Hysteresis Modeling Framework Using Bouc-Wen Model Considering Cracking Effects

Abstract

Bouc-Wen class models, i.e., hysteretic models developed based on the original Bouc-Wen model, have been extensively used to describe hysteretic behaviors in various fields of engineering. Nonetheless, it has been consistently pointed out that Bouc-Wen class models may not provide accurate predictions of the force-deformation relationships if the loads are different from those actually used to fit the model. Accordingly, in seismic performance evaluation, the application range of the equivalent-single-degree-of-freedom system employing a Bouc-Wen class model is limited to a nonlinear static seismic analysis. Applications to nonlinear time history analysis requiring a highly accurate model are considered challenging. In this thesis, a hysteresis modeling framework based on Bouc-Wen class models is developed to improve the predictive performance for general loads including seismic excitations. To this end, first, a novel Bouc-Wen model that broadens the coverage of Bouc-Wen class models to the structural elements susceptible to cracks, such as reinforced concrete, is developed. Two additional parameters are introduced to consider the effects of cracking on hysteresis. Next, a new strategy is proposed to identify the parameters of the developed Bouc-Wen model. The proposed strategy is represented by a cyclic loading history with which one can obtain simulation or experiment data for effective parameter identification. A quasi-static cyclic loading history is developed to investigate the full modes of hysteretic behaviors of the specimen. The proposed hysteresis modeling framework is demonstrated by nonlinear finite element analyses and 50 actual experimental datasets of reinforced concrete columns. The proposed model exhibits more accurate and reliable predictive performance than existing Bouc-Wen class models under various types of quasi-static loads. Comparison with finite element analysis results for El Centro and Loma Prieta earthquakes confirms that the nonlinear time history analyses employing the proposed model predict the peak displacement more accurately than existing models. Furthermore, for an RC column, the equivalent SDOF system with the proposed Bouc-Wen model shows notable matching with the finite element analysis results for six earthquakes. The proposed framework is expected to facilitate efficient nonlinear time history analyses that would provide results similar to those by time-consuming finite element analyses.