A Systematic Review of Solar Irradiance Forecasting Using Tree-Based Ensemble Learning Algorithms: Emerging Trends and Future Research Directions

Abstract

Solar irradiance is a critical factor in solar energy projects, and its unpredictability can hinder the management and efficiency of solar systems. To overcome this challenge, the development of intelligent and accurate forecasting models has become essential. These models play a key role in predicting energy production, optimizing grid operations, and balancing power supply and demand. While there has been extensive research on solar irradiance forecasting, there remains a gap in the literature regarding the analysis of tree-based ensemble learning methods. In contrast, Artificial Neural Network (ANN) models have been extensively studied. Tree-based ensemble learning models are notable for their lower computational cost and effective performance, especially when compared to the more complex structure of ANN models. This paper has three objectives. First, it provides an overview of the critical elements for developing solar irradiance forecasting models, including data characteristics, preprocessing steps, and feature selection for accurate prediction. Second, it provides a comprehensive review of the state of the art in tree-based ensemble learning methods applied to solar irradiance forecasting. Finally, the paper presents a discussion of the research findings and proposes recommendations and potential directions for future research in this area.