[S01B-03] Bayesian TV-ARX model for evaluation of girder stiffness fluctuation due to breathing crack under running train load
Keywords:High Speed Train, Forced Vibration, Nonlinear System Identification, Breathing Crack, Resonance
In order to evaluate the instantaneous stiffness decrease due to breathing cracks of deteriorated PRC girder in high-speed railways based on measured displacement response under train passages, Time Varying Auto Regression with eXogenous variables (TV-ARX) model, which can consider the fluctuation of girder frequency over time and periodic external force of passing train, and its hierarchical Bayesian estimation method are proposed in this study. Its effectiveness is performed by an application to measured PRC girder displacement responses. The results indicate that instantaneous stiffness decreases that cannot be evaluated by the continuous wavelet transform could be estimated by proposed method.