Research on Intelligent Diagnosis Method for a Mechanical Electric System Based on Decision Tree and Random Forest

In response to the complex control logic, easy occurrence of faults, and difficult diagnosis of a mechanical electric system, maintenance methods are often based on expert experience, and manual fault diagnosis is carried out through post maintenance, resulting in low diagnostic accuracy and high maintenance costs. This article collects key signals of the system, selects classification methods such as BP neural network, LVQ neural network, decision tree, and random forest to compare the diagnosis results of various types of faults, and finds decision tree combined with random forest can improve the accuracy of fault diagnosis. And a data collection system was built for real-time collection and verification, indicating that this method can accurately identify various system faults and narrow down the scope of fault diagnosis.