Page 413 - Thermal Hydraulics Aspects of Liquid Metal Cooled Nuclear Reactors

376 Thermal Hydraulics Aspects of Liquid Metal Cooled Nuclear Reactors

In France, multiscale models to ASTRID safety transients have been applied.
l
In Belgium, multiscale models to MYRRHA safety transients have been applied.
l
(In both cases, also complete CFD models of the complete primary circuit were used.)
If these models are to be used in the final safety analysis of their respective reactors,
then they will need to reach a level of verification, validation, and uncertainty quan-
tification (VVUQ) comparable with those attained with system codes. As for system
codes, the validation database of a coupled code should consist in a combination of
analytic, combined-effect, large-scale, and integral validation. This classification is
used in the overview of coupled models presented here. For convenience, this over-
view mainly covers models developed as part of the THINS and SESAME EU pro-
jects: these projects have indeed contributed to the development of a majority of the
liquid-metal multiscale coupled models in use today.

7.3.1 Analytical validation of coupling algorithms
The notion of analytic validation takes a different meaning in a coupled code than in a
system of a CFD code. The individual effects predicted by a coupled model will, by
definition, be the result of a computation by one of the codes contained in the model;
the “new” phenomena predicted by the coupled code, such as those generated by the
interaction of different scales, are usually considered “combined effects.”
On the analytic level, it is sufficient to guarantee that the coupling algorithm used is
validated at the coupling boundaries between the codes; this can be verified by con-
structing a number of analytic test cases, covering all the potential types of coupling
boundaries available to the algorithm, for which one can verify that

l mass is conserved across the boundary;
l energy fluxes are consistent between the codes at the boundaries;
l the pressure fields computed by both codes are consistent, that is, match at the boundaries.
Verifying the coupling algorithms against these test cases should be sufficient to dem-
onstrate the validity of the coupling algorithm at the analytic level.

7.3.2 Small and intermediate-scale validation

In the framework of the THINS and SESAME projects, two facilities have been
exploited to study coupled effects at a small scale:
The TALL-3D facility (Grishchenko et al., 2015) is an experiment designed at KTH to study
l
the coupling between a three-leg LBE loop and a cylindrical 3-D test section subjected to
local effects such as flow mixing, stratification, jet impingement, and heat transfer. The
facility has been specifically designed to promote interactions between these local effects
and the global loop scale: the 3-D section is placed in one of two hot legs, while the other
hot leg has a simpler pipe geometry. During loss-of-flow transients, both legs compete for
natural-convection flows, leading to strong oscillations. These oscillations are in turn
affected by local effects occurring in the 3-D section. The following coupled models of
TALL-3D are under development: