HYSTERESIS BEHAVIOUR OF REINFORCED CONCRETE BRACES AND THE SEISMIC RESPONSE OF BRACED FRAMES

Abstract

Reinforced concrete moment resisting and shear wall frames are widely used structural systems in seismically active regions. During earthquakes shear walls may develop severe diagonal cracks which are quite difficult to repair. There is a need to develop an alternative structural system. In the past few years the concept of bracing members has been extended to strengthen reinforced concrete frames. So far no study has been carried out on the cyclic behaviour of reinforced concrete bracing members. Braced concrete frames can be used with different bracing pattern and member . proportions. The strength and stiffness of bracing members could significantly affect the seismic response. It is not yet clear how to design such braced frames. The main objectives of this study were : (1) to determine experimentally, the hysteresis behaviour of axially loaded reinforced concrete bracing members. Variation in stiffness with cyclic loading and energy dissipation capacity were the parameters of main interest; (2) to develop a suitable hysteresis model for reinforced concrete bracing members; and (3) to develop an under standing of structural behaviour of reinforced concrete braced frames having different bracing patterns and member proportions. A force-controlled electro-hydraulic compression-only test cylinder plant was used to test reinforced concrete bracing members in cyclic tension and compression. Achain-pulley arrangement was devised to convert compre ssion into tension. Eight reinforced concrete bracing members were selected whose cross-section was either 150mm x 200mm or 100mm x 200mm. The (iii) length of each specimen was 2m, and length to least lateral dimension ratio was either 20 and 30 that is, slenderness ratio was 35 and 55. The longitudinal steel consisted of four bars of either 8 mm, 10 mm or 12 mm diameter. The lateral reinforcement consisted of 8 mm HYSD bar stirrups placed at a spacing of 50 mm c/c to 150 mm c/c. It was noticed in the hysteresis loops that there was considerably more strength and stiffness in compression than in tension. There was gradual degradation in stiffness with increase in number of cycles and load levels. The stiffness in tension loading and unloading cycles varied between 13% and 140% of the initial tension stiffness. The stiffness in compression loading and unloading cycles varied between 30% and 400% of the initial tension stiffness. The hysteresis loops exhibited softening behaviour at smaller loads and hardening behaviour at larger loads. The ratio of overall average stiffness in compression cycles to that in tension cycles was found to be about 2. The equivalent viscous damping coefficient varied between 1.8% and 4.7%. More energy was dissipated in tension cycles than in compression cycles on account of cracking. A multilinear hysteresis model is proposed for reinforced concrete bracing members. However, it was not possible to define its control points accurately based on the present test program. Hence another simple traingular hysteresis model is proposed. It retains three main characteristics of the reinforced concrete bracings namely, the different elastic stiffnesses in tension and compression, the stiffness degradation and the pinching-in effects. A computer subprogram is written to incorporate this model for use with the DRAIN-2D computer program. (iv) The structure selected for this study represents a 6-storey two-bay interior frame of a symmetrical building with bracing introduced in one bay. The frames were designed as having dual line of defence using the weak girder strong column philosophy. Three bracing patterns were selected con centric X, concentric K and eccentric K and designated as XI, Kl and EK1. The size of bracing members of frames XI and Kl were changed to get frames with stronger bracing members having slenderness ratio of the order of 50 and these frames were designed as X2 and K2; and to get frames with weaker bracing members having slenderness ratio more than 120 which were designated as X3 and K3. Takeda's stiffness degrading bi-linear model was used to simulate the hysteresis behaviour of concrete girders. Elasto-plastic moment-rotation model with axial force-moment interaction curve was used to simulate the inelastic behaviour of concrete columns. The triangular hysteresis model developed in this study was used to represent the cyclic behaviour of concrete bracing members. The inelastic analysis was performed by using the DRAIN-2D computer program under the May 1940 El Centro earthquake and an artificially generated earthquake. The response of these frames was analyzed for the following parameters : influence of tension stiffness of bracing members; braced concrete frame vs unbraced concrete frames; slenderness ratio of bracing members and the characteristics of earthquake motions. The seismic response of concentric X, as well as concentric and eccen tric K braced frames with bracing member having slenderness ratio of the order of 80 was found satisfactory under both the earthquakes. It may be mentioned that the inelasticity in the girders of the concentric K and (v) eccentric K braced frames occurred at the column faces and not at the connections of the bracing members with the girders. Hence, the ductility requirements in the girders and bracing members were reasonable. Very strong bracing members having slenderness ratio of the order of 50 resulted in buckling of columns of the braced frames. Hence, such bracing members are undesirable. Very slender bracing members having slenderness ratio more than 120 resulted in excessive floor displacements and member ductilities. Such bracing members may themselves collapse in the event of a severe earthquake. Thus, the concept of using bracing members of moderate slender ness only in a reinforced concrete frame appears to be promising. There is a need to carry out more tests on large size bracing members and braced frame models using displacement controlled tension-compression actuators and develop a more accurate hysteresis model. This model can be used to verify the analytical findings of this research program.