| 引用本文: | 郭红霞,陆进威,杨苹,刘泽健.非侵入式负荷监测关键技术问题研究综述[J].电力自动化设备,2021,41(1): |
| GUO Hongxia,LU Jinwei,YANG Ping,LIU Zejian.Review on key techniques of non-intrusive load monitoring[J].Electric Power Automation Equipment,2021,41(1): |
|
| 摘要: |
| 非侵入式负荷监测(NILM)技术能够利用在总线处单点测量的数据识别用户内部的负荷，是建设泛在电力物联网与透明电网的基础技术之一。在分析NILM基本实现框架和技术体系的基础上，对NILM应用亟需解决的三大关键技术问题进行综述，包括数据源选择、算法精度和可扩展性问题。在数据源选择问题上，分析并总结了低频与高频数据源在NILM中的应用，尤其是智能电表在NILM中的应用；在算法精度问题上，对现有NILM算法模型与算法评估方案进行了回顾与分析；而针对目前少有研究涉及可扩展性问题，通过联动NILM与语音识别和机器学习领域，对去噪识别与新负荷的标记和训练问题进行分析与探讨。最后对NILM的未来发展趋势与应用进行了展望。 |
| 关键词: 非侵入式负荷监测 机器学习 智能电表 泛在电力物联网 透明电网 |
| DOI：10.16081/j.epae.202011001 |
| 分类号:TM714 |
| 基金项目:广东省科技计划项目(2017B030314124)；广东大学生科技创新培育专项资金(“攀登计划”专项资金)资助项目(pdjh2019b0038) |
|
| Review on key techniques of non-intrusive load monitoring |
|
GUO Hongxia1,2, LU Jinwei1,2, YANG Ping1,2, LIU Zejian1,2
|
|
1.School of Electric Power, South China University of Technology, Guangzhou 510640, China;2.Guangdong Key Laboratory of Clean Energy Technology, South China University of Technology, Guangzhou 510640, China
|
| Abstract: |
| NILM(Non-Intrusive Load Monitoring) technology can identify the users’ internal load by using the data measured at a single point on the bus, and it is one of the basic technologies for the construction of Ubiquitous Power Internet of Things and transparent power grid. Based on the analysis of the basic implementation framework and technical system of NILM, three key technical problems that need to be solved in NILM application are summarized, including data source selection, algorithm accuracy and scalability. In terms of data source selection, the application of low frequency and high frequency data source in NILM is analyzed and summarized, especially the application of smart meter in NILM. In the aspect of algorithm accuracy, the existing NILM algorithm model and algorithm evaluation scheme are reviewed and analyzed. In view of the problem that there are few researches related to scalability, the denoising recognition and the labeling and training of new load are analyzed and discussed through combining NILM with speech recognition and machine learning. Finally, the future development trend and application of NILM are prospected. |
| Key words: non-intrusive load monitoring machine learning smart meter Ubiquitous Power Internet of Things transparent power grid |
