Validation of different numerical models with benchmark experiments for modelling microencapsulated-PCM-based applications for buildings

This paper assesses the thermal performance of four microencapsulated-PCM-based samples that can be used in the design of new thermal energy storage systems for buildings: a PCM-enhanced plasterboard; three different fin-enhanced aluminium containers filled with the PCM. The methodology covers: the thermophysical characterization of the PCM-based products; the development of charging and discharging experiments to provide data for numerical validation purposes; the numerical modelling considering five different methods dispersed in literature, and a new-improved method. The main goals are to validate the numerical predictions against reliable experimental results and to find out which method is preferable for future simulations. The 3D numerical simulations were based on the effective heat capacity (EHC), additional heat source (AHS) and enthalpy (H-P) approaches. The first two methods were implemented in ANSYS CFX (R); the other in ANSYS FLUENT (R). For the EHC method, four different ways to specify the variation of the effective specific heat with temperature were assessed. While the H-P commercial code already implemented in ANSYS FLUENT (R) was used, the other two methods implemented in ANSYS CFX (R) were developed and validated for the purpose of this study. Good agreement between numerical and experimental results was achieved for all numerical approaches. In fact, it was concluded that the six numerical methods can be used to simulate heat diffusion with solid-liquid phase-changes. However, it was verified that the EHC method with the triangular or the self-adjusted triangular profiles is preferable for both charging and discharging simulations, since it can be drawn based on few information about the PCM-based products; it ensures good predictions of both charging/discharging kinetics and stored/released energy; it can be used without compromising computation time in comparison to other methods. The EHC method with the rectangular profile, recurrently used in literature, exhibits some convergence issues whose resolution has required longer computation times.