基于判别分析的故障选线方法

Ground fault protection schemes based on discriminant analysis

为提高小电流接地故障选线方法的鲁棒性和精度,应用统计学理论并结合工程实际,借助判别分析方法将稳态故障信息和暂态故障信息进行融合处理。研究结果表明,经过训练后的判别分析模型误判率为0,利用该模型对另外4组待测样本数据进行测试,测试结果与实际情况吻合,验证了该接地故障选线方法的高效性和实用性。

In order to improve the precision and reliability of traditional earth fault feeder detection schemes for Neutral Ineffectively Earthed Power Systems （NIEPS）, on the basis of the Mahal distance discriminant theory and in combination with the project practice, a distance discriminant analysis model of earth fault feeder detection scheme is established, in which five parameters are selected as the discrimination factor. The back substitution method is introduced to verify the stability of the distance discriminant analysis model and the ratio of mistake-dis- crimination is equal to zero after the distance discriminant analysis model was trained. By means of the model, oth- er four groups of measured data are tested as testing sample, and the results are consistent with the measured data. The simulation results show that the distance discriminant analysis model has a higher accuracy and low mis-dis- crimination ratio.