Prediction of Aeroengine Remaining Useful Life Based on SE-BiLSTM

Abstract

The prediction of the remaining useful life (RUL) of aeroengines is very important in practical engineering applications. Aiming at the problem that it is difficult to accurately predict aeroengine RUL in practical engineering applications, this paper proposes a method for predicting the remaining useful life of aeroengines based on the Sequeeze-and-Excitation (SE) and bidirectional long short-term memory network (BiLSTM). First, use SE to improve the BiLSTM network structure and create a SE-BiLSTM network prediction model. Second, use the engine performance degradation data set provided by NASA to train the SE-BiLSTM network prediction model to determine the remaining useful life prediction model of the SE-BiLSTM aircraft engine. Finally, the designed SE-BiLSTM aeroengine RUL prediction model is used to conduct a prediction verification study on the aeroengine RUL. For example, for the remaining useful life prediction study of the data set FD001, the prediction results of the SE-BiLSTM model are compared with the prediction results of the BiLSTM model, the LSTMBS model and the LSTM model, and the root mean square error (RMSE) is reduced by 9.01%, 16.59% and 23.05% respectively. The research results show that the SE-BiLSTM based aeroengine remaining useful life prediction method proposed in this paper has a good predictive effect on the remaining useful life of aeroengines, and can effectively predict the RUL of aeroengines. The prediction accuracy of the SE-BiLSTM model is significantly better than the prediction accuracy of the BiLSTM model, LSTMBS model and LSTM model, and has good application prospects.