Artificial neural network modeling of the Casson fluid flow over unsteady radially stretching sheet with Soret and Dufour effects

The Dufour and Soret effects on the flow of a Casson fluid about an unsteady radially stretched sheet are analyzed. The system of nonlinear ordinary differential equations is obtained from the equations governing the flow by employing similarity transformations. The solution to these nonlinear ordinary differential equations is calculated using artificial neural networks. The trial functions employ a multilayer perceptron neural network with programmable parameters (biases and masses). In order to fulfill the governing equations, the ADAMS (adaptive moment estimation algorithm) optimization technique is used to calculate the trial solution's adjustable parameters. The results suggest that the artificial neural network-based method gives significant accuracy and that the solution's efficacy increases as the number of neurons in the neural network increases. Also, the computations of skin friction and heat transfer coefficient using the current method are compared with the values obtained by the Runge-Kutta fourth-order method. Further, the impact of relevant parameters on the physical quantities is displayed through graphs. Finally, a comparison using existing literature is made to back up our findings, and an excellent correlation is discovered, affirming our findings. According to the current computation, raising the Soret number improves the Nusselt number and drops the Sherwood number, whereas improving the Dufour number diminishes the Nusselt number and enhances the Sherwood number.

Automated summary

The Dufour and Soret effects on the flow of a Casson fluid about an unsteady radially stretched sheet are analyzed. The solution to the nonlinear ordinary differential equations is calculated using artificial neural networks. The results suggest that the artificial neural network-based method gives significant accuracy and that the solution's efficacy increases as the number of neurons in the neural network increases.