New concepts of magnetohydrodynamics and entropy rate for radiative nanofluid flow invoking artificial neural network approach

In the modern era concept of artificial neural networks is an innovative phenomenon in industrial, mechanical, pharmaceutical and automotive applications. Furthermore, the advanced computational approach based on artificial neural networks (ANNs) employing Levenberg-Marquardt algorithm (LMA) provides exceptional capabilities in accurately obtaining solutions for the complex features of thermal and solutal transport rates in nonlinear flow problems. The current analysis scrutinizes the chemically reactive MHD flow of Reiner-Rivlin nanoliquid by curved stretched surface. Artificial neural networks (ANNs) based on (LMA) is differentiated by its remarkable stability and used to compute the flow characteristic of Reiner-Rivlin nanoliquid employing validation check, regression plots (RP), mean square error (MSE), fitness curve, error histograms and comparative solution analyses. Variable fluid characteristics are under consideration. Energy expression consists of heat generation, Joule heating and thermal radiation. Buongiorno's model is utilized to discuss nanofluid features by thermophoresis and Brownian motion. First order reaction is accounted. New concept of entropy generation in reactive flow with subject to heat generation, magnetohydrodynamics and radiation is considered. Related nonlinear equations are altered into dimensionless ordinary equations by using suitable transformations. Numerical solutions of nonlinear ordinary systems are obtained using Bvp4c scheme via MATLAB. Subsequently the advanced computational technique (ANNs) based on Levenberg-Marquardt algorithm (LMA) is integrated to train the resulting datasets and facilitate predictions of advanced solutions. Analysis for liquid flow, concentration, entropy rate and temperature via pertinent variables are graphically explored. Comparative results of Bvp4c scheme and artificial neural networks (ANNs) model are also discussed.