Analysis of heat transfer to supercritical carbon dioxide by the STAR-CCM+ CFD code

The Supercritical Water Reactor (SCWR) is one of the six conceptual designs proposed for Generation IV for which a considerable amount of research and development is being spent for tackling the challenging aspects of the related technology. Among the different issues to be considered in the design of SCWRs, there are obviously the consequences of the peculiar characteristics of fluids at supercritical pressures on the heat transfer efficiency. In particular, heat transfer enhancement and deterioration are the peculiar phenomena that are known to occur when a supercritical fluid expands along a heated duct. In fact, on one side, the large increase in the specific heat experienced close to the pseudo-critical temperature strongly enhances heat transfer in that region; on the other hand, acceleration and buoyancy effect may be responsible for damping the level of turbulence in the fluid, thus dramatically decreasing the heat transfer efficiency by laminarisation effects. Experiments on heat transfer to supercritical fluids in the frame of the studies for SCWRs have mainly the purpose to acquire information about the behaviour of water. However, because of the high values of the critical pressure and temperature of water, it is sometimes convenient to make experiments on substitute fluids characterized by lower values of these parameters. Among them, carbon dioxide is a very common choice. Kim et al. (2005), in particular, performed interesting studies of heat transfer to carbon dioxide in cylindrical, square and triangular channels, highlighting the effects of deterioration which are taken as reference in this work. In the aim to assess the behaviour of CFD models in predicting these interesting phenomena, the STAR-CCM+ code was used to predict the data by Kim et al. (2005) in the frame of a training program performed at the University of Pisa and supported by the International Atomic Energy Agency. The analyses involved the use of different models and provided results of comparable quality with respect to those already obtained by other CFD codes in the past, allowing to extend previous conclusions. In particular, the tendency of available low-Reynolds k-ε models in overestimating the effects of heat transfer deterioration has been confirmed. Moreover, sensitivity analyses performed with different models and modeling parameters allowed to highlight aspects of interest for contributing drawing general conclusions on the state-of-the-art of turbulence models applicable to these challenging operating conditions.