Prediction on loads and photovoltaic output coefficients based on Informer

doi:10.3969/j.issn.2097-0706.2024.04.008

Abstract

Abstract:

A long-term time series forecast on electric loads and photovoltaic output coefficients is crucial for the layout and installed capacity design of renewable energy. A photovoltaic output coefficient can reflect the power-generating efficiency of a photovoltaic system in operation. But the inaccurate prediction on meteorological information brings challenges to the prediction on the maximum daily photovoltaic output coefficient. To overcome the problem，an envelope is constructed based on the maximum and minimal daily photovoltaic output coefficients which are calculated every 7 days. The upper and lower limits of the envelope provide a range of the factor. The envelope can overcome the uncertainties of meteorological information， so as to facilitate the prediction to be more robust and reliable. The Informer model employed as the prediction framework in this study is compared with Transformer， LSTM， and RNN models. A simulation test is carried out on a data sequence of actual electric loads and the upper and lower bounds of the photovoltaic output coefficients' envelope. The results validate the feasibility and outstanding prediction accuracy of the Informer model.

Key words: renewable energy, load prediction, photovoltaic output coefficient, Informer model, long-term time series forecasting, installed capacity planning

CLC Number:

TK01：TM615

MIAO Yuesen, XIA Hongjun, HUANG Ningjie, LI Yun, ZHOU Shijie. Prediction on loads and photovoltaic output coefficients based on Informer[J]. Integrated Intelligent Energy, 2024, 46(4): 60-67.

Figures/Tables 17

Fig.1

Fig.2

Table 1

Pearson correlation classification

类别	极低相关	低相关	中度相关	强相关	极强相关
$r x y$	[0,0.2)	[0.2,0.4)	[0.4,0.6)	[0.6,0.8)	[0.8,1.0]

Table 1

Fig.3

Fig.4

Table 2

Time Encoding

日期	时间戳
2016-12-01	[4；1；336]
2016-12-02	[5；2；337]
$⋮$	$⋮$
2022-11-30	[3；30；334]

Table 2

Fig.5

Fig.6

Fig.7

Table 3

Fig.8

Fig.9

Fig.10

Table 4

Table 5

Table 6

Table 7

References 33

[1]	YORO K O, DARAMOLA M O. CO₂ emission sources, greenhouse gases, and the global warming effect[M]. Cambridge: Woodhead Publishing, 2020:3-28.
[2]	ZOU C N, PAN S Q, HAO Q. On the connotation, challenge and significance of China's "energy independence" strategy[J]. Petroleum Exploration and Development, 2020, 47(2): 449-462. doi: 10.1016/S1876-3804(20)60062-3
[3]	高世楫, 俞敏. 中国提出“双碳”目标的历史背景、重大意义和变革路径[J]. 新经济导刊, 2021(2):4-8.
[4]	PARIDA B, INIYAN S, GOIC R. A review of solar photovoltaic technologies[J]. Renewable and Sustainable Energy Reviews, 2011, 15(3): 1625-1636. doi: 10.1016/j.rser.2010.11.032
[5]	马志程, 李维俊, 周强, 等. 含光热电站的风光联合发电系统互补效益评估[J]. 综合智慧能源, 2023, 45(2): 1-9. doi: 10.3969/j.issn.2097-0706.2023.02.001
	MA Zhicheng, LI Weijun, ZHOU Qiang, et al. Evaluation on the complementary benefit of a wind-solar combined power generation system with a photothermal power station[J]. Integrated Intelligent Energy, 2023, 45(2): 1-9. doi: 10.3969/j.issn.2097-0706.2023.02.001
[6]	龚莺飞, 鲁宗相, 乔颖, 等. 光伏功率预测技术[J]. 电力系统自动化, 2016, 40(4):140-151.
	GONG Yingfei, LU Zongxiang, QIAO Ying, et al. An overview of photovoltaic energy system output forecasting technology[J]. Automation of Electric Power Systems, 2016, 40(4):140-151.
[7]	刘洋, 于海东, 刘文彬, 等. 基于“云边端”协同的分布式光伏柔性调节体系架构[J]. 山东电力技术, 2023, 50(7):23-29.
	LIU Yang, YU Haidong, LIU Wenbin, et al. Architecture of distributed photovoltaic flexible regulation based on cloud-edge-end collaboration[J]. Shandong Electric Power, 2023, 50(7):23-29.
[8]	贺国伟, 肖峻, 曹加胜. 考虑分布式光伏消纳的大型城市配电网规划实践[J]. 供用电, 2021, 38(10):26-34.
	HE Guowei, XIAO Jun, CAO Jiasheng. Planning practice of large urban distribution network considering distributed photovoltaic consumption[J]. Distribution & Utilization, 2021, 38(10):26-34.
[9]	王鑫, 陈祖翠, 卞在平, 等. 基于粒子群优化算法的智慧微电网风光储容量优化配置[J]. 综合智慧能源, 2022, 44(6): 52-58. doi: 10.3969/j.issn.2097-0706.2022.06.006
	WANG Xin, CHEN Zucui, BIAN Zaiping, et al. Optimal allocation of a wind-PV-battery hybrid system in smart microgrid based on particle swarm optimization algorithm[J]. Integrated Intelligent Energy, 2022, 44(6): 52-58. doi: 10.3969/j.issn.2097-0706.2022.06.006
[10]	路晓敏, 张明, 邓星, 等. 考虑多重不确定性的多站融合容量优化配置方法[J]. 综合智慧能源, 2022, 44(1):31-38. doi: 10.3969/j.issn.2097-0706.2022.01.005
	LU Xiaomin, ZHANG Ming, DENG Xing, et al. Optimal capacity configuration for multi-station integration considering multiple uncertainties[J]. Integrated Intelligent Energy, 2022, 44(1): 31-38. doi: 10.3969/j.issn.2097-0706.2022.01.005
[11]	SANSA I, BOUSSAADA Z, BELLAAJ N M. Solar radiation prediction using a novel hybrid model of ARMA and NARX[J]. Energies, 2021, 14(21): 6920. doi: 10.3390/en14216920
[12]	林女贵. 基于经验模态分解和 ARIMA 模型的地区售电量预测[J]. 电器与能效管理技术, 2018(7):73-77.
	LIN Nügui. Regional electricity sales forecasting based on empirical mode decomposition and ARIMA model[J]. Electrical & Energy Management Technology, 2018(7):73-77.
[13]	魏妍萍, 王军, 李南帆, 等. 基于Markov链的园区随机功率多场景预测模型[J]. 综合智慧能源, 2023, 45(1): 14-22. doi: 10.3969/j.issn.2097-0706.2023.01.002
	WEI Yanping, WANG Jun, LI Nanfan, et al. Prediction model of stochastic power in industrial parks based on Markov chain[J]. Integrated Intelligent Energy, 2023, 45(1): 14-22. doi: 10.3969/j.issn.2097-0706.2023.01.002
[14]	章熙, 郭翔, 万会江, 等. 一种基于BP神经网络的地区用电量预测方法研究[J]. 无线互联科技, 2019, 16(17):114-116.
	ZHANG Xi, GUO Xiang, WAN Huijiang, et al. Study a BP neural network based forecasting method for regional electricity consumption[J]. Wireless Internet Technology, 2019, 16(17):114-116.
[15]	MALVONI M, DE GIORGI M G, CONGEDO P M. Data on Support Vector Machines (SVM) model to forecast photovoltaic power[J]. Data in Brief, 2016, 9: 13-16. doi: 10.1016/j.dib.2016.08.024 pmid: 27622206
[16]	于秋玲, 许长清, 李珊, 等. 基于模糊聚类和支持向量机的短期光伏功率预测[J]. 电力系统及其自动化学报, 2016, 28(12):115-118.
	YU Qiuling, XU Changqing, LI Shan, et al. Application of fuzzy clustering algorithm and support vector machine to short-term forecasting of PV power[J]. Proceedings of the CSU-EPSA, 2016, 28(12):115-118.
[17]	XIA M, SHAO H D, MA X D, et al. A stacked GRU-RNN-based approach for predicting renewable energy and electricity load for smart grid operation[J]. IEEE Transactions on Industrial Informatics, 2021, 17(10): 7050-7059. doi: 10.1109/TII.2021.3056867
[18]	田杨阳, 张小斐, 耿俊成, 等. 基于改进聚类和 LSTM 的居民小区中长期负荷预测[J]. 河南电力, 2020(S2):58-63.
	TIAN Yangyang, ZHANG Xiaofei, GENG Juncheng, et al. Medium- and long-term load forecasting of residential communities based on improved clustering and LSTM[J]. Henan Electric Power, 2020(S2):58-63.
[19]	郑真, 朱峰, 马小丽, 等. 基于TL-LSTM的新能源功率短期预测[J]. 综合智慧能源, 2023, 45(1): 41-48. doi: 10.3969/j.issn.2097-0706.2023.01.005
	ZHENG Zhen, ZHU Feng, MA Xiaoli, er al. Short-term new energy power prediction based on TL-LSTM[J]. Integrated Intelligent Energy, 2023, 45(1): 41-48. doi: 10.3969/j.issn.2097-0706.2023.01.005
[20]	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
[21]	王海燕, 刘佳康, 邓亚平. 基于预估-校正综合BP神经网络的短期光伏功率预测[J]. 智慧电力, 2023, 51(3):46-52.
	WANG Haiyan, LIU Jiakang, DENG Yaping. Short-term photovoltaic power forecasting based on predict-correct combination BP neural network[J]. Smart Power, 2023, 51(3):46-52.
[22]	王粟, 隗磊锋, 曾亮. 基于GWO-SVM与随机森林的组合光伏功率预测模型[J]. 昆明理工大学学报(自然科学版), 2021, 46(5):82-88.
	WANG Su, WEI Leifeng, ZENG Liang. A combined model for photovoltaic power forecasting based on GWO-SVM and random forest[J]. Journal of Kunming University of Science and Technology(Natural Science), 2021, 46(5):82-88.
[23]	李昭昱, 艾芊, 张宇帆, 等. 基于attention机制的LSTM神经网络超短期负荷预测方法[J]. 供用电, 2019, 36(1):17-22.
	LI Zhaoyu, AI Qian, ZHANG Yufan, et al. A LSTM neural network method based on attention mechanism for ultra short-term load forecasting[J]. Distribution & Utilization, 2019, 36(1):17-22.
[24]	SAJJAD M, KHAN Z A, ULLAH A, et al. A novel CNN-GRU based hybrid approach for short-term residential load forecasting[J]. IEEE Access, 2020,8:143759-143768.
[25]	JEBLI I, BELOUADHA F Z, KABBAJ M I, et al. Prediction of solar energy guided by Pearson correlation using machine learning[J]. Energy, 2021, 224: 120109. doi: 10.1016/j.energy.2021.120109
[26]	陈新和, 宋月新, 张立国, 等. 模型与数据混合驱动的分布式光伏超短期功率预测[J]. 供用电, 2023, 40(1):2-9.
	CHEN Xinhe, SONG Yuexin, ZHANG Liguo, et al. Ultra-short-term power prediction of distributed photovoltaic driven by model and data hybrid[J]. Distribution & Utilization, 2023, 40(1):2-9.
[27]	谢宏文, 查浩, 陈庆文, 等. 一种使用CALSTM模型进行短期光伏功率预测的方法[J]. 供用电, 2020, 37(12):61-66.
	XIE Hongwen, ZHA Hao, CHEN Qingwen, et al. A method for short-term photovoltaic power prediction using CALSTM model[J]. Distribution & Utilization, 2020, 37(12):61-66.
[28]	张磊, 王小明, 吴红斌, 等. 基于FCM和LSTM的光伏功率超短期预测[J]. 供用电, 2023, 40(1):10-17.
	ZHANG Lei, WANG Xiaoming, WU Hongbin, et al. Ultra-short-term prediction of PV power based on FCM and LSTM[J]. Distribution & Utilization, 2023, 40(1):10-17.
[29]	史静, 李琥, 李冰洁, 等. 基于长短期记忆神经网络的配电网负荷预测方法研究[J]. 供用电, 2019, 36(7):71-77,90.
	SHI Jing, LI Hu, LI Bingjie, et al. Distribution network load forecasting based on LSTM neural network[J]. Distribution & Utilization, 2019, 36(7):71-77,90.
[30]	比利时Elia公司运行数据[EB/OL].[2023-10-05]. https://opendata.elia.be/pages/home/.
[31]	邱大伟, 刘子辰, 周一青, 等. 基于 Transformer 神经网络的滚动轴承故障类型识别[J]. 高技术通讯, 2021, 31(1):1-11.
	QIU Dawei, LIU Zichen, ZHOU Yiqing, et al. A novel fault type detection method of rolling bearing using Transformer neural networks[J]. Chinese High Technology Letters, 2021, 31(1):1-11.
[32]	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]// Advances in Neural Information Processing Systems, 2017:5998-6008.
[33]	ZHOU H Y, ZHANG S H, PENG J Q, et al. Informer: Beyond efficient transformer for long sequence time-series forecasting[EB/OL].(2020-12-14)[2023-10-02].https://doi.org/10.48550/arXiv.2012.07436.

参数	数值
批次	32
编码器/解码器层数	4/4
注意力头数	8
编码器/解码器输入长度	96/413
优化器	Adam

模型	δ_MSE	δ_MAE	R²
Transformer	0.252	0.373	0.697
LSTM	0.468	0.551	0.435
RNN	0.474	0.555	0.299
Informer	0.204	0.333	0.756

模型	δ_MSE	δ_MAE	R²
Transformer	0.077	0.220	0.910
LSTM	0.108	0.335	0.781
RNN	0.213	0.378	0.725
Informer	0.075	0.203	0.923

模型	δ_MSE	δ_MAE	R²
Transformer	0.439	0.485	0.321
LSTM	0.641	0.600	0.266
RNN	0.639	0.606	0.178
Informer	0.408	0.467	0.504