Modeling and control of a hybrid PV-T collector using machine learning

Photovoltaic-thermal (PV-T) systems are expected to fulfil an increasingly vital role in future energy production. The current research endeavors to showcase machine learning modeling and control of a water-based PV-T collector. In this work, the PV-T collector is modeled using a decision tree algorithm and artificial neural network (ANN). The predicted outputs are compared with the actual outputs to validate the models. The ANN-based model performed better and proved its efficacy in training and testing. Further, various control strategies are implemented and their performance is compared. All the techniques presented are illustrated through simulation results.