Fault Diagnosis of Electric Actuator Based on Spiking Neural Network

Actuator is an important part of the control system, and its fault diagnosis is essential to guide the safe and stable operation of the control process. This paper designed a fault diagnosis method for electric actuators based on the spiking neural network(SNN) of the adaptive threshold neuron model. First, the required training Dataset and test Dataset were obtained from the established experimental platform. Through the local mean decomposition(LMD) of the Dataset, the signal was decomposed into product function components with real physical meaning, so as to obtain the main characteristics of the Dataset; then the adaptive threshold was introduced to optimize the neuron structure, and through the relative ordering learning algorithm(ROL), the accuracy of fault recognition was improved. Finally, the test Dataset was used for verification. The experimental results showed that the method made full use of the effective information of the data and could be effectively applied to the fault diagnosis of the electric control valve.