Controlling H-Mode Particle Transport with Modulated Electron Heating in DIII-D and Alcator C-Mod via TEM Turbulence
Abstract not provided
Engineering topics
Publications and source records attributed to Holland, C.
Abstract not provided
Two-dimensional (2D) turbulence simulations are reported in which the balancing of the parallel and perpendicular currents is modified by changing the axial boundary condition (BC) to vary the sheath conductivity. The simulations are carried out using the 2D scrape-off-layer turbulence (SOLT) code. The results are compared with recent experiments on the controlled shear de-correlation experiment (CSDX) in which the axial BC was modified by changing the composition of the end plate. Reasonable qualitative agreement is found between the simulations and the experiment. When an insulating axial BC is used, broadband turbulence is obtained and an inverse cascade occurs down to low frequencies and long spatial scales. Robust sheared flows are obtained. By contrast, employing a conducting BC at the plate resulted in coherent (drift wave) modes rather than broadband turbulence, with weaker inverse cascade, and smaller zonal flows. The dependence of the two instability mechanisms (rotationally driven interchange mode and drift waves) on the axial BC is also discussed.
Electron profile stiffness was studied in DIII-D L-mode discharges by systematically varying the heat flux in a narrow region with electron cyclotron heating and measuring the local change produced in {nabla}T{sub e}. Electron stiffness was found to slowly increase with toroidal rotation velocity. A critical inverse temperature gradient scale length 1/L{sub C} {approx} 3 m{sup -1} was identified at {rho}=0.6 and found to be independent of rotation. Both the heat pulse diffusivity and the power balance diffusivity, the latter determined by integrating the measured dependence of the heat pulse diffusivity on -{nabla}T{sub e}, were fit reasonably well by a model containing a critical inverse temperature gradient scale length and varying linearly with 1/L{sub T} above the threshold.
The kinetic energy transfer between shear flows and the ambient turbulence is investigated in the Experimental Advanced Superconducting Tokamak during the L-H transition. As the rate of energy transfer from the turbulence into the shear flow becomes comparable to the energy input rate into the turbulence, the transition into the H-mode occurs. As the observed behavior exhibits several predicted features of zonal flows, the results show the key role that zonal flows play in mediating the transition into H-mode.
Abstract not provided
Verification of nonlinear microstability codes is a necessary step before comparisons or predictions of turbulent transport in toroidal devices can be justified. By verification we mean demonstrating that a code correctly solves the mathematical model upon which it is based. Some degree of verification can be accomplished indirectly from analytical instability threshold conditions, nonlinear saturation estimates, etc., for relatively simple plasmas. However, verification for experimentally relevant plasma conditions and physics is beyond the realm of analytical treatment and must rely on code-to-code comparisons, i.e., benchmarking. The premise is that the codes are verified for a given problem or set of parameters if they all agree within a specified tolerance. True verification requires comparisons for a number of plasma conditions, e.g., different devices, discharges, times, and radii. Running the codes and keeping track of linear and nonlinear inputs and results for all conditions could be prohibitive unless there was some degree of automation. We have written software to do just this and have formulated a metric for assessing agreement of nonlinear simulations. We present comparisons, both linear and nonlinear, between the gyrokinetic codes GYRO[J. Candy and R. E. Waltz, J. Comput. Phys. 186, 545 (2003)] and GS2[W. Dorland, F. Jenko, M. Kotschenreuther, and B. N. Rogers, Phys. Rev. Lett. 85, 5579 (2000)]. We do so at the mid-radius for the same discharge as in earlier work [C. Holland, A. E. White, G. R. McKee, M. W. Shafer, J. Candy, R. E. Waltz, L. Schmitz, and G. R. Tynan, Phys. Plasmas 16, 052301 (2009)]. The comparisons include electromagnetic fluctuations, passing and trapped electrons, plasma shaping, one kinetic impurity, and finite Debye-length effects. Results neglecting and including electron collisions (Lorentz model) are presented. We find that the linear frequencies with or without collisions agree well between codes, as do the time averages of the nonlinear fluxes without collisions. With collisions, the differences between the time-averaged fluxes are larger than the uncertainties defined as the oscillations of the fluxes, with the GS2 fluxes consistently larger (or more positive) than those from GYRO. However, the electrostatic fluxes are much smaller than those without collisions (the electromagnetic energy flux is negligible in both cases). In fact, except for the electron energy fluxes, the absolute magnitudes of the differences in fluxes with collisions are the same or smaller than those without. None of the fluxes exhibit large absolute differences between codes. Beyond these results, the specific linear and nonlinear benchmarks proposed here, as well as the underlying methodology, provide the basis for a wide variety of future verification efforts.
Robust validation of predictive turbulent transport models requires quantitative comparisons to experimental measurements at multiple levels, over a range of physically relevant conditions. Toward this end, a series of carefully designed validation experiments has been performed on the DIII-D tokamak [J. L. Luxon, Nucl. Fusion 42, 614 (2002)] to obtain comprehensive multifield, multipoint, multiwavenumber fluctuation measurements and their scalings with key dimensionless parameters. The results of two representative validation studies are presented: an elongation scaling study performed in beam heated L-mode discharges and an electron heating power scan performed in quiescent H-mode (QH-mode) discharges. A 50% increase in the elongation {kappa} is observed to lead to a {approx}50% increase in energy confinement time {tau}{sub e} and accompanying decrease in fluctuation levels, qualitatively consistent with a priori theoretical predictions and nonlinear GYRO [J. Candy and R. E. Waltz, J. Comput. Phys. 186, 545 (2003)] simulations. However, these simulations exhibit clear quantitative differences from experiment in the predicted magnitudes and trends with radius of turbulent fluxes and fluctuation levels which cannot be fully accounted for by uncertainties due to transport stiffness. In the QH-mode study, local nonlinear GYRO simulations that neglect fast ion effects show a similar proportional response to the applied electron cyclotron heating as the experiment, but overpredict the magnitudes of transport and fluctuation levels by a factor of 10 or more. Possible sources of this overprediction, namely nonlocal effects and self-consistent fast beam ions, are identified and discussed.
The local value of a/L{sub Te}, a turbulence drive term, was modulated with electron cyclotron heating in L-mode discharges on DIII-D [J. L. Luxon, Nucl. Fusion 42, 614 (2002)] and the density and electron temperature fluctuations in low, intermediate, and high-k regimes were measured and compared with nonlinear gyrokinetic turbulence simulations using the GYRO code [J. Candy and R. E. Waltz, J. Comput. Phys. 186, 545 (2003)]. The local drive term at rhoapprox0.6 was reduced by up to 50%, which produced comparable reductions in electron temperature fluctuations at low-k. At intermediate k, k{sub t}hetaapprox4 cm{sup -1} and k{sub t}hetarho{sub s}approx0.8, a very interesting and unexpected result was observed where density fluctuations increased by up to 10% when the local drive term was decreased by 50%. Initial comparisons of simulations from GYRO with the thermal diffusivity from power balance analysis and measured turbulence response are reported. Simulations for the case with the lowest drive term are challenging as they are near the marginal value of a/L{sub Te} for trapped electron mode activity.
The experimental observations of the dynamics of the coupled drift wave turbulence (DWT)/sheared zonal flow (ZF) system in a cylindrical plasma device using a combination of Langmuir probe and fast-framing imaging measurements are reported. The results show the presence of an azimuthal ZF that exhibits low frequency (approx250 Hz) fluctuations. The envelope of the higher frequency (above 5 kHz) floating potential fluctuations associated with the DWT, the density gradient, and the turbulent radial particle flux are all modulated out of phase with the strength of the ZF. The divergence of the turbulent Reynolds stress is also modulated at the same slow time scale in a phase-coherent manner consistent with a turbulent-driven shear flow sustained against the collisional and viscous damping. The radial turbulence correlation length and cross-field particle transport are reduced during periods of strong flow shear. The results are qualitatively consistent with theoretical expectations for coupled DWT-ZF dynamics.
Frequency-resolved nonlinear internal and kinetic energy transfer rates have been measured in the Controlled Shear Decorrelation Experiment (CSDX) linear plasma device using a recently developed technique [Xu et al., Phys. Plasmas 16, 042312 (2009)]. The results clearly show a net kinetic energy transfer into the zonal flow frequency region, consistent with previous time-domain observations of turbulence-driven shear flows [Tynan et al., Plasma Phys. Controlled Fusion 48, S51 (2006)]. The experimentally measured dispersion relation has been used to map the frequency-resolved energy transfer rates into the wave number domain, which shows that the shear flow drive comes from midrange (k{sub t}hetarho{sub S}>0.3) drift fluctuations, and the strongest flow drive comes from k{sub t}hetarho{sub S}approx =1 fluctuations. Linear growth rates have been inferred from a linearized Hasegawa-Wakatani model [Hasegawa et al., Phys. Fluids 22, 2122 (1979)], which indicates that the m=0 mode is linearly stable and the m=1-10 modes (corresponding to k{sub t}hetarho{sub S}>0.3) are linearly unstable for the n=1 and n=2 radial eigenmodes. This is consistent with our energy transfer measurements.
An azimuthally symmetric radially sheared azimuthal flow is driven by a nondiffusive, or residual, turbulent stress localized to a narrow annular region at the boundary of a cylindrical magnetized helicon plasma device. A no-slip condition, imposed by ion-neutral flow damping outside the annular region, combined with a diffusive stress arising from turbulent and collisional viscous damping in the central plasma region, leads to net plasma rotation in the absence of momentum input.
The characteristics and scaling properties of the turbulence driven shear flow are investigated in a cylindrical laboratory plasma device. For a given plasma pressure, the density fluctuation amplitude and radial particle flux increase with the applied magnetic field. Strong flow shear is found to coexist at high magnetic fields (>700 G) with approx10 kHz drift wave turbulence, but not at low magnetic fields (<700 G). The absolute value of the divergence of the turbulent Reynolds stress at the shear layer is shown to increase with the magnetic field as well. For a fixed magnetic field, the shear flow is found to decrease as the discharge gas pressure is increased. The density fluctuation amplitude and divergence of the turbulent Reynolds stress also decrease with the plasma pressure. For both situations the cross phase between the radial and azimuthal components of the velocity is found to be a key factor to determine variations in the turbulent Reynolds stress at different magnetic fields and discharge pressures. The results show that the generation of the shear flow is related to the development of specific frequency components of the drift wave turbulence for a variety of plasma conditions. The linear stability analysis shows that the observed variation in the turbulence and shear flow with magnetic field is also consistent with a critical gradient behavior.
Tokamak transport modeling scenarios, including ITER [ITER Physics Basis Editors, Nucl. Fusion 39, 2137 (1999)] performance predictions, are based exclusively on reduced models for core thermal and particle transport. The reason for this is simple: computational cost. A typical modeling scenario may require the evaluation of thousands of individual transport fluxes (local transport models calculate the energy and particle fluxes across a specified flux surface given fixed profiles). Despite continuous advances in direct gyrokinetic simulation, the cost of an individual simulation remains so high that direct gyrokinetic transport calculations have been avoided. By developing a steady-state iteration scheme suitable for direct gyrokinetic and neoclassical simulations, we can now compute steady-state temperature profiles for DIII-D [J. L. Luxon, Nucl. Fusion 42, 614 (2002)] plasmas given known plasma sources. The new code, TGYRO, encapsulates the GYRO[J. Candy and R. E. Waltz, J. Comput. Phys. 186, 545 (2003)] code, for turbulent transport, and the NEO[E. A. Belli and J. Candy, Plasma Phys. Controlled Fusion 50, 095010 (2008)] code, for kinetic neoclassical transport. Results for DIII-D L-mode discharge 128913 are given, with computational and experimental results consistent in the region 0{<=}r/a{<=}0.8.
The deployment of multiple high-resolution, spatially localized fluctuation diagnostics on the DIII-D tokamak [J. L. Luxon, Nucl. Fusion 42, 614 (2002)] opens the door to a new level of core turbulence model validation. Toward this end, the implementation of synthetic diagnostics that model physical beam emission spectroscopy and correlation electron cyclotron emission diagnostics is presented. Initial results from their applications to local gyrokinetic simulations of two locations in a DIII-D L-mode discharge performed with the GYRO code [J. Candy and R. E. Waltz, J. Comput. Phys. 186, 545 (2003)] are also discussed. At normalized toroidal flux {rho}=0.5, we find very good agreement between experiment and simulation in both the energy flows and fluctuation levels measured by both diagnostics. However, at {rho}=0.75, GYRO underpredicts the observed energy flows by roughly a factor of 7, with rms fluctuation levels underpredicted by a factor of 3. Interestingly, at both locations we find good agreement in the shapes of the radial and vertical density correlation functions and in the shapes of the frequency power spectra. At both locations, the attenuation of the GYRO-predicted fluctuations due to the spatial averaging imposed by the diagnostics' spot sizes is significant, and its incorporation via the use of synthetic diagnostics is shown to be essential for quantitative comparisons such as these.
A method for measuring nonlinear energy transfer in the frequency domain using a two-field model of drift turbulence is proposed, and the theoretical motivation and experimental results are presented. The approach is based on the cross-bispectral analysis of quadratic nonlinearities in the turbulent internal and kinetic energy balance equations directly derived from the fluid plasma continuity and momentum equations. Application of the technique to data from a laboratory plasma experiment reveals the nonlinear energy transfer in weak collisional plasma drift turbulence; the results show a transfer of density fluctuation energy toward higher frequency (which correspond to smaller azimuthal spatial scales) and a transfer of kinetic energy to lower frequencies (corresponding to larger azimuthal scales)
A new procedure for calculating model coefficients from fluctuation data for fully developed turbulence is derived. This procedure differs from previous related methods in that it is constructed in a spatial rather than spectral representation. This has a number of advantages, such as reduced data set requirements, ability to represent spatially inhomogeneous systems such as the ones with curvature or zonal flows, and ability to use data from experimental diagnostics with limited spatial resolution. In this method, the model equation is represented as a linear superposition of linear and nonlinear differential operators. The coefficients of this superposition are calculated using a least-squares method. This method has been tested on simulations of fully developed two dimensional turbulence and compared to previous metho0008.