Magnetohydrodynamic free convection of nano-encapsulated phase change materials between two square cylinders: Mapping the thermal behavior using neural networks

The study focused on investigating the convective heat transfer of nano-encapsulated phase change suspensions in the presence of a non-uniform magnetic field within an annuli space between two square cylinders. The principal equations for the fluid flow and phase change heat transfer were formulated as partial differential equations and then represented into dimensionless format. The finite element method was used to solve these equations and simulate the free convection heat transfer. The effect of various factors, including Hartmann, Rayleigh, Eckert and Stefan numbers, geometry aspect ratio, nanoparticles' concentration, and fusion temperature, on the heat transfer rate was examined. A neural network was also introduced and trained to establish the connection between the control parameters (inputs) and the heat transfer rate (output). The outcomes were presented in the form of the modified Nusselt number, along with isotherms, heat capacity ratio (phase change) contours, and streamlines. The results demonstrated that the neural network could accurately predict the heat transfer rate and provide a comprehensive map of heat transfer with respect to the control parameters. NanoEncapsulated Phase Change Materials (NEPCMs) can be considered as a new type of nanofluids, in which the nanoparticle consists of a core and a shell. The core part is made of a Phase Change Material (PCM) which can undergo solid-liquid phase change at a certain fusion temperature, and absorb/release a significant amount of energy due to latent heat of the phase change.