You are on page 1of 11
Cmte Title no. 92-825 TECHNICAL PAPER Cyclic Behavior of Reinforced Concrete Cantilever Walls, Part I: Experimental Results & 9, by Kypros Pilakoutas and Amr Elnashai The results of an experimental research program concemed with the response of reinforced concrete walls under eylic lading are presented in this paper Information pertinent 10 the experimental work, including description ofthe test rig, instrumentation ofthe specimens, control proce dure of the tests, details of model manufacture, properties of materials used, and choice ofthe loading regime is first given. The description ofthe experiments is presented subsequently, together with details ofthe loading procedure and the crack development. Only brief comments on the results ‘obvained are given in this paper. Further discussion of the experimental results and comparisons benween experimental and analytical work are given in a companion paper. Keywords: ductility: earthquake-resstamt structures; loads (forces): ret forced concrete; sifness; structural design: ets: walls Field investigations following destructive earthquakes in- dicate that many structural failures and the majority of non- structural damage sustained by reinforced concrete (RC) structures are primarily due to excessive story displace- ments, Large deformations not only impose excessive ductil- ity demands on critical RC members but also generate second-order forces normally unaccounted for in design. In- comporating RC walls as lateral force-resisting elements in buildings considerably reduces story drift. Due to their high bending stiffness, RC walls attract a large proportion of the seismic forces at the critical lower stories and, as a result, re- duce the overall demand on other structural components. Present code requirements for earthquake-resistant design generally underestimate the ductility of RC walls. This is manifested in the reduced value of the behavior factors, “R” (ACI standard) or “” (Eurocode 8), which result in a higher design base shear coefficient imposed on buildings with such walls. The reason for this undue conservatism may be attrib- uted to the desire to avoid brittle modes of failure in walls, observed following destructive earthquakes. Such failures are associated with either inadequate shear capacity or du tility supply and result in the loss of the energy dissipation potential To address the preceding issues and quantify the true due: tility supply and energy absorption potential of RC walls, the ACI Structural Journal / May-June 1995 (ae vr ee Fig, I—Elevation of test-rig assembly program described in this paper was undertaken, This com- prised testing of nine walls under severe cyclic loading re- gimes, as well as analytical and design-oriented studie Results of six cyclic experiments undertaken at scale 1:2. are presented in this paper, while a full discussion of all ex- perimental data and correlation with analytical expressions are presented in References I and 2. RESEARCH SIGNIFICANCE ‘The response of reinforced concrete walls subjected to cy- clic loading regime is investigated experimentally, and some of the important behavior aspects are reported. The shear stiffness of specimens is shown to be unaffected by shear re- inforcement in excess of the one required for shear strength purposes. Dilation of the concrete member is shown to be significant during post-yield load cycling, and this can affect the strength and stiffness characteristics. EXPERIMENTAL METHODOLOGY ‘The six walls, hereafter referred to as SW4-SW9, were tested as isolated cantilever walls in a specially designed and ACI Matra Journal V.92, No.3 May-June 1998. Received bee, 193, and reviewed ter insti pubication pbc. Copy sight 1985 American Concrete ltt. All igi saved, along the making {Fee tls Pio oe om he np ores, Prem e “isi lb published nthe March: Apel 19964CY Sintra! Joura recived 27t [jpros Piakowas i a ectarer nthe department of Ci and Sacer Engineer Ing, University of Sheed UK. He received hs BSc(Eng) in Civil Engincering and PRD in Earthquake Engincerng fom Imperial Collegeof Scene, Tecnology ad ‘Medicine in 1986 and 1990, especivey. Hs racarch is primary concerned wih Sear and ductility design, repair and nonlinear behavior of reinforced concrete structures and alternative per of reinforcement. ‘Amr Blnash is Profesor of Earthquake Engineering and Head of the Engineering Seionologh and Earthquake Engineering Section a Inperal College, Landon UK. He graduated from the Coil Engineering Department at Cairo, Egy in 1977 and ‘obtined his MSc and PRD from Imperial College in 1980 ad 1986, respectively He ‘nus employed asa senor engineer i technology development at Wimpey Ofshore [Lad UK. ovo years before rerning to InperialCollege in 1986 as lecturer in frthquae eninering, Hes cure a ember of th Technical Coordinator and {Nationa Technical Contac forthe UK. onthe European Conveton of Constr: tional Stel (ECCS) and the Comité ary Interationl du Beton (CEB) seamic design subcommines. He teresa the UR representative onthe Earocode 8 (eimic ‘design cade cammie. He is alo a isting profesor atthe University of Southern California. His inerest 1s expermental and anastial response of statues to ‘anhquake loading, and he has underakn port earthquake fd investigation of Several damaging etrthes in Europe, Africa and the US DispLaceamr CONTROL evel pontine Fig. 2—Schematic representation of test-rig arrangement ‘and instrumentation used for RC wall tests constructed test rig. The setup of the testing arrangement is shown in Fig. 1, while a schematic representation is shown in Fig. 2 ‘The horizontal load was applied through a top beam de- signed to spread the load over the wall panel. The boundary conditions atthe top, as shown in Fig. 2, assume a horizontal load at a constant height. To eliminate friction between the Jack and the specimen, load was applied through solid cireu- lar rollers. The rollers were encased in a stiff collar frame with one end being adjustable for alignment purposes. The loading collar frame was free with respect to the in-plane horizontal movement, but was restrained from displacing vertically and out of plane by flat rollers reacting against two standard laboratory assembly portal frames, as shown in Fig, 2. In previous tests on reinforced concrete walls," this detail ‘was not used and, hence, lateral frictional forces at the point of application reduced the applied moment.S Model manufacture and materials ‘The model dimensions are shown in Fig. 2. The models were cast monolithically in the horizontal direction, A con- crete mix with a maximum aggregate size of 10 mm (0.39 272 Table 1—Average concrete compressive strength at the day of testing Cube | Cylinder = Specimen | Day ot | srensth sucngthf,, | Peak stn, code ‘eins mm? psi} Némm?, psi} €c22%10" Swe @ ws 369 2083 Gis) | 651) swe w 415 318 ae ss) | asi, Swe 7 96 3a6 | aur is) | 6598) Sw7 74 152 220 1950 (6555)_| Getty _ Swe 76 sa7 158 1960 78s) | eid, wo B 337 389 2300 iss) | Goi, in.) and cement content of about 300 kg/m? (18.7 Ib-mass! f¢)) was used in all walls. At least three of the cubes were tested at the end of each experiment. The strengths obtained are shown in Table 1 ‘The reinforcement characteristics and details are given in Table 2 and are shown in Fig. 3 and 4. Walls were designed in three pairs; each pair having equal flexural reinforcement but different shear reinforcement. In all cases, an attempt was made to concentrate the reinforcement in the boundary ‘elements to maximize the flexural capacity for the given area and yield stress of reinforcement. The web reinforcement ‘was kept nominal. Shear reinforcement was varied in each pair of walls to investigate the effect of various degrees of safety margins in shear. The confinement of the boundary el- cements varied as a consequence of the variation in shear re- inforcement. Instrumentation and control ‘A severe cyclic loading regime was used, at a very slow rate. At each new displacement level, two full cycles were imposed, the displacement level having been incremented by 2 mm (0.079 in.) in both directions. All specimens were test- ced to failure, Strain in the steel reinforcement was measured at a number of locations. All the force-versus-strain results obtained are given in References 1 and 2. The cracks formed ‘were marked during every cycle at every maximum displace- ‘ment level (MDL - mm). EXPERIMENTAL RESULTS Model SW4-General observations First cracking was observed at about 1 mm of displace- ment. Cracks propagated from the wall boundaries toward the centers and from the bottom upward. Near the bound- aries, cracks were nearly horizontal, while further away they were inclined to the horizontal. The inclination increased along the wall height. At the boundaries, the density of the cracks increased, while in the web the number of main cracks was limited to about three to four on each side. The propagation of these main web cracks toward the opposite side was more extensive in the lower part of the wall, as shown in Fig. 5. The same pattern of cracking continued un- til yield of the flexural reinforcement. First-yield occurred Just before MDL-6, and the crack pattem of the first post- ACI Structural Journal / May-June 1995 ‘Table 2—Reinforcement details Flexural bound Shear Boundary ay reinforce- | Flexural web | reinforcement | Boundary | clement A, Wallcode fment Agr percen| reinforcement | Ay. percent | element, in. percent _ ‘Ash, percent ‘swe 28 oar 039 110 O78 - 433) _ ws 302 oa oat oO on 236) we 2 oat O31 110 017 433) sw 302 07 039 © O78 236) ws 298 O31 oat 110 oa 433) _ sw 293 oat o3t 110 0s 433) ‘Stress Nimm® Fig, 3—Reinforcement stress-strain characteristics yield displacement, MDL-8, is presented in Fig. 5. Follow- ing yield, cracking became denser, and several boundary cracks joined with web cracks. The number of web cracks in- creased and so did their apparent inclination to the horizon- tal; the latter was due to the joining of web cracks with cracks originating higher up in the boundary elements. By MDL-16, the lower web cracks had opened up considerably ‘more than all the others. Vertical cracking also appeared near the bottom of the wall at the boundaries, at approximately the position of the main reinforcement. This indicates that there was spalling of the conerete cover at these locations. With increasing MDL, the concrete in the lower part of the wall began to show signs of deterioration, At MDL-22 and just before failure, the concrete, confined by the bottom two hoops in both boundary elements, was spalling considerably. Failure occurred in this area, during cycles at MDL-24, by ‘crushing of core concrete (Fig. 5). The load-versus-top di placement for Wall SW4 is shown in Fig. 6, Model SW5-General observations Wall SWS has more of the main reinforcement concentrat- dat the boundaries; hence, it was expected to sustain a high- cr flexural load, if shear failure did not ensue. Cracking at MDL-2 is shown in Fig. 7. Cracks propagated from the wall boundaries toward the center and from the bot- tom upward, reaching up to three-quarters of the height of the wall, In the lower half, the frequency of cracks near the boundary was higher than in the web. However, the inclina- tion of these cracks was steeper than that for Wall SW4, By MDL-8, the density of cracking in the boundaries increased considerably. The main web cracks started joining with cracks originating higher up in the boundary, and hence their ACI Structural Journal / May-June 1995 apparent inclination increased. MDL-10 (0.394 in.) was achieved in one direction at a load of 117.3 KN (26.4 kip- force). Yielding of the horizontal web (shear) reinforcement “occurred just prior to the attainment of yield in the main flex- ural reinforcement. On attempting to impose MDL-10 in the reverse direction, abrupt failure occurred at a load of about 110 kN (24.7 kip-force). At this stage, two of the main web cracks opened up significantly, as shown in Fig. 7. However, the crack that caused failure was the lower one that penetrat- ed through the compressive area. Following failure in one direction of loading, it was decid- ed to avoid cycling the wall twice at the same MDL, as shown in Fig. 8, and failure in the other direction occurred immediately after exceeding MDL-10 at a load of about 108 KN (24.3 kip-force). The last MDL imposed in both direc- tions was MDL-12. To demonstrate the effect of cycling on ‘a wall that failed in shear, displacements were increased by 2 mm (0.079 in.) in each direction. After failure, several ‘hoops and stirrups opened up and considerable opening was noticed in the middle sections of the wall. Additionally, it was observed that the wall was displacing in a rigid body mode above the main cracks. The considerable degradation of the wall at MDL-26 is shown in Fig. 8. The load-versus- top displacement diagram for Wall SWS is shown in Fig. 8. Mode! SW6-General observations Wall SW6 had identical flexural reinforcement to Wall ‘SW4 but considerably less horizontal reinforcement that was identical to that used in Wall SWS. ‘The pattern of cracking at MDL-2 is shown in Fig. 9. De- spite the initial slightly higher stiffness, cracking was as ex- tensive as in SW4. Yielding of the main flexural reinforcement started just before MDL-6, After yield, the ex- tent of cracking in the web increased considerably, as shown in Fig. 9. The main web cracks started joining with cracks originating higher in the boundary; hence, their inclination ‘was greater than that observed in SW4. Following MDL-8, cracks were considerably wider than at the corresponding MDL in SW4. However, by MDL-16 (0.63 in.), the maxi- mum load achieved was 107 KN (24.1 kip-force), which was higher than the ultimate load for SW4, Diagonal cracking propagated from the web into the opposite boundary and seemed to have caused the translation of the neutral axis very 273 unex Suppo} pup susw2a0qdsip doi ¢ars 1IMMy sns1a4 pooy——9 “tq 6. MS 01 FMS ST powzoyopum foots PII = (Pa0y9p TSU, WH = LH ‘opis woud = St aur uy swoysuoUCL ood word = 33 aso PIOA THIOL, = AKL cstoe=aK) or. sg Us core sayy ar oro apap -dareer | STTIT sage SE9LIHS C ] sun umf Pun ‘$1 FTN Fm 10M 0 wionDd yD —F ig BH sass rap mouanofury—p Bq B ACI Structural Journal / May-June 1 274 MDL2 ] FAILURE MDi-0 POST-FAILURE, ‘MDL-24 Fig. 7—Crack pattern of Wall SWS at MDL-2, failure, and post-failure | MDL-2 MDL-8 FAILURE Fig. 9—Crack pattern of Wall SW6 at MDL-2, MDL-8, and failure close to the edge. Concrete spalling was also noted atthe bot- tom end of the wall. By MDL-18, the concrete deterioration started affecting the wall strength and, although the maximum load was achieved during forward loading, less load was resisted ‘when the displacement was reversed. Opening of a stirrup in the lower level was observed at this MDL, as well as pro- ACI Structural Journal / May-June 1995 gression of the cracks through the compressed area. Crush- ing of concrete was initiated just before achieving MDL-20 in the left-hand side. The loss of strength continued during subsequent cycles, and by MDL-22 the resisted load was be- low 75 percent of the ultimate load in the forward direction as well. In this direction, a wide crack crossing through the bottom boundary area was observed. MDL-22 (Fig. 9 and 275 (KN) MDL-2 MDL-8 Fig. 1l—Crack pattern of Wall SW7 at MDL-2, MDL-8, and failure 10) was the last displacement to be imposed, since the wall ‘was considered to have failed. The load-versus-top displace- ‘ment diagram for Wall SW6 is shown in Fig. 10. Model SW7-General observations ‘Wall SW7 had most of the main reinforcement concentrat- ed at the boundaries, similar to SWS, but had a higher shear reinforcement percentage, similar to that of Wall SW4, The expected flexural strength was higher than in all other spec- imens, Cracking for MDL-2 is shown in Fig. 11. Cracks initially propagated in a similar manner to Wall SWS. Cracking was observed up to three-quarters of the height of the wall, with ‘an increased intensity near the boundary relative to the web. By MDL-4, cracks propagated to the whole wall length. In the lower half of the wall, at the boundaries, the density of the cracks increased, while in the web the number of main cracks was limited to about four to five in each direction. The inclination of these main web cracks was less than or equal 10 45 deg, and in general less steep than for SWS. At MDL- 8, the extent of cracking at the boundaries increased consid- erably, as shown in Fig. 11. The main web cracks started joining with cracks originating higher in the boundary and increased their inclination, but not to the extent observed for SWS. Following MDL-8, cracking continued to become 276 denser at the boundaries with an average inclination of more than 45 deg, at variance to Walls SW4 and SW6, where the boundaries were wider. By MDL-14, the main web cracks were wider than before and seemed to have penetrated the compressive area. ‘The load capacity peaked at MDL-18 (0.71 in.), corre- sponding to a load of 127.3 KN (28.6 kip-force). This was followed by widening of the web cracks and some loss of strength at MDL-22, in the forward direction. However, fail- ure occurred at MDL-22 (0.866 in.) by fracturing of one of the 6-mm (0.24-in.) shear reinforcement bars on the reverse virgin cycle. A second bar snapped during the completion of this MDL, and the strength was reduced to less than half. The considerable degradation of the wall at MDL-22 is shown in Fig. 11 and the load-versus-top displacement diagram for Wall SW7 is shown in Fig. 12. Model SW8-General observations ‘Wall SW8 was designed with the flexural reinforcement distributed in three layers within the edge members. Shear reinforcement was distributed according to the “Shear Resis- tance Surface” method described in Reference 1. The amount of shear reinforcement was half of that required by Eurocode 2 and was distributed only over two-thirds of the wall height, ACI Structural Journal / May-June 1995 MDL-2 4 Fig. 13—Crack pattern of Wall SW8 at MDL-2, MDL-6, and failure Cracking at MDL-2 is shown in Fig. 13. Cracks initially propagated in a manner similar to Wall SW4, Cracking was ‘apparent upto just over half the height of the wall with an in- ‘creased density near the boundaries, corresponding to the hoop reinforcement spacing. By MDL~4, cracks propagated to the whole wall length. At MDL-6, a crack originating from the top comer propagated through the web at an angle slightly higher than 45 deg to meet the web crack below, as shown in Fig. 13. After MDL-6, cracking continued to be- come denser at the boundaries, but the number of main cracks in the web stabilized to about four to five. By MDL- 12, the same pattern of cracking continued, but the main web cracks were wider than befor. By MDL-18, spalling of the concrete on the inside of the boundary element indicated the degradation of the concrete at that location. At this stage, vertical cracks that appeared by MDL-14 at the bottom end main reinforcement level were also visible. Following MDL-18, the widening of the web cracks increased, and more spalling of concrete within the ‘web took place at the intersection of the main cracks. The ‘maximum strength of 95.3 KN (21.4 kip-force) was achieved at MDL-22 (0,87 in.). Loss of the concrete cover at the ex- treme bottom parts of the wall exposed the flexural rein- forcement by MDL-24. Due to the considerable degradation, ACI Structural Journal / May-June 1995 the wall lost some strength by the second cycle of MDL-24 and at MDL-26. The test was stopped at MDL-26 as soon as the load dropped below 75 percent of the ultimate capacity, toavoid excessive damage to the specimen. The load-versus- top displacement diagram for Wall SW8 is shown in Fig. 14, Model SW3-General observations ‘Wall SW9 was designed with the flexural reinforcement

You might also like