Electricity theft detection method for distribution network CPS based on cyber and physical data

doi:10.3969/j.issn.2097-0706.2024.05.003

Abstract

Abstract:

Widespread applications of massive measurement equipment have raised frequent cyber attacks to the information side of a distribution network cyber and physical system（CPS）. The attacks aiming at encroaching on power companies' economic benefits through electricity thefts seriously endanger their interests. However， existing methods for detecting electricity theft including both data-driven algorithms and model-driven algorithms show limitations in their standalone applications， incapable of reducing false alarm rates of output results effectively. Therefore， an electricity theft detection method based on bilateral data of a distribution network CPS is proposed. Firstly， the method monitors the abnormal fluctuations of the line loss sequence measured on the physical side of the distribution station area， determining the time periods with abnormal electricity consumptions. Secondly， various electricity consumption data from users on the information side of the distribution station area are sent to the convolutional neural network on a weekly basis to investigate suspected electricity thieves. Finally， a second screening on electricity thefts is carried out on the outputs of the electricity stealing detection model based on the negative correlation between the line loss sequence in the station area and the electricity consumption data sequence of the electricity thieves obtained by the time distance weighted Pearson correlation algorithm. The operation process of this method is performed on an IEEE 33 bus system， obtaining normal and abnormal electricity consumption data for following analyses and detection model construction. The results of the comparative experiments verify that， compared to other methods， the proposed electricity theft detection method has a higher reliability and more solid criterion， which can further reduce the false alarm rate of data-driven model detection results while improving accuracy.

Key words: CPS for distribution network, electricity theft detection, cyber security, convolutional neural network, second screening, false alarm rate

CLC Number:

TK012：TM727

DU Long, SHA Jianxiu, FAN Bei, HU Jingwei, LIU Zengji. Electricity theft detection method for distribution network CPS based on cyber and physical data[J]. Integrated Intelligent Energy, 2024, 46(5): 20-29.

Figures/Tables 12

Fig.1

Table 1

Fig.2

Fig.3

Table 2

Fig.4

Table 3

Table 4

Table 5

Table 6

Table 7

Fig.5

References 27

[1]	刘晟源, 章天晗, 林振智, 等. 数据赋能低压配用电系统精益化运行的关键技术与算法[J]. 电力系统自动化, 2023, 47(3):187-199.
	LIU Shengyuan, ZHANG Tianhan, LIN Zhenzhi, et al. Key technologies and algorithms of data empowerment for lean operation of low-voltage power distribution and consumption system[J]. Automation of Electric Power System, 2023, 47(3):187-199.
[2]	易俊, 林伟芳, 任萱, 等. 新型电力系统发展水平指标体系构建及综合评价方法研究[J/OL]. 电网技术:1- 12 (2024-03-14)[2024-03-25]. https://doi.org/10.13335/j.1000-3673.pst.2023.2291.
	YI Jun, LIN Weifang, REN Xuan, et al. Research on construction of index system and comprehensive evaluation method of development level of new power system[J/OL]. Power System Technology:1-12(2024-03-14)[2024-03-25]. https://doi.org/10.13335/j.1000-3673.pst.2023.2291.
[3]	刘涛, 李伟华, 汤熠. 综合智慧能源系统典型构架网络安全防护研究[J/OL]. 综合智慧能源:1-11(2023-08-22)[2024-03-25]. http://kns.cnki.net/kcms/detail/41.1461.TK.20230821.1109.006.html.
	LIU Tao, LI Weihua, TANG Yi. Research on network security protection of typical architecture of integrated smart energy[J/OL]. Integrated Intelligent Energy:1-11(2023-08-22)[2024-03-25]. http://kns.cnki.net/kcms/detail/41.1461.TK.20230821.1109.006.html.
[4]	丁茂生, 王洪儒, 王超, 等. 信息物理视角下能源互联网可靠性评估方法综述[J]. 电网技术, 2021, 45(2):425-436.
	DING Maosheng, WANG Hongru, WANG Chao, et al. Review of reliability assessments for energy internet based on cyber physical integration[J]. Power System Technology, 2021, 45(2):425-436.
[5]	樊强, 刘东, 王宇飞, 等. 电力信息物理系统形态演进关键技术及其进展[J/OL]. 中国电机工程学报:1-13(2023-09-19)[2024-03-25]. http://kns.cnki.net/kcms/detail/11.2107.TM.20230918.1450.004.html.
	FAN Qiang, LIU Dong, WANG Yufei, et al. Key technologies and trends of cyber physical power system morphology evolution[J/OL]. Proceedings of the CSEE:1-13(2023-09-19)[2024-03-25]. http://kns.cnki.net/kcms/detail/11.2107.TM.20230918.1450.004.html.
[6]	郭少勇, 刘岩, 邵苏杰, 等. 新型电力系统数据跨域流通泛安全边界防护技术[J]. 电力系统自动化, 2024, 48(6):96-111.
	GUO Shaoyong, LIU Yan, SHAO Sujie, et al. Ubiquitous security boundary protection technology for cross-domain data circulation in new power system[J]. Automation of Electric Power System, 2024, 48(6):96-111.
[7]	伍志韬, 杜伟, 刘蕾蕾, 等. 恶意攻击下的电力耦合网络风险传播模型研究[J]. 电网技术, 2020, 44(6):2045-2052.
	WU Zhitao, DU Wei, LIU Leilei, et al. Risk propagation model of power coupled networks under malicious attack[J]. Power System Technology, 2020, 44(6):2045-2052
[8]	苏盛, 吴长江, 马钧, 等. 基于攻击方视角的电力CPS网络攻击模式分析[J]. 电网技术, 2014, 38(11):3115-3120.
	SU Sheng, WU Changjiang, MA Jun, et al. Attacker's perspective based analysis on cyber attack mode to cyber-physical system[J]. Power System Technology, 2014, 38(11):3115-3120.
[9]	殷涛, 苏盛, 刘爱国, 等. 基于向前逐步回归的高损线路窃电检测[J]. 电网技术, 2022, 46(7):2742-2749.
	YIN Tao, SU Sheng, LIU Aiguo, et al. Electricity theft detection of high-loss line with forward stepwise regression[J]. Power System Technology, 2022, 46(7):2742-2749.
[10]	Northeast Group LLC. Emerging market smart grid: Outlook 2017[EB/OL].(2017-02-22) [2024-03-25] https://www.prnewswire.com/news-releases/emerging-markets-smart-grid-outlook-2017-300411871.html.
[11]	林振智, 崔雪原, 金伟超, 等. 用户侧窃电检测关键技术[J]. 电力系统自动化, 2022, 46(5):188-199.
	LIN Zhenzhi, CUI Xueyuan, JIN Weichao, et al. Key technologies of electricity theft detection at consumer side[J]. Automation of Electric Power System, 2022, 46(5):188-199.
[12]	ISMAIL M, SHAABAN M F, NAIDU M, et al. Deep learning detection of electricity theft cyber-attacks in renewable distributed generation[J]. IEEE Transactions on Smart Grid, 2020, 11(4): 3428-3437.
[13]	邝萌. 基于Stacking集成学习的窃电检测方法研究[D]. 昆明: 昆明理工大学, 2021.
	KUANG Meng. Research on electricity theft detection method based on stacking ensemble learning[D]. Kunming: Kunming University of Science and Technology, 2021.
[14]	唐冬来, 刘友波, 熊智临, 等. 基于时空关联矩阵的配电台区反窃电预警方法[J]. 电力系统自动化, 2020, 44(19): 168-181.
	TANG Donglai, LIU Youbo, XIONG Zhilin, et al. Early warning method of electricity anti-theft in distribution station area based on spatiotemporal correlation matrix[J]. Automation of Electric Power System, 2020, 44(19): 168-181.
[15]	金晟, 苏盛, 曹一家, 等. 基于格兰杰归因分析的高损台区窃电检测[J]. 电力系统自动化, 2020, 44(23):82-89.
	JIN Sheng, SU Sheng, CAO Yijia, et al. Electricity-theft detection for high-loss distribution area based on granger causality analysis[J]. Automation of Electric Power System, 2020, 44(23):82-89.
[16]	黄刚. 数据-模型融合驱动的中低压配网窃电辨识方法[D]. 重庆: 重庆大学, 2021.
	HUANG Gang. A data model fusion driven identification method for stealing electricity in medium and low voltage distribution networks[D]. Chongqing: Chongqing University, 2021.
[17]	金晟, 苏盛, 薛阳, 等. 数据驱动窃电检测方法综述与低误率研究展望[J]. 电力系统自动化, 2022, 46(1):3-14.
	JIN Sheng, SU Sheng, XUE Yang, et al. Review on data-driven based electricity theft detection method and research prospect for low false positive rate[J]. Automation of Electric Power System, 2022, 46(1):3-14.
[18]	ZHENG Z B, YANG Y T, NIU X D, et al. Wide and deep convolutional neural networks for electricity-theft detection to secure smart grids[J]. IEEE Transactions on Industrial Informatics, 2017, 14(4): 1606-1615.
[19]	王德文, 杨凯华. 基于生成式对抗网络的窃电检测数据生成方法[J]. 电网技术, 2020, 44(2):775-782.
	WANG Dewen, YANG Kaihua. A data generation method for electricity theft detection using generative adversarial network[J]. Power System Technology, 2020, 44(2):775-782.
[20]	CUI L, GUO L, GAO L X, et al. A covert electricity-theft cyber-attack against machine learning-based detection models[J]. IEEE Transactions on Industrial Informatics, 2021, 18(11): 7824-7833.
[21]	杜章华, 苏盛, 刘正谊, 等. 基于生产经营状态识别的低误报率窃电检测二次筛查方法[J]. 电力系统自动化, 2021, 45(2):97-104.
	DU Zhanghua, SU Sheng, LIU Zhengyi, et al. Second inspection method for electricity theft detection with low false alarm rate based on identification of production and operation status[J]. Automation of Electric Power Systems, 2021, 45(2):97-104.
[22]	窦健, 刘宣, 卢继哲, 等. 基于用电信息采集大数据的防窃电方法研究[J]. 电测与仪表, 2018, 55(21):43-49.
	DOU Jian, LIU Xuan, LU Jizhe, et al. Research on electricity anti-stealing method based on power consumption information acquisition and big data[J]. Electrical Measurement ＆ Instrumentation, 2018, 55(21):43-49.
[23]	郑应俊, 杨艺宁, 舒一飞, 等. 基于边缘计算的低压用户窃电检测[J]. 电力系统自动化, 2022, 46(11):111-120.
	ZHENG Yingjun, YANG Yining, SHU Yifei, et al. Electricity theft detection for low-voltage users based on edge computing[J]. Automation of Electric Power Systems, 2022, 46(11):111-120.
[24]	JOKAR P, ARIANPOO N, LEUNG V C M. Electricity theft detection in AMI using customers' consumption patterns[J]. IEEE Transactions on Smart Grid, 2015, 7(1): 216-226.
[25]	孔祥玉, 马玉莹, 赵鑫, 等. 基于多阶段递推数据分析的低压台区窃电检测方法[J/OL]. 中国电机工程学报:1-19(2023-05-06)[2024-03-25]. https://doi.org/10.13334/j.0258-8013.pcsee.230325.
	KONG Xiangyu, MA Yuying, ZHAO Xin, et al. Detection Method of Electric Theft in Low Voltage Station Area Based on Multi-Stage Recursive Data Analysis[J/OL]. Proceedings of the CSEE:1-19(2023-05-06)[2024-03-25]. https://doi.org/10.13334/j.0258-8013.pcsee.230325.
[26]	MASSAFERRO P, DI MARTINO J M, FERNÁNDEZ A. Fraud detection on power grids while transitioning to smart meters by leveraging multi-resolution consumption data[J]. IEEE Transactions on Smart Grid, 2022, 13(3): 2381-2389.
[27]	UCI Machine Learning Repository. Electricity load diagrams[EB/OL].(2015-03-12)[2024-03-25]. https://archive.ics.uci.edu/dataset/321/electricityloaddiagrams2011201.

计量设备	采集变量
配电台区智能终端	A/B/C相电压幅值 A/B/C相电流幅值 A/B/C相及总有功、无功功率 A/B/C相及总功率因数正向、反向有功、无功总电量示值
用户智能电表	A/B/C相电压幅值 A/B/C相电流幅值 A/B/C相及总有功、无功功率 A/B/C相及总功率因数正向有功总电量示值

用户类别	判断为正常的用户	判断为窃电的用户
标签为正常的用户	TN	FP
标签为窃电的用户	FN	TP

窃电检测模型	R_A	R_recall	A_UC	R_FA
LR	0.712	0.758	0.689	0.319
RF	0.768	0.784	0.728	0.302
SVM	0.735	0.761	0.716	0.311
本文模型	0.834	0.863	0.793	0.240

窃电检测模型	R_A	R_recall	A_UC	R_FA
LR	0.748	0.769	0.705	0.251
RF	0.774	0.786	0.742	0.217
SVM	0.785	0.794	0.751	0.210
本文模型	0.896	0.917	0.874	0.132

台区	总用户	正常用户	窃电用户
1	128	128	0
2	128	116	12
3	128	116	12
4	128	128	0