Session GO12: Magnetic Confinement: Disruptions & Runaway Electrons

Live

The generation and evolution of relativistic electrons in tokamak plasmas has been the subject of extensive research due to their intrinsic interest as well as the threat they pose to reactor-scale tokamak devices. In this work we explore the efficiency through which a beam of runaway electrons can be suppressed via the injection of plasma impurities. It is found that the injection of a large quantity of impurities drastically modifies the underlying runaway electron phase space distribution. This modification is due in part to the strength of the knock-on collision term being drastically enhanced due to the increase in the number of target electrons, as well as the response of the inductive electric field to the sudden increase in runaway dissipation. The increase in the rate of secondary generation leads to an increase in the total number of runaway electrons, but with a substantially modified average energy and pitch. The final runaway distribution is ultimately determined by the interplay between enhanced dissipation/scattering of primary electrons, enhanced secondary generation, and the self-consistently evolving inductive electric field. Ongoing work is focused on evaluating the efficiency of a range of runaway termination schemes.   

Live

A runaway avalanche can result in a conversion of the initial plasma current into a relativistic electron beam in high current tokamak disruptions. We investigate the effect of massive material injection of deuterium-noble gas mixtures on the coupled dynamics of runaway generation, resistive diffusion of the electric field, and temperature evolution during disruptions in ITER-like plasmas. We explore the dynamics over a wide range of injected concentrations and find substantial runaway currents, unless the current quench time is intolerably long. The reason is that the cooling associated with the injected material leads to high induced electric fields that, in combination with a significant recombination of hydrogen isotopes, leads to a large avalanche generation. Balancing Ohmic heating and radiation losses provides qualitative insights into the dynamics, however, an accurate modeling of the temperature evolution based on energy balance appears crucial for quantitative predictions.   

Live

It has been demonstrated that resonant magnetic perturbations (RMPs) can suppress runaway electrons (REs). The RMP can introduce stochasticity in the plasma leading to enhanced transport. Munaretto \emph{et al.\ }[1] illustrates that in low-$q$ MST tokamak discharges, application of an $m=3$ RMP leads to suppression of REs, while an $m=1$ RMP does not have a strong effect. In $q(a)=2.2$ discharges that include a $q=1$ surface at $\frac{r}{a}=0.5$, nonlinear MHD modeling with NIMROD predicts that the region of imposed stochasticity, in the outer region of the plasma, is much larger with the $m=3$ RMP. We extend this work to include modeling of RE transport in the modeled magnetic topology. The role of the sawtoothing behavior of these discharges in transporting REs is also demonstrated, including modeling of the $q(a)=2.7$ discharges where the $q=1$ surface is separated from the stochastic edge. Development work to improve the performance of vacuum-field computations in vertical displacement events, in \emph{e.g.\ }NSTX, simulations will also be reported. [1] S. Munaretto et al 2020 Nucl. Fusion 60 046024   

Live

We have developed a code to simulate the signals from a new laser scattering diagnostic for DIII-D, which will measure the relativistic electron population expected in the DIII-D tokamak after disruptions triggered by argon pellet injection. Experimentally, the post disruption runaway current can be as large as 1 MA, and existing measurements suggest there exists a significant fraction of runaways, or even a bump on the tail distribution, at 6-10 MeV. The inverse scattering Compton cross section is highly peaked in the forward direction of the runaway electron beam. Relativistically upshifted scattered light is integrated over position and angles, and binned in energies, as it is collected at the detector position. We have done calculations for a variety of assumed electron distribution functions, with pitch angle distribution becoming smaller at higher energies. Argon pellets will be used to trigger disruptions, so we calculate the accompanying bremsstrahlung noise background from the runaways interacting with high-Z residual argon ions in the plasma. For a 10 joule laser, with 80 picosecond pulse width, using a 100 cm$^{\mathrm{2}}$ detector, a runaway electron population density of 10$^{\mathrm{10}}$ cm$^{\mathrm{-3}}$ at 6-10 MeV, and an argon density of 10$^{\mathrm{13}}$ cm$^{\mathrm{-3}}$, the expected signal-to-noise ratio in the 100 eV to 20 keV range will be \textgreater 200:1.   

Live

Disruption mitigation systems are among major challenges for ITER and future tokamaks. An electromagnetic pellet injection mechanism has been proposed that would offer a fast response time and high enough speed to deposit payloads in the plasma core [1]. The NSTX-U team is interested in testing this concept. In support of this interest, and to understand the underlying physics, simulations that can predict the evolving plasma in these conditions have been performed. The M3D-C1 code has recently incorporated the KPRAD radiation model and a pellet injection module [2,3]. We have performed simulations modelling single C-pellet injections in NSTX-U. To do this, a Carbon ablation model was incorporated in M3D-C1 and tested in an ASDEX-U-like discharge for which data existed [4], obtaining excellent agreement. Next, we performed a convergence study for NSTX-U covering different modelling parameters. We show the sensitivity of the induced thermal quench and other relevant quantities on the physical input parameters and the numerical resolution. [1] R. Raman et al., Nucl. Fusion 59 016021 (2019) [2] B. Lyons et al., Plasma Phys. and Contr. Fusion 61 064001 (2019) [3] N. Ferraro et al., Nucl. Fusion 59 016001 (2019) [4] V. Sergeev et al., Plasma Phys. Rep. 32 (2006) 363   

Live

The thermal quench in a class of tokamak disruptions is tied to plasma transport in 3D magnetic fields that have open field lines connecting the core to the wall surface. The parallel electron heat flux is thought to be the primary culprit for a core thermal collapse from 10 keV to 100 eVs over a ms. When the magnetic connection length becomes comparable to electron mean free path, one can have a thermal collapse dominated by parallel transport of extreme kinetic character. These bring considerable challenges not only in the physics itself but also in the choice of proper physics models. Here we performed fully kinetic 1D VPIC simulations to study thermal collapse in open field lines intersecting cold walls. We found that the electron flux is dominated by the parallel component, while the ion perpendicular flux can contribute significantly to the net ion flux towards the wall, hence has an important role in maintaining the ambipolarity. We also found that the electron and ion parallel heat fluxes deviate greatly from the collisional theory due to the non-Maxwellian distributions. This research used resources of NERSC, and was supported by DOE OFES and OASCR through the base theory and Tokamak Disruption Simulation (TDS) SciDAC project.   

Live

Two events often observed in MGI experiments are the excitation of the m=2/n=1 (m is poloidal mode number, n is toroidal mode number) magnetohydrodynamic (MHD) modes before thermal quench (TQ), and the formation of cold bubble structure in temperature distribution during TQ. However, the physics mechanisms underlying those phenomena have not been entirely clear. Recent NIMROD simulations have reproduced main features of both events and revealed potential connections in between. In particular, 3/1 and 2/1 islands form successively upon arrivals of impurity at the corresponding rational surfaces. At the interface between impurity and plasma, peaked poloidal magnetic perturbation along with a thin current sheet moves inward following the gas cold front. This eventually leads to the formation of an inner 2/1 mode in the region between q = 2 and q = 1 surfaces, which has an opposite phase to the dominant 2/1 mode at the vicinity of q = 2 surface. It is through the O-point of the inner 2/1 mode that the impurity front further penetrates inside the q = 1 surface, and enables the formation of a cold bubble at the beginning of TQ. In addition, a 1/1 mode appears on the q = 1 surface after the impurity penetration, and dominates the subsequent start phase of current quench (TQ).   

Live

The cross-machine data-driven study presented in this contribution shows clear evidence that non-disruptive data is machine-specific but disruptive data contains crucial general knowledge about disruptions, independent of the considered device. A Hybrid Deep Learning (HDL) architecture for disruption prediction is found to achieve high predictive accuracy on C-Mod, DIII-D and EAST tokamaks with limited hyperparameter tuning. Near-future burning plasma tokamaks will need to run disruption-free or with very few unmitigated disruptions, therefore successfully predicting disruptions on new tokamaks with limited disruption data from themselves will be crucial. The availability of data across different existing devices allows us to design numerical experiments to test transfer learning capabilities of the deep learning predictor. Surprisingly, it is found that the HDL algorithm achieves relatively good accuracy on EAST (AUC=0.959) when including 20 disruptive shots, thousands of non-disruptive data, and combining this with more than a thousand disruptive discharges from DIII-D and C-Mod. This holds true for all permutations of three tokamaks. These cross-machine studies are crucial to evaluate the performances of a general disruption prediction scheme and test its extrapolabilty.   

Live

The NDS (Nanometer Dust Simulation) module, which evaluates the charging, ablation and transport of the dust grains, has been developed under BOUT$++$ framework. The guiding-center orbits of dust particles are tracked in tokamak plasmas, whose parameters are obtained from BOUT$++$, a highly desirable C$++$ code package for performing parallel plasma fluid simulations with an arbitrary number of equations in 3D curvilinear coordinates. Calculations with NDS-BOUT$++$ provides results such as trajectories, distributions and evolutions of dust particles with different components, sizes, and velocities for different tokamak geometries. Understanding the submicron and nanometer phases of dust is fundamentally interesting, which helps better understand how lithium dropping improves plasma confinement and how beryllium pellet contributes to ELM control. The distribution of tungsten impurity, resulted from ablated tungsten dust grains for several typical scenarios, is also assessed. Preliminary results show in some cases, tungsten dust grains can cross the seperatrix and survive for several milliseconds before ablated completely, which will significantly contribute to core contamination.   

Live

To better understand plasma turbulence and transport dynamics at the tokamak edge region and resulting divertor power loads during the disruption thermal quench phase, we perform a series of BOUT++'s six-field drift-reduced Landau fluid turbulence simulations. In these simulations, appropriate sourcing levels of particles and energy from the core side are first determined by obtaining quasi-steady-state plasma profiles. The amplitudes of these sources are then lifted to mimic the onset of thermal quench in the disruption process. As the particle and energy influx from core to edge increases, plasma density and temperature are both elevated at the pedestal top and their radial profiles steepen, which eventually triggers an ELM-like burst that injects large amount of particles and energy outwards across the separatrix, resulting not only a sudden enhanced divertor heat flux peak but also a broadened width. LLNL-ABS-812058   

Progress in Tokamak Disruption Simulation (TDS) SciDAC Project

The Tokamak Disruption Simulation (TDS) SciDAC project aims to develop the physics basis for effective disruption mitigation. For the coupled nature of thermal quench and current quench, during which there can be a robust Ohmic-to-runaway current conversion, a key focus of TDS project is to understand the connection of thermal quench and runaway current conversion, under the mitigation approaches currently considered for ITER, which is based on impurity injection. Here we give an overview of TDS' efforts in (1) understanding the two primary channels for thermal quench: radiative cooling by externally introduced high-Z impurities and plasma transport in stochastic magnetic fields that have open field lines connection the fusion core to the divertor/first wall; (2) understanding the Ohmic-to-runaway current conversion physics in the presence of impurity injection of various compositions. To explore additional ways for disruption, particularly the runaway mitigation, TDS also has the efforts that aim to establish the physics underlying how interactions of runaways with plasma waves and 3D magnetic fields can facilitate runaway energy control and the subsequent mitigation of first wall damage.   

Runaway Electron Transport in Stochastic Toroidal Magnetic Fields

We study the transport and confinement of runaway electrons (RE) in presence of magnetic fields with perturbations producing different levels of stochasticity [L. Carbajal et al. Phys. Plasmas 27, 032502 (2020)]. We use KORC (Kinetic Orbit Runaway electron Code) for simulating the full-orbit (FO) and guiding-center (GC) dynamics of RE in perturbed magnetic fields that exhibit magnetic islands. Full-orbit effects on the RE dynamics are investigated, quantifying FO effects on RE transport by performing one-to-one comparisons between FO and GC simulations. It is found that in the absence of collisions and radiation losses, GC simulations predict twice the RE losses of FO simulations for 1 MeV, and four times the RE losses of FO simulations for 25 MeV. Similarly, GC and FO dynamics of RE moving around magnetic islands is compared, being very different in the scenario where the RE Larmor radius is about the size of the magnetic island. We also study the effects of rotation of the magnetic islands on RE confinement, finding no observable effect in the low-frequency toroidal rotation regime. These results shed some light into the potential of avoidance or mitigation mechanisms based on magnetic perturbations.