Predicting Remaining Useful Life Based on the Failure Time Data with Heavy-Tailed Behavior and User Usage Patterns Using Proportional Hazards Model

To enhance the reliability and availability of complex engineering products and reduce the service costs, nowadays some efficient maintenance programs including Condition-based Maintenance (CBM) are implemented for remote diagnostics and prognostics in industries. Prediction of Remaining Useful Life (RUL) for the unit/part is a key aspect of prognostics and has drawn increasing attention in recent years. Due to the rapid growth of cyber infrastructure and sensing technology, an abundance of data is now readily available for RUL prediction. This paper presents a methodology for predicting RUL based on the failure time data exhibiting heavy-tailed behavior in the survival function and user usage patterns through Proportional Hazards Model (PHM). A numerical study is implemented to investigate the effect of the heavy-tailed behavior on the RUL prediction error. The effectiveness of the proposed method is illustrated through a real-world example. Its performance is compared with that of Random Forests and the prediction results show that the proposed method can make accurate RUL prediction for machinery prognostics.