Page 62 - Thermal Hydraulics Aspects of Liquid Metal Cooled Nuclear Reactors

Thermal-hydraulic challenges in liquid-metal-cooled reactors 37

Development needs
With respect to SFR containment thermal-hydraulic challenges, Herranz et al. (2018) con-
clude that there are major phenomena potentially affecting the source term that are still
missing in numerical predictions, like particle generation from sodium spray fires, fission
product entrainment from sodium-concrete interaction, and thermal and chemical behavior
of fission products in sodium containing aerosols. Additionally, the key phenomena are
described by heavily parametrized models. A deep understanding of those phenomena that
would allow a mechanistic modeling is missing. Furthermore, the current experimental
database does not allow reliable determination of model parameters. New experiments
should be carefully designed to ensure key data aspects like representativity, accuracy,
and scalability.

2.4.5 Guidelines

With respect to the important topic of guidelines and quality control of numerical sim-
ulations, the following challenges have been identified:

l Verification, validation, and uncertainty quantification (see also Chapter 8)
Challenge
As for light-water reactors, the safety demonstration of LMFRs relies in large part in the
numerical simulation of various transients of interest. In order to qualify these simulations,
the numerical tools used must be checked for correctness. Their capability to correctly pre-
dict the physics of each transient must be assessed against an exhaustive experimental
database. And the uncertainties associated with the outputs of the calculation must be
quantified.
State of the art
Verification and validation methodologies and uncertainty quantification methods
have already been developed for light-water reactors. Translation to LMFRs is relatively
straightforward.
Development needs
Application requires many more experimental data to assess the current codes and models.
Two LMFR-specific points may require specific attention: firstly, the use of complex tools
(such as CFD or multiscale simulations) and, secondly, the rarity of large-scale, integral
experiments (in comparison, several large-scale loops exist for LWR LOCA analysis).
l Guidelines for CFD (see also Chapter 9)
Challenge
Performing good quality CFD simulations is not an easy task as the CFD engineer has to
judge that the results of the simulation can be trusted. While reality is continuous in space
and time, computers work in the discrete domain. This first level of approximation is
unavoidable, while other assumptions are introduced to reduce the time and effort to obtain
the desired information about the problem of interest.
State of the art
Over the years, generic best practice guidelines have been developed and documented for the
application of CFD. Most well known is the document made by ERCOFTAC (Casey and
Wintergeste, 2000).