Partial discharge fault diagnosis method based on combinational logic and optimal LS-SVM

Abstract

The invention discloses a partial discharge fault diagnosis method based on combinational logic and an optimal LS-SVM, and belongs to the field of partial discharge fault diagnosis in the high voltage field. The method comprises the following steps: firstly, establishing different partial discharge fault models, and carrying out signal acquisition on the different partial discharge fault models to obtain different partial discharge signals PDij; secondly, the information entropy H (PDij), the singular entropy TSE (PDij), the supply Kurt (PDij) and the cumulant C60 (PDij) of the partial discharge signal PDij are obtained; secondly, a partial discharge combinational logic feature extraction method is provided, and a feature parameter Fij is extracted; then, the principle of a least square support vector machine is analyzed, the least square support vector machine is improved and optimized in combination with a particle optimization algorithm, and a fault classification method based on the optimal LS-SVM is provided. And finally, a random combination feature extraction method is combined with the optimal LS-SVM classifier, training and testing of the optimal LS-SVM fault classifier are completed, and overall evaluation of a diagnosis result is given.

Description

technical field [0001] The invention belongs to the technical field of partial discharge fault diagnosis in the field of high voltage, and more particularly relates to a partial discharge fault diagnosis method based on combinational logic and optimal LS-SVM. Background technique [0002] Different types of partial discharges have different degrees of damage to power equipment, which will directly affect the maintenance and maintenance strategies. Therefore, it is particularly important to diagnose and monitor partial discharges after effectively extracting the characteristic parameters of partial discharges. With the rapid development of pattern recognition and computer technology, the partial discharge fault diagnosis method based on artificial intelligence technology has been developed and promoted to a certain extent. The current partial discharge fault diagnosis methods mainly include methods based on artificial neural network, fault diagnosis methods based on fuzzy rea...

AI Technical Summary

Problems solved by technology

However, there are still some defects in the fault feature extraction for different partial discharge types, the feature extraction effect needs to be improved, and the correct recognition rate of the fault diagnosis model needs to be optimized

The correct recognition rate of existing partial discharge fault diagnosis needs to be further improved

Images