Solar Irradiance Prediction for Optimal Photovoltaic Power Generation: A Comprehensive Review of Artificial Intelligence Replicas

Abstract

Many variables, such as the growing world population and shifting consumer tastes, might be linked to rising energy demands and power consumption. The rapid depletion of fossil fuels, the worrisome rise in air pollution, the impact of global warming on more frequent natural disasters are additional major reasons to adopt clean and renewable energy sources for generating power and improving energy efficiency.A key source of renewable energy is the generation of solar power. The potential power output of a photovoltaic system is greatly influenced by solar irradiation. Predicting energy production from solar power plants, climate modelling, weather forecasting are just a few of the crucial applications for studying and measuring solar irradiance. The many tools and techniques for estimating solar irradiance using data-driven methodologies based on machine learning and deep learning algorithms are summarised in this study. The forecasting horizons that can be employed by the algorithm depend on the input data used to “train” it. Although it has been demonstrated that these algorithms can estimate solar radiation, differences in their performance make comparing and selecting the best method an interesting job. The two most often utilised machine learning techniques in this study are artificial neural networks (ANN) and support vector machines (SVM), along with a comparison of these two techniques with Deep Learning models.