Speaker
Description
The STEP program run by UK Industrial Fusion Solutions Ltd aims to deliver a tritium self sufficient prototype fusion power plant generating 100 MW net electric power based on the spherical tokamak. The compact spherical geometry of STEP’s SPP-2 design raise significant exhaust challenges, with powers crossing the separatrix reaching Psep ≈ 110−140 MW, corresponding to Psep/R0 ≈ 25−32 MW/m. To overcome these challenges solutions, such as a double null with long out legged divertor, will be employed to achieve sufficient detachment and protect plasma facing components [1]. While these measures are predicted to achieve acceptable steady-state power to the divertor targets, transient power loads from fluctuations in core fusion/radiation power or vertical displacement-induced disconnected double null effects may lead to burn-through of the detachment front, and unacceptable power loading to material surfaces.
High-fidelity (SOLPS-ITER) simulations of transient burn-through remain prohibitively expensive to explore the wide parameter space of reactor scale devices. The main cause of this expense is the kinetic treatment of neutral species. To provide initial screening for STEP and comparison with analytical models [2], it is attractive to adopt the more computationally tractable method offered by reducing dimensionality to 1D and adopting a fluid treatment of the neutral species. In this study, we used the multi-fidelity Hermes-3 fluid code to study power transient burn-through in STEP-relevant 1D exhaust scenarios. We investigated the effect of a rapid rise in power entering the SOL (up to 4× power increase over a timescale of several miliseconds). Simulations found two regimes of front movement: slow burn-through driven by neutral ionisation, and faster compression in which charge exchange exerts a force on the neutrals, shrinking the neutral cloud. This imbalance leads to the detachment front motion tracking the rise-time of the power increase. In contrast, simulations in which the pressure imbalance was artificially removed saw much slower front motion with burn-through dominated by neutral ionisation.
As this phenomenology could be caused by a lack of cross-field transport and the fluid neutral description, we compare our 1D simulations against 2D SOLPS-ITER MAST-U simulations with kinetic neutrals. We then investigate reservoir models [3] to improve the agreement between the 1D fluid neutral and 2D kinetic neutral simulations, both in steady-state and transient power pulse scenarios.
[1]S.S. Henderson et al., Nucl. Fusion 65 016033 (2025).
[2]S.S. Henderson et al., Nucl. Fusion 64 066006 (2024).
[3]G.L. Derks et al., Plasma Phys. Control. Fusion 66 055004 (2024).