Research and Engineering Application of Non-intrusive Load Dynamic Identification Method

doi:10.11985/2023.03.033

Abstract

Abstract:

As the application scenarios of non-intrusive load monitoring technology continue to expand and load types become increasingly diversified, static identification methods based on single-layer features need to construct more comprehensive and complex features, and it is difficult to give consideration to both the accuracy and speed of load identification. A dynamic identification method based on multi-layer feature groups is proposed, which takes into account the sampling frequency, monitoring window width, computational complexity and load characteristic storage of different load features to construct a hierarchical feature group, extracts different feature groups for different load types in the load group as classification features to reduce the comprehensive extraction cost of features, and finally realizes on-demand load type recognition based on support vector machine multi-classification algorithm. Results from BLUED database simulation and actual engineering test of a factory show that the dynamic identification method based on multi-layer feature groups can increase the identification speed of most loads, while accurately distinguish among similar loads. It can give consideration to the accuracy and speed of load identification under the scenario of similar load and frequent switching, and has good engineering applicability.

Key words: Non-intrusive load monitoring, load feature hierarchy, dynamic identification, SVM multi-classification algorithm

CLC Number:

TM714

LIU Chunlei, PANG Pengfei, SHI Wenhe, KONG Linghao, HUANG Xun, QI Jun. Research and Engineering Application of Non-intrusive Load Dynamic Identification Method[J]. Journal of Electrical Engineering, 2023, 18(3): 307-314.

Figures/Tables 9

特征类型	特征名称	特征提取	采样频率/Hz	监测窗口宽度/s	计算复杂度	负荷特征存储量
稳态特征	有功功率P	$P = 1 N ∑ n = 0 N v (n) i (n)$	≈10³	>工频周期	低	小
	无功功率Q	$Q = 1 N ∑ n = 0 N v (n + N 4) i (n)$
	功率因数	$φ = P P 2 + Q 2$
	相角	$φ = arccos P P 2 + Q 2$
	电压-电流(V-I)轨迹	分别以目标设备的电压、电流瞬时值为横、纵坐标绘制图像			中	大
	电流谐波	对电流波形进行FFT变换分析	10³~10⁵		中	中
暂态特征	暂态电流峰值I_peak	$I p e a k = max {i (n), n = 1, 2, 3, ⋯, C}$	10³~10⁵	10^-2~10² (取决于实际负荷类型)	低	小
	暂态电流有效值I_rms	$I r m s = ∑ n = 0 N i (n) 2 N$
	暂态电流均值I_mean	$I m e a n = ∑ n = 0 C i (n) C$
	暂态电流方差I_var	$I v a r = ∑ n = 0 C i (n) − I m e a n 2 C$
	暂态时间T	$T = C f$
	S变换特征	对电流波形进行S变换分析			高	中
时间特征	负荷运行时段	统计负荷运行的常用时间段	10^-3~10²	10²~10⁵ (取决于用电规律)	低	小
	负荷工作周期性	统计负荷运行的时间周期
	负荷运行时长	统计负荷运行的单次时长

References 20

[1]	程祥, 李林芝, 吴浩, 等. 非侵入式负荷监测与分解研究综述[J]. 电网技术, 2016, 40(10):3108-3117.
	CHENG Xiang, LI Linzhi, WU Hao, et al. A survey of the research on non-intrusive load monitoring and disaggregation[J]. Power System Technology, 2016, 40(10):3108-3117.
[2]	李同智. 灵活互动智能用电的技术内涵及发展方向[J]. 电力系统自动化, 2012, 36(2):11-17.
	LI Tongzhi. Technical implications and development trends of flexible and interactive utilization of intelligent power[J]. Automation of Electric Power Systems, 2012, 36(2):11-17.
[3]	余贻鑫, 栾文鹏. 智能电网述评[J]. 中国电机工程学报, 2009, 29(34):1-8.
	YU Yixin, LUAN Wenpeng. Smart grid and its implementations[J]. Proceedings of the CSEE, 2009, 29(34):1-8.
[4]	余贻鑫. 智能电网的技术组成和实现顺序[J]. 南方电网技术, 2009, 3(2):1-5.
	YU Yixin. Technical composition of smart grid and its implementation sequence[J]. Southern Power System Technology, 2009, 3(2):1-5.
[5]	吴凡, 孙静. 基于多智能体技术的智能电网信息管理系统研究[J]. 现代电力, 2017, 34(2):87-94.
	WU Fan, SUN Jing. Research on smart grid information management system based on multi-agent technology[J]. Modern Electric Power, 2017, 34(2):87-94.
[6]	董朝阳, 陈莹莹, 罗逢吉. 未来主动配电网中的新型数据驱动应用:技术,展望与挑战[J]. 电力建设, 2017, 38(5):2-10.
	DONG Zhaoyang, CHEN Yingying, LUO Fengji. Innovative data-driven applications in future active distribution network:Technologies,prospect and challenges[J]. Electric Power Construction, 2017, 38(5):2-10.
[7]	张露. 非侵入式负荷监测算法研究[D]. 广州: 华南理工大学, 2019.
	ZHANG Lu. Research on non-intrusive load monitoring algorithms[D]. Guangzhou: South China University of Technology, 2019.
[8]	王晓换, 李如意, 周东国, 等. 基于决策融合的非侵入式电力负荷辨识方法及应用[J]. 电力系统保护与控制, 2016, 44(7):115-121.
	WANG Xiaohuan, LI Ruyi, ZHOU Dongguo, et al. Non-intrusive power load disaggregation method based on decision fusion and its applications[J]. Power System Protection and Control, 2016, 44(7):115-121.
[9]	张玉天, 邓春宇, 刘沅昆, 等. 基于卷积神经网络的非侵入负荷辨识算法[J]. 电网技术, 2020, 44(6):2038-2044.
	ZHANG Yutian, DENG Chunyu, LIU Yuankun, et al. Non-intrusive load identification algorithm based on convolution neural network[J]. Power System Technology, 2020, 44(6):2038-2044.
[10]	HART G W. Nonintrusive appliance load monitoring[J]. IEEE Proc., 1992(80):1870-1891.
[11]	王守相, 郭陆阳, 陈海文, 等. 基于特征融合与深度学习的非侵入式负荷辨识算法[J]. 电力系统自动化, 2020, 44(9):103-110.
	WANG Shouxiang, GUO Luyang, CHEN Haiwen, et al. Non-intrusive load identification algorithm based on feature fusion and deep learning[J]. Automation of Electric Power Systems, 2020, 44(9):103-110.
[12]	田正其, 徐晴, 李如意, 等. 一种融合时间特征的非侵入式负荷辨识决策方法[J]. 电测与仪表, 2022, 59(4):144-151.
	TIAN Zhengqi, XU Qing, LI Ruyi, et al. Non-intrusive load identification decision method based on time signatures[J]. Electrical Measurement & Instrumentation, 2022, 59(4):144-151.
[13]	黄友金, 熊炜, 袁旭峰, 等. 深度学习与幅值特征相融合的低频采样非侵入式负荷辨识算法[J]. 电力科学与工程, 2020, 36(4):10-16. doi: 1672-0792(2020)04-0010-07
	HUANG Youjin, XIONG Wei, YUAN Xufeng, et al. Non-intrusive load identification algorithm based on deep learning and amplitude feature fusion at low sampling frequency[J]. Electric Power Science and Engineering, 2020, 36(4):10-16. doi: 1672-0792(2020)04-0010-07
[14]	蔡志强. 基于智能电表的非侵入式负荷识别算法研究[D]. 天津: 天津大学, 2017.
	CAI Zhiqiang. Research on non-intrusive load identification algorithm based on smart meter[D]. Tianjin:Tianjin University, 2017.
[15]	MARTINS J F, LOPES R, LIMA C, et al. A novel nonintrusive load monitoring system based on the S-transform[C]// Optimization of Electrical and Electronic Equipment (OP-TIM),2012/13th International Conference on IEEE, 2012:973-978.
[16]	刘恒勇, 史帅彬, 徐旭辉, 等. 一种关联RNN模型的非侵入式负荷辨识方法[J]. 电力系统保护与控制, 2019, 47(13):162-170.
	LIU Hengyong, SHI Shuaibin, XU Xuhui, et al. A non-intrusive load identification method based on RNN model[J]. Power System Protection and Control, 2019, 47(13):162-170.
[17]	陈倩, 陈虎. 基于SVM的物质介电系数测量方法研究[J]. 机电信息, 2016(9):105-106.
	CHEN Qian, CHEN Hu. Research on dielectric coefficient measurement method based on SVM[J]. Mechanical and Electrical Information, 2016(9):105-106.
[18]	谭风雷, 张兆君, 吴兴泉, 等. 支持向量机优化的线性插值法在变压器油温预处理中的应用[J]. 现代电力, 2020, 37(6):591-597.
	TAN Fenglei, ZHANG Zhaojun, WU Xingquan, et al. Application of prediction accuracy interpolation method based on support vector machine optimization in transformer oil temperature preprocessing[J]. Modern Electric Power, 2020, 37(6):591-597.
[19]	林思沣, 冯德旺, 廖荣森, 等. SVM在大功率负荷下对小功率负荷识别研究[J]. 福建电脑, 2021, 37(3):69-70.
	LIN Sifeng, FENG Dewang, LIAO Rongsen, et al. Identification of small power load based on SVM with high power load[J]. Journal of Fujian Computer, 2021, 37(3):69-70.
[20]	ANDERSON K, OCNEANU A, BENITEZ D, et al. BLUED:A fully labeled public dataset for event-based non-intrusive load monitoring research[C]// Proceedings of the 2nd KDD Workshop on Data Mining Applications in Sustainability (SustKDD). ACM, 2012, 7:1-5.

特征分层	同层特征优先级
特征分层	高	中	低
第一层	P、Q	V-I	时间特征
第二层	I_p 、I_rms I_m、I_var、T	电流谐波	S变换特征

不同负荷	ω(%)				T_f(%)
不同负荷	方法①	方法②	方法③	方法①		方法②	方法③
计算机厨房顶灯办公室灯显示器浴室灯餐厅顶灯后院灯综合	92 94 94 90 92 53 54 81	98 96 92 98 98 92 94 95	92 94 94 90 92 92 94 92	100 100 100 100 100 100 100 100		456 477 430 450 457 440 442 459	100 100 100 100 100 522 510 218

指标	识别方法
指标	①	②	③
识别正确率(%) 平均识别时长/s 1 h系统监测时延/s 2 h系统监测时延/s	67.5 11.4 11 11	93.5 36.8 1 942 3 201	91 15.4 11 11