基于联合分布适配的智能电表运行状态评价

许丹; 李游; 李贺龙

doi:10.13700/j.bh.1001-5965.2020.0621

基于联合分布适配的智能电表运行状态评价

doi: 10.13700/j.bh.1001-5965.2020.0621

许丹^1, ,,
李游¹,
李贺龙²

1.
北京航空航天大学可靠性与系统工程学院, 北京 100083
2.
中国电力科学研究院有限公司, 北京 100085

基金项目:

国家自然科学基金 51875017

中央高校基本科研业务费专项资金 YWF-20-BJ-J-726

详细信息

通讯作者:
许丹, E-mail: xudan@buaa.edu.cn

中图分类号: TM744;TP312
计量
- 文章访问数: 590
- HTML全文浏览量: 227
- PDF下载量: 86
- 被引次数: 0
出版历程
- 收稿日期: 2020-11-09
- 录用日期: 2021-06-11
- 网络出版日期: 2022-04-20

Operating status evaluation of smart electricity meters based on joint distribution adaption

XU Dan^{1
, ,},
LI You¹,
LI Helong²

1.
School of Reliability and Systems Engineering, Beihang University, Beijing 100083, China
2.
China Electric Power Research Institute Co., Ltd., Beijing 100085, China

Funds:

National Natural Science Foundation of China 51875017

the Fundamental Research Funds for the Central Universities YWF-20-BJ-J-726

More Information

Corresponding author: XU Dan, E-mail: xudan@buaa.edu.cn

摘要

摘要:
针对智能电表运行状态评价中经常存在带标签的数据采集困难且不同地区的数据分布不一致的问题，在智能电表的运行状态评价领域引入迁移学习中的联合分布适配（JDA）算法。该算法寻找一个最优化的变换矩阵，使得在变换后的空间中不同地区数据的边缘分布和条件分布距离最小化。针对条件分布适配中目标域没有数据标签的问题，采用伪标签迭代的方法使得目标域伪标签不断接近真实标签。变换后空间数据中训练得到的分类模型可以运用于新的地区，实现迁移。实验结果证明了JDA算法在智能电表运行状态评价中的有效性。
- 联合分布适配(JDA) /
- 智能电表 /
- 运行状态评价 /
- 迁移学习 /
- 智能电网
Abstract:
In the operation status evaluation of smart electricity meters, it is often difficult to collect labeled data and the data distribution in different regions is inconsistent. To solve this problem, we introduce the joint distribution adaption (JDA) method in transfer learning in the field of operation status evaluation of smart electricity meters. This method tries to find an optimized transformation matrix to minimize the distance between edge distribution and conditional distribution in different regions in the transformed space. In order to solve the problem that there is no data label in the target domain when the conditional distribution adaptation is made, the JDA method uses pseudo-tag iteration method to make the target domain pseudo-label constantly approach the real tag. The classification model trained from the data in the post-transformation space can be applied to the new region to realize the transfer. The experimental results demonstrate the effectiveness of JDA method in the operation status evaluation of smart electricity meters.
- joint distribution adaption (JDA) /
- smart electricity meters /
- operation status evaluation /
- transfer learning /
- smart grid

HTML全文

图 1 研究线路

Figure 1. Procedure of study circuit

下载: 全尺寸图片幻灯片

图 2 JDA算法流程

Figure 2. JDA algorithm flowchart

下载: 全尺寸图片幻灯片

图 3 JDA算法分类误差随迭代次数变化

Figure 3. Classification error of JDA algorithm varies with number of iterations

下载: 全尺寸图片幻灯片

表 1 福建省电表数据集

Table 1. Dataset of electricity meters in Fujian province

属性	取值范围
有功准确度等级	0.5S, 1.0, 2.0
额定负载点误差测量	-0.054~0.25
最小负载点误差测量	-0.249~0.25
最大负载点误差测量	-0.265~0.25
第一次运行误差测量	0.05~0.4
第二次运行误差测量	0~0.45
第三次运行误差测量	0~0.5
运行时间	0~6.5
运行环境	1, 1.2, 2
家族缺陷	1
电量异常次数	0
时钟异常次数	0~7
运行等级	1, 2, 3

下载: 导出CSV

表 2 福建省电表等级分布

Table 2. Grade distribution of electricity meters in Fujian province

运行等级	数量	占比/%
1	13	2.56
2	418	82.44
3	76	15.00

下载: 导出CSV

表 3 基于SMOTE算法不平衡数据处理后的福建省数据分布

Table 3. Data distribution of Fujian province after imbalanced data processing based on SMOTE algorithm

运行等级	数量	占比/%
1	391	35.26
2	418	37.69
3	300	27.05
合计	1 109	100.00

下载: 导出CSV

表 4 目标域数据初始伪标签分布

Table 4. Initial pseudo-label distribution of target domain data

运行等级	数量	占比/%
1	49	8.13
2	369	61.19
3	185	30.68
合计	603	100.00

下载: 导出CSV

表 5 迁移后模型的混淆矩阵

Table 5. Confusion matrix of transferred model

混淆矩阵		预测类1	预测类2	预测类3
真实类	1	39	3	0
真实类	2	10	364	0
真实类	3	0	2	185

下载: 导出CSV

表 6 迁移后模型的主要评价指标

Table 6. Main evaluation indicators of transferred model

评价指标	结果
准确度	0.975
F₁分数	0.943

下载: 导出CSV

参考文献(18)

[1]	陈文娴. 基于深度学习的小样本异常用电数据检测技术研究[D]. 武汉: 华中科技大学, 2018. CHEN W X. Abnormal electricity utilization detection based on deep learning and few-shot learning[D]. Wuhan: Huazhong University of Science and Technology, 2018(in Chinese).
[2]	商曦文, 张颖, 吉莹, 等. 智能电表运行状态评估技术研究[J]. 自动化与仪器仪表, 2018, 11(3): 139-146. SHANG X W, ZHANG Y, JI Y, et. al. Research on smart meters' running state assessment techniques[J]. Automation & Instrumentation, 2018, 11(3): 139-146(in Chinese).
[3]	CAI H, CHEN H Q, YE X Q, et al. An on-line state evaluation method of smart meters based on information fusion[J]. IEEE Access, 2019, 1(7): 163665-163676.
[4]	LI Y F, WU H, PANG S, et al. Application of supervised machine learning algorithms in diagnosis of abnormal voltage[J]. Electrical Measurement and Instrumentation, 2016, 53(1): 58-62.
[5]	叶剑斌, 朱东升, 汪翊节, 等. 智能电能表状态检验技术研究[J]. 自动化仪表, 2020, 41(1): 55-59. YE J B, ZHU D S, WANG Y J, et al. Research on intelligent electric energy meter state inspection technology[J]. Process Aotumation Instrumentation, 2020, 41(1): 55-59(in Chinese).
[6]	祝宇楠, 徐晴, 刘建, 等. 数据挖掘在智能电能表故障分析中的应用[J]. 电力工程技术, 2016, 35(5): 19-23. doi: 10.3969/j.issn.1009-0665.2016.05.005 ZHU Y N, XU Q, LIU J, et al. Application of data mining in fault analysis of intelligent electricity meters[J]. Power Engineering Technology, 2016, 35(5): 19-23(in Chinese). doi: 10.3969/j.issn.1009-0665.2016.05.005
[7]	韩笑. 基于不确定性信息融合的智能电能表状态评价方法研究[D]. 长春: 吉林大学, 2018: 22-58. HAN X. Research on state evaluation method of smart electricity meter based on uncertainty information fusion[D]. Changchun: Jilin University, 2018: 22-58(in Chinese).
[8]	CHENG Y Y, HOU X Z, XIAO J. Design and realization of condition management system of gateway electrical[J]. Electrical Measurement and Instrumentation, 2013, 50(8): 87-92.
[9]	SHEN X, ZHAO D N, CAO M. Research on energy metering device remote online monitoring and analysis platform based on internet of thing[J]. Electrical Measurement and Instrumentation, 2015, 52(1): 35-40.
[10]	王新刚, 吴颖, 张垠. 基于数据挖掘的智能电能表在线监测方法[J]. 电测与仪表, 2016, 53(13): 65-69. doi: 10.3969/j.issn.1001-1390.2016.13.013 WANG X G, WU Y, ZHANG Y. Online monitoring method of intelligent electricity meters based on data mining[J]. Electric Measurement and Instrumentation, 2016, 53(13): 65-69(in Chinese). doi: 10.3969/j.issn.1001-1390.2016.13.013
[11]	LONG M S, WANG J, DING G, et al. Transfer feature learning with joint distribution adaptation[C]//2013 IEEE International Conference on Computer Vision. Piscataway: IEEE Press, 2013: 1-8.
[12]	CHAWLA N V, BOWYER K W, HALL L O, et al. SMOTE: Synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research, 2011, 16(1): 321-357.
[13]	BORGWARDT K M, GRETTON A, RASCH M J, et al. Integrating structured biological data by kernel maximum mean discrepancy[J]. Bioinformatics, 2006, 22(14): 49-57. doi: 10.1093/bioinformatics/btl242
[14]	FILISBINO T A, GIRALDI G A, THOMAZ C E. Support vector machine ensembles for discriminant analysis for ranking principal components[J]. Multimedia Tools and Applications, 2020, 79(5): 1-37.
[15]	PAN S J, TSANG I W, KWOK J T, et al. Domain adaptation via transfer component analysis[J]. IEEE Transactions on Neural Networks, 2011, 22(2): 199-210. doi: 10.1109/TNN.2010.2091281
[16]	GRETTON A, BORGWARDT K M, RASCH M J, et al. A kernel method for the two-sample problem[J]. Neural Information Processing Systems, 2006, 4(3): 1-10.
[17]	FAN J, KE Z T, LIU H, et al. QUADRO: A supervised dimension reduction method via Rayleigh quotient optimization[J]. Annals of Statistics, 2015, 4(43): 1498-1534.
[18]	龙明盛. 迁移学习问题与方法研究[D]. 北京: 清华大学, 2014. LONG M S. Research on transfer learning problems and methods[D]. Beijing: Tsinghua University, 2014(in Chinese).