Electric Field Signal Recognition Method of DS Switching Operations Based on Wavelet Packet Analysis and PSO-HSVM

In order to monitor the state of Gas Insulated Substation (GIS), this paper established a fault simulation experimental platform for Disconnecting Switch (DS), and collected data through a 3D radiation electric field (E-field) measurement system. For extract the feature parameters, four-layer wavelet packet decomposition was performed on the data to obtain normalized energy, and the dimension of the eigenvector was reduced by principal component analysis (PCA) algorithm. Then, the model was trained with the hybrid kernel support vector machine (HSVM) algorithm, and the parameters was optimized with the particle swarm optimization(PSO) algorithm. The result shows that compared with the traditional SVM model, the method proposed in this paper improves the diagnosis accuracy of DS defect signals.