Since 18 of December 2019 uses Nucleus credentials. Visit our help pages for information on how to Register and Sign-in using Nucleus.
10-15 May 2021
Nice, France
Europe/Vienna timezone
The Conference will be held virtually from 10-15 May 2021

Progress from ASDEX Upgrade experiments in preparing the physical basis of ITER operation and DEMO scenario development

10 May 2021, 14:25
Nice, France

Nice, France

Overview Oral Overview OV/2 Overview Magnetic Fusion


Prof. Ulrich Stroth (DeMPIPGarc)


Plasmas in the ASDEX Upgrade (AUG) tokamak can match a large number of fusion
relevant parameters simultaneously. With a tungsten wall and ITER-like
magnetic and divertor geometries, high values of the plasma $\beta$, the
normalized confinement time, Greenwald fraction, and power densities $P/R$
are reached under detached divertor conditions. The synopsis first addresses
the integration of a detached divertor into improved confinement regimes
while avoiding large ELMs. Secondly, it summarises the work relating to core
confinement and stability, and to the physical understanding required for
modelling ITER and DEMO plasmas.

Small or no ELM regimes have in common, that the H-mode transport barrier is
modified by weakly or quasi coherent modes or changes in turbulence regime
such that the peeling-ballooning (P-B) limit is not reached:

(i) The \emph{I-mode} has a number of attractive features with regard to a
reactor plasma. The characteristic weakly coherent mode is linked to bursty
transport and divertor heat loads which are, according to recent infra-red
measurements, smaller than those of ELMs but could still be a threat for the
targets [1]. Making use of AUG's flexible heating systems, realtime $\beta$
control helped to develop stationary I-mode phases with an H-factor of about
0.9. Gyro-fluid simulations indicate that the L-I transition is caused by the
stabilisation of ITG turbulence [2]. From the simulations a larger I-mode
operation window at higher $B$ field and problems in combining it with a
detached divertor would be expected.

(ii) The plasma edge of the recently discovered \emph{stationary ELM-free
H-mode} [3] is similar to Alcator C-Mod's \emph{EDA H-mode}. It avoids ELMs
by residing close to the ballooning but far from the peeling limit. The
regime is favoured by higher triangularity; it has an H-mode like pedestal,
an H-factor of above 1 and appears at high density. The transition to an ELMy
H-mode at higher heating power could be avoided by introducing radiative edge
cooling by argon seeding for powers up to 5\:MW. In both regimes, (i) and
(ii), a mode is made responsible for transport limiting the pressure gradient
and avoiding impurity accumulation.

(iii) In a similar way, but as a \emph{high-power L-mode}, a new scenario is
being developed, where radiative losses from argon in the pedestal region
keep the power flux through the separatrix below the L-H threshold value.
H-factors of 0.9--1 and a $\beta_N \approx 1.2$ were reached [4]. The core
energy increases with power; this leads to a growing H-factor in the
parameter range achieved by this high power L-mode scenario. The edge has
similarities to that in I-mode, with pedestals in electron and ion
temperatures and only a weak one in density. The divertor temperature drops
to low values and compatibility with detachment can be expected.

(iv) The active suppression of ELMs by eroding the density pedestal by means
of \emph{resonant magnetic perturbations} (RMP) is investigated in low
collisionality discharges [5]. Full suppression of ELMs is accompanied by the
onset of quasi-coherent fluctuations, radially and toroidally localised in
the pedestal. ELM suppression is maintained in a large range of heating
powers, which can be understood by a threshold behaviour of the
transport-inducing mode. These observations solve a problem of previous
models, which invoke classical radial diffusion around magnetic islands or in
an ergodised region and therefore predict a dependence of access to ELM
suppression on the edge heat flux.

(v) A \emph{H-mode regime with small ELMs} develops when the separatrix
pressure and local shear approach the ballooning limit. Small and for the
divertor benign pressure gradient relaxations modify the pedestal in the
vicinity of the separatrix, where the dimensionless parameters are DEMO-like,
leading to a P-B stable edge. At high triangularity this is the most
promising scenario at AUG to integrate high performance plasmas with
protection of the divertor even against transiently unacceptable heat loads.
When approaching the H-mode density limit a transition from drift-wave to
interchange turbulence occurs in the vicinity of the separatrix [7]. This
transition can also be caused by intense radiation losses from above the
X-point (\emph{X-point radiator}). The location of the X-point radiator can
now be actively controlled via realtime AXUV measurements and the nitrogen
seeding rate as actuator [9]. Based on this ITER-relevant scenario, a
discharge was developed without any type-I ELM and a divertor temperature
below 8 eV throughout. With 14 MW total heating power, flattop values with
H-factors of 0.9 and $\beta_N \approx 2.0$ were reached [6]. Density limit
disruptions were avoided by active control.

Where parameters of ITER or reactor plasmas cannot be met in present
tokamaks, physics models are developed to predict the performance. The
progress in integrated modelling provides increasingly validated physics
elements to be included in the new AUG flight simulator [10]. With only
global and engineering parameters as input, an integrated transport model was
able to reproduce AUG discharges without input from experimental profiles.
For this, the ASTRA code was used with a new pedestal model, that allows
simultaneous development of the kinetic profiles of core and pedestal, and a
simple SOL model, setting the boundary conditions [11]. For reactor
projections, discharges aiming at reaching reactor-relevant core transport
properties were analysed with the theory-based turbulence model TGLF. It was
shown that density peaking is mainly sustained by turbulence, where
electromagnetic effects are relevant, while the fueling profile only plays a
minor role. Because of the strong link between electron temperature and
density, steepening the electron temperature gradient in the confinement
region seems the only meaningful way to increase density peaking in a reactor

The prediction of the L-H power threshold for ITER is an important issue. In
contrast to recent observations at JET, the threshold in H plasmas did not
change when the concentration of helium was increased up to 20\:\%. According
to power balance analyses, the ion heat flux through the edge at the L-H
transition is independent of the helium concentration [13], being consistent
with the finding that neoclassical \exb\ shearing rate triggers the
transition. The impact of the isotope mass has been investigated by a new
experimental approach, which, by an increase of plasma triangularity in
hydrogen, allows core and edge effects to be consistently separated [14].
Nonlinear gyrokinetic simulations have revealed that edge turbulence in
L-mode is dominated by electron drift waves, strongly destabilized by
collisionality, stabilized by an increase of isotope mass and influenced by
electromagnetic effects, providing predicted heat fluxes which are
significantly larger in hydrogen than in deuterium, consistent with
observations [15,14].

A fusion reactor would benefit from advanced plasma scenarios. Even tiny
error fields can grow close to MHD limits, constraining $\beta_N$. CAFÉ
calculations showed that the correction of the AUG (2,1) and (3,1) field
errors can improve the achievable $\beta_N$ from 3 to 3.2--3.3 [16]. Elevated
core $q$-profiles are instrumental for advanced scenarios. IMSE measurements
of the core current profile in discharges with strong ECCD confirmed the
predicted beneficial radial current outward transport, introduced by an (1,1)
mode, as well as the threshold behavior [17]. Finally, the effect of fast
ions on core transport was studied by varying the rotational shear at
constant $T_e/T_i$ ratio. Thus the improvement of core ion confinement could
be attributed to the fast ion content while rotational shear turned out to
have little impact on it [18].

[1] Silvagni tbp [2] Manz tbs [3] Gil FEC [4] Fable tbp [5] Leuthold PhD [6] Faitsch FEC [7] Eich tbs [9] Bernert FEC [10] Fable FEC [11] Tardini FEC [12] Fable NF 2019 [13] Plank NF tbs [14] Schneider FEC [15] Bonanomi NF 2019 [17] Burckhart FEC [18] Stober FEC

Country or International Organization Germany
Affiliation MPI for Plasma Physics

Primary authors

Presentation Materials

There are no materials yet.