AN APPLICATION OF ARTIFICIAL NEURAL NETWORK TOWARD THE MATHEMATICAL MODELING OF MHD TANGENT HYPERBOLIC NANOFLUID ACROSS A VERTICAL STRETCHING SURFACE

Abstract

The Levenberg-Marquardt (LM) back propagation (BP) artificial neural networks (ANNs) (LM-BP-ANNs) procedure is used in this analysis to show the computational strategy of neural networks for the simulation of magnetohydrodynamics tangent hyperbolic nanofluid flow comprised of motile microorganism across a vertical slender stretching surface. The fluid flow were examined under the significance of chemical reaction, magnetic field, activation energy, and heat source. The modeled equations were simplified to the ordinary system of differential equations using similarity variables substitution. The Lobatto IIIA formula based on the finite difference method was employed for the nano-liquid flow problem with an accuracy up to five decimal points. The robustness of Lobatto IIIA is its straightforward execution of very nonlinear coupled differential equations. Several operations involving testing, authentication, and training were carried out by developing a scheme for different fluid problem elements using reference datasets. The accuracy of LM-BP-ANNs was tested through mean-square error, error histogram, curve fitting figures, and regression plot. Moreover, the examination of flow model factors for concentration, mass, and momentum outlines are expressed through graphs. It was perceived that the velocity field declines with the flourishing influence of the magnetic field and lessens with the upshot of Weissenberg number and power law index.