Research on pollution identification of catenary insulator based on Improved Gray Wolf algorithm

Abstract

Abstract Support vector machines are used as classifiers for the contact network insulator fouling level recognition model. To improve the accuracy of the recognition model, the search strategy of the gray wolf algorithm is improved, the population is initialized using chaotic mapping rules, and the position of individual gray wolves is updated based on a dynamic weighting strategy. The improved algorithm is applied to the kernel parameter and penalty factor seeking of SVM to find out the optimized parameters to construct the recognition model, and the recognition status of several other intelligent algorithms is compared. The experimental results show that the proposed method can effectively improve the accuracy of insulator fouling level recognition and has certain engineering application value.