Photovoltaic power prediction based on K-means clustering and the LSTM-SVR-DE model

doi:10.3969/j.issn.2097-0706.2025.02.007

Abstract

Abstract:

To improve the accuracy of photovoltaic power prediction， a combined prediction model based on Long Short-Term Memory（LSTM） neural networks and Support Vector Regression（SVR） was proposed. Both the LSTM and SVR models were used separately to predict photovoltaic power. On this basis， a Stacking ensemble strategy was employed to linearly combine the predictions of these two models， with the Differential Evolution（DE） algorithm optimizing the weight coefficients. Simulations and comparative analyses were conducted using real data from a photovoltaic power station in Ningxia. The results showed that the proposed method reduced prediction errors by approximately 70% compared to the LSTM and SVR models.

Key words: K-means clustering, LSTM neural network, support vector regression, differential evolution, photovoltaic power prediction

CLC Number:

TK519

ZHANG Yuanxi, YANG Guohua, YANG Na, LI Zhen, MA Xin, LIU Haorui, NAN Shaoshuai. Photovoltaic power prediction based on K-means clustering and the LSTM-SVR-DE model[J]. Integrated Intelligent Energy, 2025, 47(2): 71-78.

Figures/Tables 6

Fig.1

Fig.2

Table 1

Table 2

Fig.3

Fig.4

References 25

[1]	朱琼锋, 李家腾, 乔骥, 等. 人工智能技术在新能源功率预测的应用及展望[J]. 中国电机工程学报, 2023, 43(8):3027-3048.
	ZHU Qiongfeng, LI Jiateng, QIAO Ji, et al. Application and prospect of artificial intelligence technology in renewable energy forecasting[J]. Proceedings of the CSEE, 2023, 43(8):3027-3048.
[2]	汪鸿, 朱正甲, 陈建华, 等. 基于人工智能技术与物理方法结合的新能源功率预测研究[J]. 高电压技术, 2023, 49(S1):111-117.
	WANG Hong, ZHU Zhengjia, CHEN Jianhua, et al. Research on new energy power prediction based on artificial intelligence technology and physical method[J]. High Voltage Engineering, 2023, 49(S1):111-117.
[3]	王彪, 吕洋, 陈中, 等. 考虑信息时移的分布式光伏机理-数据混合驱动短期功率预测[J]. 电力系统自动化, 2022, 46(11):67-74.
	WANG Biao, LYU Yang, CHEN Zhong, et al. Hybrid mechanism-data-driven short-term power forecasting of distributed photovoltaic considering information time shift[J]. Automation of Electric Power Systems, 2022, 46(11):67-74.
[4]	陈凡, 李智, 丁津津, 等. 考虑光伏机理与数据驱动结合的短期功率预测[J]. 科学技术与工程, 2023, 23(20):8686-8692.
	CHEN Fan, LI Zhi, DING Jinjin, et al. Consider short-term power prediction combining photovoltaic mechanism and data-driven[J]. Science Technology and Engineering, 2023, 23(20):8686-8692.
[5]	李楠, 刘佳佳, 赖心怡, 等. 基于时间序列神经分层插值模型的光伏功率超短期多步预测[J]. 智慧电力, 2024, 52(4):69-77.
	LI Nan, LIU Jiajia, LAI Xinyi, et al. Ultra-short-term multi-step forecasting of photovoltaic power based on time series neural hierarchical interpolation model[J]. Smart Power, 2024, 52(4):69-77.
[6]	时珉, 许可, 王珏, 等. 基于灰色关联分析和GeoMAN模型的光伏发电功率短期预测[J]. 电工技术学报, 2021, 36(11):2298-2305.
	SHI Min, XU Ke, WANG Jue, et al. Short-term photovoltaic power forecast based on grey relational analysis and GeoMAN model[J]. Transactions of China Electrotechnical Society, 2021, 36(11):2298-2305.
[7]	周鑫, 李燕, 曾永辉, 等. 基于SARIMAX-SVR的光伏发电功率预测[J]. 电力系统及其自动化学报, 2024, 36(5):1-8.
	ZHOU Xin, LI Yan, ZENG Yonghui, et al. Forecasting of photovoltaic power generation based on SARIMAX-SVR[J]. Proceedings of the CSU-EPSA, 2024, 36(5):1-8.
[8]	殷林飞, 蒙雨洁. 基于DenseNet卷积神经网络的短期风电预测方法[J]. 综合智慧能源, 2024, 46(7):12-20. doi: 10.3969/j.issn.2097-0706.2024.07.002
	YIN Linfei, MENG Yujie. Short-term wind power forecasting method based on DenseNe convolutional neural network[J]. Integrated Intelligent Energy, 2024, 46(7):12-20. doi: 10.3969/j.issn.2097-0706.2024.07.002
[9]	袁俊球, 王迪, 谢小锋, 等. 基于广义天气分类的ICEEMDAN-LSTM网络光伏发电功率短期预测[J]. 综合智慧能源, 2024, 46(9):53-60. doi: 10.3969/j.issn.2097-0706.2024.09.007
	YUAN Junqiu, WANG Di, XIE Xiaofeng, et al. Short-term photovoltaic power forecasting based on generalized weather classification and ICEEMDAN-LSTM network[J]. Integrated Intelligent Energy, 2024, 46(9):53-60. doi: 10.3969/j.issn.2097-0706.2024.09.007
[10]	盛瑞祥, 张啸宇. 基于概率TCN-Transformer的短期光伏功率预测模型[J]. 综合智慧能源, 2024, 46(11):10-18. doi: 10.3969/j.issn.2097-0706.2024.11.002
	SHENG Ruixiang, ZHANG Xiaoyu. Short-term photovoltaic power forecasting model based on probabilistic TCN-transformer[J]. Integrated Intelligent Energy, 2024, 46(11):10-18. doi: 10.3969/j.issn.2097-0706.2024.11.002
[11]	殷豪, 李奕甸, 谢智锋, 等. 混合图神经网络和门控循环网络的短期光伏功率预测[J]. 太阳能学报, 2024, 45(3):523-532.
	YIN Hao, LI Yidian, XIE Zhifeng, et al. Short-term photovoltaic power prediction method based on mixed graph neural network and gated recurrent unit network[J]. Acta Energiae Solaris Sinica, 2024, 45(3):523-532.
[12]	李芬, 孙凌, 王亚维, 等. 基于CEEMDAN-GSA-LSTM和SVR的光伏功率短期区间预测[J]. 上海交通大学学报, 2024, 58(6):806-818. doi: 10.16183/j.cnki.jsjtu.2022.511
	LI Fen, SUN Ling, WANG Yawei, et al. Short-term interval forecasting of photovoltaic power based on CEEMDAN-GSA-LSTM and SVR[J]. Journal of Shanghai Jiao Tong University, 2024, 58(6):806-818.
[13]	黄牧涛, 邢芳菲, 陈兴邦, 等. 基于K-means聚类和极限学习机组合算法的短期光伏功率预测[J]. 水电能源科学, 2024, 42(2):217-220,216.
	HUANG Mutao, XING Fangfei, CHEN Xingbang, et al. Short-term PV power prediction based on K-means clustering and extreme learning machine combination algorithm[J]. Water Resources and Power, 2024, 42(2): 217-220,216.
[14]	唐雅洁, 林达, 倪筹帷, 等. 基于XGBoost的双层协同实时校正超短期光伏预测[J]. 电力系统自动化, 2021, 45(7):18-27.
	TANG Yajie, LIN Da, NI Chouwei, et al. XGBoost based bi-layer collaborative real-time calibration for ultra-short-term photovoltaic prediction[J]. Automation of Electric Power Systems, 2021, 45(7):18-27.
[15]	何威, 苏中元, 史金林, 等. 基于双重注意力GRU与相似修正的光伏功率预测[J]. 太阳能学报, 2024, 45(3):480-487.
	HE Wei, SU Zhongyuan, SHI Jinlin, et al. Photovoltaic power forecasting based on dual-attention-GRU and similarity modification[J]. Acta Energiae Solaris Sinica, 2024, 45(3):480-487.
[16]	王海军, 居蓉蓉, 董颖华. 基于时空关联特征与B-LSTM模型的分布式光伏功率区间预测[J]. 中国电力, 2024, 57(7):74-80.
	WANG Haijun, JU Rongrong, DONG Yinghua. Distributed photovoltaic power interval prediction based on spatio-temporal correlation feature and B-LSTM model[J]. Electric Power, 2024, 57(7):74-80.
[17]	王东风, 刘婧, 黄宇, 等. 结合太阳辐射量计算与CNN-LSTM组合的光伏功率预测方法研究[J]. 太阳能学报, 2024, 45(2):443-450. doi: 10.19912/j.0254-0096.tynxb.2022-1542
	WANG Dongfeng, LIU Jing, HUANG Yu, et al. Photovoltaic power prediction method combination solar radiation calculation and CNN-LSTM[J]. Acta Energiae Solaris Sinica, 2024, 45(2):443-450. doi: 10.19912/j.0254-0096.tynxb.2022-1542
[18]	FEI W, XUAN Z M, ZHAO Z, et al. A day-ahead PV power forecasting method based on LSTM-RNN model and time correlation modification under partial daily pattern prediction framework[J]. Energy Conversion and Management, 2020, 212:112766.
[19]	李鹏钦, 张长胜, 李英娜, 等. 改进Stacking算法的光伏发电功率预测[J]. 应用科学学报, 2022, 40(2): 288-301.
	LI Pengqin, ZHANG Changsheng, LI Yingna, et al. Photovoltaic power forecast improved Stacking algorithm[J]. Journal of Applied Sciences, 2022, 40(2):288-301.
[20]	何之倬, 张颖, 郑刚, 等. 基于极限学习机模型参数优化的光伏功率区间预测技术[J]. 上海交通大学学报, 2024, 58(3):285-294. doi: 10.16183/j.cnki.jsjtu.2022.338
	HE Zhizhuo, ZHANG Ying, ZHENG Gang, et al. Interval prediction technology of photovoltaic power based on parameter optimization of extreme learning machine[J]. Journal of Shanghai Jiao Tong University, 2024, 58(3):285-294.
[21]	ZANG H X, CHENG L L, TAO D, et al. Day-ahead photovoltaic power forecasting approach based on deep convolutional neural networks and meta learning[J]. Electrical Power and Energy Systems, 2020, 118:105790.
[22]	ZANG H X, LING L, LI S, et al. Short-term global horizontal irradiance forecasting based on a hybrid CNN-LSTM model with spatiotemporal correlations[J]. Renewable Energy, 2020, 160:26-41.
[23]	谭海旺, 杨启亮, 邢建春, 等. 基于XGBoost-LSTM组合模型的光伏发电功率预测[J]. 太阳能学报, 2022, 43(8):75-81. doi: 10.19912/j.0254-0096.tynxb.2021-0005
	TAN Haiwang, YANG Qiliang, XING Jianchun, et al. Photovoltaic power prediction based on combined XGBoost-LSTM model[J]. Acta Energiae Solaris Sinica, 2022, 43(8):75-81. doi: 10.19912/j.0254-0096.tynxb.2021-0005
[24]	AGGA A, ABBOU A, LABBADI M, et al. CNN-LSTM:An efficient hybrid deep learning architecture for predicting short-term photovoltaic power production[J]. Electric Power Systems Research, 2022, 208:107908.
[25]	SI Z Y, MING Y, YU Y X, et al. Photovoltaic power forecast based on satellite images considering effects of solar position[J]. Applied Energy, 2021, 302:117514.

模型	参数	数值
LSTM	训练次数	50
	初始学习率	0.1
	衰减率	0.8
	衰减周期	10
	隐藏单元数	16
	时间步长	12
SVR	epsilon	0.1
	gamma	50
	C	50
DE	种群规模	200
	迭代次数	500
	F	0.7
	CR	0.9

模型	评价指标
模型	e_RMSE	e_MAE	R²
1	13.668	6.314	0.89
2	17.053	8.892	0.82
3	4.325	1.858	0.98
4	6.196	1.954	0.97
5	3.972	1.721	0.99