Fault diagnosis of ship diesel power distribution system based on WOA-RF algorithm

Objectives Ship diesel power distribution system is very important for ship navigation. However, due to the harsh marine environment, its faults occur frequently. The traditional fault diagnosis methods are quite different from the requirements in terms of accuracy, robustness and reliability. Therefore, a fault diagnosis method of ship diesel power distribution system based on Whale optimization algorithm optimizes random forest algorithm (WOA-RF) is proposed. Methods The MATLAB / Simulink simulation software is used to build the ship 's diesel power distribution system model, and the fault and normal working condition data of the ship 's diesel power distribution system are collected. After normalizing the collected data, the time domain features are extracted and the important features are extracted by random forest to reduce the data dimension. The random forest model optimized by WOA is used to identify, diagnose and classify the operation data of the ship 's diesel power distribution system. In the original data set, compared with 10 different algorithms, the accuracy of WOA-RF is increased by 2.4 % at least and 30.32 % at most. In the addition of 5dB noise data, compared with seven different algorithms, the accuracy of WOA-RF is improved by 2.43 % at least and 22.91 % at most. Results The simulation results show that the WOA-RF method can identify the fault state and the normal state with 100 % accuracy. It can distinguish 12 fault types with 98.96 % accuracy. Conclusions The fault diagnosis method based on WOA-RF shows excellent accuracy and robustness in complex marine environment, and provides a reliable fault identification solution for ship power system.