Predictions of the Inelastic Seismic Response of Concrete Structures Including Shear Deformations and Anchorage Slip

The behavior of reinforced concrete frame structures under seismic loads is analyzed using a number of mathematical models of increasing levels of complexity. The first model, represented by the program DRAIN-2DX, idealizes plasticity in members as concentrated at the member ends and governed by a simple elastoplastic hysteretic rule with no strength or stiffness degradation. The second model, implemented in program IDARC2D, permits yield penetration into the member and accounts for stiffness and strength degradation as well as pinching of the hysteresis loops. The last model is based on a fiber formulation and is implemented in the program FIBERC.

Several large-scale member and frame laboratory specimens are analyzed with the three models under static and dynamic loading. It was found that in general the beamcolumn element of program DRAIN-2DX does not reproduce well the member behavior, particularly in terms of stiffness and energy dissipation. In contrast, program IDARC2D generally provides a good estimate of member and frame behavior, including pinching of the hysteresis loops, thanks to a set of parameters used to adjust the predicted response. Finally, program FIBERC simulates closely the member response when it is governed primarily by flexure and axial forces. It does not, however, provide good predictions when pinching of the hysteresis loops occurs.

To extend and improve the performance of the fiber element model implemented in program FIBERC, a simple model that accounts for the effects of shear deformations is added. Such a model is based on a truss analogy of the cracked structural members. A series of runs using some of the members and frames studied previously is performed to assess the validity of the model as well as its influence on the predicted response.

Additionally, a model that estimates the effects of anchorage slip on member deformation and stiffness is introduced into the program. Again, a series of runs is performed on the same specimens to evaluate the effectiveness of the model as well as the impact on the computed response.