Fault detection and isolation of electromechanical actuator based on SAE-BiLSTM

Abstract The control technology of electromechanical actuator(EMA) and its fault diagnosis is one the key problems multi-electric aircraft study. method based on deep learning used to diagnose isolate classic faults EMA. simulation model EMA modeling according working principle law. Four typical are studied,including return channel jam, spall, motor fault, position sensor fault. Sparse Auto Encoder (SAE) algorithm can perform adaptive extraction data, which preserves dimensionality reduction compression while preserving important features. Bidirectional Long Short Term Memory (BiLSTM) neural network effectively process time series considering both past future data during process. established simulated obtain normal faulty sets, train by SAE-BiLSTM algorithm, then trained for online diagnosis. After experiment, well complete detection isolation.