Computational Fluid Dynamics Code for Supercritical Fluids

by Ahti J. Suo-Anttila1 & Steven A. Wright2
1Computational Engineering Analysis LLC
2Sandia National Laboratories

Click Here To Access The Complete Paper
Click Here To Access The Presentation

Computational Fluid Dynamics (CFD) has become a widespread and useful engineering tool in many fields. There are a number of commercial CFD codes available and most have a blend of unique features and ease of use. However, most, if not all, of these codes are applicable to ideal gases or liquids, or both. A web search will reveal virtually no commercial CFD codes that are designed to handle real gases, such as supercritical CO2. Near the critical point fluids have large variations in density with respect to temperature, and thermophysical properties such as specific heat can vary by an order of magnitude. Taking these variations into account is difficult and explains why there are no CFD codes that offer real gas capability as a standard feature.

This paper describes the implementation of the NIST REFPROP library into a commercial code called C3D. The C3D code is a finite volume general purpose CFD code that has been used extensively in fire and combustion simulations. Fires typically have density variations of a factor of 4 or 5 over short distances, which is similar to the types of variations found in supercritical fluids. Thus the code is already capable of handling large property variations, and extending it to supercritical fluids seemed relatively straightforward.

The REFPROP implementation could only be accomplished by converting the energy equation into an enthalpy equation. Enthalpy has smooth behavior around the critical point and that behavior eliminated computational instabilities that occur when using a thermal energy equation with variable specific heats. The new C3D code with real fluid capability was applied to three example problems, the experimental flow loop of Milone, a Nuclear Reactor Application, and the Sandia Supercritical CO2 flow loop. The Code had good agreement with the Milone experimental data and that provides a partial validation of the REFPROP implementation. The reactor application demonstrates that natural circulation with S-CO2 can cool the decay heat in a shut down 400 MW core. The Sandia CO2 flow loop is being set up to demonstrate natural circulation in a 20KW flow loop. The C3D code predicts that reverse flow will occur unless the piping network is initiated with non-uniform temperatures.