基于知识流和迁移学习的负荷预测

doi:10.11959/j.issn.1000-0801.2022074

摘要/Abstract

摘要：

在万物互联、全面感知、智能决策的大数据信息化时代，大数据信息的采集、大量信号的处理等仍存在数据冗余、计算量大、成本高、不及时和无特征性的缺点。通过迁移学习方法，利用基于权重影响因子进行信息融合的知识流动体系，为物联感知系统提供协助分析并简化计算。在物联感知系统采用迁移学习加数据融合的知识流动方式，以地区用电功率部分数据做短期负荷预测的仿真计算，分析用户用电行为影响因子，训练得到影响因子最佳权重分配，为用电耗率预判提供依据。结果表明，通过该方式，能够清晰辨别用电行为特征，并根据用电特征预判用电耗能。

关键词: 迁移学习, 信息融合, 知识流, 物联感知, 负荷预测

Abstract:

In all things connected, comprehensive perception, intelligent decision-making information era of big data, in the information acquisition of big data and a large amount of signal processing, there are still large amount of data redundancy, calculation and the shortcoming of high cost, not in time, and no marked.The transfer learning was applied, and the knowledge flow system based on the weight impact factor for information fusion was integrated to assist the analysis and simplify the calculation for the IoT sensing system.The knowledge flow mode of transfer learning and data fusion was adopted in the IoT sensing system, and the simulation calculation of short-term load prediction was made based on the partial data of regional power consumption.The influencing factors of users’ electricity consumption behavior were analyzed, and the optimal weight distribution of influencing factors was obtained through training, so as to predict the power consumption rate.The results show that in this way, it can clearly identify the characteristics of electricity consumption behavior, and predict the energy consumption according to the characteristics of electricity consumption.

Key words: transfer learning, information fusion, knowledge flow, IoT perception, load prediction

中图分类号:

祝文军, 王思宁, 高晓欣, 郑倩. 基于知识流和迁移学习的负荷预测[J]. 电信科学, 2022, 38(5): 114-123.

Wenjun ZHU, Sining WANG, Xiaoxin GAO, Qian ZHENG. Load forecasting based on knowledge flow and transfer learning[J]. Telecommunications Science, 2022, 38(5): 114-123.

图/表 10

表1

图1

图2

图3

图4

图5

表2

图6

表3

图7

参考文献 35

[1]	Internet of things industry perspective[EB]. 2021.
[2]	周峰, 周晖, 刁赢龙 . 泛在电力物联网智能感知关键技术发展思路[J]. 中国电机工程学报, 2020,40(1): 70-82,375.
	ZHOU F , ZHOU H , DIAO Y L . Development of intelligent perception key technology in the ubiquitous Internet of Things in electricity[J]. Proceedings of the CSEE, 2020,40(1): 70-82,375.
[3]	张宁, 马国明, 关永刚 ,等. 全景信息感知及智慧电网[J]. 中国电机工程学报, 2021,41(4): 1274-1283,1535.
	ZHANG N , MA G M , GUAN Y G ,et al. Panoramic information perception and intelligent grid[J]. Proceedings of the CSEE, 2021,41(4): 1274-1283,1535.
[4]	汤凌晗 . 物联网信息感知与交互技术[J]. 信息记录材料, 2020,21(11): 218-219.
	TANG L H . Internet of things information perception and inte-raction technology[J]. Information Recording Materials, 2020,21(11): 218-219.
[5]	雷煜卿, 仝杰, 张树华 ,等. 能源互联网感知层技术标准体系研究[J]. 供用电, 2021,38(7): 14-20,33.
	LEI Y Q , TONG J , ZHANG S H ,et al. Research on technical standard system of energy Internet sensing layer[J]. Distribution＆ Utilization, 2021,38(7): 14-20,33.
[6]	薛澜, 姜李丹, 黄颖 ,等. 资源异质性、知识流动与产学研协同创新:以人工智能产业为例[J]. 科学学研究, 2019,37(12): 2241-2251.
	XUE L , JIANG L D , HUANG Y ,et al. Resource heterogeneity,knowledge flow and synergy innovation of industry-university-research institute:an empirical study on AI industry[J]. Studies in Science of Science, 2019,37(12): 2241-2251.
[7]	丁晟春, 刘梦露, 傅柱 . 概念设计中基于知识流的多维设计知识统一建模技术研究[J]. 数据分析与知识发现, 2018,2(2): 11-19.
	DING S C , LIU M L , FU Z . Unified multidimensional model based on knowledge flow in conceptual design[J]. Data Analy-sis and Knowledge Discovery, 2018,2(2): 11-19.
[8]	王继业, 蒲天骄, 仝杰 ,等. 能源互联网智能感知技术框架与应用布局[J]. 电力信息与通信技术, 2020,18(4): 1-14.
	WANG J Y , PU T J , TONG J ,et al. Intelligent perception tech-nology framework and application layout of energy Internet[J]. Electric Power Information and Communication Technology, 2020,18(4): 1-14.
[9]	万灿, 崔文康, 宋永华 . 新能源电力系统概率预测：基本概念与数学原理[J]. 中国电机工程学报, 2021,41(19): 6493-6509.
	WAN C , CUI W K , SONG Y H . Probabilistic forecasting for power systems with renewable energy sources:basic concepts and mathematical principles[J]. Proceedings of the CSEE, 2021,41(19): 6493-6509.
[10]	杨挺, 赵黎媛, 王成山 . 人工智能在电力系统及综合能源系统中的应用综述[J]. 电力系统自动化, 2019,43(1): 2-14.
	YANG T , ZHAO L Y , WANG C S . Review on application of ar-tificial intelligence in power system and integrated energy sys-tem[J]. Automation of Electric Power Systems, 2019,43(1): 2-14.
[11]	田楚杰 . 神经网络在时间序列与时空序列流量预测中的应用与研究[D]. 北京:北京邮电大学, 2021.
	TIAN C J . The research of temporal flow prediction and spa-tial-temporal flow prediction based on neural networks[D]. Bei-jing:Beijing University of Posts and Telecommunications, 2021.
[12]	叶琳, 杨滢, 洪道鉴 ,等. 深度学习在电力系统中的应用研究综述[J]. 浙江电力, 2019,38(5): 83-89.
	YE L , YANG Y , HONG D J ,et al. A survey of deep learning technology application in power system[J]. Zhejiang Electric Power, 2019,38(5): 83-89.
[13]	蔡秋娜, 苏炳洪, 闫斌杰 ,等. 基于参数迁移的节假日短期负荷预测方法[J]. 电气自动化, 2020,42(4): 59-60,98.
	CAI Q N , SU B H , YAN B J ,et al. Forecasting method for hol-iday short-term load based on parameter transfer[J]. Electrical Automation, 2020,42(4): 59-60,98.
[14]	孙晓燕, 李家钊, 曾博 ,等. 基于特征迁移学习的综合能源系统小样本日前电力负荷预测[J]. 控制理论与应用, 2021,38(1): 63-72.
	SUN X Y , LI J Z , ZENG B ,et al. Small-sample day-ahead power load forecasting of integrated energy system based on feature transfer learning[J]. Control Theory ＆ Applications, 2021,38(1): 63-72.
[15]	张勇, 陶一凡, 巩敦卫 . 迁移学习引导的变源域长短时记忆网络建筑负荷预测[J]. 控制与决策, 2021,36(10): 2328-2338.
	ZHANG Y , TAO Y F , GONG D W . Load forecasting of build-ings using LSTM based on transfer learning with variable source domain[J]. Control and Decision, 2021,36(10): 2328-2338.
[16]	史凯钰, 张东霞, 韩肖清 ,等. 基于 LSTM 与迁移学习的光伏发电功率预测数字孪生模型[J]. 电网技术, 2022,46(4): 1363-1372.
	SHI K Y , ZHANG D X , HAN X Q ,et al. Digital twin model of photovoltaic power generation prediction based on LSTM and transfer learning[J]. Power System Technology, 2022,46(4): 1363-1372.
[17]	杨秀, 吴吉海, 孙改平 ,等. 基于深度学习和迁移学习的公共楼宇非侵入式负荷分解[J]. 电网技术, 2022,46(3): 1160-1169.
	YANG X , WU J H , SUN G P ,et al. Non-intrusive load decom-position of public buildings based on deep learning and transfer learning[J]. Power System Technology, 2022,46(3): 1160-1169.
[18]	LEE E , RHEE W . Individualized short-term electric load forecasting with deep neural network based transfer learning and meta learning[J]. IEEE Access, 2021(9): 15413-15425.
[19]	PAN S J , YANG Q . A survey on transfer learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2010,22(10): 1345-1359.
[20]	ZHUANG F Z , QI Z Y , DUAN K Y ,et al. A comprehensive survey on transfer learning[J]. Proceedings of the IEEE, 2021,109(1): 43-76.
[21]	LU J , BEHBOOD V , HAO P ,et al. Transfer learning using computational intelligence:a survey[J]. Knowledge-Based Systems, 2015(80): 14-23.
[22]	WU Z H , JIANG H K , ZHAO K ,et al. An adaptive deep transfer learning method for bearing fault diagnosis[J]. Measurement, 2020(151): 107227.
[23]	MANDIC D P , REHMAN N U , WU Z H ,et al. Empirical mode decomposition-based time-frequency analysis of multivariate signals:the power of adaptive data analysis[J]. IEEE Signal Processing Magazine, 2013,30(6): 74-86.
[24]	聂品磊, 费东, 王宏杰 ,等. 基于 EMD-BP 神经网络的短期电力负荷预测[J]. 化工自动化及仪表, 2016,43(3): 305-307,332.
	NIE P L , FEI D , WANG H J ,et al. Short-term power load fore-casting based on EMD-BP neural network[J]. Control and In-struments in Chemical Industry, 2016,43(3): 305-307,332.
[25]	彭晋, 王昕, 孟小楠 . 多方数据融合挖掘安全与隐私保护参考架构研究[J]. 信息技术与标准化, 2021(8): 36-39.
	PENG J , WANG X , MENG X N . Research on security and pri-vacy-preserving reference architecture for multi-party data fu-sion and mining[J]. Information Technology ＆ Standardization, 2021(8): 36-39.
[26]	LIU M , LI D D , LI Q K ,et al. An online intelligent method to calibrate radar and camera sensors for data fusing[J]. Journal of Physics:Conference Series, 2020,1631(1): 012183.
[27]	李志刚 . 基于特征融合的异构数据分类方法研究[D]. 济南:齐鲁工业大学, 2021.
	LI Z G . Research on the classification method of heterogeneous data based on feature fusion[D]. Jinan:Qilu University of Technology, 2021.
[28]	任泽裕, 王振超, 柯尊旺 ,等. 多模态数据融合综述[J]. 计算机工程与应用, 2021,57(18): 49-64.
	REN Z Y , WANG Z C , KE Z W ,et al. Survey of multimodal data fusion[J]. Computer Engineering and Applications, 2021,57(18): 49-64.
[29]	VERGARA C , MARTIN R K , COLLINS P J ,et al. Multi-sensor data fusion between radio tomographic imaging and noise radar[J]. IET Radar,Sonar ＆ Navigation, 2020,14(2): 187-193.
[30]	单政博, 张勇, 吴小刚 ,等. 基于数据融合的数字化电网调度指挥和管控平台[J]. 机械与电子, 2021,39(6): 44-47.
	SHAN Z B , ZHANG Y , WU X G ,et al. Digital power grid dis-patching command and control platform based on data fusion[J]. Machinery ＆ Electronics, 2021,39(6): 44-47.
[31]	陈红, 肖建红, 刘文婕 ,等. 电能计量中的数据融合分析[J]. 集成电路应用, 2021,38(6): 42-43.
	CHEN H , XIAO J H , LIU W J ,et al. Analysis of data fusion in electric energy measurement[J]. Application of IC, 2021,38(6): 42-43.
[32]	李俊双 . 基于深度学习的数据融合在空气质量监测的研究与应用[D]. 北京:中国科学院大学(中国科学院沈阳计算技术研究所), 2021.
	LI J S . Research and application of data fusion based on deep learning in air quality monitoring[D]. Beijing:Institute of Computing Technology,Chinese Academy of Sciences, 2021.
[33]	董真杰, 郑琛瑶, 张国龙 . 不同精度数据融合的自适应加权平均法研究[J]. 舰船电子工程, 2014,34(10): 31-33,126.
	DONG Z J , ZHENG C Y , ZHANG G L . Self-adaption of weighted average research for data fusion with different preci-sion[J]. Ship Electronic Engineering, 2014,34(10): 31-33,126.
[34]	王美蕴 . 基于多无线传感器的数据融合算法[J]. 电子测试, 2021(8): 73-74.
	WANG M Y . Data fusion algorithm based on multiple wireless sensors[J]. Electronic Test, 2021(8): 73-74.
[35]	黄婷婷, 冯锋 . 无线传感器网络异构数据融合模型优化研究[J]. 计算机科学, 2020,47(S2): 339-344.
	HUANG T T , FENG F . Study on optimization of heterogeneous data fusion model in wireless sensor network[J]. Computer Science, 2020,47(S2): 339-344.

	学习方式	D _S＆D_T	T _S＆T_T
	传统机器学习	相同	相同
迁移学习	归纳式迁移学习	相同/相关	相关
	无监督迁移学习	相关	相关
	直推式迁移学习	相关	相同

分量	影响因子
IMF1/IMF2/IMF3	温度、多个规律性因子的共同制约等
IMF4/IMF5/IMF6/IMF7	规律性因子制约

评估参数	特征权重数值融合前	特征权重数值融合后
MAPE	8.04	10.60
MASE	0.99	1.41