A Method for Predicting the Residual Availability of the Drive System of a Generator Set Considering Mutation Detection

Residual availability prediction provides important predictive maintenance information for unit to reduce sudden failure and unplanned maintenance cost. This paper presents a method to predict the residual availability of the drive system. The method consists of two parts: degradation process segmentation and residual availability prediction. In the first part, the root mean square of vibration signal is selected as the target value to reflect the degraded state of transmission system. The heuristic segmentation algorithm is introduced to divide the degenerate stages. In the second part, the particle filter algorithm is used to predict the residual availability of the transmission system. The advantage of this method is that the signal can be obtained from the outside of the transmission system of the unit, without affecting the operation and physical structure of the transmission system, and the single prediction time is short and the operation efficiency is high. The ability of this method in predicting the residual availability of the transmission system is verified by the prediction results.