基于递归定量特征的变压器励磁涌流识别  

Transformer excitation inrush current recognition based on recurrence quantification characteristics

作　　者：宋晓[1] 李平[1] 徐公林[1] 李树卿[2] 郭家虎[1] 

机构地区：[1]安徽理工大学电气与信息工程学院,安徽淮南232001 [2]嘉兴电力局,浙江嘉兴314000

出　　处：《中南大学学报（自然科学版）》2013年第5期1932-1937,共6页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(50907001)

摘　　要：为避免励磁涌流对变压器差动保护的影响,进一步提高变压器差动保护动作的准确率,针对变压器励磁涌流和内部故障电流信号的非平稳特征,提出一种基于递归定量分析(RQA)的励磁涌流识别方法。首先,应用相空间重构理论,在高维相空间中研究励磁涌流和内部故障电流信号的内在规律;然后,采用RQA方法提取信号的非线性特征参数即递归率、确定率、分层度及熵。研究结果表明:励磁涌流的递归率、确定率和分层度均比内部故障电流的小,而熵比内部故障电流的大,说明励磁涌流比故障电流的规则性低,混沌性高;通过RQA的这4项特征量可以对变压器励磁涌流进行有效识别。In order to avoid the influence of excitation inrush current on transformer differential protection and further enhance the accuracy of transformer differential protection action, a method of excitation inrush current recognition based on recurrence quantitative analysis （RQA） was proposed given the nonstationary features of the transformer excitation inrush current and internal fault current signals. Firstly, the phase space theory was applied to show the inherent laws of inrush current and internal fault current signals in high-dimension phase space. Then the RQA method was used to extract the signal nonlinear characteristic parameters, i.e. recurrence rate, determinism rate, laminarization rate and entropy. The results show that the recurrence rate, determinism rate and laminarization rate of excitation inrush current are lower than those of internal fault current, but the entropy of excitation inrush current is larger, which means the regularity of excitation inrush current fault is lower than that of the internal fault current, and the chao is higher. Through the four characteristic features of RQA, the transformer excitation inrush current can be effectively identified.

关 键 词：励磁涌流 内部故障电流 递归定量分析 非线性 

分 类 号：TM72[电气工程—电力系统及自动化]