High Impedance Fault Detection Based on Arc Changing Tendency and Nonlinear Least Squares for Active Distribution Networks
Xiangxiang Wei, Xiaowei Wang, Jie Gao, Bin Wang, Dechang Yang
Abstract
When a high-impedance fault (HIF) occurs in active distribution networks, the features of the arc current are weak and can be easily confused with the features of the capacitor switching current or load switching current. Meanwhile, fault currents are easily affected by the harmonics and noises of distributed generators, which may result in low detection accuracy for existing HIF detection methods. To accurately detect high-impedance arc faults, an HIF detection method based on the arc changing tendency and nonlinear least squares (NLSs) are proposed. In the proposed method, variational mode decomposition and windowed fast Fourier transform are first employed to extract current amplitudes. Then, NLSs method is introduced to quantitatively estimate the changing tendency of the current amplitudes. Finally, an HIF detection criterion is developed based on the sign of the coefficient in the nonlinear function. Both simulation results and field test results confirm the effectiveness and applicability of the proposed HIF detection method. The inspiration for the proposed method is drawn from the physical characteristics during the initial stage of HIFs. These characteristics remain unaffected by network operation conditions. Comparative analysis reveals that the proposed method offers higher detection accuracy and reliability compared to existing conventional methods.