RF-ET-ANN: Hybrid Machine Learning Model for Forecasting Short-Term Photovoltaic Power Production

Accurately forecasting photovoltaic (PV) power generation is a challenging problem due to the non-linear and highly variable characteristics of solar data, which is strongly influenced by several interdependent factors, such as environmental conditions, system characteristics, and technical aspects. In this study, a newly proposed machine learning (ML) technique based on integrating random forest (RF), extra trees (ET), and artificial neural networks (ANN) is presented to tackle this problem with better predictive accuracy, ensuring that solar energy is used more consistently, effectively, and economically; this model is dubbed RF-ET-ANN for short. Hybridization of these three models will help capture non-linear aspects in the solar data, reducing overfitting and boosting forecasting accuracy. Four popular datasets, namely Planet_1, Planet_2, PVO, and SolarTech, are utilized to evaluate the performance of RF-ET-ANN under different environmental conditions. In addition, a grid search combined with cross-validation was employed for hyper-parameter tuning of each model separately to maximize its performance when applied to this problem. Several rival models were compared to RF-ET-ANN using a variety of performance metrics to demonstrate its efficacy and efficiency when applied to this problem. According to the experimental results, RF-ET-ANN could outperform all compared models in terms of all performance metrics on four solved datasets, making it a viable alternative for improving energy management in renewable systems and increasing grid stability.