Group theoretic thermal analysis (GTTA) of Powell-Eyring fluid flow with Identical free stream (FS) and heated stretched porous (HSP) boundaries: AI Decisions

Abstract

Artificial intelligence (AI) is used to study the thermal flow field for Powell-Eyring fluid model when free stream condition is the same as stretching at the surface. The temperature and concentration flow aspects are considered simultaneously. Both injection and suction processes are taken at surface. A mathematical model is constructed in terms of partial differential equations (PDEs). The theory of Lie symmetry is used to develop a scaling group of transformations (SGT). Later, SGT is used to reduce order of PDEs. The reduced equations are solved numerically by using shooting methodology. A neural networking model is constructed to predict skin friction coefficient (SFC) as a targeted value when Powell-Eyring and suction parameters are used as inputs. The SFC sample of 90 outcomes is separated into three bands namely training, validation, and testing with values 70% (62), 15% (14), and 15% (14), respectively. Ten neurons are taken in hidden layer. The neural networking model is trained by using Levenberg-Marquardt backpropagation. In line with regression (R) and mean square error (MSE) analysis, it is observed that SFC shows higher magnitude for positive values of Powell-Eyring and suction parameters. For both the suction and injection parameters, the fluid velocity admits a direct relation.