An online monitoring method for embedded time-gate displacement sensor operation state base on CNN and LSTM

In view of the problems that embedded time-gate displacement sensor (hereinafter referred to as time-gate), which has severe working conditions, high-precision state is difficult to maintain for a long time, and the operating state is difficult to monitor online. this paper proposes an online monitoring method of time-gate operating state, based on a combination of Convolutional Neural Network (CNN) and Long Short-Term Memory Network (LSTM). Firstly, feature information was obtained from the geometric, operating, electromagnetic simulation and environmental characteristics of the time-gate displacement sensor. Then, A time-gate operating state prediction model was constructed by feeding feature information into CNN-LSTM model and verified by test set, The CNN layers Perform feature learning and extraction and the LSTM layers Process timing data. Finally, the relevant real-time data are fed into the above operating state monitoring system, which achieve the real-time online Monitoring of the time-gate operation state. The experimental results show that, the extreme prediction errors such as abnormal and no alarm were avoided by this method in the test set, and the recognition accuracy of the state is up to 91.1% by this prediction model. Effectively guarantee the safe and stable operation of the time-gate and the equipment in which it is located.