Evaluation of a CFD-Simulation approach for modelling CO2 evaporation in microchannels in Ansys Fluent using a mixture-model

Microchannels shall be used for evaporation of CO2 in a heat-pump cycle of a hybrid PV-module with direct heat recovery (PVT-direct). In order to predict the required tube length and the expected pressure losses of the microchannel evaporator, a CFD simulation model was implemented and evaluated. The simulation shall finally help finding the adequate size of the evaporator and optimize its geometry. For the simulation the software “Fluent” was chosen using the two phase mixture-model including the Lee-model for evaporation. As the tubes in the microchannel system represent an axis-symmetric setup, a 2D-geometry is sufficiently suitable for the model and its mesh. After an initial testing of the mixture model, it could be seen that the conservation of mass and energy were not correctly calculated with the advised standard set of parameters, neither the phase change mechanism with its characteristic latent heat. In order to get reasonable results with the mixture model, the input parameters of the evaporation rate and the bubble diameter were systematically varied and validated by manual calculation based on experimental data published in the standard literature. To further improve the numerical quality and stability of the simulation, solver parameters were changed. The adjustment of discretization schemes and the under-relaxation factors were most successful. The effect of different boundary conditions for solar radiation and convective heat on the relevant evaporator surface was also investigated. The test simulations showed that the mixture model is appropriate for a low vapor volume fraction but maybe reaches its limits at higher vapor volume fractions as the flow regime changes. It needs more investigation to prove if drop-evaporation will be sufficiently represented by the chosen model. Consequently, the modelling approach must be optimized and validated for the relevant volume vapor fraction of up to 95 Vol %. This will be the task for the future. And also completely different model approaches have to be considered