Numerical modeling of 2D sutterby fluid stagnation-point flow over a stretching surface using levenberg-marquardt backpropagation.

Two-dimensional (2D) prediction of Sutterby fluid flow is important in polymer synthesis, industrial coatings, aviation and thermal regulation, and this work utilizes the Levenberg-Marquardt backpropagation (LMBP) method. The calculation of Sutterby nanofluid stagnation point problem over a stretching surface is solved by applying an artificial neural network (ANN) based on LMBP. In this case, the mathematical equations for the non-Newtonian Sutterby dynamic system are dimensionless velocity and energy equations. The key novelty in this investigation is the use of a very powerful soft-computing neural network framework for mapping this particular non-Newtonian configuration, for which the high-fidelity benchmark reference data sets are created using the classical numerical bvp4c solver. The data set is split into 76% for training the network, 13% for validating the network, and 11% for testing the network. The accuracy of the ANN-LMBP is shown by the absolute error charts, which show low mean squared error for both mass and heat transfer systems. The skin friction rises by up to 44.6289% from the angle of the inclined angle of the magnetic field parameter 30° to 90° and energy transfer decreases by 11.1167% respectively. On the other hand, the rate of energy propagation is enhanced by up to 14.4513% with the increment of the thermal radiation factor from 3.0 to 7.0. The error histograms, mean squared errors, correlation coefficients, state transition and regression metrics as a whole certify the accuracy, competency and reliability of the proposed ANN-LMBP approach.