Model parameter prediction of lumped plasticity model for nonlinear simulation of circular reinforced concrete columns

Abstract This paper presents the model parameter estimation of a lumped plasticity model to realistically simulate the nonlinear load-deformation response of circular reinforced concrete columns under cyclic lateral loading. The model parameter calibration is based on an experimental database consisting of 210 circular columns with various input parameters, such as material strength, reinforcement layout, specimen geometry, and test configuration. The model parameter values for the initial stiffness, plastic rotation capacity, moment strength, and cyclic damage parameters are calibrated to the first-cycle envelope of each set of test data. Empirical predictive equations are developed to estimate the model parameters as functions of the input parameters using four different regression techniques: stepwise, ridge, lasso, and elastic net regression. The equations developed with lasso regression provided a reasonable trade-off between the prediction error and the number of input parameters. Important input parameters that affect the model parameters are the volumetric transverse reinforcement ratio, spacing of confinement steel, and axial load. As an application of the proposed column model, seismic fragilities of bridge columns in two bridge classes with different interior bent type are compared based on an existing distributed plasticity model and the proposed lumped plasticity model. The seismic vulnerability prediction discrepancy between the proposed and existing models tend to increase as the level of ground motion intensity increases. The prediction trend between the two models may differ by bent configuration and deck mass. The proposed column model reduces the seismic vulnerability of one-column bent bridge class, but increases the vulnerability of two-column bent bridge class.