFFECTS OF WIND ON PEOPLE IN THE VICINITY OF BUILDINGS

T V LawsonDeparbnent of Aeronautical EngineeringUniversity of Bristol United Kingdom

andA D Penwarden

Environmental Design DivisionBuilding Research Establishment United Kingdom

This paper deals with some aspects of the environmental wind studies performed inthe Department of Aeronautical Engineering of the University of Bristol and in theBuilding Aerodynamics Section at the Building Research Station, Garston, Herts.Each establishment handles a large number of enquiries concerned with pedestriancomfort and safety among groups of buildings, and each has developed its own tech­niques for helping clients. These techniques are to a large extent complementary,and it is the purpose of this paper to describe them and to show how they areapplied,~ A number of other wind-dependent factors which must be considered duringthe design of buildings are also included.

III

1. INTRODUCTION

In the body of the paper the authors con­sider the two major effects that wind hasupon man in a built environment - namely,the mechanical and thermal effects. In Sec­tion 2 these are considered separately andit is concluded that, within certain limita­tions, the criteria of accepta~ility forthermal comfort will automatically be satis­fied, providing the criteria for mechanicaleffects are met. Consequently, the bulk ofthe paper considers the mechanical effectsand how best to write criteria of acceptabi­lity.

At the inception of a design project, thequestion of whether a problem will arisefrom the mechanical effects of the wind onpeople should be considered. At an earlystage when the design is only half-conceived,the assessment will be crude and based uponpast experience alone. As the design har­dens, the quality of the assessment needs toimprove until the stage is reached when suf­ficient is knmvn about the detail design fora model to be built. At this stage, a de­cision has to be taken as to whether a suffi­cient problem exists to justify the cost and

time involved in conducting a windtunnel investigation. To make fulluse of a wind tunnel investigation,all parties to the study (the plan­ning officers, the architect, theconsulting engineer, the.client andthe wind tunnel specialist) shouldmeet to determine the objectives ofthe investigation and to evaluatethe subjective terms used to definethe wind environment.

A range of assessment proceduresis thus required, starting withthose based on experience and intui­tion only, developing into quantita­tive studies based upon experienceand thence to the full wind tunnelinvestigation. It is to be hopedthat the first two stages will formpart of all studies, but that thewind tunnel investigation will onlybe made if it is concluded in anearlier stage that there could be aproblem of sufficient magnitude towarrant the time and cost involved.

This paper considers two kindsof investigation, the quantitativestudy without a wind tunnel inves-

60S

T V LAWSON AND A D PENWARDEN

tigation, Section 3, and the full wind tun­nel investigation, Section 4. In Section 3the assumptions made to allow a generalisedsolution prescribe the form of the criteriaof acceptability so that the values of thecriteria to be used are discussed in thatSection. In Section 4 describing the fullwind tunnel investigation, more latitude isavailable in the writing of criteria so thatdifferent criteria are used; these are ex­plained following a discussion of the philo­sophy of criteria of acceptability in gene­ral.

An illustrative example of each type ofinvestigation is presented in its respectiveSection.

A discussion of other related topics whichmust be borne in mind by the designer isgiven in Appendix 1, while Appendix 2 givessome background information on wind and rain.

2. EFFECTS OF WIND ON MAN

2.1 Mechanical EffectsThe most serious direct effect the wind

can have upon man is to blow him over. Thiscan cause injury and sometimes death: Forexample two old ladies died in 1972 afterbeing blown over by wind close to tall buil­dings. These are extreme occurrences andcateful design and layout of buildingsshould ensure that they do not happen.

Of lesser severity is the imposition of awind load against which man has to work ashe moves from point to point, when the windcan create annoyances such as turning hisumbrella inside out or removing his hat.

Less seriously, the wind can make stan­ding and chatting to friends an uncomforta­ble experience; it can make the opening ofa newspaper difficult and the subsequentreading laborious due to the watering orblinking of the eyes, and there are otherminor effects.

A list of these and other effects isgiven in Table 1, based on references 1 and2.

It is to prevent hazardous or unpleasantwind conditions in man-made environmentsthat criteria of acceptability have to bewritten and agreed between planning autho­rity, client, architect and consultant engi­neer. The exact form in which they arecrouched depends upon the sophistication ofthe investigation envisaged, but in generalthey will demand that a given wind speed isexceeded for less than a stated duration.Suggested values for the criteria are givenfor each of the two types of investigation

606

described in Sections 3 and 4.Table 1 ~ Effects of wind on people

BeaU1'ort Description Windspeed EffectNumber mls

0 Calm 0-0.2

Light air 0.3-1.5 No noticeablewind

COl1FORT 2 Light breeze 1.6-3.3 Wind felt onface

3 Gentle 3.4-5.4 Hair disturbedbreeze clothing flaps,

newspaper dif-f icult to read

4 Moderate 5.5-7.9 Raises dustbreeze and loose

paper, hairdisarranged

DISCOl1FORT 5 Fresh 8.0-10.7 Force of windbreeze felt On bcxly,

danger pfstumbling whenentering a win-dy zone.

INCREASING 6 Strong 10.8-13.8 Umbrellas usedbreeze with diffi-

culty, hairblown straIght,difficult towalk steadily,sideways windCoree aboutequal to for-wards Walkingforce, windnoise on earsunpleasant

7 Near gale 13.9~17.1 Inconveniencefelt whenWalk1ng

8 Gale 17.2-£0.7 Generallyimpedes pro-gress, greatdifficultywith balancein gusts

DANGER 9 Strong 20.8-24.4 People blowngale over

2.2 Thermal EffectsAn attempt to aSsess the effects of wind

upon the thermal environment of man becomesextremely complex because of the inter-relation of so many different phenomena.The property which characterises man's sub-jective description of this condition ishis skin temperature. This is controlledfirstly by his metabolic rate and his workload. It is then modified by his clothingwhich acts as an insulation between himselfand his environment. Unfortunately, itsinsulation depends, among other quantities,upon the moisture content of the clothing(supplied either from

THE EFFECTS ON WIND ON PEOPLE

-6

-4

Winterclotheswithcoat

Businesssuit

lightsummerclotheS

Nu(Je

Businesssui!

lightsummerclothes

Winterclotheswithcoat

Nude

Wind speed lm/sl

011 ! I ! ! ! I I I I I 1 I 1 I !o 1 2 3 4 5 6 7 B 9 10 11 12 13 14 15

Wind speed (m/s)

_! I ! ! I I I ! ! I I I I ! I !

80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

-2

(b) Shade

~

•IE•

(al Sunshine

301,

2'

26

24

22

20

18

16

Q12 14

•2 12

\1" 10E•~

8

6

4

2

within as sweat or without as rain) and onthe extent to which it is penetrated by thewind. Heat is also removed from the exter­nal surface of the clothing by wind, andheat is added by radiation from the sun orany other source of heat. A state of equi­librium can eventually be established inwhich all surfaces attain temperatures forwhich the heat flow throughout is consis­tent with the thermal properties of thevarious parts.

Wind appears three times in this system:firstly as a work load against which man hasto walk, secondly by removing heat from theouter surface of the clothing, and thirdlyin penetrating the clothing and removingheat directly.

In the first effect, the energy needed toprogress against the wind load can be equa­ted to that involved in walking up a hill(Reference 1). For example, walking againsta 13 mls wind is roughly equivalent to wal­king at the same speed up a slope of 1 in 10with no wind. In each case the walker hasto work harder to maintain his speed, andthe extra heat produced by his body cancause sweating and discomfort.

The second effect has been examined byseveral experimenters for a range of clothingimpervious to wind aqd rain, both in sunshineand shade. Using a relationship derived byHumphreys (Ref. 3) and modified to includethe effects of solar radiation, and inser­ting subjective comfort descriptions alsodefined by Humphreys, it is possible to showthe combination of outside air temperatureand wind speed for different clothing whichconform to the subjective descriptions ofman's warmth. Examples are reproduced fromRef. 1 in Fig. I.

The third effect is difficult to assessbecause all the experiments which have sofar been conducted have concerned specialisedclothing (military, mountaineering, polar,etc), which is impervious to wind and rain.However, it is possible to postulate thattaking account of wind penetration wouldproduce a curve similar to the dashed curvein Fig. l(b). The philosophy behind thiscurve is that the porosity of the clothingwould progressively reduce its insulationvalue as the wind speed increased. Thedashed curve represents clothing with exces­sive porosity to wind.

Fig. I - Examples of thermal comfortconditions for strolling

607

II

T V LAWSON AND A D PENWARDEN

In an investigation in which subjectswere exposed to various wind and temperatureconditions for short periods, Hunt (Ref. 2)showed that the subjective assessment ofwarmth tended towards the colder descriptionas the wind speed increased. As this ten­dency is the same as that described in Fig.when the exposure has been long enough foran equilibrium heat flow system to be esta­blished, it is reasonable to accept thesecurves.

The interesting fact from these curves isthe small variation due to the wind if aperson is wearing the correct clothes whichare impervious to wind and rain. It wouldappear therefore that, assuming a tall buil­ding complex provides shelter from directsolar radiation, a person who is comfortablywarm in a windspeed at the top of BeaufortForce 2 will not experience an unacceptablechange of warmth if the speed increases tothe top of Beaufort Force 6. It followsthat if a person has dressed in imperviousclothing and correctly for the weather(taking into account both wind speed andtemperature) in shade at home, then his ther­mal comfort will not be greatly changed whenwalking in a building complex in which thewind speeds conform with the criteria basedupon consideration of the mechanical effects.

In the rest of this paper therefore cri­teria for and descriptions of manls environ­ment will be developed entirely in terms ofmechanical effects.

3. A QUANTITATIVE STUDY WITHOUT A WINDTUNNEL INVESTIGATION(Designing for comfort in pedestrianareas)

3.1 IntroductionThis aspect of .environmental studies is

one which has received much attention in theUK. The reason for this is the considerablenumber of developments, particularly shoppingcentres, where wind has proved to be a bigproblem. The ERE approach, described indetail in Ref. 4, for the case of a tallbuilding standing in a suburban area of lowbuildings, involves a combination of humancomfort considerations, meteorological winddata, and building aerodynamics. A meanhourly wind speed of 5 mls is taken as themaximum for comfort, and the frequency withwhich this is exceeded in pedestrian areasis used as a comfort criterion. The proce­dure involves knowing the height H and widthW of the tall building, the average heighth of surrounding low buildings, the average

608

hourly wind speed for the locality, and therelevant region of increased speed due to thetall building - usually corner-steams orthrough-flow, as described below.

3.2 High speed regionsA description of the basic pattern of wind

flo'\o1 around a single tall building with a lowbuilding to wind'tolard is helpful t(,l designers.Fig. 2, based on wind tunnel tests, showsthat as the wind approaches a tall buildingit gradually diverges until, at the windwardwall, upward and downward flow can be seen.Some of the air deflected domlwards rolls upto form a vortex in front of the tall buil­ding. The vortex stretches out sideways andwraps around the building in a typical horse­shoe shape. The flow close to the windwardwall radiates from a central region at aboutthree-quarters of the building height, anda considerable quantity of air flows down­wards and outwards to be concentrated nearground level at the windward corners. Theseaccelerated airstreams pass around the cor­ners as two jets of air which penetratedownwind for a distance similar to the heightof the building. If the building is raisedabove ground on columns, or is pierced by anarcade or passageway, wind flows at highspeed through the opening.

.~8

Fig. 2 - Basic pattern of wind flow around atall building

I11III

THE EFFECTS OF WIND ON PEOPLE

This flow pattern results in three mainregions in which increased speeds are expe­rlenced by pedestrians. These vortex-flow,corner-streams, and through-flow regionsrespectively produce maximum speeds VA. at A,VB at B, and Vc at C, see Fig. 2, which varywlth the buildlng dimensions and with theheight of surrounding buildings. Fig. 3shows some typical relationships for corner­streams and through-flow found in tests atthe BRE, with speeds expressed as ratiosR

Hto the free wind speed VH at building

height H. Represent'ative values for a widebuilding, four or more times the height ofsurrounding buildings, are VA/VB = 0.5,VB/V

H= 0.95, and VC/V = 1.2, all speeds

representing mean hour¥y values at full­scale.

Fig. 4 - Tbe area around a tall buildingaffected by corner-streams

~~

~Section AA . ~I. ~~.'

~.

~R:1. ~---~ ..~~ - ~'-",--"0:. ~ ~.='-1---".

wind angles. In particular, maximum speedsVB in corner-streams vary very little withw,nd angle (typically ± 15% during rotationthrough 360 degrees), although the positionat which the maximum occurs does change.Fig. 4 sketches the area around a singletall building where conditions for pedes­trians are likely to be adversely affectedby corner-streams.

N

1

8

-

W/H " 0·5

6

Through_How VC/V/i

HI"

CorfHlr $\fcams VB/~\

5

:1/

10

08

xa:

06

04

02

043

Fig. 3 - Typical ratios~. Width of tallbuilding W, height of surrounding low buil­dings h.

These values are for wind perpendicularto the wide face of a tall building. Otherwind angles produce different $peeds in thevortex region, but maximum speeds in thecorner-streams and in through-flow are notsensitive to direction over a wide range of

3.3 Wind FrequenciesThe relative constancy of the ratio

R = VB/VH

allows a simple calculation oft*e frequency with which a mean hourly speedof 5 mls is exceeded in corner-streams. VBwill be equal to 5 m/s when VH is 5/~ m/s;that is, when the meteorological-site speedV

IOis 5/RHS. Here S is a factor Vj/V10

based on predicting the effect on tHe open­site speed at 10m in passing over urban

..609

T V LAWSON AND A D PENWARDEN

H metres 10 20 30 40 50 60 70 80 90 100

Table 2 -Variation of S (~VH/V10) with tall buildingheight H

Table 3 - Variation or cumulative frequency with V101;(1o

VlO'VlO 0 0.1 0.20.30.40.5 0.6 0.7 0.8 0.9

F'requency, % 100 95 89 81, 78 73 68 62 57 51

Vl0'VlO 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

Frequency, % 46 40 35 30 24 2016.012.710.0 7.8

Vl dVl0 2.0 2.1 2.22.32.42.5

Frequency, % 5.9 4.5 3.3 2.4 1.7 1.2

terrain; values are given in Table 2. Scould also be modified to take account oflocal hill and valley configurations. Thefrequency of speeds greater than V

IOat the

meteorological site can be found in Table 3from the ratio vIOIVIO' where VIO is thelong-term average hourly windspeed at themeteorological site, available from a windmap of the UK (Ref. 4). Thus it is neces­sary to find the frequency equivalent to aratio 5/RHSV I0 '

RH

7H8

1-20

r:1-22 0·98

0-96l:241-24 0·92

1 100'901'20

110·85HO

+121'02 0·80

13

14

15

16

Through flow ICorner streams

3

Fig. 5 - Nomogram for calculating the fre­quency of sp8eds greater than 5 m/s.

70

80

105

30

60...r100

85

70

90

50~

40

HIs80

75

20t

S vlO f% R,S R,1 1 I I

110t7

105 ,~70

130

6+, 125

100+

r60 120

110 11595+ T f

50 110

5~100

90t105

90100

85t J,.-,···/TO

rr90

10

80+ I70

'5 r2 80

1 60

75

0.60 0.73 0.82 0.89 0.94 0.991.041.081.11 1.14s

As an example, consider a building 50 mtall among suburban houses 10 m tall, in anarea where VIO is 3.9 m/s. From table 2S = 0.94 for R = 50 m, and from Fig. 3R1 = ~B/VR = 0.95 for H/h = 5.0. Then5~SV10 = 5/0.95 x 0.94 x 3.9 = 1.44. Thecorresponding frequency obtained from Table 3is 22%, and this is the frequency of speedsgre~ter than 5 m/s in the corner-streamsregion around the tall building.

Simple though this calculation is, theprocess can be made even easier by using thenomogram of Fig. 5. Lines showing the calcu­lation in the paragraph above are superimpo­sed on the nomogram.

The calculation has been applied to a num­ber of sites which have been brought to theattention of the BRE as being unpleasantlywindy. These sites are plotted in Fig. 6,which is a graph of RHS against VIO' withlines of equal frequency included. In allcases, complaints had been made about wind

610

THE EFFECTS OF WIND ON PEOPLE

conditions, and in some cases protectivemeasures were taken after consultation withBRE. The points shown as open circles referto those sites where roofs or screens havebeen installed to improve conditions. Theblack circles concern centres which wereinvestigated following complaints, but whereno action was taken. The crosses representsites where action is contemplated, but hasnot yet been taken. It can be seen thataction was taken only in cases where thefrequency was estimated to exceed 20%.

enquiries which come to the Building ResearchEstablishment. Over the past 10 years some300 such enquiries have been received fromarchitects, consulting engineers and planners.Several of these problems, particularly inthe early days of this work, have entailedwind tunnel tests, but most cases involvingtall buildings can now be adequately dealtwith by using the simple procedure above.In some cases the resulting appraisal hasled to radical changes in design which haveeliminated wind problems at a very earlystage.

Fig. 6 - Estimated frequencies of spe,edsgreater than 5 mls on windy sites.

3.4 Remedial ActionHaving had the wind problems diagnosed,

the designer needs advice on what action hecan take to improve conditons. Variousways of reducing windspeeds in pedestrianareas around tall buildings can be suggested:

At the design stage:(I) Avoid tall buildings - an obvious

remedy which is sometimes possible. In gene­ral it is buildings more than twice the heightof their neighbours which give trouble.

(2) Keep pedestrian areas away from tallbuildings - avoid the need for people towalk close to the corners of tall buildings,and particularly avoid routes beneath tallbuildings.

(3) Stand the building on a podium ­provided th~t the podium is large enough,the winds which will inevitably de deflecteddownwards by the building will be kept abovestreet .level.At the design stage or later:

(4) Build a roof over pedestrian areas ­this is a measure which has been used in manyof the cases indicated with black circles inFig. 6 (see Ref. 4 for details). The extracost may run into hundreds of thousands ofpounds, but a complete cure is possible.

(5) Install screens - floor-to-ceilingscreens, offset to form a simple labyrinth,can reduce the flow of wind beneath a buil­ding, while maintaining access for pedes­trians. Alternatively, solid screens withdoors can be used.

(6) Build canopies - model tests suggestthat a canopy at first-floor level, exten­ding some 6 to 8 metres out from the facesof a tall building, and having a parapet wall1 to 2 metres high around its periphery, willreduce speeds beneath the building to aboutone~third of their unobstructed speed. Itmay also be necessary to build screen wallsat strategic points beneath the canopy.

(7) Provide handrails - this represents

50%

407,

55,

30%

20%

--=-----,0/,5%

4 4·5\110 rn/s

353

0·2

12

04

06

08

It is concluded from this work that thefrequency of mean hourly speeds greater than5 mls is a useful, practical criterion basedon User Satisfaction. When this frequency,calculated by the method above, exceeds 20%,protective measures are likely to be needed.Between 10 and 20% complaint is likely butmay be insuffie~ent to provoke action.Below 10% conditions are likely to be satis­factory.

The procedure outlined above is used inhandling many of the environmental wind

161 '- oRemedial action taken

x Remedial action contemplated• No remedial action taken

14

lfJ 10I

0:

611

11III

T V LAWSON AND A D PENWARDEN

the minimum which can be done to assistpeople, but may be valuable to old peoplewalking up and down steps in a windy region.

(8) Use aerodynamic attachments - thisis a largely unexplored field, but it ispossible that fins, vanes, or deflectors, asused to control airflow over aircraft and indue ted air systems, could be adapted for useon buildings.

4. A FULL WIND TUNNEL INVESTIGATION

4.1 PreambleIn the situation when the building complex

consists of separate buildings of varyingheights, or when accurate measurements ofwind speed at locations in the complex arerequired, a full wind tunnel investigationmust be conducted.

4.2 Subjective Acceptability CriteriaIt is recommended that four subjective

terms be used to define the suitability ofan area. They are:-

(1) UNACCEPTABLE, when, as its nameimplies, the conditions will not be allowedunder normal circumstances.

(2) TOLERABLE, when conditions wouldcause annoyance and a search should be madefor a remedy. The possible cures shouldthen be examined in regard to aesthetics,economies or inconvenience, and a decisiontaken either to proceed with a cure which iscompatible with the basic design philosophy,or to accept the original conditions withperhaps a slight change or restriction tothe use of the area.

(3) OPEN, wheu conditions would be thesame as those in an imaginary situation withthe building complex razed to the ground,leaving an open area of limited extent. Thepurpose of this criterion is to establishthose situations for which the building orcomplex provides no shelter, but neitherdoes it aggravate the conditions.

(4) SHELTERED, when the building or com­plex provides positive shelter.

4.3 Suggested Numerical Values forthe Acceptability Criteria

The purpose of separating the subjectivecriteria from the numerical criteria is toallow a range of numerical values to beapplied to each subjective criterion atvarious parts of the site which have diffe­rent uses. This is attempted in Table 4.

Firstly, a subjective term is chosen; forinstance, UNACCEPTABLE. A series of acci­dents Or incidents are listed and the mini-

612

mum value of wind speed which could causeeach occurrence is assumed. The major diffi­culty now arises in attempting to determinethe frequency of occurrence of the incidentwhich conforms with the subjective descrip­tion.

At this point, four different probabili­ties Hill be defined:P. - the probability of the particular per­s6n being at the chosen location at the cor­rect time.P .. - the probability of the stated winds§€ed being exceeded for a given value ofthe reference wind in a sample of stationarylength.P... - the percentage of the time when the

111 d fl' I'state value 0' ve OClty at the ocatlonwill be allowed.P - the frequency of occurrence of the inci­dent (the joint probability).

As the separate probabilities can be iso­lated, the joint probability is the productof the other three, so that values have tobe ascribed to P, P. and P .. in order thatP h . 1 • 11. h . d... , t e quantlty requlred In t e Wln tun-n~tlinvestigation,·can be evaluated.

The value of P .. is the easiest tochoose. Assuming fEat the length of thesample at every location for every winddirection is greater than the 'stationarylength', it is possible to draw a uniqueprobability density curve for the readingscovering the central 98% of the readings.From a series of samples each of stationarylength, the actual value of the readings inthe largest 1% and smallest 1% differs fromsample to sample. It is therefore impossi­ble to determine a value of velocity whichwill be exceeded for less than 1% of thetime from a single sample of stationarylength. Procedures to obtain less frequentvalues are long and are described in Ref. 5.For our application, we can ignore the com­plication and accept P .. as 10-2 in allinstances. II

Values for Pi can be obtained by guess­work. For example, in the case of the oldlady riding a bicycle, the value of 10-8was determined thus: the chance of anyonebeing at such a location is less than 1 in102 . If we consider 4 in 10 2 of the popu­lation are old ladies, of whom 2.5 in 10 3ride bicycles and of them only 1 in 10 2would go out in high wind conditions, weobtain a value of 1 in 10 8 as the valuefor P .•

Th~ value of the j oint probability isharder to determine, but it is suggestedthat the values in Table 4 are reasonable.

THE EFFECTS OF WIND ON PEOPLE

T3bl~ /, - Sugs","'"'' val"oa of accep'abili,y cri,",in

If the values of Table 4 can be accepted,it is then possible to define UNACCEPTABLE,TOLERABLE for each location on the site, andthe OPEN criterion for the site as a whole.

Having determined P, P. and P,o, the valuesof P ... are obtained f6r eachlgituation.Thes~lfepresent the limiting value of theduration in which the wind speed exceeds thevalue in the critical velocity column.

17·15

iJ.85

20·75

24·45

10·75195s:43-3

231020301minutes hour

4.5 The Wind Tunnel InvestigationThe first choice to be made in a wind tun­

nel investigation is the scale of the modeland, because of the size of most wind tunnelsengaged in this work, this should be between1/150 and 1/500. Within these scales, it ispossible for the whole or lower part only ofthe atmosphere to be correctly simulated inthe wind tunnel with respect to the velocitygradient, the turbulence gradient and theenergy spectrum of the wind. This is essen­tial to the accuracy of the results and itis recommended that all clients for a windtunnel investigation require both a trace ofthe variation of wind speed with time(Fig. 8) and the energy spectrum of the refe­rence wind (with the atmospheric curve from

Fig. 7 - Variation of windspeed with avera­ging time.

ranges of the Beaufort Scale, it is obviousthat the numerical value of the boundarydiffers with the averaging time of the mea­surement. A study of the variation of windspeed with averaging time in Ref. 6 suggeststhat it is a function of the intensity ofturbulence only, and as this value differsfrom location to location in a complex, itis impossible to derive a unique relation­ship. At Bristol the boundaries of theBeaufort Scale are defined by constant velo­cities for any averaging time between onehour and ten minutes, and for shorter avera­ging times by the logarithmic variation sug­gested in Ref. 6 for an intensity of turbu­lence of 28%. These boundaries are plottedin Fig. 7 on which the values of wind speedat the boundaries are quoted for averagingtimes of 3 seconds and 15 minutes.

40~35'5

1--3029·3

~ 23-1•lt20 ,-.~ 18·4!'

13-610 force 4

93 force 35-7

forces 1.2

2 3 5 10 2030seconds

Separate hobabilitic"

._-'. PH Piii

10-8 10-2 1O~1

10'6 10-2 10- 2

11)-7 10-2 10-2

10-5 lO-l 10-2

\0- 4 10-2 2 x 10-2

;0-4 10-2 2 " 10-2

10- 3 10-2 I, x 10"2

10-4 \0- 2 ~ x 10-2

\0- 4 IO-?' t, x 10-2

10-3 10-2 /, x 10-2

10-3 10-2 I, " 10-2

10-2 2 x \0- 2

10-7. 7. x 10-2

10-2 2 x 10-2

\0- 2 I, x 10-2

10-2 6 x 10-2

10-2 5 x 10-2

CriticalVelocity

in r"ng"~ I J",ntof (he hobohhty

Bc.>ufortsoalc P

--- ------

Occurrence

!1.'!!!£.£.<,1>}i\blc

Old lady blo"" off bicycle > , 10-\1

Old lady bi""" over "hile ""tking , . 10-10

Child hlo,"", over "hit. "alking '" 10- 11

Adult blo,"", ovor "hile "alking " 10-9

Adut, "alkitlg around build ins '"'); x 10-8

Ped... oi"n prcdnt-""lk through arM~ > ; 2 x 10-8

- ...,ndins arc"., <loors '" 4 x 10-7

Toler"ble

",lult "alkiug aro\lud building. ".,

f, x 10

pede..rian prednt-".,Il< tbrollgh "rc.," '" f, x 10-8

- """ding "rcao, door. > , 4 x 10-7

- covered .'<C.15 "-,

" x 10

4.4 The Beaufort ScaleIn 1805 Admiral Beaufort encountered the

difficulty of obtaining accurate estimatesof wind velocity from his unskilled sailors.He divided wind speed into ranges, each ofwhich affected the sea or the ship in a dif­ferent way, so that the sailors could lookat sea or ship and evaluate the wind speedin terms of Beaufort Forces. The descrip­tion used would be meaningless to landlub­bers, so we have produced in Table 1 a LandBeaufort Scale in which the descriptionsrefer to occurrences readily experienced bylaymen. We propose to use this as our mea­sure of wind speed, because the division ofwind speeds which occur in building complexesinto 10 ranges produces a scale coarse enoughto yield a sensible number of categories andyet fine enough for each range to have a dif­ferent meaning.

The values of wind speed used to definethe ranges of the Beaufort Scale in Table 1are values averaged over one hour. Assumingthat the trace of wind speed against time inFig. 7 represents the boundary between two

Q£:.!.'.. - ApproprL't~.v"\\le" <ou.t be t"l<C!~..!5'~~edg0~9:.

C,',c. ,n Sou,h Sos< l> 2 2 x 10-1•-,

,n !Iorth lIest > ) 2 x 10-,Urb"n 11\ South f.aot > 3 2 x 10.',

1n north lie.' > ) 4 x 10

1I.u,al ill Sou,h Soot > 3 6 x 10-4

in 110nh lIest > 4 5 x 1O~1,I _

613

iIlIII

T V LAWSON AND A D PENWARDEN

Ref. 7 superimposed upon it), (Fig. 9) to beincluded in every report on the investiga­tion. It is not suggested that the clientshould examine these in detail; their pre­sence will demonstrate that the aerodynami­cist has given the simulation of the atmos­phere sufficient attention.

or importance to the client. At Bristol itis customary to fix a small numbered disk ateach location chosen to facilitate the exactpositioning of measurements for all winddirections. A Xerox copy of a photograph ofthe model can be used to identify locationsin the report; it is also useful to includethe direction of the 'North Wind' on thisphotograph (see Fig. 10).

Fig. 8 - Variation of reference winds peedwith time.

000 2·50 s-oo

Windtunnel time (sees)

7'60 1000

nS{n)

Fig. 9 - Comparison of turbulence spectra.

In the light of his experience, the aero­dynamicist will choose a number of locationson the model at which high wind speeds couldoccur. Included in the list of locationsstudied should be any of particular interest

Fig. 10 - Measuring points on a model.

Twelve wind directions at 30 0 intervalsare normally used. The justification forthis is twofold; firstly, that 'frequency ofwind speed and direction' Tables (an excerptis shown as Fig. 11) produced by the Meteo­rological Office, Ref. 8, are used as thesource of data on the wind, and these splitthe wind directions into 30 0 sectors •Secondly, this is considered justifiedbecause, with the correct simulation of theatmosphere, the wind direction is outside+ 150 of the nominal direction for some 30%of the time. The choice of 12 wind direc­tions is therefore a compromise which isboth practical in using data from othersources and economically viable in time andcost.

Most of the occurrences, which have beenused to determine the values of our accepta­bility criteria, are produced by short termgusts and not by steady winds blowing forlong periods. In Section 4.4, describingthe Beaufort Scale, values of wind speedaveraged over 3 seconds were added to theusual boundaries of the Beaufort Scale. Thedifficulty in that instance was that the

10-0

•••••

Curve from ESOU 74031.,!!rstrong winds

2: SOm 20 ,lm.•,. .

• .t.~......-'l\ ••

•

•

0·1 1-0"Iv (wind tunnel) m· l

01

0·01

0·3

02

614

THE EFFECTS OF WIND ON PEOPLE

relationship between wind speed and avera­ging time depends upon the intensity of tur­bulence. The same difficulty arises inevery instance when this 'correction' is per­formed. The conclusion is therefore thatgust velocities (using short averaging times)must be measured in a wind tunnel investiga­tion, in addition to values averaged overlong time intervals.

In the literature concerning wind tunnelinvestigations, as well as most loading codes(for example, Refs. 9 and 10) the referencewind speed is usually either the hourly ave­rage or the 3 second average value measuredat the height of the building. The meteoro­logical data is usually referred to what iscalled 'meteorological standard conditions'which mean values averaged over one hourmeasured 10 m above open flat level countryin the vicinity of the site. The relatingof the wind speed at building height in thetown or city to the value at meteorologicalstandard conditions is still an art andought not to be left to the layman. Conse­quently, all velocities in reports oughtto be referred to 'meteorological standardcondi tions' •

for each of 12 wind directions. In the caseof the 3 second average values, the windspeeds quoted ought to be the values whichare exceeded for 1% of the sample which mustexceed the stationary length. Fig. 12 illus­trates such a form of presentation.

4.6 The Reduction of ResultsFrom the meteorological data in the form

of frequency of wind speed and directionTables (typically illustrated in Fig. 11)and the wind tunnel results (Fig.12), theduration of time when the wind speed atevery location, averaged over 3 seconds and15 minutes, is in each range of the BeaufortScale can be computed and presented as 'Hoursof Wind' Tables, as in Fig. 13.

3D 1 601 90 1120 1150 1160 1 210 1 240 1270 I 300 . 330

AVER'Nt>TIHE ILOC,

H1N.SEC

WIN 0 D 1 , ECTION

,.

TABLE X - PERCENTAGE FREQUENCIES OF WINDS ATSELECTED SIATIONS

BIRMINGHAM (EDGDASTON) (c(mld)

Mean windPerccn"'se nllmber of hOllrs whh winds f,omspeed

kno<s m.p.h.)~OQ 0200 0500 0800 1100 I~OQ 170Q 200Q2300 26QQ 290Q320Q All directions

oiOo O~OQ 070Q

1000 1300 160Q

190Q 2ioQ2;00 280Q 3ioQ )~oo 1"Q(:>1

YEARC~lm 0.'

'·3 '·3 '.1M 4·; I.; I.; I.; I.' I., ,., '.6 3.' 3.4 ,., ,., ,,, 27.57.10 9.12 I.; '" '.6 '.0 1.4 I.' 3.' 3.1 ,., 4.0 3.1 3.' 36.0

11.16 13-18 1.0 1.0 1.6 1.0 O.~ 0.' ,., 3.1 3.1 2.9 '.0 '.4 21.617.21 19·24 0.' 0.' 0.2 0.1 0.1 0.1 OA 0.6 0.6 0.7 0.' 0.' 4.122.27 25-31 O. .. 0> O. O. O. 0.1 0.1 0.1 0.1 0.1 0.1 0.628·33 32·38 o. O. O. .. .. O. O• O.3~..<I1) 39..<16 O. O. 0.> O.~1-47 ~7.~~

~8·55 ~5-63

%-63 6~_72

>6, '"TOTAL 4.6 4.' 6.1 4.' J,8 4.' 8.5 12.4 13.0 10.2 7.9 '.6 99.7

Pc'cen'~se nllmber of ho"'s missed 0.1

Fig. 11 - Percentage frequencies of winds.

Measurements in the wind tunnel investi­gation ought to be values at every locationof wind speed averaged over both 3 secondsand 15 minutes (say) expressed as percenta­ges of the meteorological standard wind speed

1 66 55 ~ 61 ~ 76 93 73 ~ 46 ~ 29

.1.L" Y i--~}- -§~+-gt rB-l-l}· .-~~~. f-t~· ~ --leo .H $~ 7;

.!~·'---T· --j-- '''§5---l--~~' --t} 4%- if 4t- ·..t}· _..~ -3~' ..g~. _.-~~-- -16'1L __J ....-} T~~' ·15}"~~-·-~g. .~~-- .J'~._. ,Jri·- --~t- --~~. ~i···n '1~~

J1-~-T --t~t ~~- -iJ- _.~-~ .. ~~~.. -~6' ,18~- ~~ 5~ .~~ ~~ ~~~-~ --··J"'I----l· -..~~-- .-~_. )J ._.~~- -1~-~ --rid" .. ~;- ..;~ -i~· _-.§~ ~~

1.2.-. -T ... ~- -166-' ~~ ..~~. jg. -,~~j~- .. gl . ,~~. '1~' ·H .~.g .tl·15 .7 26 23 ,16 J3 J1 31 21 .16 11 27 25 ,16f:';'--')" ·..·6- '106' 10)" -73"- '''94'' --6S" 50' '--8'6 '''93" -62: -'67 -·-56' '04'

rY--T- -..-} 1}. ·_tf --£1- --~~ -i~-- ..--ji- '-~l '11- J1'" §:J ~1 jg.1.? --3 ':16- _.j~- %i·· -~~ .~~- .~ .-~§ -i:' -,~~. :1j ._~§. ~~ ~~

15"' T -.~.~- :1~-- .~~, -gg' ':1§}' H .;~ ·gl .. j} J~' R jj 3-~

.,15 3 ·u-- -;~. J~ 3f -~~. ~~ -,~~. ···K -~t --~! t .~~ .. ~}- .. _J­_1_~--"r' --H' ···B·, :id~ }~-_ ...~g .. '16~ j~ .._~~- .* -j~t .§~ -1}--"l15 .1) 33 J3 26 31 47 37 25 26 ·10 1 ,16 21 30

) 1-4 6<1 '"60 60' 69' 74- 70 53 '''32 55' 71 80' R

~. 3 -~ --~} -~_.._~. -~}. -H· --H ·li~· 'fO~ -{~- ·3~'. ~~- ...§~,1L__). "-l~ ··§z- -4~' "a6' 1a~' ··~·t· 'i~' l8~ '11~ ··l~ -6R· .4S '4j

..1_5. - J t~- TH -~3'" ...~. -l}~" ..~- ·it _.~~. -~~l "'3-~- --J?,. -;6- "~l

..15~·T ~U--l!;- ,~~" '~6 .. ~~...._g}- .. -§-g. -~""i~-"-~~ffg- "f~' . ~tJS )'- ·-~·2 .--~~. -g~ -Jg. T~ T!~ -~~- --}~ ---H' "'1~-' .g~- "~d- -.~*J5. --J - _.:~- -i4-- ..~~.. -it _.~~. -}~- -,~~" ._}~ -'it ..~ ..~~- .~ --~~.

:15, ....'3 -~.~.. ··-u "J~ ~1~ _. ~~ .. 51· --~~ .. '1~- ._~~. _. :t~-t' '~8-- -~~ "'1~1; -,- .~~ -H- ._~~- -~. --jt· ~--~o ·n· ..{~ .._~~ ....%j- --}d' '~a' "'1~

.n···T -~3"-~B- .. ~~- --6l--·,!· "~J- .:~~. ji ..Jj "1~' ,..ji- -J~" -O!}~,15 231) ,19 ·15221711 ·12,10 7.6 4 6

------y- '-2'4-- J8" )'1- '-,y' -"'63- -7r' '64' -61'"'"95' -"S9'"r-'5F -67' ')"$15 24 8 .1,1 22 ).1 16 25 29 34 ,16, ,16 16 111

Fig. 12 - Velocities as percentages of thereference velocity.

615

T V LAWSON AND A D PENWARDEN

SHEET 1•.1 TABLE OF ACC£PT,\ElILITY CRITERIA USIN(; VALUES:-

2.0% 1.0% 10.0% 5.0% 4.0% 2.0% 5.0% 4.0% 5.0110 4.0~

Fig. 14 - Table of acceptability criteria.

in which two symbols have been allocated toeach range of the Beaufort Scale for veloci­ties exceeding Force 6 (E and e) to Force 2(A and a). Any percentage value can be allo­cated to each symbol, so that the UNACCEPTA­BLE and TOLERABLE criteria for a pedestrianstand-about area become (C, b) where Te'represents a situation when the wind speed isgreater than Beaufort Force 4 for more than4% of the time, and fbI a situation in whichthe wind speed is greater than Beaufort Force3 for more than 4% of the time.

In a column at the left-hand side of thesheet next to the location number is listedthe critical letters and a quick scan acrossthe line will show in which month, if at all,the conditions at that location, exceed thecriteria for either gusts of long durationwinds. A typical sheet is shown in Fig. 14 .

'....0'."--.­_.~ ..... aB~O;Q.": b

''''0'0....~. 4.- cO~";·~ d

_.~.?__~bOL denotes ACCEPTABl.E coml1 Hons)

£ '::'G'~ e

- --" 0 " T H S

eR1T YEARcoo

ls"JAH I~;~t; MAR I~~~'> "AY I~~~~ :~~y 'v, ~?,T J~;T HOV OEC

" " " " " " " 51 5 J,~-1--s-EC, , , ' 8 , " BA : ' AA , ' ; , AA : ' ; , ·,EC, 2 , "

, , , A ,{c.~. -, , , , , •n, , -s -,., ., BA SA SA SA , , 6A SA ., ., ., ·,"c.o 5 , b b b A , , , , , , b bn, 6 , B , 8 , ,

" , 6 " so• 6

, 8 , 6 , 8 " , 6;,C~_ 7 A , , , , AA , , , . , , , ,n, , BA , b , 6 , 8 , , BA b A , , b , b A SA BA ·,0, 9 AA SA , A AA , , , , ,, , , ,.AA , , , , , ,ECc 1B AA , , , , , , , , , , , AA , , , Ab , , , ,", U 6 , SA BA SA SA SA , , AA A , 6 , b , , , 8 ,·c.!;c 6 b , 6 , 8 , . so SA b ,

• b8 , 8 b 6 , BA , 8,,, 1) " , 8 , 6 " ,. so SA 6 6 so , . , 6 , . , 8

", " SA ., 8 , ·, bA b , AA AA b A b , 8 , SA 8 ,ce, 15 ., , , 8 , SA b , bA AA BA ·, SA b , b , ·,;e, 16;e, 17 6 , , 8 , , , , SA SA 8 , 6 , ·, ., SA ., , ,C" 16 , , , , , , , , , , A A ,,", 19 A , SA b , ·, b , , A Ab , , A , AA b A0, 20 8 , , b 8 , , b ., ·, b , 8 , SA SA 8 , SA , AdCe 21 6 • , 6 , . , 8 , 8

• 8• b •• ·, 6 6 8 8 6 6

• 8C" 22 , , A A , , , , , , , , A,", 2J AA SA b , 8 , A , AA ,. , , , , , , , , b ,C"' 24 , , • ,'C, 25 6 , , b , b SA • A SA b , 8 , SA 6 , 6 , SA , b

'" 26 , , , A , , Ab , , , , , , , , , " , .0, 27 SA , ,

" , b 8 , , , , , , , , . b , , , b , 8 ,

~CC128 ,C:he 29 , , , , , , , , , , , ,Cbe )0 , , , , , , , , , , , , ,~~~I ~~

, , •aCe' )) SA , A , , , A , , 6 A " 8A 6 , " 8 , 8 A BA,,«1)4 A , , , , • , , , , , ,"C: )S , . , , , A b " , A b A 6 , b 8 , , A

", )6 SA o , 6 , " AA b , , b A b , 6 , b , b , 6 ,

", 37 SA 8 , 8 , SA " SA , , 6 , b , 8 , b A SA 8 ,,", JB SA , A SA , A B , b , 6 , 8 , "

, ," BA

n, ", A , , , , , , , , , , ,

,", 40 " B A " 8 , b • b , 6 b 8 • b 6 6 A(be i41 A , A , , A , , , , , A ,0(1 42 , ,C:b,~l~)

, , , , , , , , , A , , , A,", ----

MONTH 10LOC. B2 8) 84 85 86 87 88 89 810

29 )20.) 146.8 4J.8 9·0 1.6 .0 ·0 ·0 .029 )40.4 132.0 41.6 8.1 1.) .0 .0 .0 .0)0 26).) 159.9 82.1 15.2 2.9 .1 ·0 .0 .0)0 291.) 133.8 69.0 2).4 5.0 ·9 .0 .0 .0)1 244.0 163.9 76.1 28.0 9.6 1.8 .1 .0 .0)1 268.1 111.7 74.0 40·9 19.2 7.1 2.2 .2 .0)2 )55.6 133.6 29.6 4.5 .1 .0 ·0 .0 .0)2 429. 4 66.5 23.1 4.2 .) .0 ·0 .0 .0)) 468. :3 48.6 6.5 .0 .0 .0 .0 .0 .0)) 487.0 12.9 .4 .0 .0 .0 .0 .0 .0)4 25.3.0 161.) 17 .1 2).9 6.8 1.2 .1 .0 .0)4 282.8 111.2 17.2 36·7 12.2 2.5 ·8 .1 .0)5 410.7 99.4 12.8 .5 .0 .0 .0 .0 .0)5 480.5 .36.1 6.8 .0 .0 .0 .0 .0 .0)6 255.5 139.5 97.2 24.6 5.7 ·9 01 .0 .0)6 269.9 96.0 88.6 48·9 14.9 ).7 1.) .1 .0)7 353.0 97.6 53·8 14.9 ).8 .) .0 .0 .0)7 368.4 74.5 5).7 19.! 6.1 100 ·0 .0 .0)8 266.0 153.1 83.7 15. 2.) .5 ·0 .0 .0)8 289.3 110.7 80.8 3.3.2 7.7 1.) .4 .0 .0)9 279.1 14).4 80.2 16.8 ).5 .4 .0 .0 .0)9 321.6 10.3.6 68.7 21.4 6.) 1.5 .2 .0 .040 262.8 153.0 66.) 17 .2 Jo9 .) • 0 .0 .040 334.4 94.6 65.8 21.4 6.5 .8 '0 .0 .041 297.5 140.8 69.3 1).1 2.4 .) .0 .0 .041 330.5 10).2 72.1 12.9 J,5 100 .) .0 .042 241.7 164.3 83.6 21.9 5.2 .7 .0 .0 .042 269.9 119·0 80.2 .38.6 12.7 206 .1 .0 .04) 204.8 154.9 110.1 38.6 11.9 209 .) .0 .04) 186.7 126.4 106.5 68.5 25.) 7·1 2.5 .4 .1

" 288.0 1.31.4 64.1 )2.4 6.4 .7 .4 .1 .044 327.0 101.4 49.4 )2.9 9·8 2.) ·5 .2 .045 JOO.9 142.9 69.9 8.7 1.1 .0 .0 .0 .045 354.0 125.9 )9·2 ).7 .6 .0 .0 .0 .046 224.) 159.2 108.5 25.) 5.6 .6 .0 .0 .046 2)5.4 152.2 106.2 24.0 5.1 .6 .0 .0 .047 29).8 165.9 52.7 9.7 1.4 .0 .0 .0 .047 354.1 11).5 44.4 9.) 2.2 .0 .0 .0 .048 224.2 148.5 108.0 )2.) 9·0 1.2 .1 .0 .048 2·2).4 127.2 112.4 4).6 1).) 2.8 .7 .1 .049 )03.0 146.1 55.6 14.4 309 .4 .0 .0 .049 )67.) 94.) 4).4 13.5 4.2 .8 .p .0 .050 282.9 146.) 71.2 18.8 ).7 .6 .0 .0 .050 321.) 125.9 55.7 16.6 ).2 .8 .0 .0 .051 210.0 158.2 106.9 35.2 10.5 205 .2 .0 .051 220.0 140.8 86.6 48.7 18.7 6.2 2.2 .2 .052 393.0 97.6 2/.6 4.8 .5 .0 .0 .0 .052 442.) 62.8 1 .0 2.) .2 • 0 .0 .0 .05) 475.6 43.8 4.0 ·0 .0 .0 .0 .0 .05) 490.6 6.4 .0 .0 .0 .0 .0 .0 .054 22).4 159.9 105.7 27.8 5.7 .9 .0 .0 .054 258.5 144.8 76.6 32.9 8.2 109 .5 .0 .055 445.) 68.0 9.9 .2 .0 .0 ·0 .0 .055 478.0 17.8 1.2 ·0 .0 .0 .0 .0 .056 254.8 141.1 98.8 22.1 5.6 .9 .1 .0 .056 278.1 109.9 94.1 29·5 8.8 205 .5 .0 .0

Fig. 13 - Hours of Wind.

Although this is the basic data on whichdecisions ought to be made, it has becometoo bulky for easy digestion. Some form ofpredigestion is required, although the basicTables (such as Fig. 12) ought to be presen-ted in full in every report.

The predigestion takes the form of exami-ning each location in turn, discovering theuse to which it will be put and, using Table4, allocating to each location a percentageof time when a given wind speed can be excee-ded corresponding to the first two subjectiveassessment criteria. The uses will normallyfall into four or five groups; for instance,car parks, pavements, pedestrian areasthrough which people walk, pedestrian areasfor loitering and general access areas.

At Bristol a programme has been developed

616

..

THE EFFECTS OF WIND ON pEOPLE

It now becomes a simple matter to seewhether conditions at each location satisfythe acceptability criteria or not. Thosewhich satisfy the criteria can be forgottenwithout further comment. For those locationsat which the conditions do not meet the cri­teria, the first question is to see by howmuch they exceed the criteria. This canimmediately be seen from the 'Hours of Wind'Figure (Fig. 13). If the criteria required5% (36 hours in the month) and the hours ofwind showed that the given wind speed wasexceeded for 36.5 hours, then the conditioncould be accepted as marginal. If, however,the exceedence is of considerable duration,then either the area can be down-graded toa different use requiring less strict condi­tions, or some remedial action is required.This is discussed in the next Section.

If the planning authorities, the client,the architect and consultant engineer canbe involved in writing the criteria in thefirst instance before the investigation, thesubsequent dialogue is easier because eachhas declared his interest and position.

4.7 Cures for Problem AreasAt each location with a problem the same

procedure to suggest cures is applied. FromFig. 12 are determined those wind directionsresponsible for the high wind speeds (usuallyonly one or two in number). The relationshipbetween the offending location and the sur­rounding buildings is quickly seen in Fig. 10when the troublesome wind direction is marked.It is almost always possible now to explainthe cause of the high wind speed at the loca­tion and to suggest modifications to thebuildings which would reduce its value. Allproblem locations are considered in succes­sion.

In referring to cures, one general remarkshould be made: the use of solid screens orother such devices-merely deflects the windfrom one location to another. Care must betaken not to exchange one trouble spot foranother and unknown one. If deflectors areused, the wind tunnel investigation of thesurrounding areas ought to be repeated toensure that no new areas of high velocitywind have been created.

5. CONCLUSIONS

The authors have tried to show that tech­niques are avai1able to assist the architectto realise a design which will only beadversely affected by the wind to an accep­table extent. Although it· might not appear

obvious from the foregoing paper, they doappreciate that wind is only one of the con­straints within which the design has to grow.

6. ACKNOWLEDGEMENTS

The work described in Sections 2 and 3 hasbeen carried out as part of the research pro­gramme of the Building Research Establishmentof the Department of the Environment, and ispublished by permission of the Director.

The authors would like to place on recordtheir thanks to Miss M E Gibbs who has typedmany drafts of this paper. We are sure sheis glad that this is the last!

APPENDIX I

RELATED TOPICS OF IMPORTANCE

There follows a series of topics whereman's relation with the problem of wind envi­ronment is at secondhand. The purpose of thisSection is only to describe the problems forthe sake of completeness, and not to go intothem in any depth.

AI.I Removal of PollutantsThe desire is always to ensure that the

wind speed in the vicinity of buildings andbuilding complexes is low: the impression isgiven that zero velocity would be ideal. Butthis is often not the case where pollutantshave to be removed. The most obvious areaswhich require this ventilation are those usedby motor cars, either for parking or for wai­ting to drop or pick up passengers. Otherobvious point sources are outlets of ventila­ting systems, especially if servicing kit­chens or certain food shops, and in thesecases reingestion may also be a problem ifthe inlet and outlet of the system are loca­ted close together.

The conventional way of removing pollu­tants is to use chimneys. Tall ones can bedesigned so that they do not pollute thenear field except in conditions of completecalm or ~vhen inversions occur. Lmv chimneyson low rise buildings can cause problems ifunventilated areas are allowed to fill upwith pollutants.

It is interesting to note the slightchange in terminology: before, we were tal­king about gusts and windiness, now we arespeaking of ventilation. In both instances,we mean the movement of air masses. The newreader of this subject is recommended tostart with the light reading of Ref. II.

617

T V LAWSON AND A D PENWARDEN

AI.2 Fire RiskAn important aspect of this subject, and

one which is particularly relevant to thepresent discussion, is that of smoke disper­sal from covered shopping centres. This isa problem which must be considered when aroof is to be built to combat bad wind con­ditions in a previously open shopping centre.Vents must be provided in the roof for smokeremoval in case of fire, but careful designis needed, and the possible adverse effectof wind must be considered.

Extensive tests carried out at the FireResearch Station on a full-scale mock-up ofa covered precinct and in models (Fig. 15)have shown how the smoke spreads and whatcan be done to contain it.

Fig. 15 - Model test of smoke movement in acovered shopping precinct.

In the tests, a fuel typical of thatwhich could be found in some of the shopsand furniture stores of these shopping cen­tres (wood, polyurethane foam, polystyrenepieces and foam rubber) is ignited in a com­partment representing a shop. The shoprapidly becomes filled with black smoke, andtoxic fumes billow out through the open shopfront into the mall. In the mall the smokeforms a buoyant high-level layer whichadvances at a fast walking pace. This well­defined layer is usually destroyed when itreaches the end of the mall, with the smokebeing brought down to a low level and retur­ning towards the fire. Within about twominutes visibility is eliminated and virtualblack-out conditions exist in the mall.These tests show that in a real mall therecan be a aerious escape problem. and all theingredients for panic, combined with condi­tions which make it difficult for the firebrigade to approach the fire and extinguishit, The problem is to prevent the smokefrom travelling too f.ar along the mall. TheFire Research Station have found that fourremedial measures are necessary:

Smoke reservoirs must be created by thedesign of the mall or by installing screensunder the ceiling.

Smoke must be extracted from the reser­voirs as fast as it flows in, by mechanicalor natural ventilation.

Air must be allowed to enter the mall ata low level to replace the gases flowing out.

Sprinklers must be installed in the shopsto limit the size of the fire and the amountof smoke produced.

These remedial measures are illustrated inFig. 16. All four must be used; none willwork effectively on its own.

Smoke extraction

Screen

Fig. 16 - Measures needed to maintain access and escape routes.

618

..

THE EFFECTS OF WIND ON PEOPLE

Care has to be taken to prevent adversewind pressures from destroying the smoke­clearing action of roof-vents. This couldoccur with a vent placed at the windwardside of a tall building where wind entrycould cause smoke-logging by mixing the pre­viously stable layer of hot gases throughoutthe mall. In such cases, wind tunnel testsmay be useful in establishing the optimumpositions of roof vents.

It is important that the smoke controlmeasures listed above should be consideredat an early stage in the design of a buil­ding development as they may be difficult toincorporate as modifications later. Asrelaxations of the normal fire regulationswill be needed whenever such covered shop­ping centres are built, advice should besought at the design stage from the localauthorities, the Building Regulations sec­tion of the Department of the Environmentand the local fire authority. Some usefulreferences (12, 13, 14) are given at the endof this paper.

AI.) Wind NoiseMany elements of a building may become

noise sources during windy conditions.Berhault and Davies (Ref. 15) have investi­gated many of these sources" in a 14-storeyapartment building in Southampton. Theyfound that the wind noise spectrum inside abuilding could be conveniently divided intothree spectral bands.

Infrasound - frequencies below 15 Hz.Measurements suggested that high noiselevels could be expected in tall buildingsduring strong winds, the levels being com­parable to those naturally occurring in theincident wind. Levels might be enhanced byeddy shedding from the building or by thewakes from neighbouring buildings.

Low Frequency Sound - 15 to 200 Hz. Theacoustic behaviour in this band was found tobe dominated by room resonance effects.Room resonance can be excited by naturalwind turbulence or eddy shedding causingvibration of windows or wall panels, or bywind-driven cross-flows in lift shafts orventilating ducts. Such resonances are com­mon experience and include moaning in chim­neys. Berhault and Davies were able todemonstrate clearly the strong couplingwhich can exist between window and roomvibrations.

High Frequency Sound - 200 Hz to 10 kHz.The acoustic energy in the free wind in thisband was found to be negligible compared withthe levels observed in the building. In all

cases such noise seemed to be related tosome local mechanism excited by air flowthrough the building. Air flow throughitems such as partly-open windows, ventila­tor grilles, letter boxes and gaps arounddoors produced broad-band noise with a fewstrong discrete tones. The loud howling andwhistling emanating from these sources wasdependent on the detailed design of eachcomponent, and such details as the width ofa window sill or frame were found to havea large bearing on whether a narrow windowopening produced a significant noise levelor remained inaudible. The details givingrise to each source appeared too complex forany general conclusions to be drawn aboutthem beyond the fact that they were elimina­ted if the wind-generated internal flowswere suppressed, which was hardly a practi­cal procedure.

Berhault and Davies' results showed thatthe leeward side of the building was gene­rally some 15 dBA quieter than the windwardside. Their work is continuing with thestudy of the acoustic behaviour of letterboxes, door slits and natural ventilationgrilles, the quantitative relationshipsbetween noise levels and wind strength nothaving been established yet.

AI.4 Rain Run-Off EffectsIn the past, the texture of buildings was

such that it could absorb a large quantityof rain during the heavy part of a storm.And as wind tends to remove water from thesurfaces of the buildings in liquid formduring a storm, as well as by evaporation,with gutters to collect and remove the rainimpinging upon the roof, there was no problemfrom the rain impinging or being driven bythe wind onto the walls.

The advent of walls constructed of imper­vious materials, coupled often with architec­tural features in the form of vertical pro­jections, also often made of impervious mate­rials and forming gullies to bring the waterimpinging upon the walls rapidly to groundlevel, has introduced a new problem due tothe restricted possibility of removal of therain by the wind.

In these latter extreme cases, the mea­ningful quantity must be the volume of rainimpinging upon the face of the building/sec.,divided by the length of the wall. Coinci­dent values of wind speed and rainfall areoutlined in Section A2.2, but the calcula­tion of how much water impinges upon thevertical faces of tall buildings can only beestimated in the absence of surrounding tall

619

T V LAWSON AND A D PENWARDEN

building, using Ref. 16. Page (Ref. 17) quo­tes the results of some theoretical studiesin which the static pattern of wind veloci­ties averaged over long periods is assumedand shows impingement patterns on the buil­ding. Flower and Lawson (Ref. 18) describea method whereby the rate of impingementcould be measured in a wind tunnel experi­ment, allowing for the removal of rain fromthe wind by the surrounding buildings. Thisproblem has arisen seldom in the past andwork on remedies is scarce. However, it ispossible that it could become more importantin the future.

Al.5 Effect on Vega tat ionThis is a huge subject meriting whole

papers to itself,.ranging from the problemof windstorms on extensive plantations tothe planting of a small hedge in a suburbangarden. For the large problems, readersare referred to the Forestry Commission atAlice Holt Lodge, Wrecclesham, Farnham,Surrey, where professional advice can beobtained.

In windy areas, especially near the coast,we have all seen trees bowed over and stun­ted because of the continual battering fromthe prevailing wind, sometimes accompaniedby salt-laden spray. Few species can sur­vive at all, let alone prosper and lookattractive in this extreme climate and care­ful choice of plant. variety is essential.Even without the added salt spray, choice ofspecies, their planting and subsequent cul­tivation, particularly in their early years,are matters which require careful attention.Beginners are referred to a simple publica­tion by the Royal Horticultural Society(Ref. 19).

Certain pollutants, notably sulphur di­oxide, tend to discolour or even kill cer­tain species of trees and care should betaken not to mix the two. Concentrations ofpollutants which are scarcely noticeable tothe nose can still kill the trees, becauseof their continuous exposure to contamina­tion. This aspect of occupation is referredto in a different context in Section 4.3.

Al.6 Aesthetics - Staining due toWind-Rain Interactions

This may sound trivial, especially com­pared with talk of fire risks, but neverthe­less it is relevant to the architect when heimagines the completed building. He hastaken care in order that the building shallbe visually pleasing to the viewer and inthat care has been an appreciation of light

620

and shade as well as mass. These effectscan change during the lifetime of a buildingdue to deposits on some areas and continualwashing of others. Alkaline mortar mayresist deposition longer than the stones itbinds, producing a chequerboard effect notenvisaged initially and unattractive to theviewer. The weathering of the newer mate­rials in sun, wind and rain requires consi­deration, but guidelines in this generalpaper are impossible.

APPENDIX 2

BACKGROUND TO WIND DATA

A2 • 1 The Wind

A2.l.1 Generation of the WindWind is created in the first instance by

the differential solar heating of the earth'ssurface. Both cloud cover and the differentreflectivities of the earth's surface pro­duce a patchwork of atmospheric heating.Change of temperature affects the other twoproperties of state of the air, namely pres­sure and density. The changes in pressurein particular cause the air to move fromregions of high pressure to regions of lowerpressure and the movement of the air is cal­led wind. Due to the rotation of the earth,equilibrium considerations for the airrequire that the velocity vector at heightswell above the surface (of the order of500 m) should lie close to the isobars orlines of constant pressure.

As with the flow of any gas over a sur­face, the mean velocity decreases towardsthe surface, until, at the surface, it iszero: the region in which this change occursis called the boundary layer. This reduc­tion in velocity is caused both by the fric­tional drag of the air over the surface andthe drag of any bodies (such as trees, hed­ges, buildings, and mountains) protrudinginto the airflow. These retarding forcesare transmitted through the layer by shearforces and by the exchange of momentum dueto the vertical movement of the air. Theco-ordinated movements in these layers ofair break down into a random movement whichis called 'turbulence'. If the surface isthat of the earth, the region is called theatmospheric boundary layer: the main diffe­rence from boundary layers on any man madesurface is due to the rotation of the earth.This causes the velocity vector to slewround with height from the ground, an effectcalled the Ekman spiral, but this has little

..

Tllli EFFECTS OF WIND ON PEOPLE

relevance to the problems discussed in thispaper and will be ignored.

The speed and direction of the wind varywith time in the way shown in Fig. 8; itsmean value varies with height, type of sur­face, the averaging time and the geographi­cal location. Other variations are of asmaller order.

A2.1.2 Quantification of the WindThe mean value for the trace of wind

speed in Fig. 8 is shown by the horizontalline. If wind speed is averaged over periodsmuch shorter than the whole sample, say forexample 3 seconds, then a range of values isobtained from the trace, the largest of whichis always greater than the mean value of thewhole trace. It becomes obvious that everymeasurement of wind speed must be accompaniedby a statement of the averaging time or themeasurement is meaningless.

In calculating wind loads for use in thestructural analysis of a building, the ave­raging time is carefully chosen in the lightof the component to be stressed and theshape and size of the building complex. Thenanother time quantity is added - the ReturnPeriod. This arises from the requirementthat the component shall not fail during thelifetime of the building - say 100 years.The design wind speed defined as the maximumvalue of wind speed averaged over, say, Isecond, which will not be exceeded in 100years is required to calculate the designloads. It is difficult to quote a valuewhich will not be exceeded but, after recour­se to extreme-value analysis, it is possibleto determine a value which will be exceededonce in, say, 100 years, with a probabilityof P. Wind data in this form have been stu­died and are presented in extended form (withderivations) by the Engineering Sciences DataUnit in Ref. 6, or can be obtained in extractform in the relevant Code of Practice (theBritish Code is listed as Ref. 9).

But wind data in this form are of littleuse in solving the problems 'listed in theIntroduction: much more frequent occurrencesare important, for example the wind speedexceeded for more than 10% of the time. vllienmore regular occurrences are introduced, anew parameter, the wind direction, must beadded to the variables. It also becomesnecessary to define groupings of wind speedand direction. Wind directions in the UKare usually divided by the MeteorologicalOffice into 30-degree sectors, and windspeeds into the ranges of the Beaufort Scale.This format appears ideal for our needs,

because it produces a managable number ofdivisions (10 in wind speed and 12 in winddirection): it is also convenient to deviseprocedures which require the wind data inthis format, because the MeteorologicalOffice is the main supplier of meteorologicaldata in the UK. These data are available ona monthly as well as an annual basis (Ref. 8);a typical extract is shown in Fig. 11.

A word of warning is needed at this stageconcerning the evulation of wind speed deter.­mining th<;'!. limits of the ranges of theBeaufort Scale. The wind ~xemplified by thetrace of Fig. 8 will represent a typicalwind over an open site in one of the rangesof the Beaufort Scale. As mentioned before,this trace will be represented by differentmaximum velocities if different averagingtimes are used. Therefore, the velocitiesdefining the ranges must declare their ave­raging time. The Table in Fig. 11, as mostdata from the UK Meteorological Office, defi­nes the ranges with hourly average values.Ref. 6 shows that the ratio of maximum windspeeds for two different averaging timesdepends upon the intensity of turbulence,which is affected not only by the local typeof environment (city centre, urban, etc.)and by the height above ground of the mea­surements, but also the detailed layout ofthe buildings locally. Consequently, nogeneral relationship between the divisionsof the Beaufort Scale and the averaging timeis possible when considering wind amongstbuildings.

A2.2 RainRain in this context is not a problem in

itself: we are interested in the interrela­tion of wind and rain. This aspect of thesubj ect has been summarised by Lacy (Ref. 16),who define'S a 'Driving Rain Index 1 ~.;rhich

gives an indication of the risk that wallsmay be penetrated by rain. This index iscalculated as the product of annual averagerainfall and annual average windspeed, anda contour map in Ref. 16 shows the geographi­cal variations within the United Kingdom.Lacy also shows that wind direction is impor­tant, and presents roses of the Driving RainIndex for 23 meteorological stations, basedon simultaneous measurements of wind andrain. Further analysis shows that the higherthe windspeed, the greater the percentage ofthat windspeed's time will it rain. This isillustrated in Table 5, with values foreastern and western meteorological stationsseparated. Readers requiring more informa-

621

T V LAWSON AND A D PENWARDEN

tion on this and other aspects of the inter­relation of climate and buildings are refer­red to Ref. 20.

Table 5 - Proportion of hours during which rain falls Indtf terent wInd-speed classes.British stations, means for period 1957~6 (Ref. 20)

Wind-speed class Eastern stations Western stations(hourly averages)

m/s % %

0.5'" 2.1 5 - 6 7 - 10

2.6 - 4.1 7 - 9 9 - 13

4.6 - 6.2 10 13 - 15

6.7 - 8.2 12 - 13 15 - 21

8.8 - 10.3 15 - 16 17 - 28

10.8 - 12.4 17 - 18 . 20 - 40

12.9 - 14.4 20 - 21 25 - 44

14.9 - 16.5 27 - 56

17.0 - 18.5 36 - 62

19.1 - 20.6 37 - 75

REFERENCES

PENWARDEN, A.D. 'Acceptable Wind SpeedsIn Towns' BRE Current Paper No. 1/74,January 1974, DoE.2 HUNT J.C.R. Wind Tunnel Experiments onthe Effects of Wind on People. Unpublisheqreport to the Building Research Establishment.1974.3 HUMPHREYS M.A. 'A Simple TheoreticalDerivation of Thermal Comfort Conditions'J Inst Heat and Vent Engrs. Vol 38, No. 95(1970) .4 PENWARDEN A.D. and WISE A.F.E WindEnvironment around Buildings. BuildingResearch Establishment Report, London,IIMSO 1975.5 LAWSON T.V. 'The Measurement of ShortTerm Average Pressures in a Wind TunnelInvestigation'. Department of AeronauticalEngineering. University of Bristol ReportNo. TVL!7436, 1974.6 ESDU 'Characteristics of Wind Speed inthe Lower Layers of the Atmosphere near theGround: Strong Winds (Neutral Atmosphere)'ESDU Data Item 72026, Dec 1972.7 ESDU 'Characteristics of AtmosphericTurbulence near the Ground'.'Part 1 ­Definition and General Information'. ESDU

. Data Item 74030. 'Part 2 - Single Point Data

622

for Strong Winds (Neutral Atmosphere)' ESDUData Item 74031. (Both obtainable fromEngineering Data Sales Ltd, 34 HaymarketLondon SWIY 4HZ).8 HMSO 'Tables of Surface Wind Speed andDirection over the U~Met 0.792, 1968, IIMSO.9 BSI 'Code of Basic Data for the Designof Buildings'. Chapter V Loading. Part 2,Wind Loads. CP3 Ch V, Part 2, 1972, BSI.10 NEWBERRY C.W. and EATON K.J. Hind LoadingHandbook. Building Research EstablishmentReport. London, HMSO 1974.11 SCORER R.S. 'Air Pollution' The Common­wealth and International Library of ScienceTechnology Engineering and Liberal Studies.Pergamon Press 1972.12 HINKLEY P.L. BUTCHER E.G. and PARNELL A.Fires. in Shopping Malls. Architects Journal,28 Nov. 1973 p. 1319.13 HINKLEY P.L. Some Notes on the Controlof Smoke in Enclosed Shopping Centres.Joint Fire Research Organisation FireResearch Note 875, 1971.14 BUILDING RESEARCH ESTABLISHMENTSmoke Control in Single-storey Shopping Malls.Digest 173, 1975.15 BERHAULT J.P.A. and DAVIES P.O.A.L.Wind Noise in Buildings. Symposium on FullScale Measurement of Wind Effects on TallBuildings and other Structures. Universityof Western Ontario. London, Ontario. Canada,June 1974.16 LACY R.E. 'An Index of Exposure to Dri­ving Rain' Building Research StationDigest No. 127, March 1971.'-17' PAGE J .K. 'Five Meteorological Facets ofthe Design of Tall Buildings' InternationalAssociation of Bridge and Structural Engi­neers and the Institution of StructuralEngineers. Conference. Oxford,Sept. 1974.18 FLOWER J.H. and LAWSON T.V. 'On theLaboratory Representation of Rain Impinge­ment on BuildingA' Atmospheric Environment,Vol. 6, pp 55-60, 1972.19 SHEPHERD F.W. 'Hedges and Screens'.Wisley Handbook No. 17, The Royal Horticul­tural Society, 1974.20 LACY R.E. Climate and Building.Building Research Establishment Report(to be published).