Effect of Thermal Radiation on Electrically Conducting Nanofluid with Slip Conditions and Heat Source Using Artificial Neural Networks

In science and engineering, analytical and numerical methods have been utilized to compute the solutions of partial differential equations. Mathematicians are now employing artificial neural network (ANN) to investigate complex problem involving highly nonlinear partial differential equations. The aim of current study is to implement the Bayesian regularization methodology (BRM) to investigate the heat generation/absorption in thermally radiative nanofluid flow over stretched surface embedded in a porous medium with slip conditions. The flow governing equations have been obtained through boundary layer approximation theory. Reference data has been accumulated using the simplified form of equations with the help of “ND Solver” in Mathematica. Different scenarios have been created, and for each scenario, three different cases have been developed. Additionally, the reference data has been trained, tested, and validated using Bayesian regularization methodology in MATLAB. The reference data has been distributed in the following portions, training 71% and testing and validation 15% each, during simulation of each scenario. The proposed methodology’s effectiveness and accuracy have been validated through generated mean square error (MSE), error histograms (EH), proposed solutions, absolute errors (AE) comparative comparison plots, and regression plots. It has been observed that maximum absolute error 10 −08 found for mass suction effect. Convergence of proposed methodology has been validated with mean square error. It has been observed that with high applied magnetization force and permeability motion of fluid declines whereas opposite behavior has been observed for non uniform inertial force. It has been noted that when slip mechanism is significant, contraction has been observed in thermal boundary layer of fluid. Present outcomes of drag coefficient and heat transfer coefficient show good agreement with already published results.