Remaining Useful Life Prediction Based On Ensemble Learning

Effectively estimating remaining useful life(RUL)has essential significance for developing maintenance strategies and reducing maintenance costs.In the industry,there exists a high degree of inconsistency among the length of condition monitoring data.Thus,an ensemble framework based on convolutional bi-directional long short-term memory with multiple time windows(MTW CNN-BLSTM Ensemble)is proposed for accurately predicting RUL under this circumstance.In the training phase,multiple CNN-BLSTM base models with different time window sizes are trained to capture various temporal dependencies between features.This setting expands the time window size and reduces the training error compared to traditional static time window size approaches.In the testing phase,test units are classified and suitable base models are applied according to the length of running time.A weighted average method is exploited to aggregate base models’ outcomes.This ensemble strategy can increase the utilization rate of the test data and further enhance prediction accuracy.The effectiveness of this framework is validated and the comparison with state-of-the-art methods available has been provided.The results have shown that this framework can achieve the minimum prediction errorBased on the proposed ensemble framework,several ensemble learning techniques are adopted to improve model performance further.Firstly,a sort-based ensemble pruning method is applied to eliminate the disadvantages when choosing CNN as individual learners.Besides,a heterogeneous dynamic ensemble approach based on the similarity of health indicators is proposed.Different neural networks are reasonably selected to be individual learners according to the selected time window size,which can improve the diversity and accuracy of individual learners at the same time.After that,a denoising Auto-encoder is exploited to construct the health indicators of training and test units.For each test unit,we search for the most similar training segment and calculate ensemble weights according to the performance of each individual learner on this training segment.The experiment results show that this dynamic weighting method is better than the conventional weighted averaging method.These two improvements further enhance the prediction accuracy of the model,and the proposed ensemble framework can provide strong support for the health management and maintenance strategy development.