A CFD model of the thermal hydraulic behaviour of the HTTU annular packed bed

Abstract

Very high temperature gas-cooled reactors (VHTGRs) could play a pivotal role in the future of the energy transition away from carbon as several industries relying on high temperature process heat and steam as input. VHTGR plants offer a zero-emission thermal energy source, compared to direct heat from the combustion of coal, oil and gas or indirectly from electricity generated at a lower efficiency from the same sources. Compared to renewables, nuclear has unmatched availability for base load supply and is therefore seen as an important component towards energy sustainability (World Nuclear Association, 2024). Several VHTGR designs have been proposed based on a pebble bed core consisting of spherical fuel elements. Such pebble bed reactors (PBRs) have been built in the past showing promising results. In their design and operation, however, it was noted that there is a large degree of uncertainty around predicted and observed temperatures.

Modern modelling techniques such as computational fluid dynamics (CFD) aim to solve modelling issues of the past, but often come with a large computational expense. Integral effects tests contribute not only to the qualification of next-generation designs but also to the validation of codes including CFD models. The High Temperature Test Unit (HTTU) was an integral effects test of the thermal conduction, radiation and convection in an annular pebble bed, based on the PBMR-400 design.

This study presents two approaches to the numerical modelling of the pebble bed of the HTTU experiment in the finite volume CFD code, Star-CCM+: a coarse mesh full model of the bed using a representative porous medium, and a fine mesh model of a section of the bed with discrete geometry of resolved solid spheres and flow domain. The annular packed bed geometry was generated using a discrete element method (DEM) code ESSS Rocky, with the geometry extracted in the case of the discrete geometry, and a separate method used to obtain the radial and axial porosities used in the porous approach.

Results from experimental isothermal tests using nitrogen are used to validate the flow distribution and pressure drop of the two modelling approaches without heating. The models are then solved with forced convection, calibrated for the 1 kg/s 140 kW case and finally compared with the corresponding experimental results. A further comparison is done to other test conditions. Good agreement was found with differences less than measurement uncertainties. In the near-wall region the average relative error was found to be equal to 0.9% but in the bulk region the error reaches 38%.