Intelligence computing for the dynamic of an entropy-optimized hybrid nanofluid system under the impacts of radiation, Joule heating, and energy dissipation

Abstract

In this research paper, the main focus is to inspect Entropy generation of MHD Hybrid nanofluid (MHD-HNF) in a rotating system by considering the AI-based method of Levenberg Marquardt with Back-propagated Neural Network (LM-BPNN). The type of flow is magnetohydrodynamic so all the factors which directly affect the flow like radiation, Ohmic/Joule heating and dissipation in energy expression are discussed by the variation of dimensionless parameters. The system of PDEs presenting the entropy optimized hybrid nanofluid in the rotating system is converted into ODEs using the capability of similarity transformation then the simplified version of fluid flow governing system is further solved by using Adam numerical solver for the computation of the reference dataset with the variation of injection/suction parameter, Reynold Number, Hartmann number, rotation parameter, Brinkman number and heat generation. The validation and correctness of the considered model MHD-HNF are examined by training, testing and validation process of LM-BPNN. Regression analysis, histogram for error and mean square error (MSE) results validate the performance analysis of designed LM-BPNN solver. The assessment of LM-BPNN for velocity and temperature with absolute error analysis plotted graphically and numerically as well as sway of influential parameter on entropy generation and Bejan number analysis are also discussed.