A selective review diagnosis method for thermal fault of switch cabinet

Abstract

This invention discloses a selective verification and diagnosis method for thermal faults in switchgear. The method acquires operational monitoring data of the switchgear, sends the input features of the operational monitoring data to a first diagnostic module based on a pre-trained artificial intelligence multi-classification model, outputs the class probabilities of multiple candidate diagnostic categories, and determines the first diagnostic result and its reliability index. Then, based on the reliability index and a preset triggering criterion, it determines whether to trigger a second diagnostic module for verification. When the second diagnostic module is triggered, it verifies and diagnoses the operational monitoring data based on a mathematical model corresponding to the thermally related physical phenomena of the switchgear, obtaining a second diagnostic result. Finally, according to a fusion rule, the final diagnostic result is determined based on the first and second diagnostic results. This method improves the accuracy and reliability of switchgear thermal fault diagnosis while ensuring diagnostic efficiency.

Images