基于Mean-shift聚类的大规模新能源对电网线损影响评估方法

随着不同发电出力形式,特别是新能源出力的大量并网,现代电网呈现出电网结构不断扩大、运行控制日趋复杂的特性。为准确评估不同发电出力对电网线损影响因子大小,保障电网经济运行,提出一种基于Mean-shift聚类分析的电网线损影响因子评估方法。首先,提出一种基于误差平方和的滑动窗口半径最优选取方法;其次,采用基于高斯核函数的Mean-shift聚类分析法将电网量测数据按疏密程度聚类,提取出其主要特征;再次,通过对聚类结果进行多元线性回归,得到不同发电出力的线损影响因子。最后,通过在某省级电网上的实际应用效果,证明了该方法的有效性。

Evaluation Method of Large Scale New Energy Impact on Power Grid Line Loss Based on Mean-shift Clustering

With the large number of different power generation forms integrating to the power grid,especially the renewable energy generation,the modern power grid presents the characteristics of continuous expansion of power grid structure and increasingly complex operation control.In order to accurately evaluate the impact factors of different generation capacity on power grid line loss and ensure the economic operation of power grid,a method based on mean shift clustering analysis is proposed to evaluate the impact factors of power grid line loss.Firstly,an optimal selection method of sliding window radius based on the sum of squared error is proposed.Secondly,the mean shift clustering analysis method based on Gauss kernel function is used to cluster the measured data of power grid according to the density degree and extract its main features.Thirdly,the line loss influencing factors of different power generation output are obtained through multiple linear regression of the clustering results.Finally,the effectiveness of the method is proved by the practical application in a provincial power grid.