ABSTRACT
Most of existing studies used simplistic lumped models to model the nonlinear vibration response of structure. These models, such as mass-springdamper models and shear-building models, are based on simplifying assumptions and insufficient to predict the vibration characteristics/response of a real structures. The recently development of nonlinear finite element methods with the improvement of computer efficiency and capacity made it
1 INTRODUCTION
In recent years, research on condition assessment and damage detection of these structures has gained attention. Most of existing studies are proposed for linear structures only. However, structural failures normally occur in the nonlinear state and, the inelastic material behavior and geometric nonlinearity should be considered. Several methods such as Bayesian Approach (Yuen and Beck, 2003, Ting et al., 2006), Least squares algorithm (Xu et al., 2012, He et al., 2012), Kalman filter (Lei and Wu, 2011, Lei et al., 2012), and intelligence algorithm have been proposed and employed for damage detection in nonlinear structures. Yuen and Beck proposed nonlinear parameter identification based on spectral density approach. A stochastic model is used for the uncertain input and a Bayesian probabilistic approach is used to quantify the uncertainties in the model parameters (Yuen and Beck, 2003). Ting et al. develop a Bayesian estimation technique for rigid body dynamics parameter estimation. Bayesian regularization method is employed to ensure robustness of the algorithm for high dimensional ill-conditioned data contaminated with noisy input and noisy output data (Ting et al., 2006). Xu et al (Xu et al., 2012) proposed a nonlinear system identification method with power series polynomial model. The nonlinear restore force can be presented with power series polynomial modeling technique. The coefficient of the polynomial was identified by means of standard least-square techniques without any assumptions and prior
possible to simulate complex nonlinear behavior of large structure under earthquake input. Existing nonlinear finite element methods focus on the dynamic analysis (Taucer et al., 1991), seismic design, and reliability evaluation (Haukaas and Der Kiureghian, 2004). These studies show nonlinear finite element methods can be used to predict and model the nonlinear dynamic behavior of structures with high accuracy.
