Improving the PV Performance Using Rectangular Channels and Recurrent Neural Network

Photovoltaic systems are widely utilized due to their direct energy conversion capabilities, low maintenance, and simple installation. However, a significant portion of the energy generated by these systems is lost in the form of heat. To harness this thermal energy and maximize its utility, photovoltaic-thermal (PVT) systems have undergone development. These systems integrate the production of both thermal and electrical energy, thereby enhancing overall system efficiency and reducing emissions associated with hot water production. In this research, a cascade channels thermal module is employed alongside PV panels to facilitate cooling and further increase system efficiency. The simulations conducted encompassed parameter ranges of 600–1200 W/m<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of solar radiation, 0.06 to 0.117 kg per second of flow rates, and a coefficient of convection ranging from 5.48 to 20.7 W/m<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>.K. By implementing this cooling configuration, thermal, electrical, and overall efficiencies of 63.42%, 13.95%, and 77.21 %, respectively, were achieved. Additionally, a dynamic artificial neural network called Recurrent Neural Network (RNN) is proposed to simulate the PVT system and attain its performance. Moreover, a novel optimal construction of the RNN is obtained by using Gorilla Troops Optimizer (GTO) algorithm, which is also used for training the RNN. Consequently, a hybrid model combining RNN and GTO (RNN-GTO) is created to evaluate the performance of PVT system.