Research on Insulator Fault in Haze Days Based on Particle Swarm Optimization

In this paper, the leakage current data is collected through experiments, and the particle swarm optimization algorithm is used to analyze the development process of pollution flashover. The artificial contamination test is designed reasonably. The leakage current samples under five different pollution levels are collected, and the unbiased threshold method is selected. The leakage current signal is denoised, and the eigenvectors of the frequency components are extracted by wavelet packet analysis and empirical mode decomposition. By comparison, the energy of the leakage current increases as the degree of contamination increases. On the basic of summarizing the current research on pollution flashover mechanism, a least squares support vector machine algorithm based on particle swarm optimization of leakage current frequency component is proposed to detect the contamination degree of insulator. This method is compared with traditional particle swarm optimization algorithm. The effect of this algorithm in processing the feature quantity of frequency components extracted by EMD method is better than that of processing the feature quantity of wavelet packet analysis.