Assessment of Cfd Codes Used in Nuclear Reactor Safety Simulations

The spectacular growth in computer hardware over the last quarter century and the accompanying advances in software development have resulted in the availability of reliable numerical tools for addressing safety issues in Nuclear Power Plants (NPPs). The first step forward was undertaken in the 1970s with the development of system codes using the two-fluid model approach [1], such as RELAP-5 [2], TRAC/TRACE [3], CATHARE [4] and ATHLET [5] for example, for the analysis of primary circuit transients. Other programs, such as GOTHIC [6], GASFLOW [7], MELCOR [8] SCDAP [9] and MAAP [10] have also been written for containment and severe accident analyses, respectively. The application of Computational Fluid Dynamics (CFD) methods to problems relating to Nuclear Reactor Safety (NRS) is less well developed but is rapidly accelerating. The need to use CFD arises because many traditional reactor system and containment codes are based on a network of 1-D or 0-D volumes. It is evident, however, that the flow in such components as the upper and lower plena, downcomer and core of a Reactor Pressure Vessel (RPV) is strongly three dimensional. Natural circulation, mixing and stratification in containments is also essentially 3-D in nature, and representing such complex flows by pseudo 1-D approximations may not just be oversimplified but could even be misleading, resulting in erroneous judgments being made. One of the reasons why the application of CFD methods in NRS has been slow to establish itself is that the transient, often two-phase, phenomena associated with accident events are extremely complex. Traditional approaches using system codes have been successful because a very large database of phasic exchange correlations has been built into them. The correlations have been formulated from 1-D special-effects experiments and have been well tested. Data on the exchange of mass, momentum, and energy between phases for 3-D flows are very sparse in Following a joint OECD/NEA–IAEA–sponsored meeting to define the current role and future perspectives of the application of Computational Fluid Dynamics (CFD) to nuclear reactor safety problems, three Writing Groups were created, under the auspices of the NEA working group WGAMA, to produce state-of-the-art reports on different aspects of the subject. The work of the second group, WG2, was to document the existing assessment databases for CFD simulation in the context of Nuclear Reactor Safety (NRS) analysis, to gain a measure of the degree of quality and trust in CFD as a numerical analysis tool, and to take initiatives to extend the existing databases. The group worked over the period of 2003-2007 and produced a final state-ofthe-art report. The present paper summarises the material gathered during the study, illustrating the points with a few highlights. A total of 22 safety issues were identified for which the application of CFD was considered to potentially bring real benefits in terms of better understanding and increased safety. A list of the existing databases was drawn up and synthesised, both from the nuclear area and from other parallel, non-nuclear, industrial activities. The gaps in the technology base were also identified and discussed. In order to initiate new ways of bringing experimentalists and numerical analysts together, an international workshop -CFD4NRS (the first in a series) -was organised, a new blind benchmark activity was set up based on turbulent mixing in Tjunctions, and a Wiki-type web portal was created to offer online access to the material put together by the group giving the reader the opportunity to update and extend the contents to keep the information source topical and dynamic.