28th IAEA Fusion Energy Conference (FEC 2020)

High-power ECCD based high-performance H-modes with $q_{min} > 1$ in ASDEX Upgrade

12 May 2021, 14:00

4h 45m

Nice, France

Regular Poster Magnetic Fusion Experiments P4 Posters 4

Speaker

Joerg Stober (IPP Garching)

Description

This contribution focuses on the physics of high-beta plasmas used for advanced tokamak operation. Possible mechanisms of ITG stabilization, i.e. ExB-shear and electro-magnetic effects of high beta are separated experimentally. Sensible physics studies require $q$-profiles that are constant over several current diffusion times. In such conditions $q_{min}$ and $\rho(q_{min})$ can be enlarged using a novel technique combining central ctr-ECCD with ohmic current drive into a net off-axis co-current drive source.

Steady-state operation of a tokamak reactor is inevitably dependent on with a high fraction of bootstrap current related to high values of $\beta_{pol}$. At the same time the plasma current has to be high enough to guarantee sufficient energy confinement at the necessary density for an efficient burn. Increasing the current typically leads to reduced $\beta_{pol}$, since the stability limiting parameter is $\beta_{N}\propto\beta_{pol}*I_p$ for fixed $B_t$. The necessary parameters depend on assumptions made on technology or plasma physics advances.
Improvements in technology of super-conducting magnets may allow the connection between $\beta_{pol}$ and $I_p$ to be broken by increasing $B_t$ and thus $q_{95}$, higher plasma performance (H-factor) may allow $\tau_E$ to be raised without raising $I_p$, provided that also the density can be increased above the typical density limits which are proportional to $I_p$. Broadening the current profile towards the edge (i.e. increasing $q_{min}$) increases $\beta_{pol}$ in the central plasma and may also give rise to radial regions with reduced ion transport close to $\rho(q_{min})$, which then further increase the bootstrap current. A moderate variant of such a reactor design is the European DEMO design $[1]$, which requires $q_{95}=4.5$, $B_t=5.7$~T, $f_{BS}=0.62$, $\beta_N=3.4$, $n/n_{GW}=1.2$, H-factor=1.2. Though several, typically time evolving experimental examples of such scenarios have existed for decades, the theoretical prediction of the steady-state reactor performance based on the existing models
is rather uncertain especially for the effects of inverted q-profiles, also due to the partially heuristic nature of the related models.
ASDEX Upgrade (AUG) is particularly suited for such experiments, since it has a flat-top duration of 10~s, which is sufficiently long compared to current diffusion time of $<$~2~s, combined with strong and versatile NBI and wave heating systems and a broad set of state-of-the-art diagnostics.AUG discharges are compared to predictions with the trapped gyro-Landau-fluid code TGLF $[2]$ and local analysis with the gyro-kinetic code GENE in order to test and improve the degree of physics understanding, which is necessary for the extrapolation to larger devices.
First this contribution discusses mechanisms of steepening the central $T_i$-profile, which is a challenge for centrally dominant electron heating due to the negative effect of large ratios of $T_e/T_i$ on ITG dominated ion transport. While TGLF reproduces such steepening for strongly NBI-heated plasmas due to the strong effect of ExB shearing in this model [3], GENE does not show this strong ExB-shear dependence and relates the reduced ion transport to electro-magnetic effects driven by beta of fast ions (as found for JET [4]), but also by thermal beta [5]. These contradicting theoretical explanations have been checked experimentally replacing centrally absorbed NBI sources by ICRH, in order to vary the rotational shear, but keeping $T_e/T_i$ constant. The results are shown in figure~1 and show for constant beta that significant changes in the ExB shear only lead to small variations in $\nabla T_i/ T_i$. Indeed, while TGLF yields convincing results for the case without ICRF, it fails to reproduce the peaked central $T_i$ for the case with less shear. Actual work in progress is to get more GENE runs for a wider range of parameters and to implement a stabilizing mechanism in TGLF which shall be parametrized such that it fits the new GENE data set. More details on this approach can be found in [6].

The second part of the contribution reports on the experimental status combining broad central ctr-ECCD (where it is most effective) with ohmic current drive to study the physics of elevated and inverted current profiles at high beta. Figure 2 shows the evolution of a plasma with $q_{95}=4$, starting with a relaxed $q$-profile after an ohmic ramp-up (from [7], where also more details can be found). Note that during this discharge $q_{min}$ rises from 3s to 5s at constant beta. This carries the potential of achieving steady-state scenarios with real-time current-profile control acting on ECCD. Further optimization of the development includes the broadening of the ctr-ECCD profile enlarging $\rho(q_{min})$ as well as increasing $T_i(0)$ to 10~keV. $\beta_{pol}$ will be further increased by lowering the current, increasing $q_{95}$ from 4.0 (fig.\~2) to 4.5 as envisaged for a steady state DEMO. In parallel, the modifications to TGLF for high beta will be tested on these steep $T_i$- and $T_e$-profiles.

$[1]$ Zohm, H. et al., NF 57 (2017) 086002
$[2]$ Staebler, G.M. et al., PoP 23 (2016) 062518
$[3]$ Jian, X. et al., NF 59 (2019) 106038
$[4]$ Doerk, H. et al., NF 58 (2018) 016044
$[5]$ Citrin, J. et al., PPCF 57 (2015) 014032
$[6]$ Reisner, M. et al, H-mode WS 2019, submitted to NF
$[7]$ Stober, J. et al., PPCF 62 (2020) 024012