876 ACI Structural Journal/November-December 2000ACI Structural Journal, V. 97, No. 6, November-December 2000.MS No. 00-064 received March 14, 2000, and reviewed under Institute publicationpolicies. Copyright 2000, American Concrete Institute. All rights reserved, includ-ingthemakingofcopiesunlesspermissionisobtainedfromthecopyrightpropri-etors.PertinentdiscussionwillbepublishedintheSeptember-October2001ACIStructural Journal if received by May 1, 2001.

ACI STRUCTURAL JOURNAL TECHNICAL PAPERAn experimental investigation on the seismic behavior of reinforcedconcrete coupling beams is presented. The reinforcement layout andtheloadinghistorywerethemainvariablesofthetests.Fifteenshortcouplingbeamswithfourdifferentreinforcementarrange-mentsweretested.Theyweresubjectedtomonotonicandcyclicloading by a suitable experimental setup. All specimens were char-acterizedbyashearspan-depthratioof0.75.Thereinforcementlayoutsconsistedof:aclassicalscheme(a);diagonalschemewithoutconfiningties(b1);diagonalschemewithconfiningties(b2); and inclined bars to form a rhombic configuration (c). Con-crete compressive strengths of the specimens varied from 40 to 54MPa. Test results showed that the beams with diagonal or rhombicreinforcement layouts behaved better than beams with longitudinalarrangement of the steel bars. These results were produced by thedifferent resisting truss mechanisms that were developed in the cou-pling beams after the first cracking. The differences in energy dissi-pation were negligible between the diagonal and rhombic layouts.Therhombicarrangement,however,wasmoreadvantageousintermsofrotationalductilitycapacity,anddecayinstrengthandstiffness of the beams. Moreover, cyclic tests demonstrated that thebehavioroftherhombiclayoutwaslessaffectedbythedifferentloading histories.Keywords:cyclicloads;ductility;earthquake-resistantstructures;shearwalls.INTRODUCTIONStructuralwallshavelongbeenusedintheplanningofmultistory buildings. A reliable earthquake-resistant systemis represented by coupled shear walls, in which two verticalelements are interconnected by short and deep beams. Since1977, when the American Concrete Institute (ACI 318-771)stipulated specific rules for the design of ductile reinforcedconcrete(RC)structures,theconceptofductilitycapacityhas played an increasing role in the seismic design of struc-tures.2,3 The beams of a coupled wall are subjected to largecyclic shear deformations during an earthquake. Hence, oneof the most critical problems of these structures concerns thebrittle failure of the low slenderness coupling beams. If thisfailure is avoided, a large fraction of the input seismic energywillbedissipatedbytheseelementsthataredistributedthroughout these walls. Consequently, building safety will beimproved. This favorable behavior of coupled walls has been pre-viously demonstrated in other research.4-6In previous experimental research work, shear strength andhysteretic behavior of short reinforced concrete members un-der cyclic loads have been investigated. In the original exper-imental studies, the classical layout of the steel bars was used,whichconsistedoflongitudinalreinforcementsthatwereplaced parallel to the beam axis. In later experimental studies,differentarrangementsofmainreinforcementswerepro-posed to improve the seismic response of these short mem-bers. In 1974, Paulay and Binney first introduced the idea ofusing diagonal main reinforcements that followed the bend-ing moment distribution along the beam axis. Subsequent re-search showed that the beams reinforced with diagonal barsperformedbetterthanthosewithparallelreinforcements.7,8These conclusions were true even when the number of trans-versaltiesincreased.Recently,differentresearcherspro-posedalternativesolutionsforthearrangementofthesteelbarsintheseshortcouplingelements.9Innovativecouplingbeams with keyways and through slits along the axis of bothfacesweretested.Nevertheless,theseschemescomplicatethe construction and consequently the cost of the building. Giventhesatisfactorybehaviorofdiagonallyreinforcedcoupling beams, their use in seismic design is recommendedbyEuropeanStandards.3Inapreviouspaper,10themaindrawbacks of the diagonal arrangement were illustrated and adifferentlayoutwasproposedtoovercomethesedisadvan-tages.TegosandPenelis10proposedanarrangementofthemain steel bars with an inclination that formed a rhombic lay-out. Their results showed that diagonal and rhombic layoutsbehaved in a similar manner. By using the rhombic arrange-ment, however, the occurrence of diagonal explosive cleav-ageshearfracturewasavoided.Despitethesatisfactoryresults, direct comparisons of ultimate strength, stiffness de-cay, and energy dissipation were not presented.The results of an experimental program in which 15 shortcouplingbeamswithfourdifferentreinforcementlayoutswere tested under monotonic and cyclic shear loading are de-scribedinthispaper.Theresearchfocusedmainlyonthecomparativebehaviorsofthebeamsintermsoffailuremechanism, ultimate strength, degradation in stiffness, andhysteretic capabilities.RESEARCH SIGNIFICANCEWhencoupledshearwallsareusedinseismicdesignofbuildings, a satisfactory global response is achieved only ifcoupling beams respond in a ductile fashion. The resultsof this research show that the rhombic layout of the mainreinforcementimprovestheductilitycapacityofshortmembersundercyclicshearwithoutasignificantlossinstrengthandstiffness.Untilnow,thepracticalapplicationofthis layout in coupling beams has been limited. Serious problemswith construction and difficulties in manufacturing can occur us-ing diagonal layout with confining ties when the thickness ofthe wall is less than 250 mm. For thin coupled shear walls es-pecially,however,therhombicschemesimplifiesbuildingTitle no. 97-S89Seismic Behavior of Short Coupling Beams with Different Reinforcement Layoutsby Luciano Galano and Andrea Vignoli877 ACI Structural Journal/November-December 2000construction and reduces the costs when compared with thediagonal arrangement.EXPERIMENTAL PROGRAMSpecimensTheliteratureindicatesthatthestrengthandductilityofshort coupling beams are mainly affected by: 1) the concretestrength; 2) the shear span-depth ratio; 3) the amount and thearrangementofthemainreinforcement;4)theamountoftransversalreinforcement;and5)thetypeoftheappliedloading. In this investigation, the variables that were the ob-ject of the study were the main reinforcement layout and theloadinghistories.Therewere15shortcouplingbeamsandfourdifferentreinforcementlayoutsusedinthisseriesoftests:longitudinal(a);diagonalwithouttransversalconfin-ing ties around the main bars (b1); diagonal with transversalconfiningties(b2);andrhombiclayout(c).Figure1andTable 1 show the overall dimensions and design details ofthe specimens; two lateral stiff blocks were used to simu-latethesurroundingwallsandapplytherequiredloadinghistories. To emphasize the adverse shear effects in the in-elastic response, the shear span-depth ratio was assumed tobe constant and equal to 0.75. The cross section of all cou-pling beams was equal to 400 (height) x 150 mm (width) andthepercentageofmainreinforcementl=As/(hs)wasas-sumed to be equal to 0.524 (four steel bars of 10 mm diame-ter; refer to Fig. 1).The b2 scheme differs from b1 (Fig. 1) in the use of trans-verse ties that are around the four diagonal bars (22 ties of 6mm diameter). This corresponds to the provisions detailingof Eurocode 8. The c scheme was previously proposed10 andused here for a direct comparison with b1 and b2 by using anequalamountofmainreinforcements.Transverseverticaltiesfollowedthecapacitydesignapproach.Forthespeci-mens of series a the provisions of Eurocode 211 were used.For the series c, the suggestions given in Reference 10 for theminimumtransversereinforcementratiowereutilized.Fi-nally, for b1 and b2 schemes, two 6 mm diameter longitudi-nal bars were placed on the top and the bottom sides of thesections, and a minimum quantity of vertical ties was used(as proposed in Eurocode 8). Table 1 gives designations, re-inforcement arrangements, main (l) and transversal (v) re-inforcement ratios for all coupling beams (v was defined asthe ratio between the total volume of vertical stirrups insidethe beams and the concrete volume Vc = hslT). MaterialsThe steel used for all reinforcements consisted of hot- rolleddeformedbarsofFeB44kgrade.Themainreinforcementconsisted of four 10 mm diameter bars. Bars of 6 and 8 mmwere used for vertical ties that were placed with different spac-ing for the different series of specimens (Fig. 1 and Table 1).Stress-strain tests were undertaken on the reinforcing steel us-ing a universal testing machine. The strains were measured us-ing electric strain gages over a gage length of 5nom (nom isdefined as the nominal diameter of the bar). A total number of20 tests were performed. Average values obtained by the testswereyieldstrengthsyequalto567MPa(variationintherange 528 to 611 MPa) and ultimate strength st of 660 MPa(variation in the range 641 to 685 MPa). Prior to the fractureofthesteelbars,theminimumelongationmeasuredoveralength of 5nom was 18.3%. No significant differences in mod-ulus of elasticity of the steel bars were found and the averagemeasured value was Es = 206 103 MPa.A normal-strength commercial concrete mixture was usedand the cement was portland cement Type 425. The concretewas specified to have a 10 mm maximum aggregate size withcomplete grain size distribution (Fuller bounds) and averagewater-cement ratios (w/c) = 0.40 were used. Workability wasachieved by adding to the mixture a normal-setting water re-ducer that was approximately 1% of the cement weight. TheLuciano Galano is a researcher in the Department of Civil Engineering at the Univer-sity of Florence, Florence, Italy. He received his PhD from the University of Florence in1993. His research interests include the seismic behavior of reinforced concrete struc-tures, the structural use of high-strength concrete, and theoretical dynamics.Andrea Vignoli is an associate professor of dynamics of structures at the Universityof Florence. His research interests include experimental tests on reinforced concreteand masonry structures, high-performance concrete, and structural rehabilitation.Fig. 1Details of short coupling beams.878 ACI Structural Journal/November-December 2000specimens were cast horizontally in reusable steel forms. Theprescribednominalvalueoftheconcretecover(equalto15mm) was obtained using suitable spacers placed in the bottom(Fig. 2). The specimens were cured at a temperature of 20 Cand with relative humidity of 50% for a minimum of 28 days.Subsequently, they were placed in a covered environment thatwas open to the elements and were subjected to normal atmo-spheric conditions until the testing. To assess the seismic per-formances of the coupling beams to an age corresponding tothe real conditions of serviceability, experimental tests wereconducted after a time that varied from 4 to 5 years from thetimeofcasting.Therefore,themechanicalcharacteristicsofthe concrete were determined on cylinder cores extracted fromthe lateral blocks of each specimen after the test. Maximumcare was taken during the extraction of the cores to avoid anydisturbance. Three compression tests for each specimen wereperformed(averagedimensionsofeachcylinderwere:130mm in height by 65 mm in diameter).ForSpecimensP02,P03,P14,P15,andP16,thesecantmodulusofelasticityatastresslevelofapproximately0.3timestheultimatestrengthwasdetermined.Forthesefivespecimens, one cylinder 150 mm in height by 35 mm in di-ameter was extracted to evaluate the direct tensile strength.Table 1 gives: 1) the specific weight c; 2) the average valuesof compressive strength fcm; 3) direct tensile strength fct; and 4)secant moduli Ecs. The obtained average compressive strengthswere similar for a, b1, and c series (approximately 48 MPa) butthecompressivestrengthoftheb2serieswasless(approxi-mately42.7MPa).Theobtainedabnormalcompressivestrengthsofb2wereprobablyduetothecastingandcuringconditionsthatoccurredintheconcrete,whichhadagreatercomplexity of steel bars arrangements in the b2 specimens. Experimental procedure and instrumentationShort coupling beams were tested in a vertical plane un-derconditionsofasymmetricbendingandconstantshear(Fig. 3). Six steel rollers were used to constraint the speci-mensinafabricatedstifftestingsteelframe.Tworollers,placed laterally, were used to prevent undesirable horizontalmovements of the specimens. The other four rollers were uti-lizedtoproducethedesiredloadinghistories.Toevenoutany irregularities of the concrete, six stiff steel plates gluedtotheconcretesurfacewereused.Stiffverticalsteeltieswere used to attach the specimens to the two hydraulic actu-ators with each actuator capable of exerting a maximum loadof 350 kN. Figure 3 shows details of the experimental setupand Fig. 4 depicts a coupling beam before the test.Twodifferenttypesoftestswereconducted:monotonictests (M) in which equal and opposite rotations B1 and B2wereimposedonthelateralblocksandincreaseduntilcol-lapse, and cyclic tests (C) in which loading and reloading cy-clesofincreasingamplitudeforB1andB2wereimposed.Table 1Details of test specimensSpecimenReinforcement arrangementlvc, kN/m3fcm, MPafct, MPaEcs, MPaLoading historyP01 a 0.524 0.84 21.36 48.9 MP02 a 0.524 0.84 21.78 44.5 3.3 24,464 C1P03 a 0.524 0.84 21.37 52.4 4.1 24,357 C2P04 a 0.524 0.84 21.41 48.7 C3P05 b1 0.524 0.39 20.76 39.9 MP06 b1 0.524 0.39 21.43 46.0 C1P07 b1 0.524 0.39 21.37 54.0 C1P08 b1 0.524 0.39 21.61 53.4 C2P10 b2 0.524 0.31 21.45 46.8 MP11 b2 0.524 0.31 21.22 39.9 C1P12 b2 0.524 0.31 21.20 41.6 C1P13 c 0.524 0.55 21.37 47.5 MP14 c 0.524 0.55 21.24 45.0 3.3 23,436 C1P15 c 0.524 0.55 21.34 45.0 3.8 24,304 C2P16 c 0.524 0.55 21.38 51.1 3.5 27,342 C3Note: 1 MPa = 145 psi; 1 kN = 225 lbf; fcm = mean cylinder compressive strength; M = monotonic loading; and C = cyclic loading.Fig.2Reinforcementlayoutsb2andcofspecimensinsteel form before casting.ACI Structural Journal/November-December 2000 879Instrumentation consisted of four linear variable displacementtransducers (LVDTs) to measure vertical displacements 1,2, 3, and 4 of the lateral blocks. Additional LVDTs wereplaced to evaluate accidental movements of the specimens inthehorizontaldirection(O1andO2).Thetwodiagonaldeformations of the coupling beams on both sides werealsomeasured. Electric signals of the LVDTs and the loadvaluesP1 and P2 were continuously measured and recordedduring the tests. Due to the dimensional and machine toleranc-es of the testing setup, the rotations B1 and B2 were slightlydifferent; therefore, the following equations were used to evaluatethe rotation L and the shear V in the coupling beams (Fig. 3) =(1)the rotation of the beam axis(2)Table1andFig.5showthedifferentloadinghistoriesused for all the tests; Ly indicates the yielding rotation of thecouplingbeamsevaluatedasspecifiedlaterinthispaper.Three different types of cyclic patterns were used to investi-gatetheeffectsonstrengthandstiffness.Testswerecon-ducted applying a velocity of approximately 0.4 mm/s to theactuatorsstroke.Thetotaltestingtimeforeachspecimentested using monotonic loading was approximately 50 min.The duration of the cyclic tests varied considerably for dif-ferent specimens, from approximately 100 to 180 min.STATIC AND CYCLIC DUCTILITY IN SHORT COUPLING BEAMSThe case to be considered concerns a short beam subjectedtomonotonicincreasingsheardistortioninanRCcoupledwall. Strength and ductility of the member are related to theLL1L2+2-----------------------B1B2+2------------------------ G+ = =V P1R1 P11cl-- = =Fig. 3Loading setup.Fig. 4Experimental setup of specimen before test.Fig. 5Loading histories.880 ACI Structural Journal/November-December 2000shear load (V) rotation (L) diagram (Fig. 6(a)). A measure-ment of static ductility s is given by(3)where Ly (yielding rotation) is defined by the 3/4 rule (Fig. 6(a)),and Lu is the rotation at the point when the shear is reduced to0.85 times the peak value Vu. The ductility ratio, as suggest-ed by Eq. (3), is largely used in the literature but still has sev-eraldrawbacks.Themaindifficultyisthatthereisnodistinction between hinges with a large residual carrying ca-pacityandhingeswithnoresidualcapacityafterthebeamsoftens to 85% of the peak load. Alternative measurementsof ductility are similar to those adopted for flexural tough-ness and are based on the area under the shear load V versusL curve (Fig. 6(b)). Two ductility indexes are denoted as I5and I10; I5 is the ratio between the area OACD and the areaOAB,whereasI10istheareaOACEFdividedbytheareaOAB. Rotation Ly is defined by the 3/4 rule, as previouslygiven. For a perfectly elastoplastic material, values of I5 = 5and I10 = 10 are obtained. In cyclic tests, energy dissipationand the concept of cumulative ductility are used to assess thebehavior of an RC member under alternate loading in the in-elastic field. In the present study, the cumulative ductility in-dexcumasafunctionofthenumberofhalf-cycleswasdefinedbytheequationsgivenasfollows(themeaningofL1, Ly, and L2 is defined in Fig. 6(c))(4)(5)(6)Equation (4) applies only if L1 > Ly. Decay of the peakloads (V1, V2, V3, etc.) and of the secant stiffnesses (k1, k2,k3, etc.) are functions of the cumulative ductility and repre-sent further characteristics of the progressive damage of themembers.sLuLy-------- =cum11L1Ly-------- = =cum2cum12+ =2L2Ly------------ =ANALYSIS OF TEST RESULTSThe seismic behavior of the short coupling beams was in-vestigatedbasedonthefollowingobservedcharacteristics:(1) strength, stiffness, and static ductility as given by mono-tonic tests, (2) the degradation of strength and stiffness, and(3) energy dissipation based on the cyclic test results.Shear-rotation responsesResults of monotonic tests are summarized in Table 2. Thecomparison of the load-rotation responses of Tests P01, P05,P10, and P13 is given in Fig. 7. Shear strength Vu is identifiedin the diagrams with the maximum values of the shear forceV. The configuration of the bars with diagonal reinforce-mentlayouts(P05andP10)givehigherstrengthsandelasticstiffnessky;smalldifferencesbetweenb1andb2were caused by the different concrete compressive strength.Lower strength values were obtained with the reinforcementlayouts a and c, which corresponded to a decrease of 7.5 and17%.Thepercentagesofreductionofelasticstiffnessbe-tween diagonal and the other schemes of reinforcement weremore pronounced. Specimens P05 and P10 produced a rota-tion ductility factor s of approximately 8. This value was inaccordance with theoretical calculations for these structuralFig.6Ductilitymeasurementsincouplingbeams:(a)staticductility;(b)energyductility;and(c)cumulativeductilityincyclic tests (ky of Fig. (c) is equal to ky of Fig. (a)).Fig.7Shear-rotationresponses(monotonicload,1kN=225 lbf).(a) (b)(c)ACI Structural Journal/November-December 2000 881members. Tests of P01 and P13 specimens produced ductil-ity factor values equal to 6 and 9.3. Finally, the computed I5and I10 ductility factors were similar for all the specimens.As shown in Fig. 7, the tests were conducted until the cou-pling beams collapsed due to the fracture of one or more re-inforcingbarsorwhenthecarriedloadwasreducedmorethan2/3ofthereachedmaximumload.SpecimenP01showed the typical cracking diagonal pattern resulting fromthe shear loading, and, after the yield point, the increase ofthe shear force was negligible. The collapse occurred due tothesubsequentfailureofthestirrupsatthetopsideofthespecimenthatresultedintheabruptdecreasingofthesus-tained load (Fig. 8). Specimens P05 and P10 showed a moreductile response on failure; after the diagonal cracking, twomain vertical cracks appeared in the sections where the max-imum bending moment was located. This allowed the cou-pling beams to develop high rotations after yielding withoutsignificant loss of strength. The collapse occurred due to theinstability of the compressive strut that caused a gradual de-creasing of the shear without fracture of steel bars (Fig. 8).Finally, the test on Specimen P13 was characterized by: 1) alarge increment of carried shear after yielding; and 2) veryhighrotationvalues.Asuddenfailure,however,occurreddue to the fracture of two diagonal reinforcing bars. Beforethefailure,thespecimendevelopedtwodistinctseparatezones of diagonal cracking. This was followed by the split-tingoftheconcreteonthetopandthebottomsidesofthesections that were attached to the two lateral blocks.A comparison between experimental (Vu) and theoretical(Vu*) ultimate strengths is given in Table 2. For the a schemeof reinforcement, shear Vu* was computed based on ultimateflexural capacity using Eurocode 2; for b1 and b2 configura-tions, Vu*wascalculatedbyaddingthecontributionoftheflexural minor reinforcements to the shear force carried by themaindiagonalbarscalculated,assuggestedinEurocode8.For the c configuration, the truss model proposed in Reference10 was used. The previously given yielding strength valuessy were used for all the previously mentioned calculations.Theoretical shear values were notably less than experimentalvalues. This was especially true for the c reinforcement lay-out. These discrepancies between the theoretical and the ex-perimentalresultsweremainlyduetotheoverstrengthproduced by the hardening of the steel. Therefore, the theo-retical truss schemes used for the previsions of ultimate shearare reliable for design purposes. Strength and stiffness degradationThereductionofstrengthandstiffnessofthereinforcedconcrete members subjected to cyclic loading is significant forstructures in seismic areas. Therefore, members with signifi-cant degradation of strength and stiffness due to the impositionofseverecyclicloadingmustbeavoidedinseismicdesign.The reductions of the shear load Vi corresponding to the max-imum rotation in each half-cycle versus cumulative ductilityare shown in Fig. 9(a), (b), and (c) for the applied differentloading histories.All specimens tested using the loading history C1 sustaineda shear load that was greater than 0.8 Vu until the cum valueswere less than 20. As the cumulative ductility increased, dif-ferent behaviors were observed between Specimen P14 (c re-inforcement layout) and Specimens P11 and P12 (b2 layout);the P02 and P07 specimens (a and b1 schemes) produced in-termediateresults.Thesignificantdegradationofthestrengths of Specimens P11 and P12 that were tested with thesame loading histories was caused by the adverse effects ofthe lower concrete compressive strength of the b2 series. Dur-ingthecyclictests,anunexpectedcrushingoftheconcretestrut occurred when the imposed rotation ductility was equalTable 2Ultimate shear loads, rotations, and ductility factors dataSpecimenVu, kN Vu*, kN Ly rad 103Lu rad 103sI5I10P01 (a) 223.9 211.2 8.42 50.89 6.04 4.61 9.15P05 (b1) 239.3 198.6 8.42 65.41 7.77 4.59 9.26P10 (b2) 241.1 198.6 7.75 62.04 8.00 4.57 9.11P13 (c) 200.5 155.6 8.98 83.63 9.31 4.37 8.97Note: 1 kN = 225 lbf.(a)(b)Fig.8Stateofspecimensafterfailurestageundermono-tonic loading: (a) Specimen P01 (layout a); and (b) Speci-men P10 (layout b2).882 ACI Structural Journal/November-December 2000toapproximately5.Thecrushingoftheconcretestrutpro-duced the instability of the diagonal reinforcements. This wasfollowed by a noticeable loss in strength for cumulative duc-tility greater than 20. In spite of the absence of the square tiesaround the diagonal reinforcements, the instability phenom-enondidnotoccurduringthetestsofSpecimensP06andP07 (b1 layout). Consequently, a less pronounced decay inthe strength occurred.Specimen P02, with main reinforcement placed parallel tothebeamaxis,demonstratedratiosVi/VucomparablewithSpecimensP06andP07butwithlowervaluesinstrengthwhen the cum value was approximately 40 (Fig. 9(a)). Cumu-lativeductilityvaluesrelatedtothesignificantratioVi/Vuequal to 0.5 were approximately 40 for the b2 layout, 60 forthe a and b1 layouts, and 80 for the c layout. Similar conclu-sions can be drawn from Fig. 9 (b) and (c). With the loadinghistoryC2,SpecimensP03andP08givesimilarresponseswithin the range in which cum reached values of 60. For val-ues of cum higher than 60, Specimen P15 demonstrated betterperformances.Finally,testsonSpecimenP03revealedagreaterdecayinthestrengthcorrespondingtointermediateductility values.Figure 10(a), (b), and (c) show the decay in the stiffnessand plots the ratio ki/ky versus cumulative rotation ductility.Thesecantstiffnessvaluesineachhalf-cycleareki((Fig.6(c)). Again, the c layout of reinforcement demonstrates itssuperiority for higher values of cum. Figure 11 shows the fi-nal results of the tests on Specimens P03, P07, P12, and P16.SpecimenP12,whichcrushedduetotheinstabilityofthemain reinforcements, shows a minor spalling of concrete. TheP16specimenconcretespallednearthefourcornersoftheshortbeam,butnodegradationoccurredinthecentralpart.This behavior was due to the favorable effect of concrete con-finement produced by the rhombic configuration of the bars.Energy dissipation characteristicsThe energy dissipation of the specimens under cyclic load-ingwasdefinedastheareaenclosedbytheshear-rotationhysteresis loops. Graphs of cumulative dissipated energy Ehversus half-cycles number are plotted in Fig. 12(a), (b), and(c)forthedifferentloadinghistories.Table3presentsthevalues of Eh at 20, 30, and 40 half-cycles for C1, C2, and C3historiestogetherwiththeassociatedcumulativeductilitycum.Figure12(a)and(b)showsaslight(a)(b)(c)Fig.9Degradationofstrengthfordifferentlayoutsofreinforcement(shearViisdividedbyultimateshearstrengthVuasgiven by monotonic tests): (a) loading history C1; (b) loading history C2; and (c) loading history C3.(a)(b)(c)Fig. 10Degradation of stiffness for different layouts of reinforcement (stiffness ki is divided by stiffness ky as given by mono-tonic tests): (a) loading history C1; (b) loading history C2; and (c) loading history C3.ACI Structural Journal/November-December 2000 883butobservablesuperiorbehaviorofSpecimensP07andP08 compared with Specimens P14 and P15. This superiori-ty was due to a more acceptable form of the hysteresis loops.SpecimensP02,P03,andP04demonstratedlowerenergydissipation. This was mainly due to the pinched form of thehysteresisloops(Fig.13).Figure12(a)showsthatSpeci-mens P11 and P12 suffered failures that were caused by thestrutsinstability.Inthefirst15half-cycles,however,thedissipated quantities of energy were comparable to the ener-gies dissipated by Specimens P06 and P07. From the resultsof the experimental tests on specimens of b1, b2, and a con-figuration,nosignificantdifferencesinenergydissipationwere observed when the loading history was varied. The teston Specimen P16 (c layout) that was subjected to a C3 load-ing history produced higher quantities of Eh compared withP14 and P15 tests.CONCLUSIONSThe previously discussed results lead to the following con-clusions:1. The rhombic layout of the main reinforcements gave thehighestrotationalductilityvalues(greaterthan9.0inthemonotonictest).Valuesapproximatelyequalto8.0wereobtained using the diagonal arrangement. The rhombic lay-out,however,producedlowervaluesofstrengthwiththesamegeometricalpercentageofsteelarea.Thisreductionwas approximately equal to 17%;2. The concrete compressive strength greatly affected theseismic behavior of the coupling beams that were reinforcedwith diagonal bars (compressive strength of 48 MPa for b1specimens, and 42.7 MPa for b2 specimens). Despite the factthat the diagonal bars were confined by the square ties, spec-imens with the lower values of fcm suffered a short time fail-ure for instability of the compressive strut. Specimens of b1series with the greater values of fcm behaved better and theinstability phenomenon was avoided. This different behaviorresulted in the cyclic tests in terms of different energy dissipationthat reached average value of 59.5 kNrad for coupling beams ofb1 series, and 34.8 kNrad for coupling beam of b2 series;3. A notable superiority of the rhombic layouts in carryingthe shear load under repeated loading at high rotation ductil-ity was obtained (Fig. 9);4. Using the rhombic configuration, the occurrence of thediagonal explosive cleavage shear fracture was avoided;(a)Fig.11Stateoffourspecimensafterfailurestage(cyclictests):(a)SpecimenP03(layout a); (b) Specimen P07 (layout b1); (c) Specimen P12 (layout b2); and (d) Spec-imen P16 (layout c).Table 3Dissipated energy and cumulative ductility data in cyclic testsSpecimenReinforcement arrangementNo. of half- cyclesEh, kN rad cumP02 (C1) a 20 33.47 101.89P03 (C2) a 30 26.38 76.46P04 (C3) a 40 30.09 134.77P06 (C1) b1 20 52.39 94.84P07 (C1) b1 20 64.97 99.60P08 (C2) b1 30 61.18 71.48P11 (C1) b2 16 34.51 49.46P12 (C1) b2 16 34.99 53.90P14 (C1) c 20 65.14 94.48P15 (C2) c 30 52.72 73.99P16 (C3) c 40 79.52 115.37Note: 1 kN = 225 lbf.(b)(c)(d)884 ACI Structural Journal/November-December 20005.Comparableenergydissipationquantitieswereachieved with the diagonal and the rhombic layouts. A slightsuperiority, however, of the rhombic configuration that dis-sipatedanaveragehystereticenergyofapproximately65.8kNradwasachieved(9.6%morethantheenergydissipatedbythespecimenswiththediagonalconfigura-tion); and6. Using this studys test results, the analytical formulationfortheultimatestrengthproposedbyTegosandPenelis10wasverified.Thetestresultsindicatedthatthetheoretical(a)(b)(c)Fig. 12Cumulative dissipated energy for different layouts of reinforcement (1 kN = 225 lbf): (a) loading history C1; (b) load-ing history C2; and (c) loading history C3.Fig. 13Shear-rotation relationships (cyclic loading): (a) P02; (b) P07; (c) P12; and (d) P14.(b) (a)(c) (d)ACI Structural Journal/November-December 2000 885truss scheme used for the previsions of ultimate shear is safefor design purposes.Overall, the results of this study confirm better seismic per-formances using inclined (rhombic) layout of reinforcing barsforshortcouplingbeams.Thislayoutdemonstratedoptimalductility resources and capacity of sustained high-shear forc-esunderrepeatedloading.Thislayoutofreinforcementiseasier to fabricate than the diagonal layout that involves con-finementties.Anotherdisadvantageofthediagonallayoutwhen used in thin RC walls is the difficulties of the placementof the steel bars when the RC walls are being fabricated. Given the optimal seismic performances and the simplici-ty of detailing, it is hoped that the rhombic layout of main re-inforcements of coupling beams will be used with increasingconfidence in aseismic buildings.ACKNOWLEDGMENTSThe financial support provided by the MURST (Ministry for Universityand Scientific and Technological Research) is greatly appreciated. The au-thors thank Baraclit S.p.A. (Arezzo, Italy) that manufactured the specimenstested in this research work.NOTATIONAs=area of main reinforcementsa, b=distances of lateral rollers from axis of applied loadsc=distance between axes of applied loadsEcs=concrete secant modulus Eh=dissipated hysteretic energyEs=average modulus of elasticity of steel reinforcementsfcm=concrete compressive strength on day of beam testingfct=concrete tensile strength on day of beam testingfst=average ultimate strength of steel reinforcementsfsy=average yield strength of steel reinforcementsh=height of coupling beamsI5, I10=toughness ductility measuresky=secant stiffness at yield pointk1, k2, . . ., ki=secant stiffnesses at end of each half-cyclel=distance between extreme supports of specimenlT=length of coupling beamsM=bending momentP1, P2=loads applied to specimensR1, R2=reactions of boundary supportss=thickness of coupling beamsV=shear forceVc=concrete volume of beamVu=ultimate shear force V1, V2, . . ., Vi=shear force values at end of each half-cyclec=concrete specific weightLi=accumulated plastic rotation in each half-cyclei=accumulated rotation ductility in each half-cycle1, 2, 3, 4=displacements of specimen during testnom=nominal diameter of steel barscum =cumulative ductilitys=static rotation ductilityl, v=main and transversal reinforcement ratiosB1, B2=rotations of two lateral blocksL1, L2=right and left rotations of coupling beam axisG=rotation of coupling beam axis respect to horizontal directionL=average rotation of coupling beam axisLy=rotation at yield pointLu=ultimate rotationL1, L2, . . ., Li =rotation values at end of each half-cycleREFERENCES1.ACICommittee318,BuildingCodeRequirementsforReinforcedConcrete(ACI318-77),AmericanConcreteInstitute,FarmingtonHills,Mich., 1977, 132 pp.2.Paulay,T.,DuctilityinSeismicDesign,StructuralEngineeringInternational, No. 1, Jan. 1992, pp. 19-22.3.Eurocode8,DesignProvisionsforEarthquakeResistanceofStruc-tures, ENV 1998-1-1, 1-2, 1-3, 1994.4.Paulay,T.,TheDesignofDuctileReinforcedConcreteStructuralWalls for Earthquake Resistance, Earthquake Spectra, V. 2, No. 4, 1986,pp. 783-823.5.Paulay,T.,andSanthakumar,A.R.,DuctileBehaviorofCoupledShear Walls, Journal of the Structural Division, ASCE, V. 102, No. 1, Jan.1976, pp. 93-108.6.Mahin,S.A.,andBertero,V.V.,NonlinearSeismicResponseofaCoupled Wall System, Journal of the Structural Division, ASCE, V. 102,No. 9, Sept. 1976, pp. 1759-1780.7. Paulay, T., Simulated Seismic Loading of Spandrel Beams, Journalof the Structural Division, ASCE, V. 97, No. 9, Sept. 1971, pp. 2407-2419.8. Paulay, T., and Binney, J. R., Diagonally Reinforced Coupling Beamsof Shear Walls, Shear in Reinforced Concrete, SP-42, American ConcreteInstitute, Farmington Hills, Mich., 1974, pp. 579-598. 9.Dajun,D.;Zhengliang,C.;Shuangyi,Z.;Zhibing,Q.;Aiqun,L.;Xiaodong,X.;Jiansheng,S.;Hang,D.;Cheng,G.;andXingrong,T.,ExperimentalInvestigationonHigh-DuctilityReinforcedConcreteShear-Walls,ScuoladiSpecializzazioneinCostruzioniinC.A.,FratelliPresenti, Politecnico di Milano, V. 13, 1992, pp. 395-418.10. Tegos, I. A., and Penelis, G. Gr., Seismic Resistance of Short Col-umns and Coupling Beams Reinforced with Inclined Bars, ACI StructuralJournal, V. 85, No. 1, Jan.-Feb. 1988, pp. 82-88.11. Eurocode 2, Design of Concrete Structures, Part 1-1: General Rulesand Rules for Buildings, ENV 1992-1-1, 1992.