Runaway electron avalanche and macroscopic beam formation: simulations of the DTT full power scenario

TL;DR

This work extends the disruption RE safety analysis from the Day-0 phase to the full-power DTT regime ($I_p=5.5$ MA) using 2D non-linear MHD simulations with JOREK and STARWALL, coupled RE fluid dynamics, and an artificial TQ to seed the current quench. It reveals an avalanche gain of $G_ ext{av} \approx 1.3 \times 10^{5}$ at full power, capable of converting seed currents as small as $I_ ext{seed}/I_p=10^{-6}$ into macroscopic RE beams, up to $I_ ext{RE} \approx 3.2$ MA for large seeds and impurity levels around $N_ ext{imp} \approx 3\times10^{21}$. The simulations identify three regimes in the full-power scenario: benign impurity levels causing no significant RE growth, an avalanche-transition regime where outcomes depend on seed magnitude, and a macroscopic-beam regime where substantial RE currents form and interact with wall structures, effectively stalling the current quench. The results underscore the need for carefully balanced disruption mitigation strategies in the full-power setting and motivate future 3D studies and mitigation optimization to assess and minimize PFC damage risk.

Abstract

The transition of the Divertor Tokamak Test (DTT) facility from its initial commissioning phase (Day-0, plasma current $I_{p}=2$ MA) to the full power scenario ($I_{p}=5.5$ MA) introduces a critical shift in the dynamics of runaway electrons (REs) generation. While previous predictive studies of the low-current scenario indicated a robust safety margin against RE beam formation, this work reveals that the exponential scaling of the RE avalanche gain with plasma current severely narrows the safe operational window in the full power scenario. Using the non-linear magnetohydrodynamic code JOREK, we perform comprehensive 2D simulations of the current quench (CQ) phase of several disruption scenarios, systematically scanning initial RE seed currents and injected impurity levels. The results demonstrate that in the full power scenario, the avalanche multiplication factor is sufficiently high ($G_\text{av} \approx 1.3 \cdot 10^5$) to convert a mere 5.5 A seed current into macroscopic RE beams of $\approx 0.7$ MA when large amounts of impurities are present. For even higher RE seeds, the RE current can peak at $ \approx 3.2$ MA, constituting up to $\approx$ 80% of the total plasma current during the CQ. These findings suggest that, unlike the Day-0 phase, the disruption mitigation strategy for the full power scenario involves a careful balance between thermal load mitigation and RE avoidance, necessitating a well-chosen quantity of injected impurities. This work provides the baseline needed for future estimations of RE loads on the plasma-facing components of DTT, which will be essential for designing and positioning mitigation components like sacrificial limiters.

Figures (9)

• Figure 1: Comparison of the equilibrium profiles for electron density $n_e$, electron temperature $T_e$, FF', and safety factor $q$ (from top to bottom) as a function of the normalized poloidal flux $\psi_n$. The plots display the CREATE-NL equilibrium data (blue) and their fitted approximations (orange). The fitted $n_e$ profile was not modified, while for $T_e$, FF', and $q$, the modified MHD-stable profiles employed in the simulations are shown in green.
• Figure 2: Time evolution of the total plasma current $I_\text{p}$ (solid lines) and RE current $I_\text{RE}$ (dashed lines) for a small initial RE seed of $I_\text{seed} = 5.5$ A. The colors correspond to varying levels of injected neon impurities ($N_\text{imp}$), ranging from $1\cdot10^{20}$ to $6\cdot10^{21}$ particles. Thin brown vertical lines indicate the ATQ phases: the dotted line separates the thermal collapse from the current flattening phase, while the dashed-dotted line marks the onset of impurity injection. The transition to the self-consistent CQ is identified by the change in slope of the plasma current traces following the dashed-dotted line. Since impurities are injected at a fixed rate, higher $N_\text{imp}$ values require longer injection times, delaying the CQ onset. Despite the very small initial seed, macroscopic RE beams reaching $\approx 0.7$ MA form at high impurity levels ($N_\text{imp} \gtrsim 3\cdot10^{21}$), demonstrating the high avalanche gain of the full power scenario of DTT.
• Figure 3: Comparison of the total plasma current $I_p$ (top) and the vertical position of the magnetic axis $Z_\text{axis}$ (bottom) for the smallest-seed case ($I_\text{seed}=5.5$ A, solid thick lines) and reference simulations without REs (dashed thin lines, obtained by numerically suppressing the avalanche source term). The colors correspond to varying levels of $N_\text{imp}$, and thin brown vertical lines demarcate the ATQ phases, consistent with Figure \ref{['fig:seed_6']}. The bottom panel reveals that higher impurity levels lead to steeper vertical drifts. For this small initial seed, the RE beam takes several milliseconds to become macroscopic and deviate from the RE-free dynamics, eventually slowing the vertical displacement.
• Figure 4: Time evolution of the total plasma current $I_p$ (solid lines) and RE current $I_\text{RE}$ (dashed lines) for an intermediate initial RE seed of $I_\text{seed} = 550$ A. The colors correspond to varying levels of $N_\text{imp}$, and thin brown vertical lines demarcate the ATQ phases, consistent with Figure \ref{['fig:seed_6']}. Significant RE beam formation occurs at lower impurity thresholds compared to the 5.5 A case. The highest RE current reaches approximately $1.8$ MA (accounting for almost $60\%$ of the total current).
• Figure 5: Current decomposition for the intermediate seed scenario ($I_\text{seed} = 550$ A) with $N_\text{imp} \approx 4\cdot10^{21}$ particles. The solid ($I_p$) and dashed ($I_\text{RE}$) lines correspond to the orange traces in Figure \ref{['fig:seed_4']}, while the gold dotted line shows the total current in the confined region ($I_\text{core}$). Thin brown vertical lines indicate the ATQ phases as in previous figures, with the addition of a dashed vertical line at $t \approx 1.6$ ms marking the end of the impurity injection and the start of the self-consistent CQ for this specific scenario. The proximity of the $I_\text{core}$ and $I_\text{RE}$ traces after $t \approx 6$ ms indicates that the confined current is dominated by the RE population, implying that the thermal current ($I_p - I_\text{RE}$) corresponds largely to halo currents.