Intelligent Fault Diagnosis of Ship Electromechanical Equipment based on Improved Outlier Detection Algorithms

In order to ensure the uninterrupted transmission of power or electric energy in a ship, it is necessary to locate the fault part timely and accurately, which mainly relies on the fault classifier to identify the fault characteristics. In the process of classifier recognition, there will be some unfiltered noise signals to interfere with it and affect the classifier recognition results. In order to solve such problems, an intelligent fault diagnosis algorithm is established through the pattern recognition method described by classification Support Vector Data Description, so that machine learning technology can play an important role in the fault diagnosis of marine electromechanical equipment. The performance of the proposed algorithm can accurately identify various possible faults, and the fault diagnosis performance is better than that of the traditional ship's generator fault diagnosis model.