Multi-load day-ahead and intra-day forecasting for integrated energy systems based on feature screening

doi:10.3969/j.issn.2097-0706.2024.03.006

Abstract

Abstract:

Load forecasting is a prerequisite for guiding the scheduling and operation of integrated energy systems（IES）. In order to carry out day-ahead scheduling and intra-day operation optimization on IES more economically and efficiently， a multi-load day-ahead and intra-day forecasting method based on feature screening is proposed. Firstly， by combining three types of feature screening methods in feature engineering，input features of forecasting models are selected. The combining method simplifies the models while preserving the important features， and the input feature sets for day-ahead and intra-day forecasting models are selected respectively. Then， taking hard parameter sharing in multi-task learning， the forecasting models are established based on long short-term memory neural network， achieving information sharing among different subtasks. And the forecasting accuracies of the models are optimized through random search method. Finally， taking an industrial park in Beijing as a study case，its energy system's electricity and heat loads are analyzed，and the comprehensive accuracies of the day-ahead and intra-day forecasting reach 91.3% and 95.2%，respectively. The method provides a sound support for IES day-ahead scheduling and intra-day operation optimization. Compared with the results of forecasting without feature screening and the forecasting on a single load， the method proposed has a higher forecasting accuracy.

Key words: integrated energy system, multi-load, feature screening, day-ahead and intra-day forecasting, multi-task learning, long short-term memory neural network

CLC Number:

TK01

XU Cong, HU Yongfeng, ZHANG Aiping, YOU Changfu. 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.

Figures/Tables 19

Fig.1

Fig.2

Fig.3

Table 1

LSTM neural network parameters

参数	含义	参数	含义
$x t$	当前时刻输入值	$h t$	当前时刻隐藏层输出
$f t$	遗忘门控制信号	$h t - 1$	上一时刻隐藏层输出
$i t$	输入门控制信号	$C t$	当前时刻细胞状态
$o t$	输出门控制信号	$C t - 1$	前一时刻细胞状态
$C t ~$	信息候选状态

Table 1

Fig.4

Fig.5

Table 2

Table 3

Table 4

Table 5

Fig.6

Fig.7

Fig.8

Fig.9

Table 6

Fig.10

Fig.11

Table 7

Table 8

References 28

[1]	别朝红, 王旭, 胡源. 能源互联网规划研究综述及展望[J]. 中国工程电机学报, 2017, 37(22):6445-6462.
	BIE Chaohong, WANG Xu, HU Yuan. Review and prospect of planning of energy internet[J]. Proceedings of the CSEE, 2017, 37(22):6445-6462.
[2]	曾鸣, 王永利, 张硕, 等. 综合能源系统[M]. 北京: 中国电力出版社, 2020.
[3]	付子昱. 基于负荷预测的综合能源系统混合时间尺度优化运行[D]. 北京: 华北电力大学, 2021.
	FU Ziyu. Hybrid-time scale optimal scheduling of integrated energy systems with load forecasting[D]. Beijing: North China Electric Power University, 2021.
[4]	朱继忠, 董瀚江, 李盛林, 等. 数据驱动的综合能源系统负荷预测综述[J]. 中国电机工程学报, 2021, 41(23):7905-7923.
	ZHU Jizhong, DONG Hanjiang, LI Shenglin, et al. Review of data-driven load forecasting for integrated energy system[J]. Proceedings of the CSEE, 2021, 41(23):7905-7923.
[5]	史佳琪, 谭涛, 郭经, 等. 基于深度结构多任务学习的园区型综合能源系统多元负荷预测[J]. 电网技术, 2018, 42(3):698-706.
	SHI Jiaqi, TAN Tao, GUO Jing, et al. Multi-task learning based on deep architecture for various types of load forecasting in regional energy system integration[J]. Power System Technology, 2018, 42(3):698-706.
[6]	CHEN B Y, WANG Y F. Short-term electric load forecasting of integrated energy system considering nonlinear synergy between different loads[J]. IEEE Access, 2021, 9:43562-43573. doi: 10.1109/ACCESS.2021.3066915
[7]	王永利, 周泯含, 姚苏航, 等. 基于多能耦合机理的综合能源系统多元负荷协同预测模型[J]. 华北电力大学学报, 2022, 49(2):118-126.
	WANG Yongli, ZHOU Minhan, YAO Suhang, et al. Multi-load cooperative prediction model of integrated energy system based on multi-energy coupling mechanism[J]. Journal of North China Electric Power University, 2022, 49(2):118-126.
[8]	吕忠麟, 顾洁, 孟璐. 基于耦合特征与多任务学习的综合能源系统短期负荷预测[J]. 电力系统自动化, 2022, 46(11):58-66.
	LV Zhonglin, GU Jie, MENG Lu. Short-term load forecasting for integrated energy system based on coupling features and multi-task learning[J]. Automation of Electric Power Systems, 2022, 46(11):58-66.
[9]	金立, 张力, 唐杨. 考虑温湿指数与耦合特征的综合能源负荷短期预测[J]. 综合智慧能源, 2023, 45(7):70-77. doi: 10.3969/j.issn.2097-0706.2023.07.008
	JIN Li, ZHANG Li, TANG Yang. Short-term prediction on integrated energy loads considering temperature-humidity index and coupling characteristics[J]. Integrated Intelligent Energy, 2023, 45(7):70-77. doi: 10.3969/j.issn.2097-0706.2023.07.008
[10]	李家钊. 基于特征迁移学习与耦合关系挖掘的综合能源系统小样本日前负荷预测[D]. 徐州: 中国矿业大学, 2021.
	LI Jiazhao. Small-sample day-ahead load forecasting of integrated energy system based on feature transfer learning and coupling relationship mining[D]. Xuzhou: China University of Mining and Technology, 2021.
[11]	翟晶晶, 吴晓蓓, 王力立. 基于径向基函数神经网络的综合能源系统多元负荷短期预测[J]. 电力需求侧管理, 2019, 21(4):23-17.
	ZHAI Jingjing, WU Xiaobei, WANG Lili. Multiple short-term load forecasting in integrated energy system based on RBF-NN model[J]. Power Demand Side Management, 2019, 21(4):23-17.
[12]	马建鹏, 龚文杰, 张智晟. 基于Copula理论与KPCA-GRNN结合的区域综合能源系统多元负荷短期预测模型[J]. 电工电能新技术, 2020, 39(3):24-31. doi: 10.12067/ATEEE1903001
	MA Jianpeng, GONG Wenjie, ZHANG Zhisheng. Short-term multiple load prediction model for regional integrated energy system based on Copula theory and KPCA-GRNN[J]. Advanced Technology of Electrical Engineering and Energy, 2020, 39(3):24-31. doi: 10.12067/ATEEE1903001
[13]	何桂雄, 金璐, 李克成, 等. 基于改进 DaNN 的综合能源系统多能负荷预测[J]. 电力工程技术, 2021, 40(6):25-33.
	HE Guixiong, JIN Lu, LI Kecheng, et al. Multiple energy load forecasting of integrated energy system based on improved DaNN[J]. Electric Power Engineering Technology, 2021, 40(6):25-33.
[14]	王玥, 鞠振河, 汤梓涵. 基于GA-BP神经网络冷热电联供系统的负荷预测[J]. 电工技术, 2022(7):40-42,189.
	WANG Yue, JU Zhenhe, TANG Zihan. GA-BP neural network based load prediction of combined cooling, heating and power energy supply[J]. Electric Engineering, 2022(7):40-42,189.
[15]	陈锦鹏, 胡志坚, 陈纬楠, 等. 二次模态分解组合DBiLSTM-MLR的综合能源系统负荷预测[J]. 电力系统自动化, 2021, 45(13):85-94.
	CHEN Jinpeng, HU Zhijian, CHEN Weinan, et al. Load prediction of integrated energy system based on combination of quadratic modal decomposition and deep bidirectional long short-term memory and multiple linear regression[J]. Automation of Electric Power Systems, 2021, 45(13):85-94.
[16]	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(Nature Science Edition), 2023, 35(2):254-262.
[17]	王琛, 王颖, 郑涛, 等. 基于ResNet-LSTM网络和注意力机制的综合能源系统多元负荷预测[J]. 电工技术学报, 2022, 37(7):1789-1799.
	WANG Chen, WANG Ying, ZHENG Tao, et al. Multi-energy load forecasting in integrated energy system based on ResNet-LSTM network and attention mechanism[J]. Transactions of China Electrotechnical Society, 2022, 37(7):1789-1799.
[18]	廖清阳, 王军, 胡凯强, 等. 基于深度并行CNN-BiLSTM 的能源互联网电负荷和热负荷联合预测模型[J]. 中国测试, 2022, 48(4):146-153.
	LIAO Qingyang, WANG Jun, HU Kaiqiang, et al. Joint forecasting model of electrical load and thermal load based on deep parallel CNN-BiLSTM in energy internet[J]. China Measurement & Test, 2022, 48(4):146-153.
[19]	YAO Z J, ZHANG T Y, WANG Q M, et al. Short-term power load forecasting of integrated energy system based on attention-CNN-DBILSTM[J]. Mathematical Problems in Engineering, 2022, 1:12.
[20]	ZHENG J Y, ZHANG L Y, CHEN J P, et al. Multiple-load forecasting for integrated energy system based on Copula-DBiLSTM[J]. Energies, 2021, 14(8):2188-2201. doi: 10.3390/en14082188
[21]	王愈轩, 刘尔佳, 黄永章. 数据驱动下的综合能源系统短期多元负荷预测[J]. 计算机工程与设计, 2022, 43(5):1435-1442.
	WANG Yuxuan, LIU Erjia, HUANG Yongzhang. Data driven short-term multiple load forecasting for integrated energy system[J]. Computer Engineering and Design, 2022, 43(5):1435-1442.
[22]	LI C, LI G J, WANG K Y, et al. A multi-energy load forecasting method based on parallel architecture CNN-GRU and transfer learning for data deficient integrated energy systems[J]. Energy, 2022, 259(15):124967. doi: 10.1016/j.energy.2022.124967
[23]	LIU Y J, LI Y, LI G L, et al. Review of multiple load forecasting method for integrated energy system[J]. Frontiers In Energy Reasearch, 2023, 11:1296800.
[24]	孙庆凯, 王小君, 张义志, 等. 基于LSTM 和多任务学习的综合能源系统多元负荷预测[J]. 电力系统自动化, 2021, 45(5):63-70.
	SUN Qingkai, WANG Xiaojun, ZHANG Yizhi, et al. Multiple load prediction of integrated energy system based on long short-term memory and multi-task learning[J]. Automation of Electric Power Systems, 2021, 45(5):63-70.
[25]	吴晨, 姚菁, 薛贵元, 等. 基于MMoE多任务学习和长短时记忆网络的综合能源系统负荷预测[J]. 电力自动化设备, 2022, 42(7):33-39.
	WU Chen, YAO Jing, XUE Guiyuan, et al. Load forecasting of integrated energy system based on MMoE multi-task learning and LSTM[J]. Electric Power Automation Equipment, 2022, 42(7):33-39.
[26]	欧阳静, 杨吕, 尹康, 等. 基于ALIF-LSTM多任务学习的综合能源系统短期负荷预测[J]. 太阳能学报, 2022, 43(9):499-507. doi: 10.19912/j.0254-0096.tynxb.2020-1383
	OUYANG Jing, YANG Lv, YIN Kang, et al. Short-term load forecasting method for integrated energy system based on ALIF-LSTM and multi-task learning[J]. Acta Energiae Solaris Sinica, 2022, 43(9):499-507. doi: 10.19912/j.0254-0096.tynxb.2020-1383
[27]	章昊. 基于中、短时间尺度的春节负荷特征滚动预测[J]. 东北电力技术, 2022, 43(2):28-32.
	ZHANG Hao. Spring Festival load characteristic time-process forecasting based on medium and short time scales[J]. Northeast Electric Power Technology, 2022, 43(2):28-32.
[28]	秦爽, 朱秋成, 施宇丹. 基于MI-LSTM输电线动态参数修正的电力系统可用输电能力计算[J]. 东北电力技术, 2023, 44(11):1-6.
	QIN Shuang, ZHU Qiucheng, SHI Yudan. Available transfer capability calculation of power systems based on MI-LSTM transmission line dynamic parameter modification[J]. Northeast Electric Power Technology, 2023, 44(11):1-6.

项目	e	h	t	d	p	v	w
e	1.000	0.367	0.498	-0.242	-0.060	0.082	0.520
h	0.367	1.000	-0.348	-0.490	0.202	0.038	0.197

排名	递归特征消除法		XGBoost梯度提升树法
排名	e	h	e	h
1	w	d	w	d
2	t	t	t	t
3	p	w	v	p
4	d	p	p	v
5	v	v	d	w

参数名称	设定范围	参数名称	设定范围
epochs	10~200（间隔为10）	batch_size	4，8
num_layers	1，2，3	gamma	0.5
hidden_size	16，32，64，128	step_size	5

参数名称	值	参数名称	值
epochs	100	lr	0.1
num_layers	2	batch_size	8
hidden_size	64

指标	电负荷	热负荷	综合指标
MAE/kW	233.7	478.2	356.0
MAPE	0.084	0.090	0.087
δ/%	91.6	91.0	91.3