基于改进模糊C均值算法的电力负荷特性分类

An improved fuzzy C-means algorithm for power load characteristics classification

为了提高负荷分类的精确性和有效性,提出了将基于模拟退火遗传算法的模糊C均值(Simulated Annealing Genetic Algorithm Based Fuzzy C-Means,SAGA-FCM)算法用于电力系统负荷特性分类。SAGA-FCM算法以模糊C均值(Fuzzy C-Means,FCM)算法为基础,融合了模拟退火算法较强的局部搜索能力和遗传算法较强的全局搜索能力,克服了传统FCM算法对初始聚类中心敏感和容易陷入局部最优的问题。将其与系统聚类法、K均值(K-Means)算法和传统FCM算法分别用于电力系统负荷特性分类实验,对比分析表明了SAGA-FCM算法用于负荷特性分类的有效性和优越性。

A simulated annealing and genetic algorithm oriented Fuzzy C-Means （SAGA-FCM） algorithm is used for load classification to improve the accuracy and validity. The traditional Fuzzy C-Means （FCM） algorithm is sensitive to its initial cluster centers, and it is easy to fall into the local optimum. While SAGA-FCM algorithm integrates the strong local search ability of simulated annealing algorithm and the strong global search ability of genetic algorithm to overcome the drawbacks of traditional FCM algorithm. Meanwhile, the hierarchical clustering method, K-Means algorithm and traditional FCM algorithm are also used for power load classification. The comparative analysis from the experimental results shows that SAGA-FCM algorithm is more effective and superior than the other three algorithms.