Review of artificial neural network applications for the optimization of photovoltaic-thermal systems integrated with phase change materials

The increasing world energy demand as well as environmental issues has made significant contributions to the necessity for the development of various renewable energy technologies, particularly solar energy systems. Within these areas of research hybrid photovoltaic-thermal (PVT) collectors stand out as possible sources for the simultaneous production of electricity and heat. This paper is a review of artificial intelligence applications in PVT systems employing phase change materials (PCM). It attempts to analyze, in a systematic manner, the ways in which artificial neural networks can be used to model, predict and optimize PVT-PCM systems. In doing this literature from 2015 through to 2024 is considered, in detail, to evaluate the ANNs used and their architectural efficiencies in terms of administering the complexities of the PVT-PCM systems. It is shown in this paper that multilayer perceptrons can give values of high accuracy for stage limits when the models reach a stage steady situation, and that recurrent networks, such as Long Short-Term Memory (LSTM) models give values which can explain adequately the dynamic behaviour characteristics of the thermal responses produced in this machine. This paper also describes the key variables which contribute to the overall performance of the systems and also describes the next stage in the direction of future research requirements in order that the viability and implementation of PVT-PCM systems can be achieved.