8th International Conference on Advanced Power System Automation and Protection (APAP2019)

780 / 2019-04-28 21:37:32

A Line Fault Type Detection Method Based on Optimal Sample Training Support Vector Machine in MMC-HVDC Transmission System

EMD,fault detection，MMC-HVDC，optimization algorithm，SVM.

The research on accurately identifying the type of line fault of the MMC-HVDC system is of great significance for the fast recovery of the faulty line. However, the high-resistance ground fault has always been a difficult point of identification. In this paper, a fault detection method based on Support Vector Machine (SVM) is proposed. The empirical mode decomposition (EMD) is used to extract several high-frequency modal quantities in the fault voltage signal. The optimal weights of these extraction quantities are searched by the optimization algorithm to extract the characteristics of the original fault voltage to train the SVM classification model.This paper establishes a ±250kV MMC-HVDC simulation model to verify the identification results. The simulation results prove that the SVM model can quickly and accurately identify different types of faults at low sampling frequencies, and has high accuracy for the identification of high-resistance ground faults.