Failure Probability Analysis of Lightining Strike Transmission Line Based on Data Mining

In recent years, lightining faults occur frequently with the rapid development of power grids and the increase of strong convective weather. Aiming at the problem that the calculation results of the existing calculation method of lightning strike rate of the transmission line are quite different from the actual operation situation, this paper proposes a lightning strike risk assessment model for the transmission line based on the multi-support vector machine combined classifier. This paper evaluates the risk of lightning strikes of the tower by using the actual lightning operation data of 500kV overhead transmission line tower and counting the probability output of all SVM sub-classifiers. An adaptive genetic algorithm is proposed to optimize the parameters of the SVM classifier, which effectively improves the problem of low classification accuracy of the model due to the large penalty parameter value in the conventional genetic algorithm. The results show that the evaluation results of multi-SVM combined classifier can greatly reduce the number of towers that are misjudged as high risk under the principle of evaluating the conservativeness of lightning strike risk on transmission lines.