Analysis and identification of melting of encapsulated phase change material under impacts of pulsating elliptic heat source during nano-enriched forced convection in a ventilated cavity

Abstract

Time dependent heating and non-uniform type heating are encountered in various renewable energy systems and in different electronic cooling applications. As those systems may be integrated with phase change materials (PCMs) for energy storage or thermal management, understanding the dynamics of interaction between the heat source and PCM becomes important. In this study, in a vented chamber, the effects of employing an elliptic-type pulsating heat source on the melt dynamics of encapsulated PCM during nano-enriched convective heat transfer are investigated. The effects of the elliptical source's aspect ratio (AR = 0.5-2), pulsation amplitude (Af = 0-0.25), and non-dimensional frequency (f* = 0.01-0.5) on the melting process are analyzed using a finite element method (FEM) at a fixed Reynolds number of 800. A hybrid nanofluid consisting of water with Ag-MgO binary nanoparticles is employed in the cavity at a solid volume fraction of 2%. The heater's aspect ratio is found to affect melting by varying the vortex size in the PCMmounted zone and heating amount. The case at AR = 2 yields 17.9% and 8% melting improvements at t = 80 min in constant and pulsating heating configurations. In contrast to the constant heating scenario, using pulsation at the maximum amplitude enhances the melt fraction by 20% at t = 80 min, respectively, whereas frequency has little effect on melting. The pulsing heat source at AR = 2 with nanofluid produces the best results in terms of melting, whereas the adiabatic heat source at AR = 1 with base fluid produces the worst results. There is a 109% improvement in melting between the best and worst cases. Using the system identification, the model's outputs accurately capture the melt fraction's behavior over a range of pulsed heat source settings. For a range of aspect ratios and frequencies, the non-linear model's fit values are higher than those of the linear model.