Remaining useful life prediction framework for crack propagation with a case study of railway heavy duty coupler condition monitoring

Abstract

A general Remaining useful life (RUL) framework for predicting crack propagation of mechanical system structures based on limited condition monitoring data was proposed. For states of invisible cracks, a method that considers the common characteristics of the object was proposed by using the crack propagation stage as the delay time, combining statistical distribution and hypothesis testing. Then, for the states of visible cracks, methods based on a hypothetical distribution and support vector regression with the Kalman filter considering the current state and individual degradation characteristics were proposed. Finally, taking the condition monitoring data of the coupler body crack state as a case, the RUL prediction from the whole to the individual differences is achieved. The research results indicate that the framework achieves a competitive effect in RUL of internal crack propagation structures of railway wagons, which have important theoretical and practical value for integrity assessment and reliability life prediction.