Research on fault diagnosis method for hydraulic system of CFETR blanket transfer device based on CNN-LSTM

Abstract

Conducting fault diagnosis on the hydraulic system of the blanket transfer device Mover in the Chinese Fusion Engineering Test Reactor (CFETR) is a key technical issue that needs to be addressed urgently. In this article, a CNN (Convolutional Neural Networks)-LSTM (Long Short-Term Memory) deep learning model-based method is proposed for fault diagnosis, combining the advantages of feature extraction of the CNN model with the advantages of the LSTM model for time series data processing. Therefore, this model shows a " multi-perspective" property, greatly improving its ability to extract features from data. In the fault diagnosis experiment under the condition of four typical faults, the proposed model has the highest accuracy of 98.56% on the test set and good efficiency in computation time compared to the other three models. This method provides some insights for future research on the Prognostics and Health Management (PHM) of the Mover's hydraulic system and the CFETR's remote handling intelligent operational decision system.