Document No. :: IITK-GSDMA-Wind06-V3.0 Final Report: B - Wind Codes IITK-GSDMA Project on Building Codes An Explanatory Handbook on Proposed IS 875 (Part3) Wind Loads on Buildings-and Structures ty Dr. N.M. Bhandari Dr. Prem Krislina Dr-Krishen Kumar Department of Civil Engineering Indian Institute of Technology Roorkee Roorkee Dr. Abhay Gupta Deprtment of Civil Engineering Shri G. S, Institute of Technology and Science IndoreThis document has been developed under the project on Building Codes sponsored by Gujarat State Disaster Management Authority, Gandhinagar at Indian Institute of Technology Kanpur. The solved Examples inchided in this document are based on a draft code developed under IITK.GSDMA Project on Building Codes. The draft code is available at_bitp://wwwaicee.org/ITK-GSDMAJIITK- GSDMA him (document number IITK-GSDMA Wind02-V5.0). ‘The views and opinions expressed are those of the authors and not necessarily of the GSDMA, the World Bank, IT Kanpur, or the Bureau of Indian Standards. Comments and feedbacks may please be forwarded to: Prof. Sudhir K Jain, Dept. of Civil Engineering, IIT Kanpur, Kanpur 208016, email: [email protected]Examples on revised 1S 875 (Part 3) FOREWORD This explanatory handbook is meant to provide a supplement to the IS. 875 (Part 3) — Wind Loads on Buildings and Structure- draft revision, This work has been supported through a project entitled Review of Building Codes and Preparation of Commentary and Handbooks awarded to IIT Kanpur by the Gujarat State Disaster Management Authority (GSDMA), Gandhinagar through World Bank finances. The views and opinions expressed therein are those of the authors and not necessarily of the GSDMA, the World Bank, IIT Kanpur, or the 3ureau of Indian Standards. Prof. Ashwini Kumar (1T Kanpur), Dr. N. Lakshmanan (Structural Engineering Research Centre, Chennai), and Prof. P.N. Godbole (VNIT Nagpur) were the reviewers of the document. Prof. LM. Gupta (VNIT Nagpur) also contributed through review comments. ‘Section 1 is an introductory note to reflect the background state-of-the-art scenario of ‘wind engineering. This indeed forms the backcrop of the entire exercise of revision in hand Section 2 of this volume contains illustrative examples designed to demonstrate the various parts of the “Indian Standard LS. 875 (Part 3)-Draft Revision” dealing with wind loads on buildings and structures. The examples take the reader to a point whereby the wind load on a particular structure is computed using the code Section 3 describes some examples which are unusual fiom the point of view of determination of wind loads, and for which straightforward answers can not be had from the code. These are dealt with through qualitative discussions. The suggested are ouly indicative. IITK-GSDMA-WINDO6-V3.0 3Examples on revised 1S 875 (Part 3) CONTENTS SECTION -1 An Introductory Note SECTION -2 Ilustrative Examples (1-26) SECTION -3 Some Unusual Cases (1-8) for the Determination of wind Forces on Building/Structures IITK-GSDMA-WINDO6-V3.0 4Section - 1Exwmples on revised 1S 875 (Pat 3) AN INTRODUCTORY NOTE Li General ‘Wind has two aspects. The first — a beneficial one ~ is that its energy can be utilized to generate power, sail boats and cool down the temperature on a hot day. The other — a parasitic one — is that it loads any and every object that comes in its way. The later is the aspect an engineer is concerned with, since the load caused has to be sustained by a structure with the specified safety. All cjpil and industrial structures above ground have thus 10 be designed to resist wild loads. This introductory note is conceming the aspect of wind engineering dealing with civil engineering structures. Wind flow generation is on account of atmospheric pressure differentials and manifests itself into various forms, such as, Gales and monosonic winds Cyclones/Hurricanes/Typhoons Tomados ‘Thunderstorms Localised storms Photographs in figure-1 depict some of these storms. Friction fiom the earth's surface leads to “boundary layer’ flow, but characteristics of flow vary depending ‘upon the storm type. (a) Cyclonic Storms () Thunderstorm IITK-GSDMA-WINDO6-V3.0L O'€A-900NIM-¥WaSO-NLIT awopa3 v yo armsonng pwsOUDD =z ANI 4 OS-0F wry 0f-07 = 8 a 14 009-008 MOI — yeupdn pue uuoneab] poods pura, ures “spnoyp renuszoqumast aay Jo uo sauo]ok uo. ep ay) oxENS Afuuanbay yey} suLIOs a1UO[29 autos aar® ¢ pue 7 samaty * <[yspeay, aq ame vIpUy UP aSaraUT sOWAIR Jo ‘sayeR readde ypuya sopemoy ay pur sistoo pur spur stuoosuomr oy wedy sursoys pays Jo saddy snopsea = | aanspyy sopemsog, (9) CT CET RTT ITTExwmples on revised 1S 875 (Pat 3) Jan, to eR YQ / [7] Almost All Cyclones form in this region © Total no. of Cyclones/Year (Global) ~ 80 # North-Eastem Hemisphere the worst ht, experiences almost 45% of the Cyclones # India experiences about 6 cyclones/ycar, mostly on the East coast Figure 3 : General Distribution of Cyclones over the Globe There has been an increasing trend of natural hazard events as well as wind — induced disasters in the recent decades, as also the losses caused in such disasters. The reasous for this increasing trend can be cuumerated as below : © Population density is increasing worldwide, particularly in large cities and agglomerations. In the last twenty ‘years or so the mumber of cities with, ‘more than 1 million population may have doubled fiom about 200. Likewise mega cities with mote than 10 million population may have gone Lup from about 10 t0 25. ‘© Atthe same time the standard of living is increasing almost everywhere — a bit faster here, a bit slower there. Together with the growing population density this means an exponential increase in the concentration of values; infrastructure cost sustaining the ‘modem living standard. © Regions formerly avoided on account of their risk potential are now optllated. This applies above all to coastal regions particularly exposed to storms and storm surges, in many cases also to major earthquakes and seismic ‘waves (tsunamis). ‘Industry is also moving into extremely dangerous regions, concentrating huge economic values in highly exposed locations In many cases new building materials such as glass and plastics used for facades and roofs are far more susceptible to damage than conventional methods and materials ‘* Incertain periods, nature also becomes responsible for bigger disasters, with more dramatic trends in unleashing natural hazards, Tables 1 to 3 contain some relevant information on these events and the losses therefrom. IITK-GSDMA-WINDO6-V3.0Exwmples on revised 1S 875 (Pat 3) ‘Table 1* ‘atural Disaster Events and Economic Losses Decadewise. Decade 1960s 1970s 1980s 1990s Numbers 2 47 63 84 Economic losses BA BLS 204.2 591 (US S billion) ‘Table 2* : Data for All Catastrophes 1985 - 2000 Item World Asia-Pacific % Loss Events (Nos.) 3.220 38.6 ‘Economic Losses USS million 426.270 47.6 Insured Losses USS million 21.970 29 ‘Loss of Life (Nos.) 5,36.250 4.33.480 80.8 Table 3* : Data for All Catastrophes Vs Wind for Asia-Pacific 1985 - 2000 Tiem ‘All Catastrophes | Wind | % Loss Events nos. 3,220 1,020 | 317 ‘Economic Losses USS million 426.270 ©2120 | 146 Insured Losses US$ million 21,970 12.470 | 568 ‘Loss of Life nos. 4.33,480 60,250 | 13.9 There arc various scales on which windstonms can be mcasured — these are shown in Table—4. * “Topics”, Munich Reinsurance Company IITK-GSDMA-WINDO6-V3.0 9Exomples on yevised 1S 875 (Part 3) ‘Table 4: Various Scales for Measuring Wind Storms ‘Windstorm : Scales and Effects Bewubrt Seale Sallis Simpson Hurricane Seale BE] Devsipive | Mean wind speedat lm | Wind | SS | Descriptive Mean wind speed tem hove wurice presse ‘ea wk | Km/h | imow | Kyi a _—| Kav [foi | calm oor [01 | on | 0 [1 | Weak Torso _[s- 155 | Gs-82 1_| Light air 03-15 1-5 13 00.1 2 | Moderate 427-495 | 154-177 | 83-96 > [tight 1633 | 611 | 6 | 20-05 [3 | Strong 49.6585 178-209 | 97-113 treeze Z [Gentle | aa54 | i219 | 710 | Oris [a | Veqrationg | Sxe04 | D029 | TneIse Ineeze T | Modents | S279 | i0ae |Tiis | TOS |S |Devestting | Ses | S350 | SS tneeze 5 | Fresh e007 a3 Fujita Tornado Seale Inceze © | song | TORISS TAT | F | Dessrpive | mk | Kai | Kan treeze rem 7 [Nesrgale | Wkocira | Socal | Seas IRIS | 0 | Weak TESS To 3 [Gsle 172-207 | 6-7 _[ 34-40 | 184-268 | 1] Moderate [32 7-S0.1 6497 9” | Stong gale | 208-244 | 75-98 | 4i-a7 | 269-313 [2 | Strong 502-702 38-136 10 [Stom | 248-284 | 89-102 | 48-88 | 374-805 | 3 | Devastating | 70.3-92.1 | 254-382 | 137-179 11] Violent [385-326 | 108-117 | 86-63 | 80.6-6¢5 | 4 | Amubilaing | 922-1162 | 33%-a18 | 180-225 om 12 | Hunicane | >327 | > 18 | >t | > 666 | 5 | Dimwter | 1hasi360 | 419-403 Wind ston disasters cause —_—_ speeds, and flooding in case of cyclones. widespread damage due to high wind Figure 4 shows typical photographs. IITK-GSDMA-WINDO6-V3.0 10Exwmples on revised 1S 875 (Pat 3) 1, Water submergence; 2.3.4.5, Wall damage and roof blown away: 6. Tower collapse, 7. Failure ofa tower; 8. Uprooted tree damages nea:by building Figure 4: Typical Wind — Storm Damage 1.2 Structure of Wind Wind is a randomly varying dynamic phenomenon and a trace of velocity versus time for wind will be typically as shown in figure 5. The wind velocity V can be seen as a mean plus a fluctuating component responsible for creating ‘gustiness’. Within the earth’s boundary layer, both components not only vary with height, but also depend upon the approach terrain and topography, as seen from figure 6. While dealing with rigid strictures, the consideration of the ‘equivalent static’ wind is adequate. However, in dealing with wind-sensitive flexible structures, the consideration of the wind-energy spectrum, integral length scale, averaging time and the frequencies of the structure become important. The determination of wind velocity for a ‘certain geographical location is essentially 2 matter of statistical reduction of a given ‘measured data. On this depend the various wind zones. Another important decision involved is the averaging time. In as far as averaging time is concemed, it may be ‘anywhere from 2-3 seconds to 10 minutes to an hour. The influence of averaging time on velocity is seen in figure 7 IITK-GSDMA-WINDO6-V3.0 ittExwmples on revised 1S 875 (Pat 3) z Mean g Fluctuating 2 Wind Velocity o) riation of Wind Velocity with (a) Time (b) Height Oven Treed, Windward Speedup Over Sheltered Leewaud Sea ‘Coast Hill Crest Coast V_~ Gradient Wind Velocity Figure 6: Influence of Terrain and Topography (Typical) IITK-GSDMA-WINDO6-V3.0 12Exwmples on revised 1S 875 (Pat 3) Sa Figure 7: Ratio of probable maximum speed averaged over period ‘t’ to that averaged over one hour. 1.3. Wind Forces on Structures Most structures present ‘bluff forms’ to the wind, making it difficult to ascertain the wind forces accurately. Thus the problem of bluff-body aerodynamics remains largely in the empirical, descriptive realm of knowledge. The flow pattern and hence the wind pressures/forces change with the Reynolds uumber (Re) making the direct application of wind tunnel test results to real structures difficult. Computational methods (CEDICFE) also pose problems for the high Reynolds Number, Re encountered in practice, leading to their inadequacy. The oncoming turbulence causes fluctuations in the flow. As a result the ‘wind pressures also change with time, thus affecting the values of design wind pressures. A typical case is that of a cylinder for which pressure distribution changes with Re and surface roughness. The value of the drag coefficient, Cs, for different situations is as given in figure 8. The surface has been roughened for case © by using sand, with the grain size K. => fee Ca Figure 8: Influence of Reynolds number, corner radins, and surface roughness on drag coefficient, square to cireul sand) -ylinders (r is the corner radius; k is the grain size of IITK-GSDMA-WINDO6-V3.0 BExwmples on revised 1S 875 (Pat 3) While the separation of flow in non-bluff sharp edged bodies is well defined, the dimension of the body/building parallel to flow affects the ‘reattachment of the separated flow’ and hence the suction pressures on leeward side, Turbulence may cause an increase in the suction pressures for thin bodies and a decrease in the thick/long bodies. The buildings and other civil engineering structures are three- dimensional bodies with a large variety of shapes and have complex flow patterns and therefore varied pressure distributions, Asa result, most 3-D studies rely partially or wholly upon experiments. ‘The complexity of wind flow is not only introduced by the geometries of typical structures, but also by the characteristics of the terrain and other structures in the close vicinity. This has necessitated the determination of wind _ pressures experimentally in wind tumels using scaled models and simulated winds. Besides, the buildings are never sealed and wind pressures develop inside even in a closed building, with maximum values occuming in open buildings. These generally add to the pressures outside, Creating worst possible effect on roofs, as well as walls, Tn recent model studies, forces acting over an clement of the building have been measured by devices that automatically add pressures occurring simultaneously at several points of the element, weighted by the respective tributary areas (University of Westem Ontario). These measurements as well as results of full-scale tests have been used to develop new design load provisions for low-rise buildings that have been incorporated in design standards. It is established that local pressures can have strongly nou-Gaussian distributions, especially at comers and edges. Tests on low-ise building models have confirmed that the fluctuating part of the load can in many instances be significantly larger than the mean load and that, for any given storm, peak pressures and the ratio between mean pressures and fluctuating pressures decreases as the terrain roughness increases. 1.3.1 Wind Sensitive Structures Tall and slender structures are flexible and exhibit a dynamic response to ‘wind. Tall structures vibrate in wind due to the turbulence inherent in the wind as well a that generated by the structure itself due to separation of the flow. Thus there is a mean and a fluctuating response to the wind. Besides, the dynamic forces act not only in the direction of wind flow but also in a direction nearly perpendicular to the flow (lift forces), so that tall structures also exhibit an across-wind response. Along-wind response has a mean ‘component (time-invariant load obtained from the mean wind speed) and a fluctuating component. The latter is further expressed as a sum of background and resonant components. If the damping is small, which is usually the case, the bulk of the contribution to the dynamic response is due to the resonant portion. Spectral response curves have been developed and are used for predicting the along-wind fluctuating response of a ‘building. The approach is frequency based ‘and uses the theory of random vibrations as well as statistical considerations. In the case of line-like structures, like chimneys, the problem becomes one-dimensional. This simplification is used for analyzing tall buildings as well, which is evidently conservative. As mentioned already, tall flexible structures, exhibit an across-wind response as well This is on account of flow separation fiom the cross section of the structure, which results in vortices being shed at a given frequency. The pattem of this across-wind phenomenon is, comparatively more regular for circular sections, such as those for chimneys and IITK-GSDMA-WINDO6-V3.0 4Exwmples on revised 1S 875 (Pat 3) towers, which can undergo resonant ‘vibrations when the structural frequency matches the forcing frequency. The response is affected significantly by the turbulence content of the wind, and is larger in smoother flows. When buildings attain slender proportions — which may happen for very tall ones — the across-wind behaviour becomes important. The theoretical treatment of tall slender structures in the along-wind direction is better developed than for the across-wind direction, and for this reason it may be advisable to undertake model studies in a wind tunnel for such structures, A body or a structure, such as a building, a tower or a chimney, when SS Sec Flow around isolated building placed in a flow of air will experience pressures and forces. When one or more similar or dissimilar bodies are placed downstream or upstream of a structure, the “stand-alone” values of pressures and forces get altered. This is termed as the Interference Effect. Interference will occur inrespective of whether the bodies involved are rigid or flexible. In the former it is the ‘wake’ of one body that affects the other, while in the latter the deflections of the body may also affect the wake itself (Figue 9). The phenomenon of interference is experienced extensively in practice but is very difficult to quantify in general because of the variability of situations involved. Systematic wind ‘tunnel studies can nevertheless give some guidance, e225 = o =e Flow around two buildings Figure 9: Typical wind flow around isolated and two nearby buildings 1.4 Determination of Wind Loads It is common practice to approach engineering problems for their analysis by employing theoretical or experimental means. Most problems are tackled satisfactorily by using theoretical solutions. Experimentation is carried out only where necessary, often for physical verification of the theory cmployed. Likewise, for tackling wind engineering problems there are these two approaches. However, in wind engineering, though theoretical solutions have been developed, there is a predominant use of experimental methods. On the theoretical side, on the ‘one hand there is the combination of statistical expression of loads and structural dynamics which has been in ‘common use for the last 40 years, and, computational fluid dynamics which has evolved more recently. A rather useful tool that is developing as part of computational ‘Wind Enginccring (CWE) is the use of Axtificial Neural Networks (ANN). The first issue to be looked at is the loading itself. The last 50 years have witnessed remarkable changes in the manner of assessing wind loads in structural design (Davenport, 2001). IITK-GSDMA-WINDO6-V3.0 15