equivalent static wind loading on buildings - new perspective

Upload: 75bm73

Post on 03-Apr-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/28/2019 Equivalent Static Wind Loading on Buildings - New Perspective

    1/8

    EquivalentStaticWindLoadingonBuildings:ANewPerspective

    XinzhongChenandAhsanKareem

    NatHazModelingLaboratory,UniversityofNotreDame,156FitzpatrickHall,

    NotreDame,IN,USA

    ABSTRACT:Aframeworkforevaluatingtheequivalentstaticwindload(ESWL)foranygivenpeakresponseofbuildingscharacterizedbyuncoupledmotionsinthethreeprimarydirectionsispresented.Thisincludesanewdescriptionofthebackgroundloadingbasedonthegustloadingenvelope, whereas the resonant component is described in terms of the inertial loading. TheESWLforthetotalpeakresponseis thenexpressedas alinearcombinationof thebackgroundand resonant components. A closed-form formulation of the ESWL based on this framework

    utilizing ananalyticalwindloading modelispresented.The gust response factorsandESWLsfor various alongwind response components at different building elevations are discussed tohighlightadvantagesoftheproposedscheme.

    KEYWORDS:Windloads,Windeffects,Gustresponsefactor,Buildings,Structuraldynamics,

    Randomvibration.

    1 INTRODUCTION

    In current design practice, spatiotemporally varying wind loads on buildings are modeled as

    equivalent static wind loads (ESWLs). Traditional gust response factor (GRF) approach

    (Davenport1967)iswidelyusedinmostbuildingdesigncodesandstandardsforthealongwindresponsethatresultsinaloaddistributionsimilartothemeanwindload(e.g.,ZhouandKareem

    2001). The GRF concept has been extended to the acrosswind and torsional response

    components (Piccardo and Solari 1996; Kareem and Zhou 2002). However, the GRFs mayexhibit wide variations for different response components of a structure and may have

    significantlydifferentvaluesforstructureswithsimilargeometricprofilesandassociatedwind

    loadcharacteristics,butdifferentstructuralsystems.Fortheacrosswindandtorsionalresponses,

    which are typically characterized by the low values of mean wind loading and associated

    response, particularly, for symmetric buildings, the corresponding GRFs may not project the

    samephysicalmeaningasthetraditionalGRFforthealongwindresponse.AnESWLdescriptionbasedonthepeakdynamicpressure/windload(includingthemean

    load)hasbeenadoptedinsomebuildingdesigncodessuchasthedraftEurocode(ENV-1991),ASCE7-02 and the new Australian/New Zealand Standards (Holmes 2002). This formatdescribestheESWLasthepeakdynamicloadmultipliedbyaconstantcoefficientreferredtoasdynamicresponsefactor(DRF)(Holmes2002).InSolariandRepetto(2002),anidenticalESWLdistributionforallresponsecomponentswassuggested.Theyutilizedapolynomialexpansion,whichwasobtainedonthepremisethattheselectedESWLresultedinaccurateestimatesofalimitednumberofpre-selectedpeakresponsecomponents.

    Separation of wind loads and their effects and the associated ESWLs into background(quasi-static) and resonant components provides not only an efficient response prediction

  • 7/28/2019 Equivalent Static Wind Loading on Buildings - New Perspective

    2/8

    framework but also a physically meaningful description of the loading (Davenport 1985;Kasperski1992;Holmes2002;Isyumov1999;Zhouetal.2000;ZhouandKareem2001;ChenandKareem2001).Accordingly,theresonantESWL(RESWL)canbeexpressedintermsoftheinertialload(e.g.,Davenport1985).WhereasthebackgroundESWL(BESWL)dependsontheexternal wind load characteristics and can be determined using a Load-Response-Correlation(LRC)approach(Kasperski1992),whichprovidesamostprobableloaddistribution(Kasperski1992 and Tamura et al. 2002). Based on the BESWL and RESWL, the corresponding peakresonantandbackgroundresponsescanbecalculatedusingasimplestaticanalysis.Thesearethencombinedusingthecompletequadraticcombination(CQC)approachorthesquarerootofthesumofsquares(SRSS)schemeforthetotalpeakresponse(excludingthemeancomponent).Alternatively,anESWLforthetotalpeakresponsecanbeexpressedasalinearcombinationofthebackgroundandresonantloadingcomponents(ChenandKareem2001;Holmes2002).

    In this paper, a framework is presented for evaluating the ESWL for any given peakresponsecomponentofwind-excitedbuildingscharacterized byuncoupledmotionsinthethreeprimary directions. A new description of the BESWL is presented based on the gust loadingenvelope(peakdynamicloadingwithoutthemeancomponent).TheRESWLisgivenintermsofthe inertial load in each fundamental mode. The ESWL for the total peak response is then

    expressed as a linear combination of the BESWL and RESWL. Based on this framework, aclosed-formformulationoftheESWLusingananalyticalwindloadingmodelispresented.TheGRFsandESWLsforvariousalongwindresponsecomponentsat different building elevationsarediscussedtohighlightadvantagesoftheproposedESWLdescription.

    2GENERALFORMULATIONS

    The response of a wind-excited building with one dimensionaluncoupled mode shapesin thetwo orthogonal translational and torsional directions at a given wind speed and direction isconsidered. The wind loads per unit height at elevation z above the ground have meancomponents of )(zPx , )(zPx and )(zP , and fluctuating components of ),( tzPx , ),( tzPyand ),( tzP

    ,inthetranslationalaxesxandyandabouttheverticalaxisz.Thediscussionhereis

    focusedontheresponsewithonedimensionalinfluencefunctionsinthethreeprimarydirections.The uncoupled class of response in the three primary directions permits treatment of windloadingand building response ineach direction independently.Withoutloss ofgenerality, thefollowingdiscussionwillfocusontranslationalresponseinthexdirectionatagivenwindspeedandorientation;asimilartreatmentinotherdirectionsisimmediate.

    For aspecific response ofinterest (displacement,bending moment,shearforceand othermemberforces)atabuildingelevationz0,R(z0, t),themean(static)andbackgroundcomponentscanbecalculatedbythestaticandquasi-staticanalysis.Theresonantcomponentcanbeanalyzedusingmodalanalysisrestrictedtothefundamentalmode.Themeanresponse,rootmeansquare(RMS)backgroundandresonantresponsesandthepeakdynamicresponse(excludingthemeanresponse)areexpressedas

    ==H H

    PxxR

    H

    xx dzdzzzRzzdzzzPR xxb 0 0 2121210 ),()()(;)()( (1)

    2222

    max11

    10

    2

    0 ;)(4)()(

    )()()(

    rbxr RrRbQH

    x

    H

    xx

    R ggRfSfdzzzm

    dzzzzm

    +=

    =

    (2)

    =H H

    PxxQ dzdzfzzSzzfS xxx 0 0 21,2121),()()()( (3)

  • 7/28/2019 Equivalent Static Wind Loading on Buildings - New Perspective

    3/8

    whereH=buildingheight;x(z)=influencefunctionindicatingtheresponseR(z0,t)underunit

    load acting at the elevationz in x direction;x(z)= fundamental mode shape;f1 and 1=

    fundamentalfrequencyanddampingratio(includingaerodynamicdamping),respectively; m(z)=

    massperunitheight; ),( 21 zzR xxP and =),,( 21 fzzS xxP covarianceandcrosspowerspectraldensity

    (XPSD)betweenPx(z1,t)andPx(z2,t); =)( fSxQ

    powerspectraldensity(PSD)ofthegeneralized

    modalforce; gband gr= peak factorsfor thebackground andresonantresponses, respectively,

    typicallyranginginvaluebetween3and4.

    FollowingtheLRCapproach(Kasperski1992),theBESWLforpeakbackground

    response,bRb

    g ,isgivenby

    =H

    Px

    R

    beR dzzzRz

    gzF

    xx

    b

    b 0111 ),()()(

    (4)

    whichdependsontheinfluencefunctionoftheresponseunderconsideration.Accordingly,theBESWLhasadifferentspatialdistributionfordifferentresponsecomponents,whichmaynotbeveryattractiveforcodeapplications.

    For the purpose of simplifying the background load description, it is proposed here to

    expresstheBESWLasthegustloading envelope(GLE), )()(' zRgzFxPbebx

    = ,multipliedbya

    backgroundfactor,Bz,

    ===H

    ebxxRRRzebxzeR dzzFzBzFBzF bbbb 0'''' )()(;/);()( (5)

    where ),()( zzRzRxxx PP

    = ; 'bRb

    g = peak background response under the loading envelope that

    does notincludethe influence of loss inspatialcorrelation ofwind loadingover thebuilding

    height;Bz represents the reduction effect with respect to the responseR(z0,t) due to loss of

    correlation of wind loading. In cases where the wind loads are fully correlated, i.e.,

    )()(),( 2121 zRzRzzR xxxx PPP = , BzbecomesunityandtheBESWLsbasedontheLRCandGLE

    schemesconvergetothegustloadingenvelope, )('

    zFebx .

    TheRESWLforthepeakresonantresponse,rRr

    g ,isgivenintermsoftheinertialload:

    )(4)()(

    )()()( 11

    10

    2fSf

    dzzzm

    zzmgzF

    xQH

    x

    xrerx

    = (6)

    whichcanalsobeexpressedintermsofthedistributionofthepeakbasebendingmomentorbase shear force over the building height following the inertial load distribution. When thetorsionalresponseisunderconsideration,theRESWLisobtainedbydistributingthebasetorqueoverthebuildingheight.

    The ESWL for the total peak dynamic response, Rmax , can be provided as a linearcombinationofthebackgroundandresonantloads(ChenandKareem 2000and2001;Holmes2002):

    ( ) maxmax' /;/;)()()( RgWRgWzFWzFBWzF

    rb RrrRbberxrebxzbeR ==+= (7)

    When the peak response includes the mean component, the ESWL is given

    as )()( zFzP eRx .

  • 7/28/2019 Equivalent Static Wind Loading on Buildings - New Perspective

    4/8

    3CLOSED-FORMFORMULATION

    Forthesakeofillustration,themassperunitheight,m(z),thefirstmodeshape,x(z),andtheinfluencefunctionoftheresponseR(z0,t),x(z),areexpressedas