from：journalofConstructionalSteelResearch.Volume59,Number1,January2003CyclicbehaviorofsteelmomentframeconnectionsundervaryingaxialloadandlateraldisplacementsAbstract:Thispaperdiscussesthecyclicbehavioroffoursteelmomentconnectionstestedundervariableaxialloadandlateraldisplacements.Thebeamspecim-ensconsistedofareducedbeamsection,wingplatesandlongitudinalstiffeners.Thetestspecimensweresubjectedtovaryingaxialforcesandlateraldisplace-mentstosimulatetheeffectsonbeamsinaCoupled-GirderMoment-ResistingFramingsystemunderlateralloading.Thetestresultsshowedthatthespecim-ensrespondedinaductilemannersincetheplasticrotationsexceeded0.03radwithoutsignificantdropinthelateralcapacity.Thepresenceofthelongitudin-alstiffenerassistedintransferringtheaxialforcesanddelayedtheformationofweblocalbuckling.IntroductionAimedatevaluatingthestructuralperformanceofreduced-beamsection(RBS)connectionsunderalternatedaxialloadingandlateraldisplacement,fourfull-scalespecimensweretested.ThesetestswereintendedtoassesstheperformanceofthemomentconnectiondesignfortheMosconeCenterExp-ansionundertheDesignBasisEarthquake(DBE)andtheMaximumConsideredEarthquake(MCE).PreviousresearchconductedonRBSmomentconnections[1,2]showedthatconnectionswithRBSprofilescanachieverotationsinexcessof0.03rad.However,doubtshavebeencastonthequalityoftheseismicperformanceoftheseconnectionsundercombinedaxialandlateralloading.TheMosconeCenterExpansionisathree-story,71,814m2(773,000ft2)structurewithsteelmomentframesasitsprimarylateralforce-resistingsystem.AthreedimensionalperspectiveillustrationisshowninFig.1.Theoverallheightofthebuilding,atthehighestpointoftheexhibitionroof,isapproxima-tely35.36m(116ft)abovegroundlevel.Theceilingheightattheexhibitionhallis8.23m(27ft),andthetypicalfloor-to-floorheightinthebuildingis11.43m(37.5ft).ThebuildingwasdesignedastypeIaccordingtotherequi-rementsofthe1997UniformBuildingCode.TheframingsystemconsistsoffourmomentframesintheEast-Westdirection,oneoneithersideofthestairtowers,andfourframesintheNorth-South

direction,oneoneithersideofthestairandelevatorcoresintheeastendandtwoatthewestendofthestructure[4].Becauseofthestoryheight,thecon-ceptoftheCoupled-GirderMoment-ResistingFramingSystem(CGMRFS)wasutilized.Bycouplingthegirders,thelateralload-resistingbehaviorofthemomentframingsystemchangestoonewherestructuraloverturningmomentsareresistedpartiallybyanaxialcompression-tensioncoupleacrossthegirdersystem,ratherthanonlybytheindividualflexuralactionofthegirders.Asaresult,astifferlateralloadresistingsystemisachieved.Theverticalelementthatconnectsthegirdersisreferredtoasacouplinglink.Couplinglinksareanalogoustoandservethesamestructuralroleaslinkbeamsineccentricallybracedframes.Couplinglinksaregenerallyquiteshort,havingalargeshear-to-momentratio.Underearthquake-typeloading,theCGMRFSsubjectsitsgirderstowariab-bleaxialforcesinadditiontotheirendmoments.TheaxialforcesintheFig.1.MosconeCenterExpansionProjectinSanFrancisco,CAgirdersresultfromtheaccumulatedshearinthelink.AnalyticalmodelofCGMRFinNonlinearstaticpushoveranalysiswasconductedonatypicalone-baymodeloftheCGMRF.Fig.2showsthedimensionsandthevarioussectionsofthemodel.Thelinkflangeplateswere28.5mm-254mm(11/8-10in)andthewebplatewas9.5mm-476mm(3/8in-183/4in).TheSAP2000computerprogramwasutilizedinthepushoveranalysis[5].Theframewascharacterizedasfullyrestrained(FR).FRmomentframesarethoseframesfor1170whichnomorethan5%ofthelateraldeflectionsarisefromconnectiondeformation[6].The5%valuerefersonlytodeflectionduetobeam-columndeformationandnottoframedeflectionsthatresultinPhcbcfarE^SjilJliric.fromcolumnpanelzonedeformationPhcbcfarE^SjilJliric.r"?■EnTi^axiyrjiwi^r偵出戲iII*網(wǎng)B?4xri^Theanalysiswasperformedusinganexpectedvalueoftheyieldstressandultimatestrength.Thesevalueswereequalto372MPa(54ksi)and518MPa(75ksi),respectively.Theplastichinges’load-deformationbehaviorwasapproximatedbythegeneralizedcurvesuggestedbyNEHRPGuidelinesfortheSeismicRehabilitationofBuildings[6]asshownin.Fig.3.△ywascalcu-latedbasedr"?■EnTi^axiyrjiwi^r偵出戲iII*網(wǎng)B?4xri^P-Mhingeload-deformationmodelpointsC,DandEarebasedonTable5.4from[6]for△ywastakenas0.01radperNote3in[6],Table5.8.Shearhingeloadload-deformationmodelpointsC,DandEarebasedonTable5.8[6],LinkBeam,Itema.AstrainhardeningslopebetweenpointsBandCof3%oftheelasticslopewasassumedforbothmodels.Thefollowingrelationshipwasusedtoaccountformoment-axialloadinteraction[6]:whereMCEistheexpectedmomentstrength,ZRBSistheRBSplasticsectionmodulus(in3),istheexpectedyieldstrengthofthematerial(ksi),Pistheaxialforceinthegirder(kips)andistheexpectedaxialyieldforceoftheRBS,equalto(kips).TheultimateflexuralcapacitiesofthebeamandthelinkofthemodelareshowninTable1.Fig.4showsqualitativelythedistributionofthebendingmoment,shearforce,andaxialforceintheCGMRFunderlateralload.Theshearandaxialforceinthebeamsarelesssignificanttotheresponseofthebeamsascomparedwiththebendingmoment,althoughtheymustbeconsideredindesign.Thequalita-tivedistributionofinternalforcesillustratedinFig.5isfundamentallythesameforbothelasticandinelasticrangesofbehavior.Thespecificvaluesoftheinternalforceswillchangeaselementsoftheframeyieldandinternalfor-cesareredistributed.ThebasicpatternsillustratedinFig.5,however,remainthesame.Inelasticstaticpushoveranalysiswascarriedoutbyapplyingmonotonicallyincreasinglateraldisplacements,atthetopofbothcolumns,asshowninFig.6.AfterthefourRBShaveyieldedsimultaneously,auniformyieldinginthewebandattheendsoftheflangesoftheverticallinkwillform.Thisistheyieldmechanismfortheframe,withplastichingesalsoformingatthebaseofthecolumnsiftheyarefixed.ThebaseshearversusdriftangleofthemodelisshowninFig.7.Thesequenceof

inelasticactivityintheframeisshownonthefigure.Anelasticcomponent,alongtransition(consequenceofthebeamplastichingesbeingformedsimultaneously)andanarrowyieldplateaucharacterizethepushovercurve.Theplasticrotationcapacity,qp,isdefinedasthetotalplasticrotationbeyondwhichtheconnectionstrengthstartstodegradebelow80%[7].ThisdefinitionisdifferentfromthatoutlinedinSection9(AppendixS)oftheAISCSeismicProvisions[8,10].UsingEq.(2)derivedbyUangandFan[7],anestimateoftheRBSplasticrotationcapacitywasfoundtobe0.037rad:FyfwassubstitutedforRyFy[8],whereRyisusedtoaccountforthedi^r-encebetweenthenominalandtheexpectedyieldstrengths(Grade50steel,Fy=345MPaandRy=1.1areused).ExperimentalprogramTheexperimentalset-upforstudyingthebehaviorofaconnectionwasbasedonFig.6(a).Usingtheplasticdisplacementdp,plasticrotationgp,andplasticstorydriftangleqpshowninthefigure,fromgeometry,itfollowsthat:And:inwhichdandgincludetheelasticcomponents.Approximationsasaboveareusedforlargeinelasticbeamdeformations.ThediagraminFig.6(a)suggestthatasubassemblagewithdisplacementscontrolledinthemannershowninFig.6(b)canrepresenttheinelasticbehaviorofatypicalbeaminaCGMRF.Thetestset-upshowninFig.Thetestset-upshowninFig.8wasconstructedtodevelopthemechanismshowninFig.6(a)and(b).Theaxialactuatorswereattachedtothree2438mmx1219mmx1219mm(8ftx4ftx4ft)RCblocks.Theseblocksweretensionedtothelaboratoryfloorbymeansoftwenty-four32mmdiameterdywidagrods.Thisarrangementpermittedreplacementofthespecimenaftereachtest.Therefore,theforceappliedbytheaxialactuator,P,canberesolvedintotwoorthogonalcomponents,PaxialandPlateral.Sincetheinclinationangleoftheaxialactuatordoesnotexceed3.0,thereforePaxialisapproximatelyequaltoP[4].However,thelateralcomponent,Plateral,causesanadditionalmomentatthebeam-tocolumnjoint.Iftheaxialactuatorscompressthespecimen,thenthelateralcomponentswillbeaddingtothelateralactuatorforces,whileiftheaxialactuatorspullthespecimen,thePlateralwillbeanopposingforcetothelateralactuators.Whentheaxialactuatorsundergoaxialactuatorsundergoalateraldisplacement_,theycauseanadditionalmomentatthebeam-to-columnjoint(P-△effect).Therefore,themomentatthebeam-tocolumnjointisequalto:whereHisthelateralforces,Listhearm,Pistheaxialforceand_isthelateraldisplacement.Fourfull-scaleexperimentsofbeamcolumnconnectionswereconducted.ThemembersizesandtheresultsoftensilecoupontestsarelistedinTable2AllofthecolumnsandbeamswereofA572Grade50steel(Fy344.5MPa).Theactualmeasuredbeamflangeyieldstressvaluewasequalto372MPa(54ksi),whiletheultimatestrengthrangedfrom502MPa(72.8ksi)to543MPa(78.7ksi).Table3showsthevaluesoftheplasticmomentforeachspecimen(basedonmeasuredtensilecoupondata)atthefullcross-sectionandatthereducedsectionatmid-lengthoftheRBScutout.Thespecimensweredesignatedasspecimen1throughspecimen4.TestspecimensdetailsareshowninFig.9throughFig.12.Thefollowingfeatureswereutilizedinthedesignofthebeamcolumnconnection:TheuseofRBSinbeamflanges.Acircularcutoutwasprovided,asillustr-atedinFigs.11and12.Forallspecimens,30%ofthebeamflangewidthwasremoved.Thecutsweremadecarefully,andthengroundsmoothinadirect-tionparalleltothebeamflangetominimizenotches.Useofafullyweldedwebconnection.Theconnectionbetweenthebeamwebandthecolumnflangewasmadewithacompletejointpenetrationgrooveweld(CJP).All

CJPweldswereperformedaccordingtoAWSD1.1StructuralWeldingCodeUseofcontinuityplateswithaUseoftwosideplatesweldedwithCJPtoexteriorsidesoftopandbottombeamflanges,fromthefaceofthecolumnflangetothebeginningoftheRBS,asshowninFigs.11and12.TheendofthesideplatewassmoothedtomeetthebeginningoftheRBS.ThesideplateswereweldedwithCJPwiththecolumnflanges.Thesideplatewasaddedtoincreasetheflexuralcapacityatthejointlocation,whilethesmoothtransitionwastoreducethestressraisers,whichmayUseofcontinuityplateswithaTwolongitudinalstiffeners,95mmx35mm(33/4inx13/8in),wereweldedwith12.7mm(1/2in)filletweldatthemiddleheightofthewebasshowninFigs.9and10.ThestiffenerswereweldedwithCJPtocolumnflanges.Removalofweldtabsatboththetopandbottombeamflangegroovewelds.Theweldtabswereremovedtoeliminateanypotentialnotchesintroducedbythetabsorbywelddiscontinuitiesinthegrooveweldrunoutregions.thicknessapproximatelyequaltothebeamflangethickness.One-inchthickcontinuityplateswereusedforallspecimens.WhiletheRBSisthemostdistinguishingfeatureofthesetestspecimens,thelongitudinalstiffenerplayedanimportantroleindelayingtheformationofweblocalbucklinganddevelopingreliableconnectionperformance4.LoadinghistorySpecimensweretestedbyapplyingcyclesofalternatedloadwithtipdisplacementincrementsof_yasshowninTable4.Thetipdisplacementofthebeamwasimposedbyservo-controlledactuators3and4.Whentheaxialforcewastobeapplied,actuators1and2wereactivatedsuchthatitsforcesimulatestheshearforceinthelinktobetransferredtothebeam.Thevariableaxialforcewasincreasedupto2800kN(630kip)at0.5_y.Afterthat,thislo-adwasmaintainedconstantthroughthemaximumlateraldisplacement.maximumlateraldisplacement.Asthespecimenwaspushedbacktheaxialforceremainedconstantuntil0.5yandthenstartedtodecreasetozeroasthespecimenpassedthroughtheneutralposition[4].AccordingtotheupperboundforbeamaxialforceasdiscussedinSection2ofthispaper,itwasconcludedthatP=2800kN(630kip)isappropriatetoinvestigatethiscaseinRBSloading.Thetestswerecontinueduntilfailureofthespecimen,oruntillimitationsofthetestset-upwerereached.5.ConclusionsBasedontheobservationsmadeduringthetests,andontheanalysisoftheinstrumentation,thefollowingconclusionsweredeveloped:Theplasticrotationexceededthe3%radiansinalltestspecimens.PlastificationofRBSdevelopedinastablemanner.Theoverstrengthratiosfortheflexuralstrengthofthetestspecimenswereequalto1.56forspecimen1and1.51forspecimen4.TheflexuralstrengthcapacitywasbasedonthenominalyieldstrengthandontheFEMA-273beamcolumnequation.Althoughflangelocalbucklingdidnotcauseanimmediatedegradationofstrength,itdidinduceweblocalbucklingTheplasticlocalbucklingofthebottomflangeandthewebwasnotaccompaniedbyasignificantdeteriorationintheload-carryingcapacity.Thelongitudinalstiffeneraddedinthemiddleofthebeamwebassistedintransferringtheaxialforcesandindelayingtheformationofweblocalbuckling.However,thishascausedamuchhigheroverstrengthratio,whichhadasignificantimpactonthecapacitydesignoftheweldedjoints,panelzoneandthecolumn.Agradualstrengthreductionoccurredafter0.015to0.02radofplasticrotationduringnegativecycles.Nostrengthdegradationwasobservedduringpositivecycles.Compressionaxialloadunder0.0325Pydoesnotaffectsubstantiallytheconnectiondeformationcapacity.CGMRFSwithproperlydesignedanddetailedRBSconnectionsisareliablesystemtoresistearthquakes.彎鋼框架結(jié)點(diǎn)在軸向變化荷載和側(cè)向位移的作用下的周期性行為摘自：鋼結(jié)構(gòu)研究雜志。59卷1號，一月，2003摘要:這篇論文討論的是在變化的軸向荷載和側(cè)向位移的作用下，接受測試的四種受彎鋼結(jié)點(diǎn)的周期性行為。梁的試樣由變截面梁，翼緣以及縱向的加勁肋組成。受測試樣加載軸向荷載和側(cè)向位移用以模擬側(cè)向荷載對組合梁抗彎系統(tǒng)的影響。實(shí)驗(yàn)結(jié)果表明試樣在旋轉(zhuǎn)角度超過0.03弧度后經(jīng)歷了從塑性到延性的變化。縱向加勁肋的存在幫助傳遞軸向荷載以及延緩腹板的局部彎曲。1、引言為了評價(jià)變截面梁（RBS）結(jié)點(diǎn)在軸向荷載和側(cè)向位移下的結(jié)構(gòu)性能，對四個全尺寸的樣品進(jìn)行了測試。這些測試打算評價(jià)為舊金山展覽中心擴(kuò)建設(shè)計(jì)的受彎結(jié)點(diǎn)在滿足設(shè)計(jì)基本地震等級（DBE）和最大可能地震等級（MCE）下的性能?；谏鲜龆龅膶BS受彎結(jié)點(diǎn)的研究指出RBS形式的結(jié)點(diǎn)能夠獲得超過0.03弧度的旋轉(zhuǎn)角度。然而，有人對于這些結(jié)點(diǎn)在軸向和側(cè)向荷載作用下的抗震性能質(zhì)量提出了懷疑?？诠窵Lwim】凸?rar匚白七舊金山展覽中心擴(kuò)建工程是一個3層構(gòu)造，并以鋼受彎框架作為基本的側(cè)向力抵抗系統(tǒng)。Fig.1是一幅三維透視圖。建筑的總標(biāo)高為展覽廳屋頂?shù)淖罡唿c(diǎn)，大致是35.36m（116ft）。展覽廳天花板的高度是口弓「LLwim】凸?rar匚白七框架系統(tǒng)由以下幾部分組成：四個東西走向的受彎框架，每個電梯塔邊各一個；四個走向的受彎框架，在每個樓梯和電梯井各一個的；整體分布在建筑物的東西兩側(cè)?？紤]到層高的影響，提出了雙梁抗彎框架系統(tǒng)的觀念。通過連接大梁，受彎框架系統(tǒng)的抵抗荷載的行為轉(zhuǎn)化為結(jié)構(gòu)傾覆力矩部分地被梁系統(tǒng)的軸向壓縮-拉伸分擔(dān)，而不是僅僅通過梁的彎曲。結(jié)果，達(dá)到了一個剛性側(cè)向荷載抵抗系統(tǒng)。豎向部分與梁以聯(lián)結(jié)桿的形式連接。聯(lián)結(jié)桿在結(jié)構(gòu)中模擬偏心剛性構(gòu)架并起到與其相同的作用。通常地聯(lián)結(jié)桿都很短，并有很大的剪彎比。在地震類荷載的作用下，CGMRFS梁的最終彎矩將考慮到可變軸向力的影響。梁中的軸向力是切向力連續(xù)積累的結(jié)果。2.CGMRF的解析模型非線性靜力推出器模型是以典型的單間CGMRF模板為指導(dǎo)。模型的尺寸規(guī)格

和多個部分。翼緣板尺寸為28.5mmx254mm（11/8inx10in），腹板尺寸為9.5mmx476mm（3/8inx183/4in）。推進(jìn)器模型中運(yùn)用了SAP2000計(jì)算機(jī)程序?？蚣艿奶厣侨s束（FR）。FR受彎框架是一種由結(jié)點(diǎn)應(yīng)變引起的撓度不超過側(cè)向撓度的5%的框架。這個5%僅與梁-柱應(yīng)變有關(guān)，而與柱底板區(qū)應(yīng)變引起的框架應(yīng)變無關(guān)。模型通過屈服應(yīng)力和匹配強(qiáng)度的期望值來運(yùn)行。這些值各自為372Mpa（54ksi）和518Mpa（75ksi）。Fig.3顯示了塑性鉸的荷載-應(yīng)變行為是通過建筑物地震恢復(fù)的NEHRP指標(biāo)以廣義曲線的形式逼近的?！鱵以Eps5.1和5.2為基底運(yùn)算，如下：P-M鉸合線荷載-應(yīng)變模型上的點(diǎn)C，D和E的取值如表5.4Ay以0.01rad為幅度取值見表5.8。切變鉸合線荷載-應(yīng)變模型點(diǎn)C，D和E取值見表5.8。對于連續(xù)梁，假定兩個模型點(diǎn)B和C之間的形變硬化比有3%的彈性比。用下面的公式計(jì)算彎矩與軸向荷載之間的相互關(guān)系Mce是期望彎矩強(qiáng)度，Zrbs是RBS塑性模量，尸形是材料的屈服強(qiáng)度，P是梁中的軸向力，Pye是RBS屈服力，等于A^F^。梁的最終彎曲能力和模型的連續(xù)行見圖1。Fig.4定性的給出了側(cè)向荷載下的CGMRF中的彎矩，切應(yīng)力和正應(yīng)力的分布。其中切應(yīng)力和正應(yīng)力對梁的影響要小于彎矩的作用，盡管他們必須在設(shè)計(jì)中加以考慮。內(nèi)力分布圖解見Fig.5，可見，彈性范圍和非彈性范圍的內(nèi)力行為基本相同。內(nèi)力的比值將隨框架的屈服和內(nèi)力的重分布的變化而變化?；緝?nèi)力圖見Fig.5,然而，仍然是一樣的。rl;?IjMejrl;?IjMej-MillAFig.5所示。在四個RBS同時(shí)屈服后，發(fā)生在腹板與翼緣端部的豎向的統(tǒng)一屈服將開始形成。這是框架的屈服中心，在柱子被固定后將在柱底部形成塑性鉸。Fig.7給出了基本切應(yīng)力偏移角。圖中還給出了框架中非彈性活動的次序。對于一個彈性組成，推進(jìn)器將有一個特有的很長的過渡（同時(shí)形成塑性鉸）和一個很短的屈服平穩(wěn)階段。%=%=5賜（9-0.3塑性旋轉(zhuǎn)能力，，被定義為：結(jié)點(diǎn)強(qiáng)度從開始遞減到低于80%的總的塑性旋轉(zhuǎn)角。這個定義不同于第9段(附錄)AISC地震條款的描述。使用Eq源于RBS塑性旋轉(zhuǎn)能力被定在0.037弧度。Ff被R替代，R用來計(jì)算理論屈服強(qiáng)度與實(shí)際屈服強(qiáng)度的區(qū)別(標(biāo)號是50鋼)3.實(shí)踐規(guī)劃如圖6所示，實(shí)驗(yàn)布置是為了研究基于典型的CGMRF結(jié)構(gòu)下的結(jié)點(diǎn)在動力學(xué)中的能量耗散。用圖中所給的塑性位移，塑性轉(zhuǎn)角，塑性偏移角，由幾何結(jié)構(gòu)，有如下：=(S小p(5)和尸作=己=編⑹這里的。和Y包括了彈性組合。上述近似值用于大型的非彈性梁的變形破壞。圖6a表明用圖6b所示的位移控制下的替代組合能夠表示CGMRF結(jié)構(gòu)中的典型梁的非彈性行為。iilnS-dcm*圖8所示，建立這個實(shí)驗(yàn)裝置來發(fā)展圖6a和圖6b所示的機(jī)構(gòu)學(xué)。軸心裝置附以3個2438mmX1219mmX1219mm(8ftX4ftX4ft)RC塊。并用24個32mm徑的桿與實(shí)驗(yàn)室的地板固定。這種裝置允許在每次測驗(yàn)后換實(shí)驗(yàn)樣品。iilnS-dcm*根據(jù)實(shí)驗(yàn)布置的動力學(xué)要求，隨著側(cè)面的元件放置，軸向的元件，元件1和元件2,將釘?shù)紹和