基于灰狼优化聚类算法的日负荷曲线聚类分析

吴亚雄; 高崇; 曹华珍; 陈吕鹏; 唐俊熙; 李浩

引用本文:	吴亚雄,高崇,曹华珍,等.基于灰狼优化聚类算法的日负荷曲线聚类分析[J].电力系统保护与控制,2020,48(6):68-76.
	WU Yaxiong,GAO Chong,CAO Huazhen,et al.Clustering analysis of daily load curves based on GWO algorithm[J].Power System Protection and Control,2020,48(6):68-76

【打印本页】【下载PDF全文】【查看/发表评论】【EndNote】【RefMan】【BibTex】

←前一篇|后一篇→

过刊浏览高级检索

本文已被：浏览 6154次下载 2209次	码上扫一扫！
分享到：微信更多字体:加大+\|默认\|缩小-
基于灰狼优化聚类算法的日负荷曲线聚类分析
吴亚雄^1,2,高崇^1,2,曹华珍^1,2,陈吕鹏³,唐俊熙^1,2,李浩^1,2
(1.广东电网有限责任公司电网规划研究中心，广东广州 510030;2.广东电网发展研究院有限责任公司，广东广州 510030;3.苏州华天国科电力科技有限公司，江苏苏州 215000)

摘要:

针对模糊C-均值聚类算法(Fuzzy C-Means, FCM)应用于日负荷曲线聚类分析时存在易受初始聚类中心影响，易收敛于局部最优值以及日负荷曲线的内在特性难以通过距离得到充分反映的问题，利用日负荷特征值指标对日负荷曲线进行数据降维处理。提出了基于灰狼算法(Grey Wolf Optimizer, GWO)优化的模糊C-均值聚类算法(GWO-FCM)。该算法利用GWO为FCM优化初始聚类中心，结合了GWO的全局搜索能力和FCM的局部搜索能力。算例结果表明所提方法可有效提高日负荷曲线聚类效果，算法鲁棒性好。

关键词: 日负荷曲线聚类分析灰狼优化算法模糊C-均值聚类算法数据降维算法鲁棒性

DOI：10.19783/j.cnki.pspc.190486

分类号:

基金项目:国家自然科学基金项目资助(51777078);中国南方电网公司科技项目资助(GDKJXM20172939)

Clustering analysis of daily load curves based on GWO algorithm

WU Yaxiong^1,2,GAO Chong^1,2,CAO Huazhen^1,2,CHEN Lüpeng³,TANG Junxi^1,2,LI Hao^1,2

(1. Grid Planning & Research Center, Guangdong Power Grid Co., Ltd., CSG, Guangzhou 510030, China;2. Guangdong Power Grid Development Research Institute Co., Ltd., Guangzhou 510030, China;3. Suzhou Huatian Power Technology Co., Ltd., Suzhou 215000, China)

Abstract:

Fuzzy C-Means (FCM) is susceptible to the influence of initial clustering centers, easy to converge to local optimum values and the inherent characteristics of daily load curve can not be fully reflected through distance when applied to clustering analysis of daily load curve. In order to solve the above problems, the dimensionality of daily load curve is reduced by using the characteristic value index of daily load. A fuzzy C-means clustering algorithm based on Grey Wolf Optimizer (GWO) optimization is proposed. The algorithm uses GWO to optimize the initial clustering center for FCM, and combines the global search ability of GWO and the local search ability of FCM. The results show that the proposed method can effectively improve the clustering effect of daily load curve and has good robustness. This work is supported by National Natural Science Foundation of China (No. 51777078) and Science and Technology Project of China Southern Power Grid Company (No. GDKJXM20172939).

Key words: clustering analysis of daily load curves grey wolf optimizer fuzzy C-means data dimension reduction algorithm robustness

X关闭