Fault Diagnosis Method for an Underwater Thruster, Based on Load Feature Extraction

Targeting the problem of fault diagnosis in magnetic coupling underwater thrusters, a pattern classification method based on load feature extraction is proposed this paper. By analyzing output characteristics thrusters under typical patterns, torque model established, and two characteristic parameters are constructed to describe different patterns thrusters. Then, thruster reconstruction method, sliding mode observer (SMO), parameter identification least square (LSM), proposed. According identified parameters, support vector machine (SVM) Finally, feasibility superiority aspects verified, through comparative simulation experiments. The results show that diagnostic accuracy higher than 95% within 5 seconds fault. lowest with single failure state 96.75%, average five states 98.65%.