Abstract:
The ASTRA Critical Facility at the Kurchatov Institute in Moscow was used to perform
critical experiments to simulate the PBMR reactor core design as being developed by
PBMR (Pty) Ltd. The aim of these benchmarks is to validate the code system VSOP,
currently used for PBMR core neutronics calculations. In this work these benchmark
calculations are discussed and investigated.
Experiments performed at the facility include criticality experiments, control rod worth
measurements and reactivity measurements. To aid in the verification, the Monte Carlo
code MCNP was used to calculate some of the experiments and compared to the VSOP
results.
The flux solver in VSOP uses the finite difference diffusion method to determine the
neutron distribution in the core. The problem of modelling a highly absorbing region, such
as control rods, in diffusion calculations is well known and many methods have been
developed to accommodate the transport effects in diffusion theory. This problem is
compounded in the case of the ASTRA facility (and PBMR) due to the positioning of the
control rods in the side reflector and the associated directional dependence of the flux.
One of the methods that can be used for the control rod model is that of equivalent cross
sections. This method has been shown in the past to yield acceptable results for pebble bed
type reactors. In this work the method of equivalent cross-sections is evaluated by applying
it to calculations of control rod experiments performed at the ASTRA facility. It is shown
that results are favourable for control rods situated within the first ring of reflector blocks,
with larger error obtained for control rods situated further from the core. Additionally, a
method in which an equivalent boron concentration is used to represent the absorber
region is investigated. This is shown to be useful if applied correctly and with care,
especially in the case of differential control rod worth.
