Group theoretic thermal analysis on heat transfer coefficient (HTC) at thermally slip surface with tangent hyperbolic fluid: AI based decisions

Abstract

By following the statistics, over the last few years, the use of Artificial Intelligence conjectured with mathematical models has increased abundantly for physical problems having thermal engineering standpoints. Owning to such importance, we offer Levenberg-Marquadt-based neural networking modeling of non-Newtonian fluid flow at a flat heat-generating surface with thermal and velocity slip effects. A tangent hyperbolic fluid (THF) is considered with the transfer of heat and mass. Heat absorption and generation aspects are carried out by using an energy equation. The equations are reduced by Lie symmetry analysis and solved by shooting. The neural networking model is constructed by using 91 sample values of inputs namely power law index, Weissenberg number, heat generation parameter, and Prandtl number. 63 (70%) for training, while validation, and testing holds 14 (15%), and 14 (15%) respectively. To train the model, the Levenberg-Marquadt backpropagation technique is utilized. Regression analysis enables a satisfactory correlation between the targeted and output values of the non-magnetic Nusselt number on a flat surface for training, validation, and testing. Following the ANN results, the non-magnetic Nusselt number is found to be an increasing function of the Prandtl number, but the opposite is true for the Weissenberg number, and heat production parameter.

Keywords

• Heat transfer
• Nusselt number
• Heat generation
• Neural networks
• Tangent hyperbolic fluid
• Group theoretic method