Artificial neural network analysis of natural convection of Casson fluid flow over a curved stretching surface with viscous dissipation

Abstract

The focus of the current study is to analyze the non-similar solutions of natural convection transport of Casson fluid over a curved surface in the presence of a magnetic field, joule heating, and heat source. The flow of Casson fluid in the presence of a magnetic field has applications in bioengineering such as MRI, and artificial neural networks (ANNs) are used to optimize MRI images for better diagnostic results. Non-similar transformations are utilized for the conversion of governing equations into dimensionless partial differential equations (PDEs). The dimensionless system of partial differential equations is treated as a system of ordinary differential equations (ODEs) by utilizing the local nonsimilarity method. The solution of coupled ODEs is obtained through BVP4c. Additionally, velocity profile, temperature profile, and local Nusselt number are studied through the Levenberg-Marquardt method using an artificial neural network. A total of 400 datasets are taken, in which 70 % of this data is used for training and the remaining data is used for validation and testing for each case. The results indicate that velocity is decreasing for both Hartman number and Casson parameter. Fluid temperature is increased for the Grashof number, Casson parameter, and nonlinear radiation parameter. The comparison of artificial neural network (ANN) results with numerically computed results has also shown convergence and good accuracy of the obtained results. The ANN predicted values of skin friction and Nusselt number show stability and convergence with high accuracy.