Design of artificial neural network to predict heat and mass transfer of Casson fluid over a shrinking curve: A stability analysis

Abstract

This investigation focuses on a novel examination of the Darcy-Forchheimer model of Casson fluid under slip constraints via a penetrable shrinking curve. The significance of variable thermal conductivity, viscous dissipation, multiple slip conditions and chemical reactions are uniquely incorporated. The governing PDEs are truncated into ODEs under the implementation of similarity variables. To train the solution data for reliable predictions under changing factor values, artificial neural networks (ANNs) are created, and a dual solution analysis is also conducted through the bvp4c approach. The major findings represent that the first solution is stable and the second is unstable via temporal stability analysis. The ANN's efficacy as a quick and precise surrogate model for assessing energy transfer performance in intricate Casson fluid systems was confirmed by the close match between its prediction and the numerical data. The variations in curvature factor result in reduced shear stress and mass transportation rates. Furthermore, flow, thermal, and mass transfer are estimated using regression analysis, and the average estimation was accurately predicted by an artificial neural network model (R = 1).