Fault Diagnosis Method of Impulse Impedance Characteristic Spectrum Based on Naive Bayes Classifier

The normal operation of the grounding grid is related to the safety of transmission lines and even the power grid. Due to its buried underground location, wide distribution, and harsh operating environment, the grounding body is prone to corrosion. Therefore, the corrosion problem of the grounding grid has always been a problem that troubles the power sector. Therefore, in-depth research on the methods of grounding grid corrosion diagnosis is of great significance for ensuring the safe operation of transmission lines. To effectively improve the accuracy of detecting the location and type of corrosion faults in substation grounding grids, a grounding grid fault diagnosis method based on Naive Bayes Classifier (NBC) impulse impedance feature spectrum is proposed. First, CDEGS software simulation is used to obtain the corresponding spectrum maps under various fault conditions, and then the waveform fitting degree and amplitude difference between each frequency band and the standard spectrum map are input into the Naive Bayes Classifier as characteristic parameters, and the expected results are obtained. The test results show that the identification accuracy of the model for fault types is as high as 93.33%, and can accurately distinguish Single point of failure and regional fault.