基于多尺度能量统计和小波能量熵测度的电力暂态信号识别方法

A Study of Electric Power System Transient Signals Identification Method Based on Multi-scales Energy Statistic and Wavelet Energy Entropy

电力暂态信号的检测与分类在电力系统暂态保护、电能质量分析等诸多领域得到广泛的应用,暂态信号的有效特征提取是信号分类识别的基础和前提条件。在介绍了电力系统暂态信号的多尺度表示及定义用于特征提取的2种小波分析处理方法(能量统计分析和小波能量熵分析方法)的基础上,建立了单相接地短路、开关操作、电容投切、一次电弧和雷击等5种暂态类型的EMTDC仿真模型,利用能量统计分析和小波能量熵分析方法给出了5种暂态信号在不同状态下随尺度的分布规律。仿真结果表明,部分暂态信号的多尺度能量统计特征和小波能量熵测度的分布表现出一定规律性,是一种有效的特征提取方法,具有一定的可聚类和信号识别能力。该方法为故障信号的分类提供了新的思路。

Detection and classification of electric transient signal were applied in many fields, such as transient protection and power quality analysis; besides picking-up available characteristics played an important role in the filed of transient signal identification. Based on multi-scales expression and the definition of two post-analysis methods （energy statistic and wavelet entropy）, the EMTDC simulated models were built about five transient signals including single phase grounding, switch operating, capacitor switching, primary arc and lightning strike, the distribution rule along with scales of the five transient signals was given when they happened in different conditions. The results indicated that there were some laws based on multi-scales energy statistic and entropy measure, and it is able to cluster. Meanwhile, the analysis methods offered a new thought to fault classification.