| dc.description.abstract | This paper presents a numerical study of the phenomena that occur in the use of PCM (Phase Change Materials) packed in spherical cavities. The numerical simulation was performed with commercial software ANSYS-FLUENT. The numerical model is two-dimensional and consists of the mass conservation equations of momentum, energy and volume fraction, modeled by enthalpy-porosity technique. The computational mesh is the hexahedral type with refinements in regions with the highest thermal and fluid dynamic gradients. The numerical model was validated with experimental and numerical results of literature. It was studied melting cases of PCM RT 35, RT 55 and RT 82 in spherical cavities with constant temperature on the wall. The diameters of the spheres were 40, 60, 80 and 100 mm and temperatures imposed on the wall were 10, 20 and 30°C, above the melting temperature of the PCM. It was presented results of melt fraction and heat flow of 36 cases studied. Research shows that PCM with similar properties have the same behavior for melt fraction and heat flow, regardless of its melting temperature. The temperature rise in the sphere wall is responsible for higher heat flows and by decreasing the melt fraction obtained in time. The best percentage reduction in the melting time was obtained with 20°C of temperature differences. Research also shows that the increase in diameter does not influence the heat flux, but increases the PCM melting time. This melting time can be related to the sphere characteristic length, regardless of the temperatures imposed on the wall. The studies also show that the largest quantity of thermal storage through the liquid PCM is obtained by combining the larger diameters with higher temperatures. The main parameters involved in the phase change process are correlated through numbers of Fourier, Stefan, Grashof and Prandtl for the calculation of the liquid fraction as a function of time. To calculate the liquid fraction as a function of time through a correlation, the parameters involved were calculated using the Fourier, Stefan, Grashof and Prandtl numbers. | en |
