Thermal-Hydraulics and Dynamics of Supercritical CO2 Reactor

by T. Gallaway, S. P. Antal and M. Z. Podowski
Rensselaer Polytechnic Institute

Click Here To Access The Complete Paper

Fluids at supercritical pressures are considered to be very promising as coolants for Generation IV reactors, including the supercritical water cooled reactor (SCWR) and the CO2 cooled reactor. However, several unresolved issues must still be addressed in order to fully understand the effect of variable properties of supercritical fluids on the performance and safety of future reactor designs.

The objective of this paper is to investigate various aspects of fluid property variations at supercritical pressures and their impact on heat transfer in heated channels. Channels of two geometries have been analyzed: a cylindrical tube heated from outside and an annulus heated from inside and insulated from outside. Both kinematic and thermal aspects of turbulence have been investigated using a High-Reynolds k-ε model and a Low-Reynolds k-ε model. It has been shown that both models yield similar results for the velocity distribution across a heated channel. On the other hand, it has also been demonstrated that a standard model of the near-wall heat convection, based on the High-Reynolds k-ε model, yields erroneous results for fluids having the molecular Prandtl number not equal to one. A new model has been proposed, based on the thickness of the laminar thermal sublayer. It has been demonstrated that the predicted velocity and temperature profiles away from the heated wall, obtained from both models of turbulence, agree well with each other, thus proving the consistency of both models. Also, the results of predictions are in a good agreement with the experimental data, although some differences can still be noticed. Since the High-Reynolds k-ε model does not allow for directly calculating the near-wall velocity or temperature distributions, the use of averaged fluid properties is necessary in this region. On the other hand, since the Low-Reynolds k-ε model extends to a distance very close to the wall, its predictions are much better without any averaging.