谐振接地系统暂稳态信息智能融合选线方法 the transient and steady-state information fusion method for single phase grounding line detection in resonant grounding system.pdfVIP

第41 卷 第7 期 电力系统保护与控制 Vol.41 No.7 2013 年4 月1 日 Power System Protection and Control Apr.1, 2013 谐振接地系统暂稳态信息智能融合选线方法 郭谋发，林妙玉，杨耿杰 (福州大学电气工程与自动化学院，福建 福州 350108) 摘要：谐振接地系统单相接地故障特征信息微弱，现有采用单一故障特征信号的选线方法难于实现可靠选线。综合考虑暂态 零序电流能量比、零序有功功率比、负序电流变化量比等三种单一故障特征选线方法的互补性，针对谐振接地系统提出采用 多人工神经网络对单相接地暂稳态信息进行融合的选线方法。区分稳定接地及非稳定接地，分别采用不同的融合方案。大量 仿真验证该选线方法受接地位置、接地发生时刻、接地过渡电阻大小等因素的影响较小，适用于各种单相接地情况，具有较 高的选线鲁棒性。 关键词：谐振接地；单相接地；暂稳态信息；智能融合选线；人工神经网络 The transient and steady-state information fusion method for single phase grounding line detection in resonant grounding system GUO Mou-fa, LIN Miao-yu, YANG Geng-jie (College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108, China) Abstract: The feature information of single phase grounding fault occurring in resonant grounding system is weak. The existing line selection based on single-fault-characteristic is difficult to achieve reliable line detection. The complementarity of three kinds of line selection methods based on single-fault-characteristic such as energy ratio of zero-sequence current, zero-sequence active power ratio and variation ratio of negative-sequence current is comprehensively considered. A new method using multiple neural networks to fuse transient and steady-state characteristics is proposed for fault line detection in the resonant grounding system. Different integration programs have been taken respectively to distinguish the differences between stable and unstable ground fault. Lots of simulation results verify that the proposed method f