| 193 | 4 | 97 |
| 下载次数 | 被引频次 | 阅读次数 |
非侵入式工业负荷识别技术能在不改变现有电路结构的前提下对电力系统中的负荷组成进行识别，从而获取不同工业用户用电信息，对于工业园区的用户优化用电计划、节能降本具有重要意义。为解决目前针对工业负荷的非侵入式识别准确率低且实用性差等问题，提出一种基于小波随机森林(wavelet transform-randomized forest, WRF)的非侵入式多目标工业负荷识别方法。通过分析工业负荷波形特性，选取有效特征变量，使用小波变换提取变量的局部暂态和稳态特征，使用随机森林对所有特征进行多目标分类匹配，实现工业负荷识别。为验证所提出方法的有效性，将WRF算法的识别效果与目前常用的方法进行了对比，并对其鲁棒性做了进一步分析，结果表明所提出方法具有更高的准确性与更强的实用性。
Abstract：The non-intrusive industrial load identification technology can identify the load composition in the power system without changing the existing circuit structure. It′s significant for different industrial users to obtain the power consumption information through NILM, so as to optimize the power consumption plan and reduce costs. In order to solve the problems of low accuracy and poor practicability of the current non-invasive identification of industrial loads, this paper proposes a non-invasive multi-objective industrial load identification method based on wavelet transform-randomized forest(WRF). Firstly, the waveform characteristics of industrial load are analyzed, and effective characteristic variables are selected. Then, the local transient and steady-state characteristics of the variables are extracted by wavelet transform. Finally, the random forest is used to perform multi-objective classification and matching on all the characteristics to realize industrial load identification. In order to verify the validity of the proposed method, this paper has a work to compare the recognition effect of the WRF algorithm with the currently commonly used methods, and further analyzes its robustness. The results show that the proposed method has higher accuracy and stronger practicability.
[1] 李言瑞，安岩，孙善星，等.欧洲能源危机原因分析及对我国核电产业发展的启示[J].产业与科技论坛，2022,21(8):13-15.
[2]杨昆.贯彻“四个革命、一个合作”能源安全新战略,推动电力工业高质量发展[J].当代电力文化, 2019(7):12-15.
[3] LIU H. Non-intrusive load monitoring[M]. Beijing:Science Press,2020.
[4]余贻鑫,刘博,栾文鹏.非侵入式居民电力负荷监测与分解技术[J].南方电网技术, 2013, 7(4):1-5.
[5]祁兵,董超,武昕,等.基于DTW算法与稳态电流波形的非侵入式负荷辨识方法[J].电力系统自动化, 2018, 42(3):70-76.
[6]刘炜,谭兴,周克,等.基于SAGA-FCM算法的非侵入式负荷监测方法[J].现代电子技术, 2019, 42(23):72-76.
[7]孙毅,崔灿,陆俊,等.基于差量特征提取与模糊聚类的非侵入式负荷监测方法[J].电力系统自动化, 2017, 41(4):86-91.
[8]魏恩伟,李伟华,张之涵,等.基于改进匈牙利算法的非侵入式负荷匹配方法[J].电测与仪表, 2019, 56(22):58-64.
[9]李云峰,李欣,刘三丰,等.基于改进最近邻算法的非侵入式电器负荷识别[J].信息技术与网络安全, 2019, 38(7):84-87.
[10]周任飞,汤鹏飞,刘三丰,等.基于卷积神经网络的非侵入式负荷识别研究[J].信息技术与网络安全, 2019, 38(8):64-68.
[11]杨福生.小波变换的工程分析与应用[M].北京:科学出版社,1999.
[12]李宏博.基于小波分析的变压器异常特征量的提取[J].科技创新与应用, 2019(20):7-8.
[13]陈韬伟,辛明.基于小波变换的雷达辐射源信号特征提取[J].信息与电子工程, 2010, 8(4):436-440.
[14] BELGIU M, DRǎGU?L. Random forest in remote sensing:a review of applications and future directions[J]. ISPRS journal of photogrammetry and remote sensing, 2016, 114:24-31.
[15] LIANG J, NG S K K, KENDALL G, et al. Load signature study:Part I:Basic concept, structure, and methodology[J]. IEEE transactions on power Delivery, 2009, 25(2):551-560.
[16]史秋亮,林江.基于小波包分解与能量特征提取的相关分析法[J].声学与电子工程, 2010(4):18-20.
[17] SHI Qiuliang, LIN Jiang. Correlation analysis method based on wavelet packet decomposition and energy feature extraction[J]. Acoustics and Electronic Engineering, 2010(4):18-20.
[18] PHAM V L, WONG K P. Wavelet-transform-based algorithm for harmonic analysis of power system waveforms[J]. Generation, Transmission and Distribution, IEE Proceedings-, 1999, 146(3):249-254.
[19] BUTUCEA C, NDAOUD M, STEPANOVA N A, et al. Variable selection with Hamming loss[J]. The Annals of Statistics, 2018, 46(5):1837-1875.
基本信息:
DOI：10.13928/j.cnki.wrahe.2022.S2.055
中图分类号:TM73;TP181
引用信息:
[1]赵志刚,冯忠义,王咏欣,等.基于小波随机森林的非侵入式工业负荷识别及其鲁棒性研究[J].水利水电技术(中英文),2022,53(S2):265-270.DOI:10.13928/j.cnki.wrahe.2022.S2.055.
基金信息:
国网山西省电力公司科技项目资助“基于互联网平台的电力用户用能特性数据挖掘及用能服务策略研究”(5205J020000W)