基于粒计算和双尺度相似性的负荷曲线集成聚类算法

doi:10.12204/j.issn.1000-7229.2022.05.013

电力建设 ›› 2022, Vol. 43 ›› Issue (5): 117-126.doi: 10.12204/j.issn.1000-7229.2022.05.013

基于粒计算和双尺度相似性的负荷曲线集成聚类算法

孙园¹, 李秋雨², 黄冬梅²(), 孙玉芹¹, 胡安铎², 孙锦中²

1.上海电力大学数理学院,上海市 201306
2.上海电力大学电子与信息工程学院,上海市 201306

收稿日期:2021-07-22 出版日期:2022-05-01 发布日期:2022-04-29
通讯作者: 黄冬梅 E-mail:dmhuang_dl@163.com
作者简介:孙园(1980),男,博士,副教授,主要研究方向为数据分析挖掘和建模。
李秋雨(1995),男,硕士研究生,主要研究方向为电力大数据处理技术。
孙玉芹(1971),女,博士,教授,主要研究方向为组合数学与图论。
胡安铎(1983),男,博士,讲师,主要研究方向为电力时空信息技术。
孙锦中(1981),男,博士,讲师,主要研究方向为机器学习。
基金资助:
国家自然科学基金项目(11871377);国家自然科学基金项目(12071274);上海市科委地方院校能力建设项目(20020500700)

Clustering Ensemble Model Based on Granular Computing and Dual-Scale Similarity

SUN Yuan¹, LI Qiuyu², HUANG Dongmei²(), SUN Yuqin¹, HU Anduo², SUN Jinzhong²

1. College of Mathematics and Physics, Shanghai University of Electric Power,Shanghai 201306,China
2. College of Electronic and Information Engineering, Shanghai University of Electric Power,Shanghai 201306, China

Received:2021-07-22 Online:2022-05-01 Published:2022-04-29
Contact: HUANG Dongmei E-mail:dmhuang_dl@163.com
Supported by:
National Natural Science Foundation of China(11871377);National Natural Science Foundation of China(12071274);Local College Capacity Building Project of Shanghai Municipal Science and Technology Commission(20020500700)

摘要/Abstract

摘要：

电力负荷曲线聚类通常依靠负荷形态差异和负荷数值差异对负荷曲线进行分类。提出了一种基于粒计算和双尺度相似性的集成聚类算法,采用以欧氏距离和皮尔森相关系数作为相似性度量的K-means算法生成基聚类,再通过粒度距离度量基聚类间的相似性,从而选择部分基聚类参与集成,最后生成相似度矩阵并采用层次聚类获得最终聚类结果。算例结果表明,该算法能够克服传统负荷聚类算法只能从数值相似性或形态相似性上单一的度量负荷曲线相似性的局限,可以显著提高电力负荷曲线聚类的质量。

关键词: 负荷聚类, 粒计算, 双尺度相似性, 相似度矩阵, 集成算法

Abstract:

Power load profile clustering usually classifies load profiles according to shape difference and numerical difference of the curves. In this paper, an ensemble clustering algorithm based on granular computing and dual-scale similarity is proposed. The K-means algorithm, which takes Euclidean distance and Pearson correlation coefficient as similarity measures, is used to generate base clustering. Then the part of base clustering is selected to participate in the ensemble algorithm through granular computing. Finally, the similarity matrix is generated and the hierarchical clustering is used to obtain the final clustering. The result of the experiment shows that, the proposed algorithm can overcome the limitation that the traditional load profile clustering can only measure the load similarity from the value or shape, and significantly improves the quality of power load profile clustering.

Key words: load profile classification, granular computing, dual-scale similarity, similarity matrix, clustering ensemble algorithm

中图分类号:

TM714

孙园, 李秋雨, 黄冬梅, 孙玉芹, 胡安铎, 孙锦中. 基于粒计算和双尺度相似性的负荷曲线集成聚类算法[J]. 电力建设, 2022, 43(5): 117-126.

SUN Yuan, LI Qiuyu, HUANG Dongmei, SUN Yuqin, HU Anduo, SUN Jinzhong. Clustering Ensemble Model Based on Granular Computing and Dual-Scale Similarity[J]. ELECTRIC POWER CONSTRUCTION, 2022, 43(5): 117-126.

图/表 11

图1 基于粒计算和双尺度相似性的聚类集成算法流程

Fig.1 Flow chart of clustering ensemble model based on granular computing and dual-scale similarity

图2 谱聚类算法聚类结果

Fig.2 Clustering results of spectral clustering

图3 本文算法聚类结果

Fig.3 Clustering results of the proposed method

表1 不同算法的聚类指标

Table 1 Clustering indices of different algorithms

表2 聚类指标稳定性分析

Table 2 Stability analysis of clustering indices

图4 基聚类个数H对聚类结果的影响

Fig.4 The influence of base clustering number on clustering results

图5 挑选出的基聚类个数对聚类结果的影响

Fig.5 The influence of the number of selected base clusters on the clustering results

表3 不同比例的基聚类选取方案的聚类指标对比

Table 3 Clustering index comparison of base clustering selection schemes with different proportion

图6 粒度大小对聚类结果的影响

Fig.6 The influence of granularity on clustering results

表4 基聚类比对不同数据集上的影响

Table 4 The influence of base clusters’ proportion on different data sets

表5 算法在不同数据集上的最优参数

Table 5 The optimal parameters of the algorithm on different data sets

参考文献 20

[1]	杨颖, 刘友波, 黄媛, 等. 电能替代下考虑需求响应的综合园区风光储配置方法[J]. 电力建设, 2021, 42(10): 9-18.
	YANG Ying, LIU Youbo, HUANG Yuan, et al. Allocation of wind-solar-battery under electric energy substitution in integrated park considering demand response[J]. Electric Power Construction, 2021, 42(10): 9-18.
[2]	仲立军, 周晓琴, 顾曦华, 等. 基于负荷分解技术的小区电能替代需求分析[J]. 电力系统保护与控制, 2021, 49(17): 147-153.
	ZHONG Lijun, ZHOU Xiaoqin, GU Xihua, et al. Non-intrusive load monitoring using low sampling-rate data for electric energy substitution demand analysis[J]. Power System Protection and Control, 2021, 49(17): 147-153.
[3]	张杰, 刘洋, 李文峰, 等. 基于两阶段数据增强和双向深度残差TCN的用户负荷曲线分类方法[J]. 电力建设, 2022, 43(2): 89-97.
	ZHANG Jie, LIU Yang, LI Wenfeng, et al. Power load curve identification method based on two-phase data enhancement and bi-directional deep residual TCN[J]. Electric Power Construction, 2022, 43(2): 89-97.
[4]	HUA W Q, MO L F. Clustering ensemble model based on self-organizing map network[J]. Computational Intelligence and Neuroscience, 2020, 2020: 2971565.
[5]	PARMAR D, WU T, BLACKHURST J. MMR: An algorithm for clustering categorical data using rough set theory[J]. Data & Knowledge Engineering, 2007, 63(3): 879-893. doi: 10.1016/j.datak.2007.05.005 URL
[6]	CHEN Y W, TANG S Y, BOUGUILA N, et al. A fast clustering algorithm based on pruning unnecessary distance computations in DBSCAN for high-dimensional data[J]. Pattern Recognition, 2018, 83: 375-387. doi: 10.1016/j.patcog.2018.05.030 URL
[7]	TRUONG H Q, NGO L T, PEDRYCZ W. Granular fuzzy possibilistic C-means clustering approach to DNA microarray problem[J]. Knowledge-Based Systems, 2017, 133: 53-65. doi: 10.1016/j.knosys.2017.06.019 URL
[8]	WANG X, YU F S, ZHANG H X, et al. Large-scale time series clustering based on fuzzy granulation and collaboration[J]. International Journal of Intelligent Systems, 2015, 30(6): 763-780. doi: 10.1002/int.21726 URL
[9]	黄冬梅, 葛书阳, 胡安铎, 等. 采用中心优化和双尺度相似性度量的改进K-means负荷聚类方法[J]. 电力系统及其自动化学报, 2021, 33(12): 93-100.
	HUANG Dongmei, GE Shuyang, HU Anduo, et al. Improved K-means load clustering method based on center optimization and double-scale similarity measurements[J]. Proceedings of the CSU-EPSA, 2021, 33(12): 93-100.
[10]	陈烨. 电力系统负荷曲线聚类与负荷模型聚合的方法研究[D]. 杭州: 浙江大学, 2018.
	CHEN Ye. Research on load profile classification and load model aggregation methods for power system[D]. Hangzhou: Zhejiang University, 2018.
[11]	张斌, 庄池杰, 胡军, 等. 结合降维技术的电力负荷曲线集成聚类算法[J]. 中国电机工程学报, 2015, 35(15): 3741-3749.
	ZHANG Bin, ZHUANG Chijie, HU Jun, et al. Ensemble clustering algorithm combined with dimension reduction techniques for power load profiles[J]. Proceedings of the CSEE, 2015, 35(15): 3741-3749.
[12]	徐胜蓝, 司曹明哲, 万灿, 等. 考虑双尺度相似性的负荷曲线集成谱聚类算法[J]. 电力系统自动化, 2020, 44(22): 152-160.
	XU Shenglan, SI Caomingzhe, WAN Can, et al. Ensemble spectral clustering algorithm for load profiles considering dual-scale similarities[J]. Automation of Electric Power Systems, 2020, 44(22): 152-160.
[13]	曾博, 张建华, 丁蓝, 等. 改进自适应模糊C均值算法在负荷特性分类的应用[J]. 电力系统自动化, 2011, 35(12): 42-46.
	ZENG Bo, ZHANG Jianhua, DING Lan, et al. An improved adaptive fuzzy C-means algorithm for load characteristics classification[J]. Automation of Electric Power Systems, 2011, 35(12): 42-46.
[14]	CHOKSI K A, JAIN S, PINDORIYA N M. Feature based clustering technique for investigation of domestic load profiles and probabilistic variation assessment: Smart meter dataset[J]. Sustainable Energy, Grids and Networks, 2020, 22: 100346. doi: 10.1016/j.segan.2020.100346 URL
[15]	XU L, DING S F. A novel clustering ensemble model based on granular computing[J]. Applied Intelligence, 2021, 51(8): 5474-5488. doi: 10.1007/s10489-020-01979-8 URL
[16]	LIANG J Y, WANG J H, QIAN Y H. A new measure of uncertainty based on knowledge granulation for rough sets[J]. Information Sciences, 2009, 179(4): 458-470. doi: 10.1016/j.ins.2008.10.010 URL
[17]	LIANG J Y, LI R, QIAN Y H. Distance: A more comprehensible perspective for measures in rough set theory[J]. Knowledge-Based Systems, 2012, 27: 126-136. doi: 10.1016/j.knosys.2011.11.003 URL
[18]	CHICCO G, NAPOLI R, PIGLIONE F. Comparisons among clustering techniques for electricity customer classification[J]. IEEE Transactions on Power Systems, 2006, 21(2): 933-940. doi: 10.1109/TPWRS.2006.873122 URL
[19]	林顺富, 顾乡, 汤继开, 等. 基于稀疏自动编码器神经网络的负荷曲线分类方法[J]. 电网技术, 2020, 44(9): 3508-3515.
	LIN Shunfu, GU Xiang, TANG Jikai, et al. Power load profile classification method based on neural network of sparse automatic encoder[J]. Power System Technology, 2020, 44(9): 3508-3515.
[20]	李阳, 刘友波, 刘俊勇, 等. 基于形态距离的日负荷数据自适应稳健聚类算法[J]. 中国电机工程学报, 2019, 39(12): 3409-3420.
	LI Yang, LIU Youbo, LIU Junyong, et al. Self-adaptive and robust clustering algorithm for daily load profiles based on morphological distance[J]. Proceedings of the CSEE, 2019, 39(12): 3409-3420.

基于粒计算和双尺度相似性的负荷曲线集成聚类算法

Clustering Ensemble Model Based on Granular Computing and Dual-Scale Similarity

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 20

相关文章 2

编辑推荐

Metrics

本文评价

[1]	黄冬梅, 林孝镶, 胡安铎, 孙锦中. 基于卷积自编码器的日负荷深度嵌入聚类方法[J]. 电力建设, 2021, 42(1): 132-138.
[2]	彭勃，张逸，熊军，董树锋，李永杰. 结合负荷形态指标的电力负荷曲线两步聚类算法[J]. 电力建设, 2016, 37(6): 96-102.