The High Temperature Gas cooled Reactor (HTGR) is an improved, gas cooled
nuclear reactor. It was chosen as one of the candidates of generation IV nuclear
plants [1]. The reactor can be shut down automatically because of the negative
reactivity feedback due to the temperature's increasing in designed accidents. It is
graphite moderated and Helium cooled. The residual heat can be transferred out of
the reactor core by inactive ways as conduction, convection, and thermal radiation
during the accident. In such a way, a fuel temperature does not go beyond a limit
at which major ssion product release begins.
In this thesis, the coupled neutronics and
uid mechanics code MGT-3D used
for the steady state and time-dependent simulation of HTGRs, is enhanced and
validated [2]. The
uid mechanics part is validated by SANA experiments in steady
state cases as well as transient cases. The fuel temperature calculation is optimized
by solving the heat conduction equation of the coated particles. It is applied in
the steady state and transient simulation of PBMR, and the results are compared
to the simulation with the old overheating model. New approaches to calculate
the temperature prole of the fuel element of block-type HTGRs, and the calculation
of the homogeneous conductivity of composite materials are introduced. With
these new developments, MGT-3D is able to simulate block-type HTGRs as well.
This extended MGT-3D is used to simulate a cuboid ceramic block heating experiment
in the NACOK-II facility. The extended MGT-3D is also applied to LOFC
and DLOFC simulation of GT-MHR. It is a
uid mechanics calculation with a
given heat source. This calculation result of MGT-3D is veried with the calculation
results of other codes. The design of the Japanese HTTR is introduced. The
deterministic simulation of the LOFC experiment of HTTR is conducted with the
Monte-Carlo code Serpent and MGT-3D, which is the LOFC Project organized by
OECD/NEA [3]. With Serpent the burnup of the reactor core is calculated starting
from the rst loading. From this calculation the nuclide inventory is obtained
and the result is interfaced to MGT-3D. The steady state and time dependent
calculation is conducted with MGT-3D. The elapsed time and peak power level
at the occurrence of the re-criticality of the LOFC experiment are compared to
the simulation results. Up to now, these results meet the experimental values best
compared with the other participants.
Dunfu Shi