A model for degradation prediction with change point based on Wiener process

Prediction of degradation paths is important to condition-based maintenance (CBM). To address the detection of the existence of a sudden change point in a degradation path, a piecewise model based on Wiener process with a linear drift is proposed. Two different degradation drifts are introduced in the model as hidden states. Based on the Bayesian theorem, the likelihood function is given and the algorithm of parameter estimation is presented subsequently. Meanwhile, the Kalman filter and the smoother algorithm are used to estimate the hidden states. And we detect the change point of the two-stage degradation according to the concordance correlation coefficient. To validate the proposed method, a numerical simulation and a case study of bearing are presented at last.