Research on multi-scale load prediction in parks based on CNN-LSTM-Self attention

doi:10.3969/j.issn.2097-0706.2025.02.008

Abstract

Abstract:

Accurate load prediction is critical for improving the energy efficiency and profitability of zero-carbon smart parks. However， the application of conventional load prediction techniques faces two main challenges which are the difficulty in obtaining hourly numerical weather forecast data and the need for predictions across different time scales. In the absence of weather forecast data， a method using Convolutional Neural Networks（CNN） was proposed to extract the coupled spatial features between multiple loads. The reconstructed features were input into a Long Short-Term Memory（LSTM） network to extract temporal features of the load， followed by the application of a self-attention mechanism to enhance the model's ability to extract feature information. A fully connected network was then employed for load prediction， resulting in a multi-variable-load， multi-time-scale prediction model based on CNN-LSTM-Self attention. A case study of a park was used to predict its cooling， heating， and electrical loads for the next 1 hour， 1 day， and 1 week. Experimental results showed that the CNN-LSTM-Self attention model outperformed the CNN， LSTM， and CNN-LSTM models in terms of prediction accuracy across multiple time scales. Specifically， the CNN-LSTM-Self attention model showed a more significant advantage in predicting the 1-hour load， with the mean absolute percentage error（MAPE） of cooling， heating， and electrical load predictions improved by 16.25%， 19.16%， and 10.24%， respectively， compared to the CNN-LSTM model.

Key words: zero-carbon smart park, load prediction, multi-time scale, convolutional neural network, long short-term memory neural network, self-attention mechanism

CLC Number:

TM76

YANG Lanqian, GUO Jinmin, TIAN Huili, HUANG Chang, LIU Min, CAI Yang. Research on multi-scale load prediction in parks based on CNN-LSTM-Self attention[J]. Integrated Intelligent Energy, 2025, 47(2): 79-87.

Figures/Tables 13

Fig.1

Fig.2

Fig.3

Fig.4

Table 1

Table 2

Table 3

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Fig.10

References 29

[1]	陈艳波, 张宁, 李嘉祺, 等. 零碳园区研究综述及展望[J]. 中国电机工程学报, 2024, 44(14): 5496-5517.
	CHEN Yanbo, ZHANG Ning, LI Jiaqi, et al. Review and prospect of zero carbon park research[J]. Proceedings of the CSEE, 2024, 44(14): 5496-5517.
[2]	刘天阳, 高亚静, 谢典, 等. 功能型零碳园区建设路径分析[J]. 综合智慧能源, 2023, 45(8): 44-52. doi: 10.3969/j.issn.2097-0706.2023.08.006
	LIU Tianyang, GAO Yajing, XIE Dian, et al. Analysis on the construction path of functional zero-carbon parks[J]. Integrated Intelligent Energy, 2023, 45(8): 44-52. doi: 10.3969/j.issn.2097-0706.2023.08.006
[3]	葛磊蛟, 李京京, 李昌禄, 等. 面向零碳园区的综合能源系统优化运行技术综述[J]. 电网技术, 2024, 48(5): 1821-1835.
	GE Leijiao, LI Jingjing, LI Changlu, et al. Overview of integrated energy system optimal operation technology for zero-carbon parks[J]. Power System Technology, 2024, 48(5): 1821-1835.
[4]	吉博文, 吴张傲. 指数平滑法在电力系统负荷预测中的应用[J]. 科技创新与应用, 2018, 8(30): 173-174.
	JI Bowen, WU Zhang'ao. Application of exponential smoothing method in power system load forecasting[J]. Technology Innovation and Application, 2018, 8(30): 173-174.
[5]	PARREÑO S J E. Forecasting electricity consumption in the Philippines using ARIMA models[J]. International Journal of Machine Learning and Computing, 2022, 12(6): IJMLC.2022. 12.6.1112.
[6]	罗权. 基于自适应卡尔曼滤波在气象影响下负荷预测[J]. 计算机测量与控制, 2020, 28(1): 156-159, 165.
	LUO Quan. Short-term load forecasting under meteorological influence based on adaptive Kalman filter[J]. Computer Measurement & Control, 2020, 28(1): 156-159, 165.
[7]	安颖坤, 朱永丹. 基于线性回归法和指数平滑法对电力负荷的预测[J]. 电力设备管理, 2021(5): 177-179.
	AN Yingkun, ZHU Yongdan. Power load forecasting based on linear regression method and exponential smoothing method[J]. Electric Power Equipment Management, 2021(5): 177-179.
[8]	CHAPAGAIN K, KITTIPIYAKUL S. Short-term electricity demand forecasting with seasonal and interactions of variables for thailand[C]// 2018 International Electrical Engineering Congress (iEECON). IEEE, 2018: 1-4.
[9]	AMIN M A A, HOQUE M A. Comparison of ARIMA and SVM for short-term load forecasting[C]// 2019 9th Annual Information Technology, Electromechanical Engineering and Microelectronics Conference (IEMECON). IEEE, 2019: 1-6.
[10]	高明, 郝妍. 基于BiLSTM网络与误差修正的超短期负荷预测[J]. 综合智慧能源, 2023, 45(1): 31-40. doi: 10.3969/j.issn.2097-0706.2023.01.004
	GAO Ming, HAO Yan. Ultra-short-term load forecasting based on BiLSTM network and error correction[J]. Integrated Intelligent Energy, 2023, 45(1): 31-40. doi: 10.3969/j.issn.2097-0706.2023.01.004
[11]	CHAFAK T, NUR S, CENK G, et al. Short term load forecasting based on ARIMA and ANN approaches[J]. Energy Reports, 2023, 9(S3): 550-557.
[12]	庄立生. 融合气象特征的BP神经网络电力系统短期负荷预测[J]. 山东电力技术, 2023, 50(11): 51-59.
	ZHUANG Lisheng. Short-term load forecasting for power systems based on BP neural networks[J]. Shandong Electric Power, 2023, 50(11): 51-59.
[13]	AGUILAR M E, ANTONIO N. Short-term electricity load forecasting with machine learning[J]. Information, 2021, 12(2): 50.
[14]	COSTAS M C, VILLANUEVA D, OLLER P E, et al. Load forecasting with machine learning and deep learning methods[J]. Applied Sciences, 2023, 13(13): APP13137933.
[15]	IOANNIS P, ELENI V, PAPAKOSTAS GEORGE A. A survey on deep learning for building load forecasting[J]. Mathematical Problems in Engineering, 2022, 2022: 1008491.
[16]	ZABOLI A, KASIMALLA S R, PARK K, et al. A comprehensive review of behind-the-meter distributed energy resources load forecasting: Models, challenges, and emerging technologies[J]. Energies, 2024, 17(11): en17112534.
[17]	陈胜, 刘鹏飞, 王平, 等. 基于LSTM人工神经网络的电力系统负荷预测方法[J]. 沈阳工业大学学报, 2024, 46(1): 66-71. doi: 10.7688/j.issn.1000-1646.2024.01.12
	CHEN Sheng, LIU Pengfei, WANG Ping, et al. Load forecasting method of power system based on LSTM artificial neural network[J]. Journal of Shenyang University of Technology, 2024, 46(1): 66-71. doi: 10.7688/j.issn.1000-1646.2024.01.12
[18]	满达, 张卓凡, 张金金, 等. 基于LSTM的高校建筑电力负荷预测方法[J]. 建筑电气, 2021, 40(11): 58-63.
	MAN Da, ZHANG Zhuofan, ZHANG Jinjin, et al. Power load forecast method for university buildings based on LSTM[J]. Building Electricity, 2021, 40(11): 58-63.
[19]	张腾达, 李琦, 陈波. 基于LSTM的热力站短期热负荷预测研究[J]. 计算机仿真, 2022, 39(9): 507-512.
	ZHANG Tengda, LI Qi, CHEN Bo. Research on short-term heat load forecasting of thermal power station based on LSTM[J]. Computer Simulation, 2022, 39(9): 507-512.
[20]	ZANG H X, XU R Q, CHENG L L, et al. Residential load forecasting based on LSTM fusing self-attention mechanism with pooling[J]. Energy, 2021, 229: 120682.
[21]	SONG J C, ZHANG L Y, XUE G X, et al. Predicting hourly heating load in a district heating system based on a hybrid CNN-LSTM model[J]. Energy and Buildings, 2021, 243: 110998.
[22]	RAFI S H, DEEBA S R, HOSSAIN E. A short-term load forecasting method using integrated CNN and LSTM network[J]. IEEE Access, 2021, 9: 32436-32448.
[23]	GUO Y X, LI Y, QIAO X B, et al. BiLSTM multitask learning-based combined load forecasting considering the loads coupling relationship for multi energy system[J]. IEEE Transactions on Smart Grid, 2022, 13(5): 3481-3492.
[24]	田浩含, 张智晟, 于道林. 基于改进LSTM的区域综合能源系统多元负荷短期预测研究[J]. 电力系统及其自动化学报, 2021, 33(9): 130-137.
	TIAN Haohan, ZHANG Zhisheng, YU Daolin. Research on multi-load short-term forecasting of regional integrated energy system based on improved LSTM[J]. Proceedings of the CSU-EPSA, 2021, 33(9): 130-137.
[25]	张文栋, 刘子琨, 梁涛, 等. 基于CNN-LSTM的综合能源系统负荷预测模型[J]. 重庆邮电大学学报(自然科学版), 2023, 35(2): 254-262.
	ZHANG Wendong, LIU Zikun, LIANG Tao, et al. Load prediction model of integrated energy system based on CNN-LSTM[J]. Journal of Chongqing University of Posts and Telecommunications (Natural Science Edition), 2023, 35(2): 254-262.
[26]	LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[27]	GRAVES A. Supervised sequence labelling with recurrent neural networks[M]. Berlin: 2012.
[28]	徐聪, 胡永锋, 张爱平, 等. 基于特征筛选的综合能源系统多元负荷日前-日内预测[J]. 综合智慧能源, 2024, 46(3): 45-53. doi: 10.3969/j.issn.2097-0706.2024.03.006
	XU Cong, HU Yongfeng, ZHANG Aiping, et al. Multi-load day-ahead and intra-day forecasting for integrated energy systems based on feature screening[J]. Integrated Intelligent Energy, 2024, 46(3): 45-53. doi: 10.3969/j.issn.2097-0706.2024.03.006
[29]	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017, 30: 5998-6008.

参数	数值
卷积、池化核大小	3
卷积、池化步幅	1
CNN卷积层层数	1
CNN池化层层数	1
CNN输出通道大小	32
LSTM层数	2
LSTM隐藏单元数	64

输入变量	输出变量
电、冷负荷	夏季冷负荷
电、冷、热负荷	电负荷
电、热负荷	冬季热负荷

时间尺度	输入变量	输出变量
1 h	前4 h负荷数据	未来1 h负荷数据
1 d	前2 d负荷数据	未来1 d负荷数据
1 周	前1 周负荷数据	未来1 周负荷数据