AI-heat transfer analysis of casson fluid in uniformly heated enclosure with semi heated baffle

Abstract

The heat transfer in Casson fluid with natural convection claims various applications namely thermal regulation in biological systems, solar collectors, polymer processing, and geothermal applications to mention just a few. Owing to such motivation, we have offered artificial intelligence-based solution outcomes for heat transfer aspects in Casson fluid flow in a partially heated square enclosure with free convection effect. The semi-heated triangular baffle is installed at the center of the cavity. The bottom and right walls have the same amount of heat. The left wall of the cavity is taken cold and the top wall is taken insulated. The surface of triangular baffle and cavity walls are carried with non-slip condition. Finite element method (FEM) with hybrid meshing is used to solve the developed flow equations. AI-based neural networks model is used to examine the variation in Nusselt number for the involved flow parameters. MSE=2.15008e^-6, 5.81476e^-5, and 3.51888e^-4 for training, validation, and testing respectively, suggesting good model performance on Nusselt number data along the bottom and vertical walls. We have observed that the heat transfer coefficient improves as Rayleigh and Prandtl numbers increase. We believe that the present AI-based outcomes will be helpful for predicting natural convection phenomena subject to thermal engineering standpoints.