Woa-svm Based Fault Line Selection Method for Small Current Grounding

Abstract

In medium and low voltage power grid, the neutral point is generally ungrounded or resonant grounding. In this process, when a fault occurs, the current is very small, a small current grounding system is generated. After failure, it can continue to operate for l-2h according to regulations, but long-term operation will cause danger. Because of this, many line selection methods have been born, but these methods have some disadvantages, such as inaccurate line selection or many limitations. These algorithms have some disadvantages, These algorithms have some shortcomings. In order to overcome these shortcomings, woa-svm is applied to small current grounding fault line selection. In this paper, whale optimization algorithm and support vector machine are combined to find the optimal parameters of SVM by using the characteristics of whale algorithm, and then it is applied to small current grounding fault line selection. In this process, the transient zero sequence current generated by the fault of small current system is used as the original data, the features in the fault data are extracted by wavelet transform, and then the extracted feature data are input into WOA-SVM for training to obtain the optimal SVM model. Finally, the test data are used for testing. From the results, the algorithm has high accuracy in line selection, is not affected by grounding resistance and fault phase angle, and avoids the disadvantage that GA, PSO and other optimization algorithms are easy to fall into local optimization