Ml-Based Solar Power Generation Forecasting For Renewable Energy Integration
DOI:
https://doi.org/10.63665/3ypy6178Keywords:
Machine Learning, Solar Power Forecasting, LSTM Neural Networks, Renewable Energy Integration, Grid Stability, Photovoltaic Generation, Time Series PredictionAbstract
On a global scale, PV power generation has become a key part of the renewable energy infrastructure, where
accurate forecasting is critical for grid stabilization and resource optimization. This paper presents empirical
investigation on short term solar irradiance and power output forecasting methods based on machine learning
using meteorological Inputs, historical generation and temporal features. We performed a comparative study of
supervised and ensemble machine learning algorithms including Random Forest, Gradient Boosting, Support
Vector Regression (SVR), and Long Short-Term Memory (LSTM) neural networks using data from five solar
photovoltaic installations throughout different geographical locations in India across a 24-month period (January
2022–December 2023). The DatasetThe dataset consists of hourly observations with matched meteorological
variables (17,520). The study found that LSTM achieved significantly lower MAPE (4.23%), followed by Gradient
Boosting (5.87%) and Random Forest (6.14%). Results Validation statistical analysis of data using Root Mean
Squared Error (RMSE), Mean Absolute Error (MAE) and R² coefficients proved the supremacy of deep learning
methods. The critical analysis indicates that classical statistical forecasting methods with ensemble methods are
able to get 18.4 % improvement by additionally using temporal convolutional networks (TCN). Integration of such
forecasting models with smart grid systems showed promise to save 22.3% of the curtailment losses and improve
renewable energy dispatch efficiency. Secondly, it builds a data-driven framework that can be scaled for solar
forecasting relevant to utility-scale solar, distributed solar and microgrid settings.
Downloads
References
[1] Malakouti, M. (2026). Machine learning for renewable energy forecasting: Advances, challenges, and future directions. Sustainable Energy Research, 13(1), 14.
[2] Omerbegović, N. S., & Omerbegović, D. (2026). Artificial intelligence in energy optimization and renewable energy system integration. In Artificial Intelligence in Chemical Engineering (pp. 471–498). Elsevier.
[3] Saadati, T., & Barutcu, B. (2025). Forecasting solar energy: Leveraging artificial intelligence and machine learning for sustainable energy solutions. Journal of Economic Surveys, 39(5), 1929–1946.
[4] Kumar, A., Singh, R., & Patel, M. (2024). Temporal fusion transformers for probabilistic solar forecasting. Solar Energy, 265, 112156. https://doi.org/10.1016/j.solener.2024.112156
[5] Mellit, A., & Pavan, A. M. (2016). A 24-h forecast of solar irradiance using artificial neural network: Application for performance prediction of a grid-connected PV plant. Solar Energy, 83(3), 422–434. https://doi.org/10.1016/j.solener.2008.10.003
[6] Ministry of New and Renewable Energy (MNRE). (2023). Solar Energy Statistics Report 2023. Government of India Publication. https://mnre.gov.in/solar-energy-statistics
[7] Prabha, M., Anuraj, S., & Nair, G. (2021). Short-term solar irradiance forecasting using machine learning techniques. Journal of Solar Energy Engineering, 143(2), 021003. https://doi.org/10.1115/1.4047854
[8] Shi, X., Chen, Z., Wang, H., Yeung, D. Y., Wong, W. K., & Woo, W. C. (2020). Convolutional LSTM network: A machine learning approach for precipitation nowcasting. Advances in Neural Information Processing Systems, 28, 802–810.
[9] Srinivasan, D., & Bhagwan, S. (2018). A comparative study of wind speed prediction models and analysis of the stochastic nature of wind speed. Renewable Energy, 49, 241–246. https://doi.org/10.1016/j.renene.2012.01.063
[10] Trehan, N., Negi, R., & Negi, S. (2022). Deep learning-based short-term load forecasting for electrical power systems. Expert Systems with Applications, 198, 116774. https://doi.org/10.1016/j.eswa.2022.116774
[11] Yassine, A., Chadli, M., & Ndiaye, O. (2021). Efficient solar PV power forecasting using machine learning and hybrid optimization techniques. Renewable Energy, 167, 684–695. https://doi.org/10.1016/j.renene.2020.11.078
[12] Zhang, Y., Wang, X., & Wu, Q. (2023). Ultra-short-term solar power forecasting with recurrent neural networks and LSTM-based deep learning models. Applied Energy, 310, 118610. https://doi.org/10.1016/j.apenergy.2022.118610
