Fault diagnosis based on multi-sensor information fusion using LSTM and D-S evidence theory

The aero-engine is so complex that make it very difficult to diagnosis faults. The fusion diagnosis is a hot research pot, and may give a new way to solve the above problems. LSTM neural network and D-S evidence theory methods are studied in detail, and a novel fault diagnosis method based on sensor information fusion is constructed. Taking C-Mapss turbofan engine simulation fault data as an example, the sensor data is classified, and the output result of LSTM neural network is used as input to construct a recognition framework of DS evidence fusion for fault diagnosis. The results show that this method can help avoiding false alarms to a certain extent, improving the accuracy of fault diagnosis.