祁兵,李鹏云,李彬
QI Bing, LI Pengyun, LI Bin

• 1:华北电力大学电气与电子工程学院，北京 102206

摘要(Abstract)：

在电力行业精细化发展的背景下，需求侧用户终端可调节潜力分析以及源网荷柔性互动研究已经成为迫切需求。针对上述问题，提出了基于边缘计算模式下的负荷可调节潜力计算方法，该方法适用于电力系统数据量大、调节潜力影响因素众多的情况。首先基于边缘计算技术，提出融合边缘计算的智能分析系统；其次，通过多维度考虑可调节潜力影响因素情况下，采用长短期记忆（Long Short-Term Memory，LSTM）神经网络算法对台区内电气参数数据进行分析建模，进而利用所建模型移植到智能配变终端中，对台区负荷可调节潜力进行分析计算；最后，通过某地区电力实际负荷数据进行仿真验证，分析不同方法、不同维度下电力负荷可调节潜力，在改进以往分析方法的同时，为后续需求侧负荷研究提供参考。
Under the background of the refined development of the power industry, the analysis of the adjustable potential of user terminals on the demand side and the research on the positive interaction of source network and load have become an urgent need. This paper proposes a calculation method of load adjustable potential based on edge calculation mode, which is suitable for power system with large amount of data and many influencing factors of regulating potential. Firstly, the structure frame of platform load edge calculation model is proposed for edge computing technology. Secondly, the LSTM neural network algorithm is adopted to analyze and model the electrical parameter data in the station area by considering the influencing factors of adjustable potential in multidimensional, and then the established model is transplanted to the intelligent transformer distribution terminal to analyze and calculate the adjustable potential of load in the station area. Finally, through the simulation validation of the power load data in a certain area, the size of the adjustable potential under different methods and different dimensions is obtained, and the analysis method is greatly improved.

关键词(KeyWords)： 源网荷;可调节潜力;边缘计算;LSTM;智能配变终端
source network load;adjustable potential;edge computing;LSTM;intelligent distribution terminal

Abstract:

Keywords:

基金项目(Foundation): 北京市重点实验室开放基金资助“需求侧多能互补优化与供需互动技术”（YD80-21-001）

作者(Author): 祁兵,李鹏云,李彬
QI Bing, LI Pengyun, LI Bin

DOI: 10.19929/j.cnki.nmgdljs.2022.0094

参考文献(References)：

• [1] 杨旭英,周明,李庚银.智能电网下需求响应机理分析与建模综述[J].电网技术, 2016, 40(1):220-226. YANG Xuying, ZHOU Ming, LI Gengyin. Demand re-sponse mechanism analysis and modeling review under smart grid[J]. Power System Technology, 2016, 40(1):220-226.
  [2] 闫华光,陈宋宋,杜重阳,等.智能电网电力需求响应标准体系研究与设计[J].电网技术, 2015, 39(10):2685-2689. YAN Huaguang, CHEN Songsong, DU Chongyang, et al. research and design of power demand response stan-dard system for smart grid[J]. Power System Technolo-gy, 2015, 39(10):2685-2689.
  [3] 王蓓蓓.面向智能电网的用户需求响应特性和能力研究综述[J].中国电机工程学报, 2014, 34(22):3654-3663. WANG Beibei. A review of customer demand response characteristics and capabilities for smart grid[J]. Pro-ceedings of the CSEE, 2014, 34(22):3654-3663.
  [4] 吴震宇.考虑净负荷峰谷差随机性的储能参与电网调峰配置与调度研究[D].长沙:长沙理工大学, 2019.
  [5] LI. L. Cloud Computing Data Center Structure Based on Internet of Things and Its Scheduling Mechanism[C]//IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA). Dalian, China, 2020:633-636.
  [6] K. Dolui, S. K. Datta. Comparison of edge computing implementations:Fog computing, cloudlet and mobile edge computing[C]//Mandat International. Global Internet of Things Summit (GIoTS). Switzerland, 2017:1-6.
  [7] 朱玉杰.电采暖负荷可调节能力评估与集群控制策略研究[D].吉林:东北电力大学, 2019.
  [8] HAN N H, GONG T R, GONG F X, et al. Evaluation and Cluster Control of Electric Heating Flexible Load Adjustable Capacity[C]//Sun Yat-sen University. Manage-ment Science Informatization and Economic Innovation Development Conference (MSIEID). Guangzhou, China:Sun yat-sen University, 2020:425-428.
  [9] 孙功伟,周克楠,马喆非,等.考虑柔性负荷的智能电网调度策略研究[J].国外电子测量技术, 2020, 39(3):40-46. SUN Gongwei, ZHOU Kenan, MA Zhefei, et al. Re-search on scheduling strategy of smart power grid with flexible load[J]. Foreign Electronic Measurement Tech-nology, 2020, 39(3):40-46.
  [10] 谈竹奎,汪元芹,赵菁,等.实时电价下用户侧电力需求响应模型优化策略及数字仿真[J].电力大数据, 2020, 23(3):10-18. TAN Zhukui, WANG Yuanqin, ZHAO Jing, et al. Opti-mization strategy and digital simulation of user-side power demand response model under real-time electric-ity price[J]. Power Systems and Big Data, 2020, 23(3):10-18.
  [11] 任惠,陆海涛,卢锦玲,等.考虑信息物理系统耦合和用户响应差异的负荷聚合商需求响应特性分析[J].电网技术, 2020, 44(10):3927-3936. REN Hui, LU Haitao, LU Jinling, et al. Analysis of demand response characteristics of load aggregators considering information physical system coupling and user response difference[J]. Power System Technology, 2020, 44(10):3927-3936.
  [12] 孙毅,毛烨华,李泽坤,等.面向电力大数据的用户负荷特性和可调节潜力综合聚类方法[J/OL].中国电机工程学报:1-14.[2021-12-08].https://doc.taixueshu.com/journal/20210559zgdjgc.xb.html. SUN Yi, MAO Yehua, LI Zekun, et al. A Comprehen-sive Clustering Method of User Load Characteristics and Adjustable Potential Based on Power Big Data[J/OL]. Proceedings of the CSEE:1-14[2021-12-08]. https://doc.taixueshu.com/journal/20210559zgdjgc.xb.html.
  [13] YU L, QIU X W, KONG X J, et al. Research and improvement of integrated debugging method for intelli-gent distribution terminal[C]//IEEE. China International Conference on Electricity Distribution, Shanghai, 2012:1-4.
  [14] 宋耐超,史雷敏,王小雷,等.基于边缘计算模式的台区线损率计算方法[J].上海电力大学学报, 2020, 36(5):471-476, 510. SONG Naichao, SHI Leimin, WANG Xiaolei, et al. Method for calculating transformer line loss based on edge calculation mode[J]. Journal of Shanghai Universi-ty of Electric Power, 2020, 36(5):471-476, 510.
  [15] 曹渝昆,巢俊乙,王晓飞.基于LSTM神经网络的风机齿轮带断裂故障预测[J].电气自动化, 2019, 41(4):92-95. CAO Yukun, CHAO Junyi, WANG Xiaofei. Prediction of wind turbine gear belt breakage based on the LSTM beural betwork[J]. Electrical Automation, 2019, 41(4):92-95.
  [16] M. Hajiaghayi, E. Vahedi. Code Failure Prediction and Pattern Extraction Using LSTM Networks[C]//2019 IEEE Fifth International Conference on Big Data Com-puting Service and Applications (Big Data Service). Nework, CA, USA, 2019:55-62.
  [17] K. Goswami, A. Ganguly, A. K. Sil. Day Ahead Forecasting and Peak Load Management using Multi-variate Auto Regression Technique[C]//Jadavpur Univer-sity. 2018 IEEE Applied Signal Processing Confer-ence (ASPCON) Kolkato, India:Jadavpur University, 2018:279-282.
  [18] NIU P X, WANG Q, LI J C. Shortterm load forecasting model using support vector machine based on artifi-cial neural network[C]//2005 International Conference on Machine Learning and Cybernetics, Guangzhou, China, 2005:4260-4265.
  [19] ZHOU K, ZHOU Y, DENG K. A circular recursion algorithm power load forecasting model and application[C]//Changsha University of Science&Technology. 2019 IEEE 3rd Conference on Energy Internet and Energy System Integration (EI2), Changsha, China:Changsha University of Science&Technology, 2019:2870-2873.
  [20] YAO W Y, HUANG P, JIA Z X. Multidimensional LSTM Networks to Predict Wind Speed[C]//China Uni-versity of Geosciences 201837 th Chinese Control Con-ference (CCC). Wuhan:China University of Geosciences, 2018:7493-7497.
  [21] 吴定安,钟建伟,王新磊,等.主成分分析和长短期记忆网络的电力负荷预测[J].物联网技术, 2021, 11(8):47-51. WU Dingan, ZHONG Jianwei, WANG Xinlei,et al. Power load forecasting based on principal component analysis and long short-term memory network[J]. Inter-net of Things Technologies, 2021, 11(8):47-51.
  [22] HU L, ZHANG L, WANG T, et al. Short-term load forecasting based on support vector regression consider-ing cooling load in summer[C]//2020 Chinese Control And Decision Conference (CCDC). Northeastern Universi-ty, 2020:5495-5498.