A comparative study on nonlinear models for performance-based earthquake engineering

Performance-based earthquake engineering requires a large number of nonlinear dynamic analyses to statistically assess the performance of frame structures. The complexity and high computational demand of such procedures, however, has hindered its use in practice. The objective of this study is to evaluate the performance of three numerical models with varying computational demand levels. Two nonlinear models with different complexities and one linear model with a concentrated plasticity approach were used to evaluate a reinforced concrete frame. The accuracy of the calculated responses was assessed using the experimental results. A total number of 126 dynamic analyses were performed to derive fragility curves. The nonlinear models calculated significantly more accurate structural responses than the more-commonly used plastic-hinge model. The model preparation and result acquisition times were found to comprise a significant portion of the total computational demand of each model. An overview of the performance-based modeling processes and the critical points for minimizing the computational demand while retaining the calculation accuracy are also presented.