In case of high impedance fault in distribution network, the resistance of grounding medium is large, the fault current is much smaller than the load current, and the fault characteristics are not obvious. Therefore, the existing current or voltage protection devices are difficult to operate, so it is of great significance to find a method to identify them accurately and quickly. Firstly, the zero sequence current of fault line is analyzed in time-frequency domain by continuous wavelet transform (CWT), and the time-frequency matrix in the characteristic frequency band is selected as the identification characteristic quantity. Then, the improved LeNet-5 network is used to identify the identification characteristic quantity to determine whether HIF occurs. A large number of simulations are carried out on PSCAD/EMTDC. The final results show that the proposed method can accurately identify HIF and various disturbances in different neutral grounding systems, and has strong anti noise ability.
王康, 高伟, 杨耿杰. 基于高频分量的配电网高阻接地故障识别[J]. 电气技术, 2022, 23(2): 61-66.
WANG Kang, GAO Wei, YANG Gengjie. High impedance fault detection for distribution networks based on high-frequency components. Electrical Engineering, 2022, 23(2): 61-66.
2022, Vol. 23 
