Neural networking-based analysis of heat transfer in MHD thermally slip Carreau fluid flow with heat generation

Abstract

The formulation of heat transfer in non-Newtonian fluid models remains a topic of great interest for researchers. The ultimate flow differential equations in this direction are non-linear and hence difficult to solve analytically. Therefore, we offer state of the art determination of film effectiveness (heat transfer coefficient) by using artificial intelligence (AI). To be more specific, the Carreau fluid model is formulated at a flat surface along with thermal slip, heat generation, velocity slip, chemical reaction, and magnetohydrodynamics (MHD) effects. The developed equations are reduced by the application of the Lie symmetry approach and solved by the shooting method. The artificial neural networking model is built by using 132 sample values of the Nusselt number (NN) as an output. Porosity parameter, Prandtl number, magnetic field parameter, and heat production parameter are all inputs. 92 (70 %) is designated for training, whereas 20 (15 %) is designated for validation and testing. The Levenberg-Marquardt algorithm is used to train the neural network. The constructed artificial neural networking (ANN) is best to predict the NN at a flat surface. It is observed that for large Prandtl numbers, the magnitude of the Nusselt number is greater for porous surface, but the converse is true for magnetic field parameter.

Keywords

• Heat transfer
• Carreau fluid
• Thermal slip
• Heat generation
• Heat transfer coefficient
• Artificial intelligence