Open AccessSolar Energy Forecasting using Machine Learning Techniques for Enhanced Grid Stability
Author(s) -
Attuluri R Vijay Babu,
N Bharath Kumar,
Rajanand Patnaik Narasipuram,
Soundhar Periyannan,
Alireza Hosseinpour,
Aymen Flah
Publication year - 2025
Publication title -
ieee access
Language(s) - English
Resource type - Magazines
SCImago Journal Rank - 0.587
H-Index - 127
eISSN - 2169-3536
DOI - 10.1109/access.2025.3574093
Subject(s) - aerospace , bioengineering , communication, networking and broadcast technologies , components, circuits, devices and systems , computing and processing , engineered materials, dielectrics and plasmas , engineering profession , fields, waves and electromagnetics , general topics for engineers , geoscience , nuclear engineering , photonics and electrooptics , power, energy and industry applications , robotics and control systems , signal processing and analysis , transportation
The increasing integration of solar photovoltaic (PV) systems into modern energy grids presents significant challenges due to the intermittent and weather-dependent nature of solar energy generation. Accurate short-term forecasting is essential to ensure grid stability and optimize energy resource allocation. This study proposes a comprehensive data-driven framework for solar energy forecasting using multiple machine learning (ML) techniques, including Multiple Linear Regression, Ridge, Lasso, Decision Tree Regression, Support Vector Regression, and ensemble-based models such as Random Forest, AdaBoost, Bagging, and Gradient Boosting Regressors. The framework incorporates advanced feature engineering using high-resolution meteorological and solar geometric parameters-such as relative humidity, temperature, cloud cover, zenith angle, azimuth, and angle of incidence-to enhance model accuracy. Historical solar power and weather datasets were used to train and evaluate the models across multiple performance metrics. Among the models, the Gradient Boosting Regressor demonstrated the best performance, achieving an R 2 of 0.827, RMSE of 399.44, and MAE of 253.62, marking a significant improvement over baseline models. The study also evaluates model robustness and discusses feature relevance, hyperparameter optimization strategies, and deployment considerations for real-time grid operations. These findings provide practical insights for stakeholders aiming to implement intelligent solar forecasting systems in smart grid environments, thereby contributing to enhanced energy management and grid resilience.