Dynamics of stratified-convected Eyring-Powell nanoliquid featuring chemically reactive species and Ohmic dissipation: Application of Levenberg-Marquardt artificial neural networks(ALM-ANNs)

Abstract

Magnetic nanoliquids are pondered cutting-edge heat transference liquids, representing a modern generation because of their aptitude to function as smart liquids. The deployed magnetism externally exerts a readily manageable impact, escalating their adaptableness and practicality in various applications. In this analysis, electro magnetized Powell-Eyring rheological nanoliquid is modeled considering stretching cylinder based buoyancy-driven flow. Transport expressions include thermal radiation, Brownian diffusion, chemical reaction, thermophoresis, and thermal source effect. Prandtl theory of boundary-layer (BL) assists to develop the governing problems. Appropriate transformations are deployed to obtain the dimensionless problems which are then computed numerically through the novel Levenberg-Marquardt artificial neural networks(ALM-ANNs). It is worth to mention that the close agreement between the error analysis of the reference dataset and the proposed datasets roughly from E−10 to E−03, further justifies the approaches utilized in this study. Besides the excellent measures of performance in terms of MSE are achieved at level 5.77E−11, 4.41E−11, 2.09E−12, 2.16E−11, 1.03E−11, 1.65E−11, 2.00E−11, and 1.91E−11against 272, 416, 266, 391, 142, 366, 425, and 508 epochs for first instance of all eight various scenarios. The acquired outcomes are compared with alternate analytical (homotopy) scheme and good agreement is witnessed. Furthermore it is noticed that as the values of Pr,RandS1 increase in size, so does the EPFM temperature profile θ, and the evaluation of AE is between E-04 to E-10.