Health Assessment and RUL Prediction of Aeroengine Under Multiple Working Conditions

Abstract

AbstractHealth assessment and remaining useful life (RUL) prediction are the focus of Prognostics Health Management (PHM). Highly accurate assessment and prediction affects the effectiveness of maintenance decision. Aeroengine is the key equipment of aircraft, and its health state and remaining life need special attention. However, complicated working conditions restrict the accuracy of prediction, and there are few researches on solving the engine health state under multiple conditions. Therefore, this paper proposes a method of engine health assessment and life prediction under multiple working conditions. Firstly, the degradation sensitive parameters are selected based on fitting function method. Then, based on the Maximum Mean Difference (MMD) method, the integrated multi-cycle health index is constructed. Finally, Gated Recurrent Unit (GRU) and segmental fusion mechanism are established to predict the remaining useful life under multiple conditions. The model is verified by C-MAPSS dataset, the average prediction accuracy of engine is improved by 13.07% on average compared with several mainstream deep learning models.KeywordsMultiple conditionsHealth assessmentLife predictionAeroengine