关键词: 注意力机制, 深度学习, 故障诊断, 液压支架

Abstract: In view of the high fault concealment of hydraulic support and the lack of effective analysis and mining of a large amount of historical detection data, a fault diagnosis method of hydraulic support based on Fusion CNN Transformer (FCT) is proposed. This method can give full play to the advantages of CNN's extraction of local features and transformer's recognition of global information, and extract more useful features hidden in the data to realize fault diagnosis of hydraulic support. The diagnostic performance of the proposed method is verified by experiments with the historical data of hydraulic support in Shanxi Xiegou Coal Mine. The results show that compared with WDCNN model, transformer model and BiLSTM, the proposed method has a good fault diagnosis effect for hydraulic support, and the fault diagnosis accuracy rate reaches 99.59%. It provides a theoretical basis for determining the fault of hydraulic support and has certain engineering application value.