摘要
电压暂降源定位对协调电压暂降引起的供用电双方纠纷等具有关键性作用。提出了一种暂降源定位的优化综合判据法。首先,在分析介绍暂降源定位现有的四种典型方法及其判据的基础上,利用115组变电站暂降记录数据,对各方法效果进行了统计分析。然后基于四种典型方法的定位判据,建立电压暂降源定位综合判据线性模型,并结合线性支持向量机建模方法,利用50组记录数据,通过机器学习求解出线性模型的最优系数。最后采用另外65组记录数据对优化综合判据模型进行验证测试。测试结果显示,与现有的单一特征判据方法相比,该法具有良好的定位性能,不仅具有较高的定位正确率,且有较好的鲁棒性,能在加一定白噪声条件下正确定位出暂降源的上游或下游位置。该方法可应用于各类电能质量监测系统。
Voltage sag source location is significant for the customers and suppliers to solve the issue between them. An optimization comprehensive criterion for voltage sag source location is proposed. First, four typical methods of sag source location are introduced, their criterions are analyzed, and statistical analysis of location result is carried out using 115 records of voltage sags from some substations. Then a linear model of optimization comprehensive criterion is constructed based on the criterions of the four location methods, and parameter estimation of the linear model is realized with the linear support vector machine (SVM). Optimal parameters of the model are gotten by machine learning of linear SVM with 50 records of voltage sags. Finally, experiment test for optimization comprehensive criterion is carried out using other 65 records of voltage sags. The results show that, comparing with single criterion location method, the proposed method has good performance of correctness and robustness. The proposed method could locate sag source correctly even under some white noise, which can be applied to various power quality monitoring system.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2013年第5期66-71,共6页
Power System Protection and Control
基金
国家自然科学基金项目(51107120
51147010)
浙江师范大学计算机软件与理论省级重中之重学科开放基金~~
关键词
电压暂降源定位
优化综合判据
最优系数
线性支持向量机
鲁棒
voltage sag source location
optimization comprehensive criterion
optimal parameters
linear SVM
robustness