An ensemble learning-based prognostic approach with degradation-dependent weights for remaining useful life prediction

Remaining useful life (RUL) prediction is crucial for the implementation of predictive maintenance strategies. While significant research has been conducted in model-based and data-driven prognostics, there has been little research reported on the RUL prediction using an ensemble learning method that combines prediction results from multiple learning algorithms. The objective of this research is to introduce a new ensemble prognostics method that takes into account the effects of degradation on the accuracy of RUL prediction. Specifically, this method assigns an optimized, degradation-dependent weight to each learner (i.e., learning algorithm) such that the weighted sum of the prediction results from all the learners predicts the RULs of engineered systems with better accuracy. The ensemble prognostics method is demonstrated using two case studies. One case study is to predict the RULs of aircraft bearings; the other is to predict the RULs of aircraft engines. The numerical results have shown that the predictive model trained by the ensemble learning-based prognostic approach with degradation-dependent weights is capable of outperforming the original ensemble learning-based approach and its member algorithms.