Artificial Neural Network Based Prediction of Thermal Behavior of Electromagnetohydrodynamic Copper–Aluminum Oxide–Silver /Water Ternary Nanofluid Over a Surface

The article investigates the ANNs approach for behaviour, its prediction and comparative study of Electromagnetohydrodynamic (EMHD) copper (Cu)–aluminium oxide (Al₂O₃)–silver (Ag)/water (H₂O) ternary nanofluid as well as single (Ag/water) and hybrid (Al 2 O 3 +Ag/water) nanofluid with respect to variations in magnetic parameter, volume of nanoparticles and Prandtl number. The non-linear ODEs are obtained from nonlinear PDEs through similarity transformations. Initially, the equations are evaluated by bvp4c in MATLAB software for the behaviour, heat transfer rate and drag force of nanofluids on the surface. Later, adopted a machine learning technique feed forward ANN (Artificial Neural Network) for computing the prediction of skin friction and rate of heat transfer. In this process, the 80% of bvp4c computational data is utilised for training, with the remaining data used for case-specific validation and testing, also used LRT (Linear regression technique) for validation of ANN results. Finally, the obtained performance measures including MSE, RMSE, correlation coefficient (R), and sensitivity analysis confirmed the excellent predictive capability and robustness of the ANN model. The results demonstrate excellent agreement between predicted and numerical solutions, while the error histogram indicates that most prediction errors are concentrated around zero. Moreover, ANN helps to forecast the Nusselt number and skin friction with the variation of parameters and observed that compared to LRT, the ANN results are more accurate. As the effect of magnetic field, ternary, hybrid and single nanofluids velocities are high and temperature is low due to enhanced Lorentz force, which enhances the fluid stream and reduces the thermal boundary layer thickness. The current analysis is very applicable in the areas of electromagnetic cooling systems, metallurgical process, solar energy collector, and chemical reactors where the magnetically controlled nanofluid flow is an important concept. Moreover, the created model of Artificial Neural Network does offer a sound predictive model of estimating skin friction and heat transfer rates that can assist in optimizing and designing complex thermal engineering systems.