Industry Analysis Method of Substation Integrated Load Based on Dynamic Time Warping

doi:10.16628/j.cnki.2095-8188.2021.02.006

Abstract

Abstract:

The low accuracy of the similarity measurement method used by the traditional clustering algorithm,results in the low quality of the clustering algorithm and the low fitting degree of the comprehensive load daily load curve in the analysis process of the integrated load industry of the substation.Time-regulated analysis method of substation comprehensive load industry is proposed.Perform the K-means clustering algorithm based on dynamic time normalization on the daily load curve set,and extract the cluster center curve reflecting the industry load characteristics.Based on the dynamic time normalization,the comprehensive load daily load curve of the substation is simulated according to the cluster center curve.Together,complete the analysis of the composition ratio of the substation industry.The results show that:the clustering center and substation industry composition ratio obtained by the method is consistent with the actual situation,and the clustering quality and accuracy of fitting the comprehensive load curve of the substation are better than the traditional method.

Key words: dynamic time warping, clustering, comprehensive load, industry composition proportion

CLC Number:

TM714

YANG Jun, HUANG Jiyuan, HU Xiangwei, LI Rensheng, WU Donglin, ZENG Linjun. Industry Analysis Method of Substation Integrated Load Based on Dynamic Time Warping[J]. LOW VOLTAGE APPARATUS, 2021, 0(2): 31-36.

Figures/Tables 7

References 19

[1]	JI Y, BUECHLER E, RAJAGOPAL R. Data-driven load modeling and forecasting of residential appliances[J]. IEEE Transactions on Smart Grid, 2020,11(3):2652-2661. doi: 10.1109/TSG.5165411
[2]	陈建枫, 范兵, 郭敏, 等. 基于用电特性的分类用户优化峰谷电价机制研究[J]. 电器与能效管理技术, 2017(4):52-58.
[3]	PAIDI E S N R, NECHIFOR A, ALBU M M, et al. Development and validation of a new oscillatory component load model for real-time estimation of dynamic load model parameters[J]. IEEE Transactions on Power Delivery, 2020,35(2):618-629. doi: 10.1109/TPWRD.61
[4]	刘洋, 许立雄, 王剑. 计及数据类别不平衡的海量用户负荷典型特征高性能提取方法[J]. 中国电机工程学报, 2019,39(14):4093-4104.
[5]	LI G, WANG Y, ZHANG L, et al. Similarity measure for time series based on piecewise linear approximation[C]// 2009 International Conference on Wireless Communications & Signal Processing, 2009: 1-4.
[6]	王潇笛, 刘俊勇, 刘友波, 等. 采用自适应分段聚合近似的典型负荷曲线形态聚类算法[J]. 电力系统自动化, 2019,43(1):110-121.
[7]	蒋雯倩, 李欣然, 钱军. 改进FCM算法及其在电力负荷坏数据处理的应用[J]. 电力系统及其自动化学报, 2011,23(5):1-5.
[8]	陈建枫, 范兵, 朱曈彤, 等. 海量数据支撑下的崇明地区电网无功负荷分布时空特性分析[J]. 电器与能效管理技术, 2018(8):65-69,78.
[9]	PARK K J, SON S Y. A novel load image profile based electricity load clustering methodology[J]. IEEE Access, 2019,7(99):59048-9058. doi: 10.1109/Access.6287639
[10]	石亮缘, 周任军, 李娟, 等. 基于时间序列相似性度量的新能源-负荷特性指标[J]. 电力自动化设备, 2019,39(5):75-81.
[11]	尤锭军, 余嘉宽, 徐政, 等. 基于改进K-均值的面向风光海的配电网典型运行场景生成[J]. 电器与能效管理技术, 2019(1):25-32.
[12]	宋军英, 崔益伟, 李欣然, 等. 基于欧氏动态时间弯曲距离与熵权法的负荷曲线聚类方法[J]. 电力系统自动化, 2020,44(15):87-98.
[13]	沈颖, 李昌. 基于边缘计算网关的企业用电行为可视化分析[J]. 电器与能效管理技术, 2020(7):81-86.
[14]	岳伟, 黄钰强, 徐佳宁, 等. 锂离子混合型电容器在线参数辨识方法研究[J]. 电器与能效管理技术, 2020(5):106-110.
[15]	刘洋, 许立雄, 王剑. 计及数据类别不平衡的海量用户负荷典型特征高性能提取方法[J]. 中国电机工程学报, 2019,39(14):4093-4104.
[16]	LI C, CAI W, YU C, et al. Electricity consumption behaviour analysis based on adaptive weighted-feature K-means-AP clustering[J]. IET Generation,Transmission & Distribution, 2019,13(12):2352-2361.
[17]	LIANG H, MA J. Develop load shape dictionary through efficient clustering based on elastic dissimilarity measure[J]. IEEE Transactions on Smart Grid, 2021,12(1):442-452. doi: 10.1109/TSG.5165411
[18]	丁纪明, 段浩伟. 交直流配电网关键设备可靠性评估研究[J]. 电器与能效管理技术, 2020(1):51-57,77.
[19]	陈刚, 何森, 周剑, 等. 多级协同的精细化负荷用户属性模型[J]. 广东电力, 2019,32(2):87-94.

算法种类	迭代次数	运算时间/s	DBI指标
算法种类	迭代次数	运算时间/s	I₁	I₂
传统算法	12	1.36	0.81	0.43
本文算法	9	1.51	0.75	0.37

算法种类	欧氏距离拟合度	DTW拟合度
传统算法	0.523 3	0.723 4
本文算法	0.452 6	0.672 7