Abstract：In order to solve the problem that when the single-phase ground fault happens in the mine underground electric power network, as the characteristics of signal by sampling with energy loss and affect the accuracy of the line selection results, this paper proposed that using the rough set theory to enhance the sampling data at first, then using complex wavelet transform the feature signal which were disposed, the normalized transient component will be extracted as the input vector of the neural network’s training and testing; the smoothing factor of generalized regression neural network was optimized globally and the optimal model of the fault line selection was established. Simulation experiments show that this method has the advantages of quick training speed, low misjudgment rate, and can satisfy the underground electric power network for fault diagnosis efficiency and accuracy requirements.
孟宪敬. 煤矿井下供电网单相接地故障诊断方法[J]. 辽宁工程技术大学学报(自然科学版), 2017, 36(1): 83-86.
MENG Xianjing. A method of diagnosing single-phase grounding fault of
coal mine electric power network. Journal of LNTU.Natural Science, 2017, 36(1): 83-86.