The high-temperature gas-cooled reactor (HTGR), or very-high-temperature reactor (VHTR), have been selected by the UK government as priorities for our nation’s next nuclear systems. The UK has a long history and experience of operating gas-cooled rectors, and the UK Government sees HTGRs as an important contributor to the Net Zero 2050 target. The new generation of nuclear reactors, termed ‘Gen IV’, are much more challenging to design and operate as they use helium as a coolant rather than carbon dioxide, making them more efficient in terms of energy and hydrogen production. Hydrogen is considered an increasingly important energy source for the nation’s future. It will be used to generate electricity, used in industrial applications, and to heat homes and businesses. It should help to make electricity cheaper, more resilient to power cuts, and have a positive impact on the UK economy.

The main issue with this new generation of nuclear reactors is that they operate at temperatures that can reach 950°C, making design, build, operation, and safety measures even more critical. The successful deployment of HTGR technology requires an in-depth understanding of its reactor physics, particularly the coolant flow dynamics and heat transfer processes within the reactor core, as well as their impacts on the reactor efficiency and structural integrity. Achieving this relies on thermal-hydraulic analyses, which are currently limited by the lack of simulation tools capable of providing high-fidelity simulations of the entire reactor behaviour.

Traditional simulation tools, such as system codes and subchannel codes, have been used in thermal-hydraulic analyses of nuclear reactors for more than half a century. However, they provide only limited insight into complex 3-D phenomena that happen within the reactor system. Advanced traditional Computational Fluid Dynamics (CFD) methods, whilst capable of resolving 3-D physics in as much detail as required by first principle physical laws, are slow and extremely costly for large-scale  applications, and sometimes too demanding for the current generation of supercomputers.

Gen IV reactors. The modelling is transforming the design phase, as it can be applied to simulate the entire reactor core rather than one small element of it, making for more global and performant engineering  and design. In addition, the lighter computational load means that results can be produced in a reasonable timescale (days rather than weeks) and a more cost-effective way. Using traditional high-fidelity models, it can take weeks or even months to simulate a small part of the reactor on the current generation of supercomputers, at far from industrial scale.