Speaker
Description
Time-dependent plasma boundary simulations can be an important contributor to power exhaust control design in ITER and future fusion reactors, but remain computationally challenging due to coupled fluid-kinetic plasma edge models, strong nonlinearities, and different timescales involved. Most SOLPS-ITER studies have focused on stationary power exhaust scenarios, with only a few attempts to use the code in time-dependent mode, with recent examples in [1,2]. This is partly because an experimental validation of the code in dynamic mode is missing. The present contribution addresses this gap.
Dedicated L-mode unseeded discharges (#42650/1: Pohm = 0.5-0.7 MW, Ip = 0.8 MA, Bt = -2.5 T) were performed in ASDEX Upgrade to enable dynamic validation. Starting from an attached state, a step increase in fuelling drives a transient towards detachment, with the target temperature dropping rapidly. In the meantime, the upstream density rises and the upstream temperature decreases. The transient lasts ~800 ms, with the target profiles responding faster than the upstream ones, likely due to puffing in the private flux region (PFR).
Recently developed Advanced Fluid Neutral (AFN) models [3,4] allow for an accurate fluid description of hydrogenic neutrals in high collisionality regions. This means that on reactor scale devices, such as ITER, initial assessments of detachment timescales can be done with B2.5, the fluid solver of SOLPS-ITER, avoiding the large computational cost of the kinetic module Eirene. Therefore, the initial validation effort focuses on simulations using AFN.
The initial and final states are first matched separately as steady-state simulations using the same set of diffusion coefficients. The transient is simulated, starting from the attached state, by increasing the deuterium puff at the PFR boundary. While the simulation qualitatively reproduces the dynamic evolution observed in the experiments, the simulated timescales are roughly a factor of 2-3 faster. Several differences between model and experiment may explain this discrepancy: the puff is imposed on the PFR boundary, not at the true location; a perfect gas step is used, whereas in reality the transition is smoother; a fluid description of neutrals might not suffice. The impact of these limitations on the simulated timescales is discussed.
[1] J. Lore et al., Nucl. Fusion 63 (2023) 046015.
[2] J. S. Park et al., Nucl. Fusion 64 (2024) 076036.
[3] N. Horsten et al., Nucl. Fusion 57 (2017) 116043.
[4] W. Van Uytven et al., Nucl. Fusion 62 (2022) 086023.