Multi-scale Thermal-hydraulic Analysis of Pwrs Using the Cupid Code

Recent advances in computational fluid dynamics (CFD) codes along with high performance computing hardware made the codes more useful in many engineering applications. For instance, single-phase CFD codes have long been used to predict the fluid flow in fuel channels of a pressurized water reactor (PWR) [1, 2]. Considering the current affordable number of computation mesh is a little below one billion, one fuel assembly of a PWR is now being simulated using a single-phase RANS model. The whole core of a PWR, which requires hundreds of billions of meshes for realistic simulation, is expected to be analyzed using CFD over the next decade. A two-phase flow also appears in many PWR transient and accident analyses. However, the two-phase CFD simulation has been challenging up to now mainly because of insufficient physical models for a two-phase flow in the CFD scale. Most of the current physical models for a two-phase flow have been developed for system codes, such as MARS [3], REALP5 [4] and CATHARE [5], which adopt onedimensional governing equations. In addition, the numerical solution methods of current CFD codes are sometimes not efficient in dealing with two-phase flows because interface mass, momentum and energy transfer terms are usually treated in an explicit manner as non-linear source terms in the governing equations. A two-phase CFD code, CUPID [6-9], has been developed at KAERI for the analysis of transient two-phase flows in nuclear reactor components. The CUPID code employs a two-fluid three-field model. The governing equations are discretized using the finite volume method (FVM) with unstructured grids, and solved using a semiimplicit numerical method. This is useful for the analysis of transient two-phase flows where the numerical solution hardly converges using the SIMPLE-based implicit method. The numerical method of the CUPID code has been verified against a set of singleand two-phase test problems. Validation of the CUPID code has also been carried out using experiment data. In this paper, two validation cases are introduced. The downcomer boiling (DOBO) [10] experiment performed at KAERI is analyzed to validate the boiling heat transfer model in open media. Then a test from STERN laboratory [11] simulating the moderator flow in a Calandria vessel is assessed using the porous media model of the CUPID code. For more practical engineering applications, the concept of “multi-scale” analysis has been proposed [12,13] by adopting the combined use of different scale computational tools, such as system and CFD codes, since direct use of two-phase CFD codes for nuclear reactor system analysis requires huge computational cost. The multi-scale analysis KAERI has developed a two-phase CFD code, CUPID, for a refined calculation of transient two-phase flows related to nuclear reactor thermal hydraulics, and its numerical models have been verified in previous studies. In this paper, the CUPID code is validated against experiments on the downcomer boiling and moderator flow in a Calandria vessel. Physical models relevant to the validation are discussed. Thereafter, multi-scale thermal hydraulic analyses using the CUPID code are introduced. At first, a component-scale calculation for the passive condensate cooling tank (PCCT) of the PASCAL experiment is linked to the CFD-scale calculation for local boiling heat transfer outside the heat exchanger tube. Next, the Rossendorf coolant mixing (ROCOM) test is analyzed by using the CUPID code, which is implicitly coupled with a systemscale code, MARS.