Similarity-based residual useful life prediction for partially unknown cycle varying degradation

Similarity-based approach is a popular data-driven prognostic method for Residual Useful Life (RUL) Prediction. The principle of this approach is based on the “similarity” between the monitored part, i.e., the sample whose the RUL has to be predicted and degradation reference trajectory patterns (or known library of a priori degradation functions). The challenge addressed in this paper concerns the RUL estimation of a test sample using degradation observations depending on acquisition time and reference dataset information, built on the knowledge of endurance degradation data. The “similarity” coefficient is estimated here using a mapping function between the aperiodic time degradation function and the known test cycle functions. Our approach is evaluated on the very classical Virkler crack-growth measurements benchmark [14] and the experimental results show that it is efficiency and promising in the context of railways maintenance.