Optimizing solar collector efficiency and safety: A comparative thermal analysis of non-toxic hybrid nanofluid mixtures using machine learning

Ethylene glycol is extensively used in solar energy systems because of its thermo-physical properties; however, its toxicity presents health and environmental risks. To overcome this, non-toxic solutions such as propylene glycol or water-ethylene glycol blends are promoted, keeping system efficiency while enhancing safety and sustainability. This study proposes the integration of advanced machine learning (ML) and artificial intelligence (AI) with computational fluid dynamics (CFD) for the thermal analysis of a mixture comprising three distinct base fluids: Ethylene Glycol (EG)-water, Propylene Glycol (PG)-water, and EG with hybrid nanoparticles, aimed at minimizing toxicity and production costs in solar collector energy systems. The effect of non-Fourier heat flux on the Blasius-Rayleigh-Stokes variable (BSRV) flow of a hybrid nano-fluid across a plate is investigated numerically for this purpose. Hyper-parameter optimization is performed for four alternative AI training methods to determine the best suitable choice. Whereas for numerical simulation, the Keller-Box method (KBM), a modified finite difference methodology, is employed. Regression scores of 1 indicate an impeccable correspondence between numerical information and the predictions. Conclusively, a comparative analysis is presented to support our claim, which states that by using combination of PG-Water, similar heat transfer rate can be achieved, which is less harmful and also cost effective.