TWO-PHASE CFD: THE VARIOUS APPROACHES AND THEIR APPLICABILITY TO EACH FLOW REGIME

Abstract

Two-phase Computational Fluid Dynamics (CFD) or CMFD (Computational Multi-Fluid Dynamics) is being applied to some nuclear reactor thermalhydraulic investigations. The NURESIM and NURISP projects of the European 5th and 6th Framework programs made some pioneering efforts to start some applications to a few selected reactor issues such as the Critical Heat Flux or the Pressurized Thermal shock. The extreme variety of flow configurations in steam-water two-phase flow makes the modeling issue very complex. Myriads of model options are available in the vast domain of CMFD, with various treatments of turbulence and of interfaces; the methods for choosing model options depending on the application have to be clarified. The purpose of this paper is to list and classify model options, to discuss some conditions and limits of applicability of the various options, and to identify the modeling needs for closure of the system of equations. The proposed classification of the modeling approaches is based on the space and time filtering or averaging of basic equations and on the number of phases and fields that are distinguished. The various flow processes that must be modeled by closure relations are identified for each type of approach. The applicability of each of these methods to each flow regime is defined and the present degree of maturity of the models is evaluated. Considering only Eulerian approaches used in an open medium, it is shown that five main approaches are considered: the RANS approach, the pseudo-DNS approach, and three types of space-filtered methods. Some of these methods can be applied only to a reduced number of flow regimes. Attention is drawn on the need to specify the model options, to guaranty the consistency between options, and to ensure completeness of closure terms. © 2011 by Begell House, Inc.