Artificial neural network surrogate for Ellis hybrid nanofluid with Soret-Dufour effects and Joule heating: Enhance heat and mass transfer

Abstract

The suggested artificial neural networks surrogate model for Ellis’s hybrid nanofluid with Dufour-Soret characteristics and Joule heating under local thermal non-equilibrium conditions has numerous applications in advanced energy and thermal systems. It can increase mass and heat transmission in electronic cooling systems, nuclear reactors, geothermal reservoirs, and microfluidic devices. Furthermore, it is beneficial for optimizing systems involving electrically conducting fluids, such as MHD power generation, solar thermal collectors, and medicinal heat treatments, where accurate prediction of complicated non-linear transport phenomena is required. To examine how an Ellis hybrid nanofluid’s Darcy-Forchheimer flow is affected by LTNE across a spinning disk. The key aim of this study is to provide a new mathematical intelligence approach of the AI-based intelligent Levenberg–Marquardt technique under the influence of artificial neural networks (ILMT-ANNs) for optimizing Soret-Dufour effects and Joule heating in conjunction with Magneto-Marangoni phenomena. For ILMT-ANNs, the numerical data-sheet is separated into 90.8% training, 4.1% testing, and 5.1% validation. The estimated solution is analyzed, and its evaluation with a numerical solution using bvp4c is described. Regression investigation, error histogram, and fitness curves based on mean squared error (MSE) are used to verify the effectiveness and consistency of ILMT-ANNs.