A Vibration Diagnosis Approach to Structural Fault

Abstract

In this paper the sensitivity of structural vibration parameters to fault is analyzed. The frequency response function, which is sensitive to structural fault and able to reflect structural dynamical properties, is chosen as a parameter for fault diagnosis. The least square identification method, Kalman filtering method, and the adaptive filtering method are adopted to diagnose structural fault. The results of simulation tests show the diagnosis approaches presented in the paper are feasible and accurate. The model testing results indicate that the structural fault can be detected even when the model stiffness is reduced to a rate of 5 percent.