Extended Kalman Filtering for Remaining-Useful-Life Estimation of Bearings

Condition-based maintenance, which includes both diagnosis and prognosis of faults, is a topic of growing interest for improving the reliability of electrical drives. Bearings constitute a large portion of failures in rotational machines. Although many techniques have been successfully applied for bearing fault diagnosis, prognosis of faults, particularly predicting the remaining useful life (RUL) of bearings, is a remaining challenge. The main reasons for this are a lack of accurate physical degradation models and limited labeled training data. In this paper, we introduce a data-driven methodology, which relies on both time and time-frequency domain features to track the evolution of bearing faults. Once features are extracted, an analytical function that best approximates the evolution of the fault is determined and used to learn the parameters of an extended Kalman filter (KF). The learned extended KF is applied to testing data to predict the RUL of bearing faults under different operating conditions. The performance of the proposed method is evaluated on PRONOSTIA experimental testbed data.