SEISMIC PERFORMANCES OF REINFORCED CONCRETE FRAMES UNDER LOW INTENSITY EARTHQUAKE EFFECTS

Transcription

1 th World Conerene on Earthquake Engineering Vanouver, B.C., Canada August -, 00 Paper No. 0 SEISMIC PERFORMANCES OF REINFORCED CONCRETE FRAMES UNDER LOW INTENSITY EARTHQUAKE EFFECTS Bing Li and Tso-Chien Pan SUMMARY In a region o low to moderate seismi risk and low wind speed, suh as Singapore and Malaysia, buildings with relatively weak lateral strutural resisting system are likely to represent a large portion o the building inventory. Many buildings appear to be o sot story strutures. Although ground motions, due to long distane earthquakes entred in Sumatra, have ourred in Singapore and Malaysia there has been no reord o earthquake damage in this region. However the reinored onrete design ode, BS 80 [], used in Singapore and Malaysia does not speiy any requirement or seismi design or detailing o reinored onrete strutures. The main objetive o this paper is to strengthen the need to look into the seismi perormane o some typial existing and prospetive reinored onrete rame strutures designed to BS 80 [] in Singapore under low seismi loading. The perormane o the strutures is heked through a non-linear dynami analysis. INTRODUCTION Reinored onrete rame strutures are very ommon in a region o low to moderate seismiity, and are the predominant strutural system in Singapore and Malaysia. They are usually quite sti in one diretion and lexible in the other. Buildings in this region are usually designed without onsideration o seismi loading. The lak o seismi onsiderations resulted in non-seismi reinoring details that are in sharp ontrast to those used in modern seismi design. Thereore, it is o great onern that the strength, dutility, and energy dissipation apaity o these rame strutures may not be adequate to sustain earthquake-indued loads due to the lak o reinorement details in this type o strutures. This paper presents results rom the seismi assessment o a six-storey reinored onrete moment resisting rame, whih was designed based on BS 80 [], and the test results o the interior joints. An analytial and an experimental investigative program was undertaken to determine whether the strutures, as built in Singapore, ould omply with the stated (perormane) design objetives. The experimental program tested the post-yield behavior o a representative subassembly o the rame designed to assess not only the post-yield behavioral harateristis o the subassembly bur also the ability o nonompliant (rom a ode perspetive) aspets o the rame to perorm aeptably when the rame was subjeted to signiiant postyield story drits. Assoiate Proessor, Nanyang Tehnologial University, Singapore 9798 Proessor, Nanyang Tehnologial University, Singapore 9798

2 EXPERIMENTAL PROGRAM A subassembly (beam-olumn) test program was developed to assess the signiiane o the pereived onerns on the post-yield behavior o buildings designed to BS 80 []. Two ull-sale models o the prototypial subassembly were onstruted and tested by Li et al [, ] in the heavy struture laboratory o the Nanyang Tehnologial University. The details o the beam-olumn speimens are shown in Figure. To ensure that ritial stresses would be the same in the both the prototype and model subassembly, model reinorement was sized so that beam shear stress, bond stress in the lap splie. And joint shear stress levels would be the same. The beam-olumn joints were tested subjeted to quasi-stati load reversals that simulated earthquake loading. Figure Beam-olumn joints detail Speimen behaviors The measured horizontal story shear ore versus horizontal displaement hysteresis loop o Speimen A is shown in Figure. In addition, the theoretial ideal story horizontal load strength P i when the beam plasti hinges were developed and the theoretial stiness K theoretial are also shown in Figure. The theoretial initial stiness K theoretial o Speimen A is.7 kn/mm, assuming that the eetive moment o inertia o the beams and the olumns is 0.I g, and that the deormation due to joint shear distortion ontributes 0% o the total horizontal displaement, where I g is the moment o inertia based on the unraked gross onrete area. In the loading to ±0.P i, lexural raks were initiated in both the olumns and the beams. No rak was observed within the joint ore region. In the loading to ±0.7P i, diagonal raks were initiated within the joint ore region. In the olumn side ae, a ew lexural raks were initiated aompanied by a ew diagonal raks. Some pinhing was observed in the hysteresis loop (see Figure ). In the loading to a displaement dutility ator o, diagonal tension raks in the joint ore region extended and the number o those raks inreased rapidly. At this stage, more pinhing o the hysteresis loop was observed due to the ormation o the diagonal tension raks within the joint ore region, and the ormation o splitting raks along the olumn longitudinal bars, plus bond deterioration

3 along the beam bars and the olumn bars. In the irst positive yle o loading to a dutility ator o, the maximum horizontal load strength o. kn, whih was equal to the ideal story horizontal load strength o Speimen A, was reahed at a orresponding story drit angle o about %. A maximum nominal horizontal shear stress in the joint ore o 0.8 or 0. was obtained in the irst positive yle o loading to a dutility ator o, where is the measured ompressive ylinder strength o onrete. Within the joint ore region, a ew diagonal raks opened widely. Bond splitting raks along the olumn longitudinal bars extended opened wide and onneted with the joint diagonal tension raks. In the negative loading yle, the measured maximum horizontal load strength is 9. kn, whih did not reah the ideal load strength o Speimen A. The hysteresis loops were signiiantly pinhed due to severe bond deterioration along the beam and olumn bars and the joint diagonal tension raking. In the seond positive yle o loading to a dutility ator o, severe strength and stiness degradation due to joint diagonal tension raking and bond deterioration along both the beam bars and olumn bars were observed. The measured maximum horizontal load strength was kn, whih is only equal to 7% o that measured in the irst positive yle. Story shear ore (kn) Story shear ore versus Storey drit -.0% -.0% 0.0%.0% Speimen A Pi=. kn Theoretia l st i ne ss µ = µ = -0 Pi=-.kN -00 µ = -0 µ = Horizontal displaement (mm) Story shear ore (kn) Story shear ore versus Storey drit -.0% -.0% 0.0 %.0 % Speimen A Pi = 0. kn Theoretial stiness µ = µ = µ = Pi = -0. kn µ = Horizontal displaement (mm) Figure The story shear ore versus the horizontal displaement relationship or Speimens A and A The measured horizontal story shear ore versus the horizontal displaement hysteresis loop o Speimen A is shown in Figure. Also, the theoretial ideal story horizontal load strength P i when the beam plasti hinges were developed and the theoretial stiness K theoretial are also shown. Firstly, in the loading to ±0.P i, lexural raks were initiated in both the olumns and the beams. A ew diagonal tension raks were observed within the joint ore region, and there was a small joint shear distortion and expansion observed. In the loading to ±0.7P i, a large number o diagonal raks were initiated within the joint ore region, while the joint shear distortion and expansion inreased rapidly. No obvious pinhing was observed in the hysteresis loop. As the loading reahed a displaement dutility ator o, with the opening o diagonal tension raks in the joint ore region, the joint distortion and expansion ontinued to inrease. The lexural raks in the beams opened widely and were aompanied by wide lexure-shear raks, and the onrete in the beam ompression zone began to be rushed. In the olumn, although no lexural raks appeared, there were a ew diagonal tension raks extending rom the joint ore region

4 into mainly the upper part o the olumn. In the irst yle with a dutility ator o, the maximum horizontal load strength o. kn, whih was larger than the ideal story horizontal load strength o Speimen A, was reahed at a orresponding story drit angle o.0%. At this stage, some pinhing o the hysteresis loop was observed due to the ormation o the diagonal tension raks within the joint ore region and the large lexural raks at the olumn aes. In the loading to a dutility ator o, within the joint ore region diagonal raks inreased rapidly, and opened wider. These raks inally onneted with the bond splitting raks along the olumn main bars, and at the same time the joint distortion and expansion inreased rapidly. The maximum joint distortion observed was 0.9%, and the maximum joint expansion was. mm. Beam lexural raks opened wider espeially at the olumn aes, and in the beam ompression zones muh more onrete was rushed and spalled. A maximum nominal horizontal shear stress in the joint ore o 0. or 0. was obtained. In the irst positive yle o loading to dutility ator o, the maximum horizontal load strength o 7. kn, whih was greater than the ideal story horizontal load strength o Speimen A, was reahed at a orresponding story drit angle o about.0%. Bond deterioration was obvious in the bottom beam bars where bond stresses dereased rapidly in the loading to dutility ator o. No bond deterioration was observed in the olumns and this was due mainly to the low stress in the olumn main bars. In the seond yle o loading to a dutility ator o, the maximum horizontal load strength measured in the positive loading yle and negative loading yle were 8.0 kn and. kn, respetively. These were about 7% and 78% o those measured in the irst loading yle. RESPONSE OF A SIX-STOREY BUILDINGS A typial six-storey reinored onrete moment resisting rame was onsidered or the present study. The elevation and the plan o rame are shown in Figure. The eets o seismi ation were onsidered in both the strong and weak diretions. The rame was designed or ombined gravity and lateral loads in aordane with the Singapore Loading Code, and their strutural members were proportioned and detailed aording to BS 80 []. The typial beam-olumn joints in the strong and weak diretions o the rame are already shown in Figure, representing the joint regions in the rame. Torsional and P-delta eets were not onsidered in the design. In non-linear dynami analysis two ground motions are seleted rom the earthquake database system as the input ground motions or the rame. Sine in Singapore, not until reently has attention been drawn to the saety o buildings during an earthquake beause o the inreased numbers o tremors generated by the long distane Sumatra earthquake. Up to now, ew ground motions have been reorded; thereore two extensively used earthquake time-history reords have been seleted or this study. One is the 90 El-Centro reord (NS omponent) and the other is the 977 Buharest reord. Aording to the studies onduted by Pan and Sun [], 0.g may be taken as a reditable peak ground aeleration, whih may our in Singapore due to the long distane Sumatra earthquake, thus the seleted two reords were both saled down to 0.g to represent the reditable earthquake attaking in Singapore. A omputer program ommonly reerred to as RUAUMOKO [] was used to study the analytial behavior o this building in the inelasti range. This program is being developed at the University o Canterbury, New Zealand, and represents the state o the art in this area. Many hysterisis models have been proposed in the previous studies or reinored onrete strutural members. However, the hoie o a partiularly hysterisis model in the analysis depends on the atual design and detailing o the members. In this study, it is assumed that insuiient transverse reinorement has been provided or the strutural members so that the stiness and strength deterioration due to shear or bond loss are more signiiant. A bi-linear hysteresis model is thereore used to express the moment-urvature hysteresis loops o the olumns. For beams, the pinhing model is hosen. Fators ontrolling the unloading and reloading stiness were

5 seleted to make the hysteresis loop as thin as possible. The damping was represented using the Rayleigh damping model, and it is expressed as a linear ombination o the mass and stiness matries. The ombination oeiients are seleted to give % o ritial damping in the irst two modes o vibration. The lumped nodal weights are determined assuming the average weight o loor to be KPa. Figure Reinored onrete rame RESULTS OF NON-LINEAR DYNAMIC ANALYSIS Maximum Interstorey Drit Ratio Interstorey drit ratio is onsidered as the primary global perormane parameter. Figure shows the maximum interstorey drit ratios observed in the strong and weak diretion o the rame, respetively. A relationship between the desired overall seismi perormane and the maximum transient drit speiied by SEAOC-99 [] is also inorporated into the igure to better understanding o the building perormane. It an be seen that or the strong diretion rame only moderate damage may be aused while or the weak diretion rame severe damage may our.

6 Rotational Dutility Demands at Member Levels Rotational dutility is deined by the ratio o the maximum rotation at the end o a member to the yield rotation as ollows: θ θ plasti µ θ = max = + () θ θ y y Negligible Light Moderate Severe Negligible Light Moderate Severe Storey level Buharest 977 El-Centro 90 Storey level Buharest 977 El-Centro Storey drit ratio (%) Storey drit ratio (%) Figure Inter-storey drit Figure Rotational dutility

7 In whih θ max and θ represent the maximum and the yield rotations at the end o a member. y Beause the inelasti lexural deormation o a beam is assumed to be onentrated at the ends, the plasti rotations our at the plasti hinges. The rotational dutility demands at eah end o the elements are evaluated or two diretion rames and two input ground aelerations desribed. Figure shows the distributions o rotational dutility demands in two diretion rames, respetively. It an be observed that the inelasti deormations o rames subjeted to the saled Buharest earthquake are widespread at all storey levels. However, in the ase o the saled El-Centro earthquake, it an be seen, that the dutility demand was muh less. Joints behavior Beam-olumn joints are oten the weakest links in a strutural system. For lightly reinored beams, or with olumns with high axial ore levels, joint raking may not develop and the joint ailure may be judged aording to the prinipal tension stress [7] as ollows: dt = + v jh 0.9 () For beam-olumn joints with high shear stress levels, premature ailure o diagonal ompression strut tends to our and the joint ailure may be judged aording to the prinipal ompression stress (Comite Euro-International 997 [8]) as ollows: d = + + v jh 0. () For joints with prinipal tension stress greater than 0.9 and prinipal ompression stress less than 0., ailure may due to joint shear, bond slip, and lexural dutility (Comite Euro-International 997). The maximum prinipal tension and ompression stresses o interior and exterior joints in two diretion rames are shown in Figure. Aording to the results, in the strong diretion rame exterior joints are ritial, and joint ailure was predited to our. In the weak diretion rame both interior and exterior joints are ritial, the prinipal tension stresses o whih are ar beyond the ailure line. Base shear The push-over analysis shows that, or the strong diretion, the struture ollapsed when the base shear attained 0.9% o the total weight o the rame, where the global dutility o the rame was.. While or the weak diretion, the struture ollapsed when the base shear attained.% o the total weight o the rame, where the global dutility o the rame was.9.

8 El -Centro 90 Buharest Joint tensi on stress level ( ) Joint tensi on stress l evel El-Centro 90 Buharest Joint ompression stress level El -Centro 90 ( ) (a) strong diretion Joint ompression stress level Buharest Joint tension stress level ( ) Joint tension stress level El-Centro 90 Buharest Joint ompre s sion stress level ( ) (b) weak diretion Figure Joint stresses Joi nt ompression stress level

9 Base shear (kn) Base shear (kn) Global dutility Global dutility (a) Strong diretion Figure 7 Base shear (b) Weak diretion DISCUSSION Type text immediately below subheadings For the strong diretion rame the maximum joint shear stresses in interior joints obtained rom time history analysis were. MPa (0. or 0.08 ) and.0 MPa (0. or 0.0 ) or the El-Centro and Buharest earthquake, respetively. For the weak diretion rame the maximum joint shear stresses in interior joints obtained rom time history analysis were.7 MPa (0.8 or 0. ) and.7 MPa (0.9 or 0.7 ) or the El-Centro and the Buharest earthquake, respetively. So ompared to the experimental data, during a maximum redible earthquake, whih may our in Singapore, or interior beam-wide olumn joints loated in the lower part o the weak diretion wide-olumn rame, beam bar bond slip and joint shear ailure may our. The joint may undergo severe strength degradation where maximum inter-storey drit ratio is attained. While or the joints in strong diretion, there will not be any damage. CONCLUSIONS Based on the experimental results on two prototypial beam-olumn joint subassemblies, the joints experiened signiiant strength and stiness degradation due to yli loading. Suh behavior ould have deleterious eets on the drit o moment-resisting rames designed aording to BS 80. Nonlinear dynami analysis has shown that a low intensity earthquake might ause the rames to generate a maximum inter-storey drit ratio o about %. The ritial ailure mehanism o the rame is a hybrid mehanism inluding beam and olumn side-sway mehanism and beam-olumn joint ailure. A more ritial aspet in shear was ound in the beam-olumn joints. Relatively large joint shear input during the low to moderate earthquakes indiate that the joints o the struture ould suer severe diagonal tension raking and the strength o the struture is likely to be governed by the joint shear ailure mode. REFERENCES. British Standards, Strutural Use o Conrete BS 80, Part, ode o pratie or design and onstrution; 997.

10 . Li, B., Wu, Y.M., and Pan, T.C., Seismi Behavior o Non-seismially Detailed Interior Beam- Wide Column Joints Part : Experimental Results and Observed Behavior, ACI Strutural Journal, V. 99, No., Nov-De 00.. Li, B., Wu, Y.M., and Pan, T.C., Seismi Behavior o Non-seismially Detailed Interior Beam- Wide Column Joints Part : Theoretial Comparisons and Analytial Studies, ACI Strutural Journal, V. 00, No., Jan-Feb 00.. Pan, T.C., and Sun, J. (99), Historial Earthquake Felt in Singapore, Bull. Seism. So. Amerian, 8, pp Carr, A.J., RUAUMOKO- Inelasti Dynami Analysis Program. The University o Canterbury- Department o Civil Engineering, Christhurh, New Zealand, 99.. SEAOC, Vision 000 Perormane Based on Seismi Engineering o Buildings. The Strutural Engineers Assoiation o Caliornia, San Franiso, USA, Priestley, M.J.N and Calvi, G.M. (99), "Toward a Capaity Design Assessment Proedure or Reinored Conrete Frames", Earthquake Spetra, EERI, Vol. 9, No., pp Comite Euro-International du Beton: Seismi Design o Reinored Conrete Strutures or Controlled Inelasti Response, Thomas Telord Ltd, 997, pp.0-.