摘要
根据电感电流与电容电流的关系,提出一种基于2阶累加生成相关性的谐振接地系统故障选线方法。单相接地故障发生后,首先,奇异值分解两周期内的原始暂态零序电流,得到一系列特征分量,并计算各特征分量与原始暂态零序电流的相关系数;然后,小波阈值算法处理相关系数大于阈值的特征分量,并重构以得到纯净的暂态零序电流;其次,为凸显数据形态和表征信号的内在发展趋势,2阶累加生成(2-AGO)方法处理各线路暂态零序电流信号获得本征趋势分量,并对其进行自然对数变换。最后,计算各线路综合相关系数,并选取最小综合相关系数对应的线路为故障线路。理论分析和仿真结果均证明了所提选线方法的准确性,可靠性和强抗噪性。
According to internal relation of inductive and capacitive currents, a novel fault detection method based on 2-order accumulated generating operation （2-AGO） correlation analysis for resonant earthed system was proposed in this paper. When fault occurred, firstly, singular value decomposition was introduced to decompose transient zero sequence current （TZSC）, a series of characteristic components were obtained, and then correlation coefficients between characteristic components and TZSC were calculated. Secondly, if values of correlation coefficients were bigger than threshold, wavelet threshold de-noising algorithm was used to dispose the characteristic components, and the obtained characteristic components were reserved to reconstruct TZSC. 2-AGO algorithm was used to process de-noised TZSC to obtain intrinsic trend functions （1TFs） and process them with Napierian logarithm conversion. Finally, comprehensive correlation coefficients were evaluated, and the line with minimum comprehensive correlation coefficients was selected as fault line. Reliability, effectiveness, adaptability and strong noise immunity of the proposed method are verified with simulation results and theoretical analysis.
出处
《电网技术》
EI
CSCD
北大核心
2017年第5期1674-1682,共9页
Power System Technology
关键词
故障选线
奇异值分解
相关系数
小波阈值消噪
2阶累加生成
本征趋势分量
fault line detection
singular value decomposition
correlation coefficients
wavelet threshold de-noising algorithm
2-AGO method
intrinsic trend function
