Application of artificial intelligence brain structure-based paradigm to predict the slip condition impact on magnetized thermal Casson viscoplastic fluid model under combined temperature dependent viscosity and thermal conductivity

One of the main reasons for the popularity of ANNs is their wonderful capability to handle very complex and nonlinear mathematical problems. It can, therefore, provide a very valuable computational framework wide in applications, from biotechnology to biological computation and fluid dynamics. In this article, computational ANN paradigms are utilized in the analysis of boundary layer flow and heat transfer of magnetized Casson fluid over a nonlinearly stretching elastic sheet under velocity slip conditions. Casson fluid behavior under the influence of magnetohydrodynamics with the influence of heat generation, temperature-dependent dynamic viscosity, variable thermal conductivity, and viscous dissipation are considered in this study. With the help of a similarity transformation, the complex partial differential equations governing the flow and energy take on the form of nonlinear ordinary differential equations. These resulting nonlinear ordinary differential equations are solved by using the bvp4c solver in MATLAB. The numerical solutions generated from the bvp4c solver for the problem under consideration are used to develop the reference dataset for the anticipated radiative Casson fluid Levenberg-Marquardt backpropagation neural network (LMT-ABPNN). Finally, developed dataset features were applied to the artificial intelligence-based LMT-ABPNN procedure to validate the numerical results for radiative Casson fluid. This LMT-ABPNN is trained, tested, and validated in predicting the approximate numerical results for RCF under various conditions. The proposed LMT-ABPNN performance validation is carried out based on mean squared error fitness, error histograms, and regression analysis. Results obtained for regression metrics, absolute error, MSE and error histogram plots through the LMT-ABPNN architecture do confirm the superior performance attained here. From the results, the study raises close concordance with respect to reference data. The low MSE indicates that the model is fit to predict with good accuracy, but the minimal absolute error close to zero proves the prediction accuracy of the developed approach at its best.