基于微分熵与RQA的电能质量扰动分析

基于非线性混沌和相空间重构理论,将电能质量扰动信号序列重构到高维相空间,进行递归图(RP)分析。采用微分熵法对电能质量信号进行相空间重构,避免分别求取嵌入维数和延迟时间的不一致性;引入递归定量分析(RQA)进行扰动的定量分析,克服传统特征提取方法对过程平稳的严格要求。利用能够表征信号发散程度的RQA参数-确定率(DET)和分层率(LAM)组成电能扰动信号识别的特征向量,根据不同电能质量扰动信号各自的分布情况,来区分不同的电能质量扰动信号。通过对6种电能质量扰动信号进行实验分析,结果表明:该方法不仅能够很直观地识别电能质量扰动信号,还能利用RQA的特征量对信号进行具体的定量分析,为电能质量扰动分析提供了高效、直观的方法。

Power quality disturbance analysis based on differential entropy and recurrence quantification analysis

In the light of the non liner theory of chaos and phase space reconstruction, the signal sequence of power quality disturbance is reconstructed into high dimensional phase space, so as to be analyzed by recursive figure (recurrence plot, PR).The differential entropy method was introduced to reconstruct power quality signal, avoiding the inconsistency caused by obtaining embedding dimension and delay time separately. The recurrence quantification analysis method (recurrence quantification analysis, RQA) was used to quantitatively analyze perturbing signal, so as to overcome the strict requirements on stationary processes by traditional feature extraction methods. The feature vectors of power disturbance signal were consisted by determined rate (determinism, DET) and hierarchical rate (laminarity, LAM), which represented the degree of signal divergence. Different disturbance signals were distinguished depending on the distribution of respective feature vector. Experiments on 6 power quality disturbance signals showed that the presented method could recognize disturbance signal intuitively, as well as quantitatively analyze signals by the RQA feature amount, providing an efficient and intuitive way for the power quality disturbance analysis.