Improving the Short-Term Photovoltaic Power Forecasting Accuracy Based on Deep Learning Hybrid Model

Abstract

The increasing penetration of renewable resources in smart energy systems has increased the need for accurate forecasting of photovoltaic (PV) power to reduce uncertainty and improve operational planning. Rapid changes in atmospheric conditions, especially cloud movement, cause severe fluctuations and pose a significant challenge to PV power forecasting in the short term. In this paper, an integrated Deep Learning-based framework for short-term PV power forecasting is proposed that models the relationships among sky conditions, solar irradiance, and PV power. First, the cloud movement is forecasted based on the Optical Flow algorithm using sky images. Then, the forecasted images are fed into a ResNet50 network to estimate the solar irradiance. Finally, the forecasted solar irradiance, meteorological parameters, and historical PV power data are fed into the CNN-BiGRU-Attention model, whose hyperparameters are optimized using Bayesian Optimization to forecast the PV power for the next hour. The performance of the proposed model was evaluated under different atmospheric conditions and compared with other benchmark models. The results showed that the proposed model achieves high forecasting accuracy, obtaining a R² of 0.987 and RMSE, MAE, and MAPE error values of 0.231 Kw, 0.147 Kw, and 8.69%, respectively, and outperforming benchmark models.

Keywords

Photovoltaic power forecasting / Deep learning / Gated recurrent units / Convolutional neural network / Solar irradiance forecasting / Sky images

References

[1] Krishnan N, Kumar KR, Inda CS . How solar radiation forecasting impacts the utilization of solar energy: A critical review. J. Clean. Prod. 2023, 388, 135860. DOI: 10.1016/j.jclepro.2023.135860

[2] Jalali Z, Moghaddas-Tafreshi SM. Optimal Design and Simulation of a Rooftop Photovolataic System in Faculty of Engineering Guilan University. In Proceedings of the 2023 10th Iranian Conference on Renewable Energy & Distributed Generation (ICREDG), Shahrood, Iran, 15-16 March 2023; pp. 1-5. DOI: 10.1109/ICREDG58341.2023.10092071

[3] Tzoumanikas P, Nikitidou E, Bais AF, Kazantzidis A. The effect of clouds on surface solar irradiance, based on data from an all-sky imaging system. Renew. Energy 2016, 95, 314-322. DOI: 10.1016/j.renene.2016.04.026

[4] Aicardi D, Musé P, Alonso-Suárez R . A comparison of satellite cloud motion vectors techniques to forecast intra-day hourly solar global horizontal irradiation. Sol. Energy 2022, 233, 46-60. DOI: 10.1016/j.solener.2021.12.066

[5] Aguiar LM, Pereira B, Lauret P, Díaz F, David M . Combining solar irradiance measurements, satellite-derived data and a numerical weather prediction model to improve intra-day solar forecasting. Renew. Energy 2016, 97, 599-610. DOI: 10.1016/j.renene.2016.06.018

[6] Caldas M, Alonso-Suárez R . Very short-term solar irradiance forecast using all-sky imaging and real-time irradiance measurements. Renew. Energy 2019, 143, 1643-1658. DOI: 10.1016/j.renene.2019.05.069

[7] Trigo-González M, Cortés-Carmona M, Marzo A, Alonso-Montesinos J, Martínez-Durbán M, López G, et al. Photovoltaic power electricity generation nowcasting combining sky camera images and learning supervised algorithms in the Southern Spain. Renew. Energy 2023, 206, 251-262. DOI: 10.1016/j.renene.2023.01.111

[8] Logothetis SA, Salamalikis V, Wilbert S, Remund J, Zarzalejo LF, Xie Y, et al. Benchmarking of solar irradiance nowcast performance derived from all-sky imagers. Renew. Energy 2022, 199, 246-261. DOI: 10.1016/j.renene.2022.08.127

[9] Yu Z, Tan Z, Ma S, Yan W . Nowcast for cloud top height from Himawari‐8 data based on deep learning algorithms. Meteorol. Appl. 2023, 30, e2130. DOI: 10.1002/met.2130

[10] Terrén-Serrano G, Martínez-Ramón M . Detection of clouds in multiple wind velocity fields using ground-based infrared sky images. Knowl.-Based Syst. 2023, 274, 110628. DOI: 10.1016/j.knosys.2023.110628

[11] Kosmopoulos P, Kouroutsidis D, Papachristopoulou K, Raptis PI, Masoom A, Saint-Drenan YM, et al. Short-Term Forecasting of Large-Scale Clouds Impact on Downwelling Surface Solar Irradiation. Energies 2020, 13, 6555. DOI: 10.3390/en13246555

[12] Mondragón R, Alonso-Montesinos J, Riveros-Rosas D, Bonifaz R . Determination of Cloud Motion Applying the Lucas-Kanade Method to Sky Cam Imagery. Remote Sens. 2020, 12, 2643. DOI: 10.3390/rs12162643

[13] Niu Y, Song J, Zou L, Yan Z, Lin X . Cloud detection method using ground-based sky images based on clear sky library and superpixel local threshold. Renew. Energy 2024, 226, 120452. DOI: 10.1016/j.renene.2024.120452

[14] Al-lahham A, Theeb O, Elalem K, Alshawi TA, Alshebeili SA . Sky Imager-Based Forecast of Solar Irradiance Using Machine Learning. Electronics 2020, 9, 1700. DOI: 10.3390/electronics9101700

[15] Kumari P, Toshniwal D. Deep learning models for solar irradiance forecasting: A comprehensive review. J. Clean. Prod. 2021, 318, 128566. DOI: 10.1016/j.jclepro.2021.128566

[16] Feng C, Zhang J . SolarNet: A sky image-based deep convolutional neural network for intra-hour solar forecasting. Sol. Energy 2020, 204, 71-78. DOI: 10.1016/j.solener.2020.03.083

[17] Nie Y, Paletta Q, Scott A, Pomares LM, Arbod G, Sgouridis S, et al. Sky image-based solar forecasting using deep learning with heterogeneous multi-location data: Dataset fusion versus transfer learning. Appl. Energy 2024, 369, 123467. DOI: 10.1016/j.apenergy.2024.123467

[18] Ansong M, Huang G, Nyang’onda TN, Musembi RJ, Richards BS . Very short-term solar irradiance forecasting based on open-source low-cost sky imager and hybrid deep-learning techniques. Sol. Energy 2025, 294, 113516. DOI: 10.1016/j.solener.2025.113516

[19] Khan AW, Duan J, Nawaz F, Lu W . Novel hybrid BiLSTM-BiGRU and CNN-LSTM architectures for enhanced solar irradiance forecasting in semi-arid climates. Int. J. Electr. Power Energy Syst. 2025, 172, 111380. DOI: 10.1016/j.ijepes.2025.111380

[20] Ukwuoma CC, Cai D, Bamisile O, Yin H, Nneji GU, Monday HN, et al. An attention fused sequence-to-sequence convolutional neural network for accurate solar irradiance forecasting and prediction using sky images. Renew. Energy 2024, 237, 121692. DOI: 10.1016/j.renene.2024.121692

[21] Rajagukguk RA, Lee H . Sky-image-based sun-blocking index and PredRNN++ for accurate short-term solar irradiance forecasting. Build. Environ. 2025, 274, 112429. DOI: 10.1016/j.buildenv.2024.112429

[22] Jonathan AL, Cai D, Ukwuoma CC, Nkou NJJ, Huang Q, Bamisile O . A radiant shift: Attention-embedded CNNs for accurate solar irradiance forecasting and prediction from sky images. Renew. Energy 2024, 234, 121133. DOI: 10.1016/j.renene.2024.121133

[23] Tiwari S, Sabzehgar R, Rasouli M . Short Term Solar Irradiance Forecast based on Image Processing and Cloud Motion Detection. In Proceedings of the 2019 IEEE Texas Power and Energy Conference (TPEC), College Station, TX, USA, 7-8 February 2019; pp. 1-6. DOI: 10.1109/TPEC.2019.8662134

[24] Bouzerdoum M, Mellit A, Massi Pavan A. A hybrid model (SARIMA-SVM) for short-term power forecasting of a small-scale grid-connected photovoltaic plant. Sol. Energy 2013, 98, 226-235. DOI: 10.1016/j.solener.2013.10.002

[25] Rajabalizadeh S, Tafreshi SMM . A Practicable Copula-Based Approach for Power Forecasting of Small-Scale Photovoltaic Systems. IEEE Syst. J. 2020, 14, 4911-4918. DOI: 10.1109/JSYST.2019.2961661

[26] Isapour E, Moghaddas Tafreshi SM . A Hybrid Model Based on Neural networks to Predict the Power Output of Photovoltaic systems. In Proceedings of the 2019 Iranian Conference on Renewable Energy & Distributed Generation (ICREDG), Tehran, Iran, 11-12 June 2019; pp. 1-6. DOI: 10.1109/ICREDG47187.2019.194227

[27] Abed JN, Abdoos AA, Ebadi A . A new hybrid intelligent method for short-term photovoltaic power forecasting based on combination of ELM, VMD and PCA. Electr. Power Syst. Res. 2026, 250, 112104. DOI: 10.1016/j.epsr.2025.112104

[28] Cui C, Wu H, Jiang X, Jing L. Short- and medium-term forecasting of distributed PV output in plateau regions based on a hybrid MLP-FGWO-PSO approach. Energy Rep. 2024, 11, 2685-2691. DOI: 10.1016/j.egyr.2024.01.080

[29] Cecaj A, Lippi M, Mamei M, Zambonelli F . Comparing Deep Learning and Statistical Methods in Forecasting Crowd Distribution from Aggregated Mobile Phone Data. Appl. Sci. 2020, 10, 6580. DOI: 10.3390/app10186580

[30] Wang H, Yi H, Peng J, Wang G, Liu Y, Jiang H, et al. Deterministic and probabilistic forecasting of photovoltaic power based on deep convolutional neural network. Energy Convers. Manag. 2017, 153, 409-422. DOI: 10.1016/j.enconman.2017.10.008

[31] Huang CJ, Kuo PH . Multiple-Input Deep Convolutional Neural Network Model for Short-Term Photovoltaic Power Forecasting. IEEE Access 2019, 7, 74822-74834. DOI: 10.1109/ACCESS.2019.2921238

[32] Zang H, Cheng L, Ding T, Cheung KW, Liang Z, Wei Z, et al. Hybrid method for short‐term photovoltaic power forecasting based on deep convolutional neural network. IET Gener. Transm. Distrib. 2018, 12, 4557-4567. DOI: 10.1049/iet-gtd.2018.5847

[33] Qing X, Niu Y . Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM. Energy 2018, 148, 461-468. DOI: 10.1016/j.energy.2018.01.177

[34] Luo X, Zhang D, Zhu X . Deep learning based forecasting of photovoltaic power generation by incorporating domain knowledge. Energy 2021, 225, 120240. DOI: 10.1016/j.energy.2021.120240

[35] Wang K, Qi X, Liu H. A comparison of day-ahead photovoltaic power forecasting models based on deep learning neural network. Appl. Energy 2019, 251, 113315. DOI: 10.1016/j.apenergy.2019.113315

[36] Ma Y, Huang Y, Yuan Y . The short-term forecasting of distributed photovoltaic power considering the sensitivity of meteorological data. J. Clean. Prod. 2025, 486, 144599. DOI: 10.1016/j.jclepro.2024.144599

[37] Wang H, Zhang Y, Liang J, Liu L . DAFA-BiLSTM: Deep Autoregression Feature Augmented Bidirectional LSTM network for time series prediction. Neural Netw. 2023, 157, 240-256. DOI: 10.1016/j.neunet.2022.10.009

[38] Guo X, Mo Y, Yan K . Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods. Sensors 2022, 22, 9630. DOI: 10.3390/s22249630

[39] Lee W, Kim K, Park J, Kim J, Kim Y . Forecasting Solar Power Using Long-Short Term Memory and Convolutional Neural Networks. IEEE Access 2018, 6, 73068-73080. DOI: 10.1109/ACCESS.2018.2883330

[40] Elizabeth Michael N, Mishra M, Hasan S, Al-Durra A . Short-Term Solar Power Predicting Model Based on Multi-Step CNN Stacked LSTM Technique. Energies 2022, 15, 2150. DOI: 10.3390/en15062150

[41] Guo F, Yang C, Xia D, Xu J . Short-Term Prediction of Photovoltaic Power Based on Improved CNN-LSTM and Cascading Learning. Energy 2025, 122, 1975-1999. DOI: 10.32604/ee.2025.062035

[42] Xu X, Guan L, Wang Z, Yao R, Guan X . A double-layer forecasting model for PV power forecasting based on GRU-Informer-SVR and Blending ensemble learning framework. Appl. Soft Comput. 2025, 172, 112768. DOI: 10.1016/j.asoc.2025.112768

[43] Guo S, Fan H, Huang J . Ultra-short-term PV power prediction based on an improved hybrid model with sky image features and data two-dimensional purification. Energy 2025, 331, 137018. DOI: 10.1016/j.energy.2025.137018

[44] Xia Y, Su Y, Yu X, Wang Z, Su J, Zhu S, et al. Ultra‐Short‐Term Power Forecasting of Single‐Axis Tracking PV System Based on Irradiance Transformation and CNN‐BiGRU Model. IET Renew. Power Gener. 2026, 20, e70214. DOI: 10.1049/rpg2.70214

[45] Zhang X, Gao R, Zhu C, Liu C, Mei S . Ultra-short-term prediction of regional photovoltaic power based on dynamic graph convolutional neural network. Electr. Power Syst. Res. 2024, 226, 109965. DOI: 10.1016/j.epsr.2023.109965

[46] Hendrikx NY, Barhmi K, Visser LR, de Bruin TA, Pó M, Salah AA, et al. All sky imaging-based short-term solar irradiance forecasting with Long Short-Term Memory networks. Sol. Energy 2024, 272, 112463. DOI: 10.1016/j.solener.2024.112463

[47] Li H, Wu W, Chen W, Zhang M . RTI-Net: Physics-informed deep learning for photovoltaic power forecasting. Renew. Energy 2026, 256, 124152. DOI: 10.1016/j.renene.2025.124152

[48] Zhang S, Dong Z, Qi D, Wang M, Xu Z, Han Y, et al. Intra-Hour PV Power Forecasting Technique Based on Total-Sky Images. CSEE J. Power Energy Syst. 2026, 12, 210-219. DOI: 10.17775/CSEEJPES.2022.07770

[49] Ruan G, Chen X, Li Y, Lim EG, Fang L, Jiang L, et al. SkyNet: A Deep Learning Architecture for Intra-hour Multimodal Solar Forecasting with Ground-based Sky Images. Renew. Energy 2026, 256, 124354. DOI: 10.1016/j.renene.2025.124354

[50] Jittratorn N, Chi W, Huang CM, Yang HT . Ultrashort-term PV power forecasting based on weather classification and dynamic cloud tracking: Deterministic and probabilistic modeling. Energy Convers. Manag. X 2026, 30, 101713. DOI: 10.1016/j.ecmx.2026.101713

[51] Zhang Q, Zeng Z, Zhou Q, Ding X, Dong Z, Tian M, et al. Multimodal bidirectional MAMBA framework for ultra-short-term photovoltaic power forecasting using sky images. Renew. Energy 2026, 256, 124361. DOI: 10.1016/j.renene.2025.124361

[52] Walker J, Gupta A, Hebert M . Dense Optical Flow Prediction from a Static Image. In Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, Chile, 7-13 December 2015; pp. 2443-2451. DOI: 10.1109/ICCV.2015.281

[53] LeCun Y, Bengio Y, Hinton G . Deep learning. Nature 2015, 521, 436-444. DOI: 10.1038/nature14539

[54] He K, Zhang X, Ren S, Sun J. Deep Residual Learning for Image Recognition. In Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27-30 June 2016; pp. 770-778. DOI: 10.1109/CVPR.2016.90

[55] Hochreiter S, Schmidhuber J . Long Short-Term Memory. Neural Comput. 1997, 9, 1735-1780. DOI: 10.1162/neco.1997.9.8.1735

[56] Cho K, van Merrienboer B, Gulcehre C, Bahdanau D, Bougares F, Schwenk H, et al. Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), Doha, Qatar, 25-29 October 2014; pp. 1724-1734. DOI: 10.3115/v1/D14-1179

[57] Qin Y, Song D, Chen H, Cheng W, Jiang G, Cottrell GW . A Dual-Stage Attention-Based Recurrent Neural Network for Time Series Prediction. In Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence, Melbourne, Australia, 19-25 August 2017; pp. 2627-2633. DOI: 10.24963/ijcai.2017/366

[58] Victoria AH, Maragatham G . Automatic tuning of hyperparameters using Bayesian optimization. Evol. Syst. 2021, 12, 217-223. DOI: 10.1007/s12530-020-09345-2

[59] Yuzer EO, Bozkurt A . Deep learning model for regional solar radiation estimation using satellite images. Ain Shams Eng. J. 2023, 14, 102057. DOI: 10.1016/j.asej.2022.102057