UUV Fault Diagnosis Model Based on Support Vector Machine

Aiming at the most common power system faults in the UUV ancillary system, in order to diagnose related faults in a timely and accurate manner and prevent the occurrence of faults as early as possible, in this paper, we analyze four common faults based on the characteristic of the UUV power system, and obtain the fault mode of each fault. Then, according to the failure mode, the power system operation model and the failure model of the power system are established in Simulink under the Matlab platform, and four types of typical faults of the UUV power system are reproduced through the combination of experiment and simulation, and then all the necessary follow-up diagnosis process is obtained including training and validation data sets. Finally, the fault diagnosis model of UUV power system based on support vector machine algorithm is adopted, corresponding to different faults, we use genetic algorithms to optimize the selected diagnostic parameters and complete the fusion training verification process for multi-source parameter sets. The experimental results show that when the power system has a small number of fault samples, the use of support vector machine algorithm for fault diagnosis has good adaptability, and the diagnosis results have higher accuracy.