MULTI-TIMESCALE LONG-TERM SOLAR IRRADIATION PREDICTION METHOD FOR PHOTOVOLTAIC INSURANCE

Du Mingcong; Li Xinli; Yang Guotian; Gu Guogang; Lin Bin

doi:10.19912/j.0254-0096.tynxb.2025-1137

PDF(1436 KB)

Acta Energiae Solaris Sinica ›› 2026, Vol. 47 ›› Issue (12) : 205-216. DOI: 10.19912/j.0254-0096.tynxb.2025-1137

MULTI-TIMESCALE LONG-TERM SOLAR IRRADIATION PREDICTION METHOD FOR PHOTOVOLTAIC INSURANCE

Du Mingcong, Li Xinli, Yang Guotian, Gu Guogang, Lin Bin

Author information +

History +

Abstract

This paper proposes a long-term solar irradiation prediction method based on adaptive signal decomposition and intelligent joint prediction. Firstly, a recursive decomposition strategy is designed to adaptively decompose the original sequence into a series of stationary sub-sequences at different timescales, i.e., intrinsic mode functions （IMFs）, so as to reduce the difficulty of prediction. According to the characteristics of each decomposed sub-sequence, the seasonal autoregressive integrated moving average （SARIMA） model and the bidirectional long short-term memory （BiLSTM） network are adopted for prediction respectively. In particular, three multi-head attention modules capturing monthly, seasonal and annual features are connected to the output layer of BiLSTM to extract multi-timescale characteristics. Finally, the final prediction result is reconstructed by integrating the prediction outputs of all sub-sequences. The proposed method is validated using actual solar irradiation data from a certain area in Inner Mongolia. Experimental results show that the proposed method achieves the best performance in terms of evaluation metrics including mean squared error （MSE） and coefficient of determination （R²）. Specifically, the R² value reaches 97.92%, representing improvements of 1.38 and 5.70 percentage points over the STL-EMD-VMD fixed decomposition strategy and the single-timescale attention-based BiLSTM model, respectively. Compared with the Informer algorithm, the proposed method yields a 2.23 percentage point improvement. In addition, the proposed joint prediction method outperforms the multi-timescale BiLSTM prediction model alone by 1.46 percentage points.

Key words

solar irradiation / time series analysis / prediction models / multi-timescale / bidirectional long short-term memory network / autoregressive integrated moving average model

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

Du Mingcong, Li Xinli, Yang Guotian, Gu Guogang, Lin Bin. MULTI-TIMESCALE LONG-TERM SOLAR IRRADIATION PREDICTION METHOD FOR PHOTOVOLTAIC INSURANCE[J]. Acta Energiae Solaris Sinica. 2026, 47(12): 205-216 https://doi.org/10.19912/j.0254-0096.tynxb.2025-1137

References

[1] 艾琳, 司俊龙, 陈喜军. 2024年中国光伏发电行业发展现状与展望[J]. 水力发电, 2025, 51(6): 1-5.
Ai L, Si J L, Chen X J.Status and prospect of China’s photovoltaic power development in 2024[J]. Water Power, 2025, 51(6): 1-5.
[2] 国家能源局. 2025年光伏发电建设情况[EB/OL]. (2026-02-21)[2026-07-05]. https://www.nea.gov.cn/20260305/4216fa1274bd4b7f8da92090ba3999aa/c.html.
National Energy Administration. Development of photovoltaic power generation in2025[EB/OL]. (2026-02-21)[2026-07-05]. https://www.nea.gov.cn/20260305/4216fa1274bd4b7f8da92090ba3999aa/c.html.
[3] 马溪原, 郭晓斌, 周长城, 等. 电网公司投资分布式光伏发电系统的典型运营模式分析[J]. 南方电网技术, 2018, 12(3): 52-59.
Ma X Y, Guo X B, Zhou C C, et al.Typical operation mode analysis for distributed photovoltaic generation system invested by power grid corporation[J]. Southern Power System Technology, 2018, 12(3): 52-59.
[4] 王媛. 浅析新时代下我国新能源保险的发展前景: 以光伏发电保险为视角[J]. 纳税, 2018, 12(8): 167, 170.
Wang Y. Analysis on the development prospect of new energy insurance in China in the new era: from the perspective of photovoltaic power generation insurance[J]. Tax Paying, 2018, 12(8): 167, 170.
[5] 许光清, 陈晓玉, 刘海博, 等. 气候保险的概念、理论及在中国的发展建议[J]. 气候变化研究进展, 2020, 16(3): 373-382.
Xu G Q, Chen X Y, Liu H B, et al.The conception and theory of climate insurance and its enlightenment to China[J]. Climate Change Research, 2020, 16(3): 373-382.
[6] 孙映萱, 杨苹, 孙涛, 等. 基于云-辐照度-功率耦合机理的超短期光伏功率预测[J]. 电力建设, 2026, 47(3): 12-23.
Sun Y X, Yang P, Sun T, et al.Ultra-short-term photovoltaic power prediction based on cloud-irradiance-power coupling mechanism[J]. Electric Power Construction, 2026, 47(3): 12-23.
[7] 兰昆, 吴战波, 赵泽妮, 等. 基于多因子的太阳辐照度预测方法研究进展[J]. 太阳能学报, 2024, 45(5): 593-601.
Lan K, Wu Z B, Zhao Z N, et al.Advance in prediction methods for solar irradiance based on multiple-factors[J]. Acta Energiae Solaris Sinica, 2024, 45(5): 593-601.
[8] Bao Z Y, Gong S W, Chen Y H, et al.Distributed photovoltaic index insurance premium prediction method based on ARIMA model[J]. Journal of Electrotechnology, Electrical Engineering and Management, 2023, 6(5): 74-79.
[9] 宋晓通, 卢艺玮, 师芊芊, 等. 复杂天气条件下光伏电站太阳辐射量短期预测[J]. 科学技术与工程, 2024, 24(30): 12985-12995.
Song X T, Lu Y W, Shi Q Q, et al.Short-term prediction of solar radiation for photovoltaic power stations under complex weather conditions[J]. Science Technology and Engineering, 2024, 24(30): 12985-12995.
[10] Kiasari M, Aly H H.Evaluating solar power forecasting robustness: a comparative analysis of XGBoost, RNN, KNN, RF, and LSTM with emphasis on lagged steps, sensitivity, and cross-validation techniques[C]//2024 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE). Kingston, ON, Canada, 2024: 686-692.
[11] 马燕峰, 蒋云涛, 郝毅, 等. 基于非线性自回归神经网络的GHI预测[J]. 太阳能学报, 2019, 40(3): 733-740.
Ma Y F, Jiang Y T, Hao Y, et al.GHI forecast based on nonlinear autoregressive neural network[J]. Acta Energiae Solaris Sinica, 2019, 40(3): 733-740.
[12] 李超, 涂腾, 彭勋辉, 等. 融合DT-BO-GRU的中长期光伏功率滚动预测模型[J]. 太阳能学报, 2025, 46(5): 275-284.
Li C, Tu T, Peng X H, et al.Medium-and long-term rolling prediction model of PV power fused with DT-BO-GRU[J]. Acta Energiae Solaris Sinica, 2025, 46(5): 275-284.
[13] Fungtammasan G, Koprinska I.Convolutional and LSTM neural networks for solar power forecasting[C]//2023 International Joint Conference on Neural Networks (IJCNN). Gold Coast, Australia, 2023: 1-7.
[14] Kumar Y, Vaithiyanathan D, Verma M.Solar energy forecasting: a comparison on Regression and Artificial Neural Network models[C]//2024 First International Conference on Innovations in Communications, Electrical and Computer Engineering (ICICEC). Davangere, India, 2024: 1-6.
[15] 安文含, 刘建华, 刘吉营. 太阳辐照度机器学习模型预测能力分析[J]. 煤气与热力, 2025, 45(1): 26-31.
An W H, Liu J H, Liu J Y.Analysis of predictive ability of machine learning models for solar irradiance[J]. Gas & Heat, 2025, 45(1): 26-31.
[16] Singh C, Garg A R.Time series modeling techniques for solar irradiance forecasting: a comparative study[C]//2024 IEEE Third International Conference on Power Electronics, Intelligent Control and Energy Systems (ICPEICES). Delhi, India, 2024: 907-912.
[17] Zhang H J, Li B, Su S F, et al.A novel hybrid transformer-based framework for solar irradiance forecasting under incomplete data scenarios[J]. IEEE Transactions on Industrial Informatics, 2024, 20(6): 8605-8615.
[18] Boutahir M K, Farhaoui Y, Azrour M, et al.Effect of feature selection on the prediction of direct normal irradiance[J]. Big Data Mining and Analytics, 2022, 5(4): 309-317.
[19] Maciel J N, de Souza Campoi G, Zalewski W, et al. Optimizing solar irradiance prediction: feature selection for all-sky image processing using a hybrid prediction method[J]. IEEE Latin America Transactions, 2025, 23(1): 50-57.
[20] Mabodi R, Hammujuddy J.Solar irradiance forecasting for informed solar systems design and financing decisions[J]. SAIEE Africa Research Journal, 2024, 115(3): 99-108.
[21] Khalil A A, Mandour M A, Ali A.The benefits of social insurance system prediction using a hybrid fuzzy time series method[J]. PeerJ Computer Science, 2024, 10: e2500.
[22] 刘文博, 于连飞, 谢冬梅, 等. 基于多尺度特征融合的时间序列长期预测模型[J]. 计算机应用, 2024, 44(11): 3435-3441.
Liu W B, Yu L F, Xie D M, et al.Long-term prediction model of time series based on multi-scale feature fusion[J]. Journal of Computer Applications, 2024, 44(11): 3435-3441.
[23] 潘建, 汪绪豪. 融合多尺度特征和注意力机制的时间序列预测模型[J]. 计算机科学, 2026, 53(2): 180-186.
Pan J, Wang X H.Time series forecasting model integrating multi-scale features and attention mechanism[J]. Computer Science, 2026, 53(2): 180-186.
[24] Yang T B, Huang L S, Fu P, et al.Combined multi-component composite time series power prediction model for distributed energy systems based on STL data decomposition[J]. Measurement, 2025, 253: 117299.
[25] 陈小泽, 王忠静, 刘丹, 等. 径流序列两段动态分解-预测-重构中长期预报模型[J]. 水利学报, 2025, 56(7): 933-944.
Chen X Z, Wang Z J, Liu D, et al.A two-stage dynamic decomposition-prediction-reconstruction model for medium-long term runoff forecasting[J]. Journal of Hydraulic Engineering, 2025, 56(7): 933-944.
[26] 张露, 理华, 崔杰, 等. 基于稀疏指标的优化变分模态分解方法[J]. 振动与冲击, 2023, 42(8): 234-250.
Zhang L, Li H, Cui J, et al.An optimal variational mode decomposition method based on sparse index[J]. Journal of Vibration and Shock, 2023, 42(8): 234-250.
[27] Huang X Q, Li Q, Tai Y H, et al.Hybrid deep neural model for hourly solar irradiance forecasting[J]. Renewable Energy, 2021, 171: 1041-1060.
[28] 鄢伟安, 付晟, 刘卫东, 等. 基于STL-EMD-SSA-LSTM模型的光伏中期输出功率预测研究[J]. 工业工程与管理, 2024, 29(6): 93-103.
Yan W A, Fu S, Liu W D, et al.Prediction of mid-term photovoltaic output power based on STL-EMD-SSA-LSTM model[J]. Industrial Engineering and Management, 2024, 29(6): 93-103.
[29] Wang M H, Meng Y, Sun L, et al.Decomposition combining averaging seasonal-trend with singular spectrum analysis and a marine predator algorithm embedding Adam for time series forecasting with strong volatility[J]. Expert Systems with Applications, 2025, 274: 126864.
[30] Farhangi A, Bian J, Huang A, et al.Correction to: AA-forecast: anomaly-aware forecast for extreme events[J]. Data Mining and Knowledge Discovery, 2024, 38(2): 812.
[31] Yan Z H, Lu X H, Wu L F, et al.Enhancing short-term solar radiation forecasting with hybrid VMD and GraphCast-based machine learning models[J]. Expert Systems with Applications, 2025, 285: 128042.
[32] Mirjalili S, Lewis A.The whale optimization algorithm[J]. Advances in Engineering Software, 2016, 95: 51-67.
[33] Baloch M, Honnurvali M S, Kabbani A, et al.An intelligent SARIMAX-based machine learning framework for long-term solar irradiance forecasting at Muscat, Oman[J]. Energies, 2024, 17(23): 6118.
[34] Zhuang Z Y, Wang H Z, Yu C L.Prediction of short-term solar irradiance using the ProbSparse attention mechanism for a sustainable energy development strategy[J]. Sustainability, 2025, 17(3): 1075.
[35] Ma X, Wang J W, Huang J, et al.Multi-grained adaptive informer for multi-step solar irradiance forecasting[J]. Expert Systems with Applications, 2025, 282: 127412.