Nuclear thermal hydraulics is the key process in nuclear reactors to move heat
generated by nuclear fission in the reactor core to downstream processes such as
electricity generation, heat for industrial applications and potentially the
generation of hydrogen. Its understanding is fundamental for the safe operation of
nuclear reactors, which will play an increasingly important contribution in our
drive to net-zero.
One of the key unsolved problems is the treatment of uncertainties that exist in
the prediction of nuclear thermal-hydraulic phenomena, which has complex, coupled
and non-linear behaviour. Rigorous uncertainty quantification (UQ) is increasingly
vital due to the sector’s stringent safety and regulatory demands. Effective UQ
improves confidence in simulations, supports decision-making, and reduces
excessive conservatism while maintaining safety.
Traditional Monte Carlo (MC) approaches to UQ employ random sampling of the
parameter space to propagate input parameter uncertainties. However, MC methods
are computationally expensive as statistical convergence is slow and the number of
required samples is, therefore, generally large. This challenge is particularly
severe in Computational Fluid Dynamics (CFD), where a single high-fidelity
simulation can take days or more to run on high-performance computing (HPC)
systems.
To mitigate this cost, alternative UQ methods are typically employed. One common
approach is surrogate modelling, where a limited ensemble of high-fidelity
simulations trains an approximate model for rapid evaluations. These surrogates
provide fast evaluations, but are limited to the accuracy of the surrogate.
This PhD focuses on an alternative method which does not rely on surrogates, known
as deterministic sampling (DS) [1]. In DS, the parameter space is sampled
deterministically with a limited ensemble. The weighting of each sample is
determined such that the moments of the input distributions are respected. The
method has demonstrated orders-of-magnitude improvements in statistical
convergence over MC without surrogate limitations.
In this project, we aim to explore the use of DS in nuclear thermal hydraulics,
assessing the impact of ensemble choice and size upon the efficiency and accuracy
of DS. We aim to develop the DS method in a setting where the ensemble is
incrementally augmented to assess statistical convergence, with prior samples
contributing and appropriate weightings and cutoff determined.
This novel approach brings value in so far that it aims to combine a rigorous,
academic approach aiming to quantify the level of approximation made with a more
application-orientated methodology making it applicable in practice for real-world
applications. This has the potential to make a considerable impact in the adoption
of UQ in the nuclear industry, enabling the enhanced regulatory requirements in
safety cases to be satisfied for new reactors.
The successful applicant will be passionate about computational modelling, fluid
dynamics, and advanced UQ techniques. You will work at the intersection of HPC,
probabilistic analysis, and nuclear thermal hydraulics engineering. You will have
a 2:1 or higher degree, ideally with a master’s at merit or higher (or
international equivalent) in a relevant topic such as maths, physics, or
engineering.
This project is part of the SATURN Centre for Doctoral Training and is co-funded
by the National Nuclear Laboratory (NNL) and Framatome. You will, therefore, work
closely with industry partners, ensuring your research has practical impact on
reactor safety and efficiency.
As well as benefitting from the University of Manchester’s expertise in UQ, the
industrial supervision contribution in this project will enable you to interact
with top players in industrially-based thermal hydraulics in France and the UK. As
a result, the project may include short-term secondments and visits to UKNNL and
Framatome.
Eligibility
Applicants should have, or expect to achieve, at least a 2.1 honours degree or a
master’s (or international equivalent) in a relevant science or discipline.
How to apply
Please complete the SATURN CDT Application Enquiry Form | UoM - Pre-Application
We strongly recommend you contact the project supervisor after completing the form
to speak to them about your suitability for the project. You can find their
details on the PhD projects | EPSRC Centre for Doctoral Training in Skills And
Training Underpinning a Renaissance in Nuclear (SATURN)
If your qualifications meet our standard entry requirements, the CDT Admissions
Team will send your enquiry form and CV to the named project supervisor.
Informal enquiries can be made to Alex Skillen (alex.skillen@manchester.ac.uk)
Our application process can also be found on our website: Apply | EPSRC Centre for
Doctoral Training in Skills And Training Underpinning a Renaissance in Nuclear
(SATURN). If you have any questions, please contact SATURN@manchester.ac.uk
Equality, diversity and inclusion is fundamental to the success of The University
of Manchester, and is at the heart of all of our activities. We know that
diversity strengthens our research community, leading to enhanced research
creativity, productivity and quality, and societal and economic impact.
We actively encourage applicants from diverse career paths and backgrounds and
from all sections of the community, regardless of age, disability, ethnicity,
gender, gender expression, sexual orientation and transgender status.
We also support applications from those returning from a career break or other
roles. We consider offering flexible study arrangements (including part-time: 50%,
60% or 80%, depending on the project/funder).
Saturn_Nuclear_CDT
Funding Notes
This project is part of the SATURN Centre for Doctoral Training and is co-funded
by the National Nuclear Laboratory (NNL) and Framatome. Successful applicants will
receive a tax-free stipend of £20,780 (2025/26 rate)
References
[1] P. Hessling, "Deterministic Sampling for Propagating Model Covariance," in
SIAM/ASA Journal on Uncertainty Quantification, 2013.
|
