Intelligent computing system-based Bayesian regularization backpropagation for viscous dissipative transport of a ferrofluid (Fe3O4) model

Abstract

This article analyzed the viscous dissipative transport of a ferrofluid (Fe3O4) model (VDFM) on a moving surface affected by a thermal deposition and magnetic dipole. The new intelligent computingbased Bayesian Regularization technique with the back propagated neural network (BRT-BPNN) is used to analyze the VDFM problem. The Lobatto IIIA technique is applied to solve the ODEs for obtaining the reference dataset for the designed BRT-BPNN solver. This dataset facilitates to calculate the estimated solution of VDFM with the help of a designed solver. In order to compare the produced results with the reference results, the training, testing, and validation of the BRT-BPNN model are assessed in the generated scenarios. Several measures, including mean square error (MSE), error histogram (EH), and regression plots, are used for the fluidic system convergence analysis to gauge how effectively the BRT-BPNN infrastructure model is working. The flow effects on the velocity profile f(χ), and temperature distribution θ(χ) are examined. The results obtained by fine-tuning of the dimensionless parameters of interest i.e. magnetic dipole (α), Eckert number (Ec), ferromagnetic interaction parameter (β), constant parameter (ξ), radiation parameter (Rd), and Prandtl number (Pr) are also discussed with the help of different plots and tables.