A model development of high-efficiency photovoltaic thermal collectors using artificial neural network to improve solar thermal radiation

The continuous rise in global energy demand and the thermal degradation of photovoltaic modules under high solar irradiation have created a strong need for advanced photovoltaic Thermal (PV/T) systems with improved heat removal and energy conversion efficiency. Conventional working fluids exhibit limited thermal conductivity, which restricts the overall performance of PV/T collectors and shortens module lifespan. To address this challenge, the present study investigates the use of a ternary hybrid nanofluid based on Carboxymethylcellulose (CMC)–water containing ZrO 2 , Cu, and Al 2 O 3 nanoparticles, combined with the Cattaneo–Christov thermal relaxation model to capture non-Fourier heat transport effects. The model incorporates the influence of Magnetohydrodynamic (MHD) flow, solar thermal radiation, and internal heat generation. The governing nonlinear partial differential equations were transformed and solved numerically using the using the Laplace–Hankel transform method, with grid-independence and convergence verification. The Artificial Neural Network (ANN) demonstrated excellent predictive capability, with high correlation coefficients for the training, validation, and testing datasets. The model was trained using a comprehensive dataset generated from the numerical Laplace–Hankel Transform Method (LHTM) solution and optimized with the Levenberg–Marquardt backpropagation algorithm. The strong agreement between predicted and target values, together with low error metrics, confirms the reliability and generalization ability of the proposed ANN model. The results reveal that, at a total nanoparticle volume fraction of [Formula: see text], the ternary hybrid Nanofluids increases the average Nusselt number by 34.8% and the heat transfer rate by 29.6% compared with the CMC–water base fluid. The inclusion of thermal radiation enhances the temperature distribution by 27.4%.