基于PRONY特征提取的谐振接地电网故障选线相关分析新方法

A New Correlation Analysis Approach to Fault Line Selection in Resonant Earthed System Using Feature Extraction by Prony

用各馈线故障零序电流故障发生后第1个周波采样数据对应减去故障前1个周波数据,再对应减去故障后第10个周波数据,得到零序电流的暂态纯故障分量.设计一个具有相位自动补偿功能的数字低通滤波器,对零序电流的暂态纯故障分量进行滤波,用以提高Prony的拟合效果.对滤波后各馈线零序电流的暂态纯故障分量进行Prony拟合,提取其主振荡频率信息.利用故障馈线与健全馈线的零序电流在主振荡频率处相关性最弱的特征,对故障零序电流的主振荡频率分量在一定数据窗下进行综合相关度分析,选择综合相关系数最小对应的馈线为故障馈线.大量仿真结果表明该方法选线准确、可靠.

A new approach is proposed to acquire the transient pure fault component of zero modulus current, which using after - fault sampled - data of the first cycle to subtract separately the first cycle sampled - data pre - fault and the tenth cycle data after - fault. A digital low - pass filter with ability of automatic phase compensa- tion is designed to filter the pure fault component in order to improve the effectiveness of Prony algorithm. The method employs the Prony algorithm to analyze and transform the pure fault component after filtering. Prony algorithm can be adopted to extract the information about the harmonic component of pure fault component by fit to a set of exponential components. The analysis of integrated correlation degree is performed according to the harmonic components of each feeder based on the characteristic that the correlation degree between fault line and faulffree line at harmonic frequency is minimal. Simulation results show that the method is effective and reliable.