Predicting Xenon Oscillations in PWRs using Intrusive Reduced Order Modelling

Licentiate thesis, 2023

The increasing presence of intermittent energy sources in the Swedish electricity grid necessitates
a transition of Swedish nuclear reactors from constant base load operation to load-following
mode. However, such changes in power can induce xenon oscillations, a phenomenon
that poses operational challenges and the risk of fuel damage. Xenon oscillations occur due
to the decay characteristics of iodine-135 and xenon-135 produced in the fission process,
exhibiting a periodicity of 15 to 30 hours. Detecting these oscillations proves challenging
as they may result in localized power variations while the overall power of the reactor core
remains relatively stable.
This thesis aims to develop a computationally efficient and transparent model capable of
predicting the susceptibility of nuclear reactors to unstable xenon oscillations. Two models are
created and assessed: a simple physics-transparent model based on a one-group homogeneous
core representation, and a more involved model, which incorporates two energy groups and
a heterogeneous spatial discretization with nodal resolution.
Comparative analysis of the models reveals notable disparities in predicting instabilities related
to xenon oscillations. The number of energy groups emerges as the primary factor
contributing to the discrepancies observed. Moreover, spatial resolution is critical in capturing
eigenmode coupling when spatial offsets exist in the equilibrium neutron flux distribution.
It is demonstrated that the latter model indicates a higher level of system instability concerning
xenon oscillations.
The findings underscore the significance of considering both spatial and energy resolution to
accurately assess the stability of the system.