Numerical Simulation of Reinforced Concrete Columns Under Biaxial Cyclic Bending and Variable Axial Load

Abstract eng:
Numerical simulation of reinforced concrete structures is important in the assessment of nonlinear behavior of new and existing structures susceptible to earthquake loads. Columns are subjected in the general case to biaxial bending and varying axial loads during an earthquake and the interaction between these efforts has a strong influence in the nonlinear response, modifying the strength, stiffness and ductility of these elements. Several numerical models exists nowadays in literature for nonlinear dynamic analysis of beam-column elements that take into account the interaction between bending moments and axial load with different degrees of accuracy and robustness. The best way to test them is by contrast of numerical results with experimental data, but this is not an easy task mainly because laboratory tests involving cyclic biaxial bending with varying axial load are rare in literature. In this work a flexibility-based element with fiber discretization of the cross section on each integration point of the GaussLobato integration rule is used in order to reproduce part of an experimental campaign performed by other authors. EulerBernoulli’s hypothesis is used to calculate deformations of the fibers and material nonlinearity is reproduced by uniaxial laws at each fiber of the cross section. Modified versions of Mander’s constitutive law and Menegotto-Pinto’s law are used for concrete fibers and reinforcement steel fibers respectively. With this model the nonlinear behavior of four identical cantilever columns subjected to different paths of imposed horizontal displacements that lead to biaxial cyclic bending in conjunction with variable axial load is simulated numerically. Results of the hysteretic behavior obtained by the numerical model and the experimental test data are overlaid and compared. Comparison shows that the numerical model used in the simulation is able to reproduce the maximum strength and the dissymmetry caused by variations of the axial load, also the general behavior and interaction among cyclic moments and axial load is reproduced. Differences appear in the post-peak response especially softening is quicker in the simulation than the registered response of the tested specimen. Also the peak displacement values obtained by the model are smaller than the measured one. Discrepancies are explained by the limitations and hypothesis of the element and constitutive laws used. Finally further developments to improve the accuracy are proposed.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket