Prognosis of Remaining Life of Rolling Element Bearings by using Artificial Neural Network Technique

摘要

Rolling bearing failure is by and large probabilistic in nature and very difficult to capture the pattern of it by a single model. Techniques available in open literature for bearing health monitoring focuses on the identification of faults which generally have taken place at the final stage of bearing failure. At this stage, very limited magnitude of remaining bearing life exists. Hence, in the proposed technique, a composite approach has been adopted for prognosis of remaining life of bearing which is comprised of early fault identification and ANN model for predicting the span of early fault detection to final failure.