Session JO5: MF: DIII-D

\textbf{Overview of Recent DIII-D Experimental Results}

Recent DIII-D experiments contributed to the ITER physics basis and to physics understanding for future devices. RMP ELM suppression and density pump-out thresholds are matched by non-linear two-fluid TM-1 MHD modeling. Deuterium pellet fueling particle source maintains higher pedestal pressure than gas through Te profile changes. Stabilizing effect on 2/1 NTM islands seen from deep high field side pellet fueling. Changes in ideal and resistive stability well below the high-pressure external kink limit impact limits for disruption induced MHD. Local non-linear gyrokinetic simulations match inferred increased diffusion of impurity species correlated with transition from ITG to mixed ITG/TEM modes. Heat pulse propagation data and simulations show connection between millimeter-scale turbulence and deposition profile broadening of electron cyclotron waves. H-mode injection of isotopically enriched B correlated with reduced wall fueling and impurity concentrations similar to boronization. Observed and simulated tungsten erosion and deposition patterns suggest ExB transport effects dominate over parallel force balance effects.   

Testing the DIII-D Co/Counter Off-Axis NBI Performance for Accessing High Beta Steady State Scenarios

The DIII-D tokamak has recently undergone a major upgrade to orient a second neutral beamline for off-axis injection to broaden current and pressure profiles for high $\beta_N$ steady-state scenarios. Predictive integrated modeling of high beta steady-state scenarios on DIII-D indicate that the additional off-axis power will achieve fully non-inductive operation at $\beta_N$ =4.4 with $q_{95}$=6.9. In this presentation, we will present the testing of the neutral beam performance, and show how time-dependent application of the heating and current drive systems are predicted to achieve the steady state target in DIII-D. The upgraded beam is permanently off-axis with a downward angle of 18.55 degrees and toroidally steerable for injection in either the co-current (20.5 degrees from radial) or counter-current (19.5 degrees from radial) direction. Initial experiments have been performed that test the performance of the neutral beam by using visible imaging of the neutral beam shape, short and long neutral beam pulses for beam-target neutron production and fast ion confinement. Results of these initial checkout experiments that establish the absolute injection geometry and power will be presented, along with TRANSP predictions for achieving low net torque operation.   

Impact of EC location and timing on the stability and performance of the zero torque ITER Baseline Scenario in DIII-D

Scans of EC deposition at zero input torque in recent ITER Baseline Scenario (IBS) Demonstration discharges show that power deposition in the region of the q$=$2 surface is prone to causing (not suppressing) 2/1 modes, due to its impact on the local Te$_{\mathrm{ped}}$. The maximum stable Te$_{\mathrm{ped}}$ is inversely proportional to li, which points to a first order dependence of the stability on the global current profile (J) shape. The local minimum in J near q$=$2 is higher later in the shot, when li is lower, and the equilibrium can sustain a higher Te$_{\mathrm{ped}}$ without crossing the stability boundary. Local T$_{\mathrm{e}}$ impacts both the bootstrap current and the resistivity, therefore both the outer and inner layer physics, affecting the $\Delta $' and the $\Delta $' critical for instability. An in-shot dynamic scan of EC deposition from core to edge decreases H$_{\mathrm{98y2}}$ by 17{\%} and $\tau_{\mathrm{E}}$ by 30{\%}, due to loss of heating efficiency. This calls into question the compatibility of direct EC stabilization with achieving ITER's performance goals. 0-D simulations show that the zero torque IBS shots with core ECH project to the ITER goals (Q$=$10, P$_{\mathrm{fus}}=$550 MW, with heating power P$_{\mathrm{heat}}=$70 MW, compatible with the ITER hardware), and indicate the trade-offs between density, field, pressure and gain.   

Non-ideal contributions to the stability of low-torque ITER baseline discharges

DIII-D experiments and simulations provide new insights into disruption inducing tearing modes, showing how these correlate with increased plasma response to magnetic probing and stability calculations. The dependencies of the plasma response on normalized internal inductance $\ell_\mathrm{i}$ and pressure $\beta_\mathrm{N}$ are qualitatively consistent with ideal MHD, although the measurements indicate weaker stability than the simulations predict. This result is unexpected in light of similar comparisons made previously in strongly rotating discharges, wherein ideal MHD predicted poorer stability than implied by the measurements, and better agreement was obtained with simulations including drift-kinetic modifications. Calculations of the classical $(m,n)=(2,1)$ tearing index $\Delta'$ also exhibit sensitivities to $\beta_\mathrm{N}$ and \ell_\mathrm{i}$ consistent with the measured response. These results pose a challenge to low rotation reactor regimes like the ITER baseline, but also provide a foundation and sensing technique to anticipate and optimize stability.   

Physical Effects of Rotation and Rotational Shear on Quiescent H-mode

Quiescent H-mode (QH) is an attractive naturally ELM-stable regime that operates at ITER relevant collisionality and good confinement. Experimentally, strong NBI torque is usually required to excite the EHO that regulates the QH edge. ExB rotation and shear were found to destabilize the EHO in M3D-C1 linear simulation [Chen NF 2017]. New linear MINERVA-DI modelling suggests an optimal window of ion diamagnetic drift to utilize this effect. Nonlinear NIMROD simulation show the rotation is essential for the EHO saturation. On the other hand, wide-pedestal QH (WPQH) can operate at low rotation with pedestal ExB shear lower than ELMy H-mode while maintaining H98\textgreater 1. This may ease the concern that the low ExB shear in the ITER pedestal may be insufficient to excite EHOs or to suppress ion-scale turbulence to achieve high pedestal for good confinement. Low- to intermediate-k turbulence co-located with flattened pedestal profiles arise in WPQH. It is posited that high edge turbulence resulting from weak ExB shear, relaxes the pedestal gradients, and enables a wider and higher pedestal. WPQH also exhibits improved confinement with increasing ECH power, and switching from LSN to DN shape. Turbulence suppression by increased ExB shear inboard of pedestal top plays an important role in these observations.   

The role of resonant field penetration in ELM suppression and density pump-out in the DIII-D tokamak

Recent nonlinear two-fluid MHD modeling using the TM1 code demonstrates quantitative agreement with ELM suppression and density pump-out by RMPs observed in DIII-D. We find that the formation of magnetic islands at the top and bottom of the pedestal can account for both ELM suppression and density pump-out observed in DIII-D. For low collisionality plasmas with n = 2 RMPs, simulations show that the penetration of RMP at the pedestal foot drives magnetic islands at low amplitude (dB/B = 2.E-5), which flattens the local density and lowers the density at the pedestal top. Comparisons with DIII-D experiments indicate that the formation of magnetic islands at the pedestal foot is the dominant contributor to density pump-out prior to ELM suppression. Stronger RMPs cause further density pump-out and, eventually, formation of magnetic islands at the top of pedestal that can suppress ELMs near the DIII-D experimentally observed threshold dB/B = 2.E-4. A scaling law is derived for the field penetration threshold at the pedestal top, which is consistent with DIII-D experiments. The predicted threshold for forming the necessary magnetic islands in ITER should be lower than present devices due to the much lower plasma flow velocity expected in the ITER pedestal.   

Fast rampdown and disruption avoidance studies on DIII-D and EAST

Improved standard and emergency shutdown methods have been developed in DIII-D over a large piggyback experiment varying shutdown techniques, with measured improvement in disruptivity rates. The experimental survey used the shutdown phase of \textgreater 1000 plasmas in the '17-'19 DIII-D campaigns. The disruptivity of single-null plasmas was minimized with relatively fast I$_{\mathrm{p}}$ ramp-down rates of 2--3 MA/s while maintaining neutral beam heating comparable to the radiated power for the majority of shutdown. Transitioning to a limited shape for shutdown further reduced disruptivity to \textless 10{\%} compared to the DIII-D historical rate of 28{\%}. Emergency shutdown of DIII-D ITER Baseline Scenario plasmas after locked modes is a special challenge. All of 46 such attempts with continued diverted topology disrupted before reaching safe normalized currents (\textasciitilde 0.3 for ITER). However, 2 of 3 limited emergency shutdowns did avoid disruption, motivating further testing in the '19 campaign. Experiments on the EAST tokamak have likewise identified robust, fast, ramp-down techniques. These included diverted and limited shutdowns up to 0.7 MA/s with sustained lower hybrid power. In the `19 DIII-D campaign, the shutdown study will be expanded to develop and rigorously test disruption avoidance techniques, with a special focus on tearing and locked modes, and initial results will be presented.   

Shattered Pellet Injection (SPI) assimilation and superposition on DIII-D

Shattered Pellet Injection (SPI) has been chosen as the baseline disruption mitigation system for ITER. However, many questions remain regarding its operation, particularly under the presently envisaged operating scenario where several SPIs may need to be superimposed in order to inject massive quantities of deuterium prior to the thermal quench for runaway electron suppression. Simultaneous injection of two shattered pellets exhibit a reduction in the pre-thermal quench time (time from when SPI fragments reach the plasma edge until the start of the thermal quench), relative to similar single SPI mitigated shutdowns. Despite the decreased time to assimilate the injected impurities, the electron density increased by approximately a factor of two with the addition of multiple pellets but is highly sensitive to the time between injections. A maximum density increase is found when both pellets arrive at the plasma prior to the start of the TQ. The radiation cooling code KPRAD has been modified to simulate SPI shutdowns, with the ability to incorporate arbitrary pellet composition mixtures and multiple SPIs into the same discharge. Simulations agree well with experimental measurements, reproducing the electron density increase in dual SPI experiments.   

Towards Reactor-Relevant Runaway Electron Dynamics: High Temperature Formation and High Current Instability

Access to pre-disruption electron temperatures ($T_e$) in excess of 10 keV and post-disruption runaway electron (RE) currents ($I_{RE}$) approaching 1 MA allow novel observation of reactor-relevant RE dynamics, such as $\gt$ 80\% conversion of thermal to RE current and prompt termination of the RE beam via global current-driven instability. A dramatic increase in RE production with rising pre-disruption $T_e$ is observed, with efficiency increasing from 20\% in conventional $ \approx 2$ keV plasmas to 80\% in $ \approx 8$ keV plasmas. Above 10 keV, evidence supports the formation of sub-MeV RE beams. RE quantities and distribution functions measured via bremsstrahlung will be compared to model predictions across $T_e$, with implications for accurate prediction of RE formation. When post-disruption $I_{RE}$ approaches 1 MA and edge safety factor $q_a$ crosses 2, current-driven kink instabilities appear with an Alfvenic growth time and ultimately cause complete loss of the RE population without regeneration. Modeling of the critical instability amplitude for total RE orbit loss is in agreement with external magnetic measurement. Implications for RE mitigation from current-driven instability are mixed, with higher peak heat loading yet lower total energy deposition expected.   

Validation of the TGLF-EP$+$Alpha critical-gradient model of energetic particle transport in DIII-D

The TGLF-EP$+$Alpha [1] critical-gradient model of energetic-particle (EP) transport is here validated against five DIII-D H-mode scenarios, predicting how Alfv\'{e}n eigenmodes driven by beam ions \'{e}adially flatten the EP profile. The critical gradient comes from the TGLF [2] gyro-Landau fluid code, optimized for EP-AE physics, automated, and highly parallelized by the TGLF-EP wrapper code. TGLF-EP$+$Alpha is fully physics based, requiring only experimental equilibrium and beam source as inputs. It is computationally inexpensive enough to perform extensive scoping studies needed for scenario optimization. Cases show observed neutron production near classical down to an 80{\%} deficit. TRANSP simulations using the EP diffusion coefficient predicted by TGLF-EP$+$Alpha find neutron deficits within 20{\%} of experimental observations in applicable cases. [1] He Sheng et al, \textit{Phys. Plasmas} \textbf{24}, 072305 (2017) [2] G. M. Staebler et al, \textit{Phys. Plasmas} \textbf{14}, 55909 (2007)   

On the Role of Microtearing Turbulence in DIII-D High Bootstrap Current Fraction Plasmas

We report the first direct comparisons of microtearing turbulence simulations to experimental measurements in the DIII-D tokamak. Previous studies of high bootstrap fraction plasmas carried out in DIII-D with large radius internal transport barriers (ITBs) have found that while the ion energy transport is neoclassical, the electron transport remains anomalous, and not well-described by existing quasilinear transport models. The large value of normalized pressure gradient in these plasmas is shown to stabilize drift-wave and kinetic ballooning mode instabilities in the ITB, but destabilize the microtearing mode (MTM). Nonlinear gyrokinetic simulations of the ITB region performed with CGYRO demonstrate that the MTM are robustly unstable, and capable of reproducing the inferred electron energy transport within experimental uncertainties. The predicted transport levels are found to be most sensitive to the magnetic shear rather than the temperature gradients. Extrapolation to an ITER steady-state scenario suggests that whether MTMs or drift-wave transport dominates will depend sensitively on the balance between lower collisionality and high safety factor expected in these scenarios.   

Role of Poloidal E$\times$B Convection on Divertor Heat Transport in DIII-D

Simulations for DIII-D high confinement mode plasmas with the fluid code UEDGE show a strong role of poloidal $\mathbf{E}\times\mathbf{B}$ drifts on the divertor heat transport. These findings challenge the paradigm of conduction-limited scrape-off layer (SOL) transport, which is a standard assumption in analytic calculations of divertor power exhaust, such as in [1]. While simulations without drifts are well aligned with this assumption, simulations with drifts show that the poloidal $\mathbf{E}\times\mathbf{B}$ flow dominates the divertor heat transport in both attached and detached conditions. This study has identified, for the first time, the important contribution of $\mathbf{E}\times\mathbf{B}$ drifts to the strong convective heat flow in detached conditions in DIII-D previously reported in [2]. The impact of the convective $\mathbf{E}\times\mathbf{B}$ flows on the radiated power density scaling in the divertor will also be discussed. [1] R.J. Goldston, et al. Plasma Phys. Control. Fusion 59 (2017) 055015 [2] A.W. Leonard, et al. Phys. Rev. Lett. 78 (1997) 4769 – 4772   

The key role of ExB drifts in W impurity transport and redeposition in the DIII-D divertor

Mixed-material DIVIMP-WallDYN modelling, now incorporating ExB drifts, is presented that is consistent within a factor of 2 with tungsten (W) erosion and deposition patterns observed during L-mode experiments in DIII-D with toroidally symmetric W-coated tiles installed in the carbon (C) divertor. It is demonstrated that ExB drifts are required to reproduce the experimental observations, and that the spatial structure of modelled divertor poloidal ExB drifts correlates with boundaries of the observed deposition/erosion regions. With attached L-mode conditions and ion grad-B drift out of the divertor, W and C co-accumulation is observed over a band located \textasciitilde 5-8 cm outboard of the outer-strike-point (OSP) W source, but little W is observed closer to the OSP. Time-dependent simulations with modified ExB impurity drifts (set to 60{\%} of the calculated value) quantitatively reproduce these features, including depth-resolved W/C ratios, within a factor of 2 over \textasciitilde 110 seconds of plasma exposure. These simulations indicate that ExB transport can dominate over parallel force balance for W in the divertor region. This represents the first self-consistent modeling of global W redeposition in a C divertor with W targets.   

SOLPS-DIVIMP modeling of the impact of divertor closure on tungsten erosion and transport in DIII-D

Predictive modeling of the Small Angle Slot (SAS) divertor in DIII-D with toroidal tungsten rings in different poloidal locations is conducted to evaluate the impact of divertor closure on W impurity sourcing and transport. It is found that the closed slot structure of SAS results in denser and colder plasmas relative to the open lower divertor leading to lower sputtering rates, as well as stronger divertor screening associated with shorter ionization mean free paths and stronger friction force towards the target. Compared with the lower open divertor, the SAS divertor achieves 1 to 2 orders of magnitude smaller W erosion and leakage for similar upstream conditions. The enhanced redeposition and more favorable net force balance reduce both the effective W erosion and the divertor leakage from the slot bottom W source compared to the far-target W sources. With fixed upstream conditions, when the outer strike point (OSP) is placed on the slot bottom W ring, the W net erosion and leakage fraction decreases by an order of magnitude and over 60{\%}, respectively, relative to the case with the OSP on the W ring at the outer baffle. These simulations utilize the DIVIMP code for W erosion, transport and deposition/redeposition, with background plasma solutions provided by the SOLPS code. The effect of drifts on W migration in the SOL region will also be reported.   

Experimental evidence of local ExB drift in the divertor plasma influencing the upstream density profile

Local flattening of the radial profile of the upstream density near the magnetic separatrix has been observed in DIII-D H-mode plasmas with unfavorable Bt. The density plateau is correlated with a double-peaked density profile near the divertor plate, connected via 2D ExB drifts. Simulations show that near the sheath entrance the strong radial electric field induced by the radial temperature gradient leads to a supersonic parallel flow. To maintain the parallel pressure balance, the increased dynamic pressure causes a strong static pressure drop along the field line, thus generating a valley in the density profile near the plate. In addition, the poloidal electric field resulting from the static pressure loss enhances the density in the far SOL, leading to the observed double-peak density profile. Upstream of the divertor target, drifts lead to a reversal of the total poloidal flow in the main SOL, which causes formation of the density plateau. The density plateau is correlated with divertor conditions and can be greatly enhanced with peak density in the main SOL comparable to that at the pedestal top.