Home / Current Issue / Paper 1714508
AI in Solar Forecasting: Advanced Machine Learning Techniques for Photovoltaic Power Prediction
Subject area: Science,Engineering and Technology · Area of research: Machine Learning
DOI: https://doi.org/10.64388/IREV9I8-1714508
Abstract
Solar energy forecasting is critical for grid stability and renewable energy integration. This paper reviews artificial intelligence techniques applied to solar forecasting, focusing on advances from 2023-2026. We examine deep learning architectures including LSTM networks, CNNs, Transformer-based models, and hybrid approaches. Analysis of 242 studies reveals that hybrid CNN-LSTM models achieve 15-30% MAE reductions compared to standalone models. Deep learning excels for short-term predictions, while ensemble approaches benefit day-ahead forecasts. Key challenges include data quality, computational complexity, and model generalization. This review synthesizes methodologies, performance metrics, and future directions including transfer learning and physics-informed neural networks
Keywords
CNN, Deep Learning, Hybrid Model, LSTM, Photovoltaic Power Prediction, Renewable Energy, Solar Forecasting
How to cite this paper
@article{1714508,
author = {Archana. P. Haral , Manisha. R. Shiledar, Suriyakala A V},
title = {AI in Solar Forecasting: Advanced Machine Learning Techniques for Photovoltaic Power Prediction },
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {8},
pages = {1589-1592},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1714508.pdf},
abstract = {Solar energy forecasting is critical for grid stability and renewable energy integration. This paper reviews artificial intelligence techniques applied to solar forecasting, focusing on advances from 2023-2026. We examine deep learning architectures including LSTM networks, CNNs, Transformer-based models, and hybrid approaches. Analysis of 242 studies reveals that hybrid CNN-LSTM models achieve 15-30% MAE reductions compared to standalone models. Deep learning excels for short-term predictions, while ensemble approaches benefit day-ahead forecasts. Key challenges include data quality, computational complexity, and model generalization. This review synthesizes methodologies, performance metrics, and future directions including transfer learning and physics-informed neural networks},
keywords = {CNN, Deep Learning, Hybrid Model, LSTM, Photovoltaic Power Prediction, Renewable Energy, Solar Forecasting},
month = {February},
doi = {https://doi.org/10.64388/IREV9I8-1714508}
}