| 摘要: |
| 针对复杂配电网中缺少有效划分非线性负荷谐波污染责任和不能准确确定主谐波源位置的状况，提出了一种基于信息熵的多谐波源定位方法。通过支持向量机进行谐波分离，获得测量量的各次谐波数据。接着利用B样条法对测量数据进行概率密度估计，从而计算出母线电压与各支路电流的互信息值，根据互信息值的大小来确定主谐波源位置。通过对典型配电系统在ATP(选择性暂态程序)软件中仿真分析，所提方法具有定位准确、能够进行实际谐波责任区分的优点。 |
| 关键词: 互信息 B样条 支持向量机 谐波分离 多谐波源定位 |
| DOI：10.7667/j.issn.1674-3415.2015.14.006 |
| 投稿时间：2014-10-17修订日期：2015-01-18 |
| 基金项目: |
|
| Research on multiple harmonic sources localization based on information entropy |
| WANG Yunjian,CHEN Feng |
| (School of Electrical Engineering and Automation, Henan Polytechnic University, Jiaozuo 454000, China) |
| Abstract: |
| As there is lack of effective division for liability of harmonic pollutions caused by non-linear loads and there is no accurate localization of harmonic sources in complex distribution network, the method based on information entropy for localization of multiple harmonic sources is proposed. All the harmonic data measured is gained by harmonic separation using support vector machine and then probability density is estimated by using the B-spline method, thus, the value of mutual information regarding bus bar voltage and current of each branch can be figured out and the position of main harmonic sources can be confirmed according to the value of mutual information. Through simulation analysis of typical power distribution system in alternative transient process (ATP) software, the result shows that the method proposed has the advantage of accurate location and is able to divide the real liability of harmonic pollutions. |
| Key words: mutual information B-spline functions Support Vector Machine (SVM) harmonic separation multiple harmonic sources localization |
