UKF BASED PARAMETER IDENTIFICATION AND APPLICATION OF DUAL-AXIAL COUPLED BOUC-WEN MODEL

A Bouc-Wen model is commonly used in the simulation of structural nonlinear hysteretic responses. It has the advantage that the hysteretic curve is continuously differentiable while it is capable of taking into account the degradation and pinching effect etc. In this study, a lateral-rotational coupled Bouc-Wen model appropriate for the nonlinear analysis of one-way plan-asymmetric structures is proposed based on the dual-axial coupled Bouc-Wen model. The parameters of the coupled Bouc-Wen model are identified through Unscented Kalman Filter (UKF). Then, by using the equivalent dual-degree of the freedom (EDDOF) model of the one-way plan-asymmetric structures, the feasibility of the UKF in the structural system identification through nonlinear response signal analysis is studied. It is demonstrated that UKF is well applicable to the identification of the state and parameters of a coupled Bouc-Wen model with high efficiency and reliable accuracy. The applicability and reliability of the proposed method is further verified by the asymmetric structural seismic response analysis. The proposed model can be generally rewritten into a state-space model which applications can be extended to research fields such as the monitoring and control, reliability analysis and stochastic analysis of structures.