A Comprehensive Review of the Application of Machine Learning in Fabrication and Implementation of Photovoltaic Systems

Abstract

Solar energy is a promising source of renewable energy, but its low efficiency, instability, and high manufacturing costs remain a big challenge. Recently, machine learning (ML) techniques have gained attention in the photovoltaic (PV) sector because of advances in computer power, tools, and data creation. The existing ML techniques used for fabrication and the different operational procedures of the PV sector have shown very impressive results with a higher degree of accuracy and precision. While previous studies have discussed ML techniques for PV fabrication or operational procedures, there is a lack of end-to-end research that covers the entire process from fabrication to implementation. In this paper, we present a comprehensive review of the application of ML in the field of solar energy, focusing on the development of new materials, enhancement of solar cell efficiency, implementation, and integration with the system, including fault detection, sizing, control, forecasting, management, and site adaptation. We evaluated more than 100 research articles, a significant proportion of which were published in the past three years. In our study investigating ML implementation in solar cell fabrication, we discovered that the Random Forest (RF), Linear Regression (LR), XGBoost, and Artificial Neural Network (ANN) algorithms are the most commonly employed techniques. Our findings demonstrate that XGBoost exhibits superior performance in optoelectronic prediction, while RF, LR, and ANN algorithms are better suited for predicting electrical parameters. Moreover, our analysis indicates recent ML research in this field explicitly emphasizes perovskite solar cells (PSCs). This work also discusses the challenges, directions, insights, and potential applications of ML for future PV system analysis.