Remaining Useful Life Prognostics for the Rolling Bearing Based on a Hybrid Data-Driven Method

Abstract

Rolling bearing is the core part of rotating mechanical equipment, so developing an effective remaining useful life (RUL) prognostics method for rolling bearing is of necessity to guarantee the reliable operation of mechanical equipment. The relevance vector machine (RVM) is one of substantially used methods for RUL prognostics of rolling bearing. However, the accuracy generated by RVM drops rapidly in the long-term prognostics. To remedy this shortcoming of RVM, a novel hybrid method combining grey model (GM), complete ensemble empirical mode decomposition (CEEMD) and RVM is put forward. In the hybrid prognostics framework, the GM is applied to gain a `raw' prediction result and produce an original error sequence. Subsequently, a new smoother error sequence reconstructed by CEEMD method is used to train RVM model, by which the future prediction error applied to correct the raw prediction results of GM is projected. The experimental results demonstrate the satisfactory prognostics performance.