Session PO4: Research in Support of ITER

Superposition of dual shattered pellet injections for disruption mitigation

Experiments on DIII-D have injected multiple shattered pellets at different toroidal locations for the first time, as is planned for the ITER disruption mitigation system. Systematically varying the relative timing of the two pellets demonstrate that simultaneous pellets may mitigate less effectively than a single pellet injecting similar neon quantities. Thermal quench (TQ) radiation fractions measured near the injection are reduced with the dual pellets, possibly as a result of a more rapid shutdown due to a broader impurity distribution over multiple flux tubes. However, radiation measured away from the injection does not share this trend, indicating asymmetries may exist. Mitigation of current quench (CQ) loads is also similarly reduced in the dual pellet cases, consistent with the observed reduction in TQ mitigation. However, pellets entering the plasma after (or during) the TQ initiated by the other pellet, can still contribute to CQ mitigation.   

Predicting the necessary impurity quantities for thermal quench mitigation in ITER disruptions

During a thermal quench (TQ) in a $15$ MA scenario in ITER, heat fluxes can significantly exceed melt limits. It is planned to inject neon (Ne) atoms to radiate $>90\%$ of the thermal energy in this scenario. However, an upper-limit on the current quench rate prohibits the use of more than $4\times10^{22}$ Ne atoms. Previous simulation work finds the required Ne quantity for ITER TQ mitigation is marginal with respect to this limit, however a cross-machine validation of this simulation is missing, and thus its accuracy is unknown. Scaling laws and 0D and pseudo-1D models of varying complexity are developed to compare estimates of the required Ne quantity for TQ mitigation. The physics considered in these models includes atomic radiation with non-coronal equilibrium effects, the scaling of the TQ duration with the minor radius, dilution cooling, and radial localization of the impurities. These models are compared with published data on the "minimum necessary injected impurity quantity" for TQ mitigation. The published data are converted to quantities assimilated before the TQ, using measured efficiencies and 1D simulations of gas flow in injection pipes. Implications on the required Ne quantity for ITER TQ mitigation will be discussed.   

Fast Time Response Electromagnetic Particle Injection System for Disruption Mitigation

Predicting and controlling disruptions is an urgent issue for ITER. In this proposed method, a radiative payload consisting of micro spheres of Be, BN, B, or other acceptable low-Z materials would be injected inside the q$=$2 surface for thermal and runaway electron mitigation. The radiative payload would be accelerated to the required velocities (0.2 to \textgreater 1km/s) in an Electromagnetic Particle Injector (EPI). An important advantage of the EPI system is that it could be positioned very close to the reactor vessel. This has the added benefit that the external field near a high-field tokamak dramatically improves the injector performance, while simultaneously reducing the system response time. A NSTX-U / DIII-D scale system has been tested off-line to verify the critical parameters - the projected system response time and attainable velocities. Both are consistent with the model calculations, giving confidence that an ITER-scale system could be built to ensure safety of the ITER device. This work is supported by U.S. DOE Contracts: DE-AC02-09CH11466, DE-FG02-99ER54519 AM08, and DE-SC0006757.   

Shattered Pellet Injection Simulations With NIMROD

Shattered Pellet Injection~(SPI) will be the Disruption Mitigation System in ITER. SPI propels a cryo-pellet of high-Z and deuterium into a sharp bend of the flight tube, shattering the pellet into a plume of shards. These shards are injected into the plasma to quench it and mitigate forces and heat loads that may damage in-vessel components. We use NIMROD to perform 3-D nonlinear MHD simulations of SPI to study the thermal quench. This work builds upon prior Massive Gas Injection~(MGI) studies by Izzo\footnote{V.~A.~Izzo, et al. NF(55) 073032}. A Particle-in-Cell~(PIC) model is implemented to mimic the shards, providing a discrete moving source. Observations indicate that the quench proceeds in two phases. Initially, the outer plasma is shed via interchange-like instabilities while preserving the core temperature. This results in a steep gradient and triggers the second phase, an external kink-like event that collapses the core. We report on the radiation efficiency and toroidal peaking as well as fueling efficiency and other metrics that assess the efficacy of the SPI system.   

Plasma response control using advanced feedback techniques

Recent DIII-D experiments show that a new, advanced algorithm improves resistive wall mode (RWM) stability control in high performance discharges using external coils. DIII-D can excite strong, locked or nearly locked kink modes whose rotation frequencies do not evolve quickly and are slow compared to their growth rates. Simulations and experiments have shown that modern control techniques will perform better, using 77{\%} less current, than classical techniques when using coils external to the vacuum vessel for RWM feedback. ITER will have to deal with such modes, especially in steady state operation, and it is unclear whether or not rotation alone will be sufficient to counteract these modes. VALEN models the perturbed magnetic field from a single MHD instability and its interaction with surrounding conducting structures as a series of coupled circuit equations. RWM feedback based on VALEN outperformed a classical control algorithm using external coils to suppress the normalized plasma response to a rotating n$=$1 perturbation applied by internal coils over a range of frequencies. Work supported by the U.S. DOE under DE-FC02-04ER54698 and DE-FG02-04ER54761.   

Development of a real-time system for ITER first wall heat load control

The steady state heat flux on the ITER first wall (FW) panels are limited by the heat removal capacity of the water cooling system. In case of off-normal events (e.g. plasma displacement during H-L transitions), the heat loads are predicted to exceed the design limits (2-4.7 MW/m$^2$). Intense heat loads are predicted on the FW, even well before the burning plasma phase. Thus, a real-time (RT) FW heat load control system is mandatory from early plasma operation of the ITER tokamak. A heat load estimator based on the RT equilibrium reconstruction has been developed for the plasma control system (PCS). A scheme, estimating the energy state for prescribed gaps defined as the distance between the last closed flux surface (LCFS)/separatrix and the FW is presented. The RT energy state is determined by the product of a weighted function of gap distance and the power crossing the plasma boundary. In addition, a heat load estimator assuming a simplified FW geometry and parallel heat transport model in the scrape-off layer (SOL), benchmarked against a full 3-D magnetic field line tracer is also presented.   

Comprehensive Assessment of Damage Effects during Transient Events in ITER

During abnormal operations in tokamaks, the incident particle and heat fluxes during disruptions and ELMs are quickly generate a secondary plasma composed mainly from divertor plate materials. This mini-plasma will then absorb the incoming disruptive plasma and convert its energy to intense radiation fluxes to nearby components. HEIGHTS simulations showed significant increase of radiation fluxes and components heat load for high-Z (i.e., tungsten) generated secondary plasma. These radiations can seriously damage hidden components such as umbrella tubes and dome structure. In fact, simulation showed that during longer disruption times the evolving mini-plasma can damage parts of Be first wall. We have enhanced previous HEIGHTS models and implemented efficient models for photon and particle transport in evolving secondary plasma during transients. HEIGHTS now simulates full 3D detail ITER geometry to assess various damage of these components during plasma instabilities. HEIGHTS predicted again, for the first time, details of heat loads and temperatures evolution of divertor and nearby components including first wall. Current ITER divertor design needs to be modified to withstand the damage produced from disruptions. A single unmitigated disruption event can cause serious damage to components that were not directly exposed to disruptions including dome, stainless steel tubes, and parts of Be first wall.   

Prototype testing of the ITER Toroidal Interferometer and Polarimeter (TIP) on DIII-D

A 10.6 micron CO2 laser based ITER TIP system has been designed and tested for density measurements on DIII-D. Features include vibration compensation using a 5.22 micron Quantum Cascade Laser, real-time measurements at 1 kHz with \textless 1{\%} error at expected ITER operating densities, 500 kHz bandwidth density fluctuation measurements, active feedback alignment with auto signal recovery capabilities, fringe skip correction using polarimetery measurements, and a novel three-frequency heterodyne technique with real-time digital phase demodulation. A 120 m path length laboratory prototype was used to test components, demonstrate active feedback alignment capabilities, and determine noise floor capabilities. Phase errors of 1.5 degrees for the interferometer and 0.06 degrees for the polarimeter have been demonstrated for 1000 seconds. The system is now installed on the DIII-D tokamak, using a geometry and path length similar to that planned for ITER and has successfully demonstrated the ITER requirements for accuracy and time resolution.   

Divertor Heat Flux Control with 3D Stochastic Magnetic Fields during ELM Suppression.

Experiments in DIII-D have been performed to modify the divertor heat and particle flux pattern during suppression of ELMs with resonant magnetic perturbation (RMP) fields. In this work, we assessed the impact of small current modulations in a subset of DIII-D I-coils on pedestal profiles, transport and stability as well as divertor conditions. Different I-coil subset ramps were performed allowing for a slow transition of the divertor footprints from n$=$3 to n$=$2 and n$=$1 distributions. We obtained long periods of RMP ELM suppression with slow I-coil quartet ramps. Strong divertor particle flux splitting was observed in these discharges as well as modulation of the divertor heat flux due to changes in toroidal spectrum of applied perturbation. Experimental results are compared to the TRIP3D modeling and to linear M3D-C1 simulations to understand the role of the plasma response on quantitative predictions of the divertor flux splitting.   

Enhancement of Helium exhaust by resonant magnetic perturbations in DIII-D

Clear evidence of enhanced He exhaust during RMP ELM suppression has been obtained for the first time in a series of lower single null H-mode discharges with and without RMP in DIII-D. During RMP, reduced midplane He density measurements from CER and faster neutral He decay times after a 100ms He puff provided evidence for faster outward transport. Additionally, during RMP, neutral He pressure in the lower pumping plenum increased, while D$_{\mathrm{2}}$ pressure was similar to the no RMP case. A spectrometer viewing the divertor shelf in the scrape off layer measured consistently increased He-I light during RMP ELM suppression. These two measurements indicate an improved retention of He in the unconfined region, which is important for enhanced He removal. Consequently, the effective helium confinement time, $\tau $*$_{\mathrm{p,He}}$, measured for conditions in this work was reduced by \textgreater 35{\%} when RMP ELM suppression was obtained.   

Expansion of Parameter Space for Wide-Pedestal, Quiescent H-mode Plasmas in DIII-D

QH-mode is an attractive operating regime for future devices since it has excellent energy confinement time and operates without ELMs at zero net NBI torque. The recently discovered wide pedestal QH-modes exhibit an increase in the pedestal pressure height and width and increased global energy confinement associated with a bipolar change in the ExB shear [1]. Experiments in 2017 have investigated the underlying physics by expanding the parameter space for wide pedestal QH-mode. It has now been created in LSN discharges with dRsep $\ge $ -1.5 cm and sustained with dRsep $\ge $ -3.5 cm; previous experiments required balanced DN plasmas or USN with dRsep $\le $ 1 cm. A wide range of NBI torque was used to sustain the wide pedestal, from -2.5 Nm (counter-Ip) to 1.9 Nm (co-Ip); this greatly exceeds the ITER-equivalent torque range. The wide pedestal state with broadband MHD can be created directly from standard ELM-free conditions without the presence of the coherent EHO in plasmas using -1.5 Nm NBI torque. The time integrated torque required to reach and sustain the wide pedestal state has been reduced by over 90{\%}, limited by core tearing modes. This is an important step on the road to use QH-mode in future devices with much lower equivalent NBI torque. [1] K.H. Burrell et al, Phys. Plasmas 23, 056103 (2016)   

Non-linear MHD simulations of pellet triggered ELM for ITER plasma scenarios

The non-linear MHD simulations with the JOREK code have been performed to study the dependence of the pellet size required to trigger an ELM in ITER plasma, and also the dependency of the threshold on the pedestal plasma pressure when the pellet is injected. Based on the observation that the pedestal pressure leading to spontaneous ELM triggering is 150 kPa by JOREK simulation, pedestal pressure of 75 kPa and 112.5 kPa have been studied. The JOREK simulation results show that it is necessary to increase the pellet size by a factor of 1.5 of the number of particles in the pellet to trigger ELMs for a pedestal pressure of 75 kPa compared to 112.5 kPa in ITER 15MA/5.3T plasma. In these simulations it has also been found that the magnitude of the ELM energy loss is strongly correlated with the pedestal plasma pressure rather than with the size of the pellet that is required for triggering. The JOREK simulation shows the toroidally asymmetric profile of the heat flux on the outer divertor target due to the pellet triggered ELM which is consistent with the experiment observation of JET. The work contributes the estimation of the requirement of the pellet injection condition to control ELMs in ITER 15MA operation scenarios.   

Pellet Fueling of ELM Mitigated ITER Baseline Scenario Plasmas on DIII-D*

ITER has been designed to employ HFS pellet fueling to operate at high density using pellets that produce less than a 10{\%} density perturbation and that penetrate to just beyond the top of the pedestal. DIII-D ITER-like plasmas with ELM mitigation provided by either pellet ELM pacing or RMPs have been fueled with small shallow penetrating HFS pellets. In DIII-D, large natural ELMs dramatically reduce the effective fueling efficiency of the HFS pellets to nearly 0 within the time scale of a few large ELMs. In the ELM paced plasmas, the pellets trigger ELMs that are no more intense than the paced ELMs and the resulting fueling efficiency is \textgreater 80{\%}. In RMP ELM mitigated plasmas, the efficient fueling from HFS pellets results in pedestal collisionality increases that can reduce the ELM mitigation. New experiments have shown that edge ECH can decrease pedestal collisionality and will be combined with the HFS pellets to better mimic the ITER fueling scenario. These fueling results from pellet pacing and RMP ELM mitigation will be presented and implications for ITER discussed.   

Exploring an Alternate Approach to Q$=$10 in ITER

The ITER Research Plan envisions a stepwise approach in B and I due to heating system constraints to the objective of 500 MW fusion power with Q$=$10 for \textgreater 300 s, but always reaching $q_{\mathrm{95}}=$3 at each B. An alternate approach goes directly to 5.3 T, then raises I. This approach reduces disruption risk because $q$ is higher and perhaps the goal is realized at lower I. DIII-D experiments explored this path with co-NBI heating and NBI heating with 0 Nm applied torque. For the first time, stable plasmas in the ITER shape (including aspect ratio) at the ITER baseline scenario conditions ($q_{\mathrm{95}}\approx $3, $\beta _{\mathrm{N}}\approx $2) have been obtained with 0 Nm applied torque. At equivalent currents to 9-17 MA in ITER ($q_{\mathrm{95}}\approx $5.7-2.8), the maximum stable $\beta $ and the $\tau_{\mathrm{E}}$ have been measured as a function of applied torque. The equivalent $\beta $ for 500 MW of fusion power is obtained at about 13.5 MA for 0 Nm, indicating significant stability margin. However, confinement is less than predicted by the H-mode scaling at 15 MA because linear confinement scaling with I is not seen above 12.5 MA at all levels of applied torque, indicating this is not due to ExB shearing effects. These results indicate that the risk of operation in ITER at low $q_{\mathrm{95}}$ and specifically at 15 MA may not be warranted.   

Safety factor profiles from spectral motional Stark effect for ITER applications

Depositions on the first mirror and multiple reflections on the other mirrors in the labyrinth of the optical system in the motional Stark effect (MSE) diagnostic for ITER are regarded as one of the main obstacles to overcome. One of the alternatives to the present-day conventional photoelastic-modulation-based MSE principles is the spectroscopic analyses on the motional Stark emissions where either the ratios among individual Stark multiplets or the amount of the Stark split are measured based on precise and accurate atomic data and models to ultimately provide the critical internal constraints in the magnetic equilibrium reconstruction. Equipped with the PEM-based conventional MSE hardware since 2015, the KSTAR MSE diagnostic system is capable of investigating the feasibility of the spectroscopic MSE approach particularly via comparative studies with the PEM approach. Available atomic data and models are used to analyze the beam emission spectra with a high-spectral-resolution spectrometer with a patent-pending dispersion calibration technology. Experimental validation on the atomic data and models is discussed in association with the effect of the existence of mirrors, the Faraday rotation in the relay optics media, and the background polarized light on the measured spectra.