TimeGAN-based photovoltaic power prediction method under extreme weather events

doi:10.3969/j.issn.2097-0706.2025.09.006

Abstract

Abstract:

Accurate prediction of photovoltaic power generation under extreme weather events is crucial for ensuring energy supply and grid stability. However， the suddenness of such weather events leads to scarce historical data from photovoltaic power stations， making it difficult to effectively predict photovoltaic power under extreme weather conditions. To address this issue， a prediction method based on Time-series Generative Adversarial Networks（TimeGAN） was proposed to augment limited historical data. The method captured the complex temporal dependencies between photovoltaic power and weather conditions. Based on the limited historical data from photovoltaic power stations， the TimeGAN model generated realistic time-series data to simulate the occurrence of extreme weather events， and subsequently conducted photovoltaic power prediction. The experimental results showed that compared to traditional GAN for small sample augmentation， the TimeGAN-augmented prediction results demonstrated better fitting performance. After 25% data augmentation， the Mean Absolute Error（MAE） decreased by 1.14 MW， and the Root Mean Square Error（RMSE） decreased by 1.09 MW. After 50% data augmentation， the MAE decreased by 1.08 MW， and the RMSE decreased by 0.99 MW. These results indicated significant improvements in prediction accuracy.

Key words: TimeGAN, extreme weather, small sample augmentation, photovoltaic power prediction, time series

CLC Number:

TK51：TM761

SUN Shiqi, MA Gang, XU Wenjun, LI Hao, MA Jian. TimeGAN-based photovoltaic power prediction method under extreme weather events[J]. Integrated Intelligent Energy, 2025, 47(9): 51-59.

Figures/Tables 13

Table 1

Fig.1

Fig.2

Fig.3

Fig.4

Table 2

Fig.5

Fig.6

Fig.7

Table 3

Fig.8

Table 4

Table 5

References 24

[1]	张冬冬, 单琳珂, 刘天皓. 人工智能技术在风力与光伏发电数据挖掘及功率预测中的应用综述[J]. 综合智慧能源, 2025, 47(3): 32-46. doi: 10.3969/j.issn.2097-0706.2025.03.004
	ZHANG Dongdong, SHAN Linke, LIU Tianhao. Review on the application of artificial intelligence in data mining and wind and photovoltaic power forecasting[J]. Integrated Intelligent Energy, 2025, 47(3): 32-46. doi: 10.3969/j.issn.2097-0706.2025.03.004
[2]	龙小慧, 秦际赟, 张青雷, 等. 基于相似日聚类及模态分解的短期光伏发电功率组合预测研究[J]. 电网技术, 2024, 48(7): 2948-2957.
	LONG Xiaohui, QIN Jiyun, ZHANG Qinglei, et al. Short-term photovoltaic power prediction study based on similar day clustering and modal decomposition[J]. Power System Technology, 2024, 48(7): 2948-2957.
[3]	卓振宇, 张宁, 谢小荣, 等. 高比例可再生能源电力系统关键技术及发展挑战[J]. 电力系统自动化, 2021, 45(9): 171-191.
	ZHUO Zhenyu, ZHANG Ning, XIE Xiaorong, et al. Key technologies and developing challenges of power system with high proportion of renewable energy[J]. Automation of Electric Power Systems, 2021, 45(9): 171-191.
[4]	邓芳明, 刘涛, 王锦波, 等. 基于地基云图与气象因素多模态融合的光伏功率预测方法[J/OL]. 中国电机工程学报, 2025: 1-14(2025-02-21)[2025-03-20]. https://kns.cnki.net/KCMS/detail/detail.aspx?filename=ZGDC20250220002&dbname=CJFD&dbcode=CJFQ.
	DENG Fangming, LIU Tao, WANG Jinbo, et al. Research on photovoltaic power prediction based on multimodal fusion of ground cloud map and meteorological factors[J/OL]. Proceedings of the CSEE, 2025: 1-14(2025-02-21)[2025-03-20]. https://kns.cnki.net/KCMS/detail/detail.aspx?filename=ZGDC20250220002&dbname=CJFD&dbcode=CJFQ.
[5]	张元曦, 杨国华, 杨娜, 等. 基于K-means聚类的LSTM-SVR-DE光伏功率组合预测[J]. 综合智慧能源, 2025, 47(2): 71-78. doi: 10.3969/j.issn.2097-0706.2025.02.007
	ZHANG Yuanxi, YANG Guohua, YANG Na, et al. Photovoltaic power prediction based on K-means clustering and the LSTM-SVR-DE model[J]. Integrated Intelligent Energy, 2025, 47(2): 71-78. doi: 10.3969/j.issn.2097-0706.2025.02.007
[6]	王玉庆, 徐飞, 刘志坚, 等. 基于动态关联表征与图网络建模的分布式光伏超短期功率预测[J]. 电力系统自动化, 2023, 47(20): 72-82.
	WANG Yuqing, XU Fei, LIU Zhijian, et al. Ultra-short-term power forecasting of distributed photovoltaic based on dynamic correlation characterization and graph network modeling[J]. Automation of Electric Power Systems, 2023, 47(20): 72-82.
[7]	臧海祥, 程礼临, 刘玲, 等. 基于数据驱动的太阳辐射估计和预测研究与展望[J]. 电力系统自动化, 2021, 45(11): 170-183.
	ZANG Haixiang, CHENG Lilin, LIU Ling, et al. Research and prospect for data-driven estimation and prediction of solar radiation[J]. Automation of Electric Power Systems, 2021, 45(11): 170-183.
[8]	马原, 张雪敏, 甄钊, 等. 基于修正晴空模型的超短期光伏功率预测方法[J]. 电力系统自动化, 2021, 45(11): 44-51.
	MA Yuan, ZHANG Xuemin, ZHEN Zhao, et al. Ultra-short-term photovoltaic power prediction method based on modified clear-sky model[J]. Automation of Electric Power Systems, 2021, 45(11): 44-51.
[9]	张筱辰, 朱金大, 杨冬梅, 等. 基于t-SNE流形学习与快速聚类算法的光伏逆变器故障预测技术[J]. 中国电力, 2020, 53(6): 41-47.
	ZHANG Xiaochen, ZHU Jinda, YANG Dongmei, et al. Photovoltaic inverter fault prediction technology based on t-SNE manifold learning and fast clustering algorithm[J]. Electric Power, 2020, 53(6): 41-47.
[10]	梁志峰, 秦放, 崔方. “6·21”日环食对光伏发电及电网运行影响分析[J]. 电力系统自动化, 2021, 45(7): 1-7.
	LIANG Zhifeng, QIN Fang, CUI Fang. Impact analysis of annular solar eclipse on June 21, 2020 in China on photovoltaic power generation and power grid operation[J]. Automation of Electric Power Systems, 2021, 45(7): 1-7.
[11]	刘洪波, 王铎皓, 石鹏, 等. 基于图神经网络的多光伏场站出力短期时-空预测[J]. 电网与清洁能源, 2025, 41(1): 89-96.
	LIU Hongbo, WANG Duohao, SHI Peng, et al. Short term time-space prediction of multi-photovoltaic plant output based on graph neural network[J]. Power System and Clean Energy, 2025, 41(1): 89-96.
[12]	LIU W C, MAO Z Z. Short-term photovoltaic power forecasting with feature extraction and attention mechanisms[J]. Renewable Energy, 2024, 226:120437.
[13]	ELISSAIOS S, EVANGELOS S, EFSTATHIOS S, et al. Short-term photovoltaic power forecasting using meta-learning and numerical weather prediction independent long short-term memory models[J]. Renewable Energy, 2023, 216:118997.
[14]	殷豪, 张铮, 丁伟锋, 等. 基于生成对抗网络和LSTM-CSO的少样本光伏功率短期预测[J]. 高电压技术, 2022, 48(11): 4342-4351.
	YIN Hao, ZHANG Zheng, DING Weifeng, et al. Short-term prediction of small-sample photovoltaic power based on generative adversarial network and LSTM-CSO[J]. High Voltage Engineering, 2022, 48(11): 4342-4351.
[15]	肖白, 黄钰茹, 姜卓, 等. 数据匮乏场景下采用生成对抗网络的空间负荷预测方法[J]. 中国电机工程学报, 2020, 40(24): 7990-8001, 8236.
	XIAO Bai, HUANG Yuru, JIANG Zhuo, et al. The method of spatial load forecasting based on the generative adversarial network for data scarcity scenarios[J]. Proceedings of the CSEE, 2020, 40(24): 7990-8001, 8236.
[16]	顾荣直, 田心如, 禹梁玉, 等. 江苏寒潮天气过程风险预评估方法研究[J]. 气象学报, 2024, 82(2): 247-256.
	GU Rongzhi, TIAN Xinru, YU Liangyu, et al. Methodology in pre-assessment of the cold surge induced risks in Jiangsu province of China[J]. Acta Meteorologica Sinica, 2024, 82(2): 247-256.
[17]	王丽婕, 张青山, 郝颖, 等. 基于气象数据外推法和显著性分析的光伏自适应功率预测模型[J]. 太阳能学报, 2025, 46(2): 317-325.
	WANG Lijie, ZHANG Qingshan, HAO Ying, et al. Photovoltaic adaptive power prediction model based on meteorological data extrapolation and significance analysis[J]. Acta Energiae Solaris Sinica, 2025, 46(2): 317-325.
[18]	郑珂, 王丽婕, 郝颖, 等. 基于数据集蒸馏的光伏发电功率超短期预测[J]. 中国电机工程学报, 2024, 44(13): 5196-5208.
	ZHENG Ke, WANG Lijie, HAO Ying, et al. Ultra-short-term prediction of photovoltaic power based on dataset distillation[J]. Proceedings of the CSEE, 2024, 44(13): 5196-5208.
[19]	彭曙蓉, 陈慧霞, 孙万通, 等. 基于改进LSTM的光伏发电功率预测方法研究[J]. 太阳能学报, 2024, 45(11): 296-302.
	PENG Shurong, CHEN Huixia, SUN Wantong, et al. Research on photovoitaic power prediction method based on improved lstm[J]. Acta Energiae Solaris Sinica, 2024, 45(11): 296-302.
[20]	卫志农, 马智刚, 陈胜, 等. 考虑光伏随机性的交直流混合配电网鲁棒机会约束安全域模型[J]. 中国电机工程学报, 2024, 44(6): 2208-2220.
	WEI Zhinong, MA Zhigang, CHEN Sheng, et al. Robust chance-constrained security region model of AC/DC hybrid distribution network considering the uncertainty of photovoltaic generation[J]. Proceedings of the CSEE, 2024, 44(6): 2208-2220.
[21]	CHEN C C, CHAI L, WANG Q L. Research on stacking ensemble method for day-ahead ultra-short-term prediction of photovoltaic power[J]. Renewable Energy, 2025, 238:121853.
[22]	孔令国, 王嘉祺, 韩子娇, 等. 基于权重调节模型预测控制的风-光-储-氢耦合系统在线功率调控[J]. 电工技术学报, 2023, 38(15): 4192-4207.
	KONG Lingguo, WANG Jiaqi, HAN Zijiao, et al. On-line power regulation of wind-photovoltaic-storage-hydrogen coupling system based on weight adjustment model predictive control[J]. Transactions of China Electrotechnical Society, 2023, 38(15): 4192-4207.
[23]	YANG M, JIANG Y, ZHANG W, et al. Short-term interval prediction strategy of photovoltaic power based on meteorological reconstruction with spatiotemporal correlation and multi-factor interval constraints[J]. Renewable Energy, 2024, 237:121834.
[24]	ZHAO Y, WANG B, WANG S, et al. Photovoltaic power generation power prediction under major extreme weather based on VMD-KELM[J]. Energy Engineering, 2024, 121(12):3711-3733.

区间	云量	相对湿度	环境温度	风速	气压	辐照度
0.1	0.312 4	0.225 2	0.344 3	0.381 4	0.183 1	0.684 2
0.2	0.363 3	0.258 1	0.471 4	0.502 1	0.208 7	0.736 4
0.3	0.495 2	0.287 4	0.572 3	0.592 1	0.216 4	0.771 9
0.4	0.567 1	0.291 5	0.593 6	0.645 2	0.263 2	0.819 2
0.5	0.600 4	0.345 3	0.645 3	0.697 2	0.298 8	0.832 3
0.6	0.628 4	0.378 9	0.691 4	0.739 3	0.313 1	0.858 2
0.7	0.661 1	0.482 1	0.725 6	0.754 5	0.328 4	0.890 1
0.8	0.699 8	0.531 4	0.748 4	0.788 2	0.357 9	0.901 3
0.9	0.725 3	0.559 0	0.761 2	0.791 7	0.389 1	0.910 4
1.0	0.746 7	0.574 1	0.777 5	0.813 1	0.393 5	0.923 7

模型	参数	数值
TimeGAN	θ	2
	λ	0.1
	逻辑框架数量	3
	学习率1	0.001
	学习率2	0.005
	批量大小	56
CNN-LSTM	学习率	0.009
	隐藏层	134
	迭代次数	100
	批量大小	128
	梯度阈值	1

数据量	MAE/MW	RMSE/MW	R²
无扩增	7.82	10.96	0.763
扩增25%	5.40	5.81	0.873
扩增50%	4.09	4.98	0.891

数据扩增	MAE/MW	RMSE/MW	R²
无扩增	7.15	10.60	0.796
25%	4.26	4.72	0.878
50%	3.11	3.99	0.912

数据扩增方式	准确度	精度	召回率
GAN	96.09±0.79	92.93±2.76	74.75±0.63
TimeGAN	98.16±0.27	98.93±0.46	66.63±1.50