This work investigates unsteady heat and mass transfer in a squeezing channel filled with a magnetohydrodynamic Eyring–Powell non-Newtonian fluid, incorporating the effects of thermal radiation, chemical reaction, and internal heat generation or absorption. The investigation is significant for understanding complex flow behaviors in industrial and engineering applications where magnetic forces, non-Newtonian fluids, and wall motion interact. Nonlinear equations have been solved using MATLAB. The squeezing motion of the channel walls, together with magnetic forces and non-Newtonian fluid characteristics, produces complex flow behavior marked by velocity suppression in one region of the channel and enhancement in another. Thermal analysis reveals that stronger squeezing and internal heat generation elevate the temperature field, whereas thermal radiation promotes heat dissipation. In addition, an increase in the Schmidt number reduces the concentration distribution due to weakened mass diffusion. To complement the mathematical modeling, an Artificial Neural Network (ANN) framework is employed to capture the nonlinear relationships between governing parameters and the resulting velocity, temperature, and concentration profiles. The complete dataset for all parameters was divided into training, validation, and testing sets, with 70%, 15%, and 15% of the data allocated to each set, respectively. The ANN demonstrates strong predictive capability, as evidenced by low mean squared error values and close agreement between predicted and reference data across training, validation, and testing stages. The combined physical and data-driven analysis provides a comprehensive understanding of the parametric influences governing MHD Eyring–Powell squeezing flows and highlights the potential of ANN-based approaches for analyzing complex nonlinear heat and mass transfer phenomena relevant to advanced engineering and industrial applications. This study’s findings contribute to the analysis of nonlinear heat and mass transfer in advanced engineering and industrial applications, offering a new perspective compared to earlier literature.
Ganji et al. (Wed,) studied this question.