Neural network-based analysis for radiative entropy induced rheological material

• Artificial neural networks (ANNs) analysis for MHD rheological material is addressed. • ANN is proposed to address numerous features of fluid flows like analysis, testing, authentication and training. • Entropy generation rate with heat generation and radiation impacts is addressed. • Heat generation, Chemical reaction and thermal radiation features are discussed. Attention here is focused to address the artificial neural networks study for magnetohydrodynamic flow of second grade material by stretched surface. Furthermore, employing advanced artificial neural networks based computational techniques such as multi-layer neural networks provide exceptional capabilities in precisely capturing the complex nature of thermal and solutal transport in liquid flow problems. Energy expression comprises heat generation and thermal radiation. Entropy generation rate with heat generation and radiation impacts is under consideration. Isothermal chemical reaction of first order is taken. Dimensionless ordinary systems are developed. The obtained nonlinear expressions are numerically computed by Bvp4c via MATLAB and then advanced artificial neural networks computational algorithm is employed to train the obtained datasets to improve predictive capabilities for advanced solutions. Velocity, entropy rate, concentration and thermal distribution against sundry variables are examined. Artificial neural networks is used to optimize interesting quantities data through training, validation, and testing in order to verify validation of given data. Additionally, the comparison of Bvp4c method and artificial neural networks algorithm is given. Here we concluded that an increase in liquid flow via larger material variable is witnessed whereas decreasing impact holds for magnetic field parameter. An increase in entropy generation through heat generation variable is detected. Larger Schmidt number corresponds to concentration reduction. Through comparative study it is noticed that numerical computation through ANNs model aligns well with Bvp4c results.