Artificial Neural Network Analysis for Carbon Nanotubes-Based Nanofluid Flow Over Exponentially Stretching Sheet

The concern of our present investigation is to develop and validate multilayer feed-forward neural network model to forecast local Nusselt number and skin friction number for convective heat transfer of electrical magneto hydrodynamics (EMHD) flow of water-based nanofluid over an exponentially stretching sheet involving single-walled carbon nanotubes with an impact of thermal radiation. The artificial neural network model developed is a function of various pertinent parameters. The volume fraction of nanoparticle has been varied gradually from 0.0 to 0.20. The weights and bias of the constructed neuromorphic model have been adjusted by Levenberg–Marquardt learning algorithm using datasets obtained from solving the governing equations by implementing finite difference scheme. The bvp4c function of MATLAB has been utilized for this purpose. Statistical accuracy analysis validated that the consequences obtained from the postulated backpropagation neural network model are in remarkable agreement with the numerical results.