基于改进K-Means算法的电动汽车充电负荷特性分析

Analysis of Charging Load Characteristics of Electric Vehicles Based on Improved K-Means Algorithm

电动汽车充电行为具有较大的随机性,一定程度上影响电网的稳定运行和规划。为更准确地分析电动汽车充电负荷的特性,提出一种基于改进K-Means算法的聚类分析方法。针对K-Means算法在初始聚类中心选取上的随机性和不稳定性,首先利用Mini Batch K-Means算法的随机抽样能力优化初始聚类中心的选择,随后结合K-Means算法进行迭代优化,有效解决K-Means算法聚类结果不稳定的问题。以云南某城市充电桩负荷数据进行算例分析,结果表明,所提算法相比传统方法相比能更加准确地对多个不同负荷特性的用户进行分类,从而更有效地指导有序用电管理策略的制定。

The charging behavior of electric vehicles has significant randomness,which to some extent affects the stable operation and planning of the power grid.To more accurately analyze the characteristics of electric vehicle charging load,a clustering analysis method based on improved K-Means algorithm is proposed.In response to the randomness and instability of the initial cluster center selection in the K-Means algorithm,the Mini Batch K-Means algorithm’s random sampling ability is first used to optimize the determination of the initial cluster center.Then,the K-Means algorithm is combined for iterative optimization,effectively solving the problem of unstable clustering results in the K-Means algorithm.Taking the load data of charging stations in a city in Yunnan as an example,the results show that the proposed algorithm can more accurately classify users with different load characteristics compared to traditional methods,thus more effectively guiding the formulation of orderly electricity management strategies.