A New Remaining Useful Life Prediction Approach Based On Wiener Process With An Adaptive Drift

Remaining useful life prediction is a key issue in prognostics and health management. A degradation model is presented in this paper for online remaining useful life prediction utilizing a Wiener-process-based model with an adaptive drift parameter. This model is different from other Wiener-process-based model in that the drift parameter is updated iteratively when new monitored data come. With the adaptive parameter, online remaining useful life can be estimated. Kalman filter is used to perform the parameter adaption, and for the prior knowledge of the drift parameter, we take other historical degradation data from a population into account to estimate. An extensive numerical investigation is provided to substantiate the superiority of the proposed model compared with the non-adaptive model. The results show that our developed model can provide better residual life estimation accuracy.