All opinions expressed herein are those of the authors and should not be reproduced, quoted in publications, or
used as a reference without the author’s consent.
Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the U.S. Department of Energy.
Published by Fusion Energy Division, Oak Ridge National Laboratory
Building CR 5600 P.O. Box 2008 Oak Ridge, TN 37831-6169, USA
Editor: James A. Rome Issue 135 February 2012
E-Mail: jamesrome@gmail.com Phone (865) 482-5643
On the Web at http://www.ornl.gov/sci/fed/stelnews
Spatiotemporal structure of
the turbulence-flow interaction
at the L-H transition in TJ-II
Introduction
The H-mode confinement regime has been extensively
studied since its discovery in ASDEX [1]; however, the
physical mechanism triggering the H-mode has still not
been clearly identified. Bifurcation theory models, based
on the coupling between turbulence and flows, describe
the L-H transition as passing through an intermediate,
oscillatory transient stage with a characteristic predatorprey
relationship between turbulence and zonal flows [2].
This intermediate oscillatory transient stage has been seen
in the L-H transition experiments in TJ-II [3] as well as in
other devices. In these experiments, as in the predator-prey
theory model [2], only the temporal dynamics of the turbulence-
flow interaction is studied.
However, as recently pointed out [4], the spatial evolution
should also be taken into account as a necessary step in
developing a model for the L-H transition model.
The present work, a summary of the results recently
reported in Ref. [5], addresses this fundamental issue from
an experimental point of view for the first time.
Experimental Results
The experiments have been carried out in the TJ-II stellarator.
A two-channel Doppler reflectometer is used to measure
the radial electric field Er and density fluctuations, ñe
at two radial positions simultaneously with good spatial
and temporal resolution [6]. As reported in Ref. [3], close
to the L-H transition threshold pronounced oscillations in
both Er and ñe are measured at the plasma edge, just inside
the Er shear position. Depending on the NBI heating
power and magnetic configuration, the oscillations can
endure throughout the NBI heating phase of the discharge.
Simultaneous with the onset of the oscillations, there is a
drop in the H signals and an increase in diamagnetic
energy, Wdia, and ne. The oscillations appear as changes in
the intensity and frequency of the Doppler reflectometry
spectra and show a predator-prey relationship between turbulence
and flows, with the flow—the predator—following
the turbulence—the prey—with a phase delay of 90º in
a limit-cycle way [3].
The radial profile of Er, shown in Fig. 1, changes from
rather flat in L-mode to sheared during the oscillating
phase. The Er oscillation amplitude is about 1 kV/m close
to the Er shear and increases gradually as inner radial positions
are probed. As a consequence, the Er well of about
10 kV/m measured at the maximum of the oscillations
shrinks in each limit-cycle and an inner shear layer develops
at  0.75 (see blue profile in Fig. 1).
In this issue . . .
Spatiotemporal structure of the turbulence-flow
interaction at the L-H transition in TJ-II
The spatiotemporal structure of the turbulence-flow
interaction has been measured at the L-H transition in
TJ-II plasmas. The temporal dynamics of the interaction
displays an oscillatory behavior with a characteristic
predator-prey relationship. The spatiotemporal
evolution of this oscillation pattern shows both radial
outward and inward propagation velocities of the turbulence-
flow front. The results indicate that the edge
shear flow linked to the L-H transition can behave
either as a slowing-down, damping mechanism of outward
propagating turbulent-flow oscillating structures,
or as a source of inward propagating turbulence-flow
events.......................................................................1
W7-X status report
The five torus modules are in place, and two modules
have their ports installed. Interior component installation
and connections between each torus module are
under way. ................................................................4
Stellarator News -2- February 2012
The measurement of the radial propagation characteristics
of the oscillation pattern allows extension of the previous
temporal characterization to a spatiotemporal one. An
example is shown in Fig. 2, which displays the time evolution
of both (a), rms(ñe) and (b) Er measured simultaneously
at  = 0.8 and  = 0.75. In addition, the time
evolution of Er shear is shown in Fig. 2(c).
The relation between Er shear and rms(ñe) showing a
limit-cycle behavior is shown in Fig. 2(d). Pronounced
changes in Er shear appear linked to the oscillations in
rms(ñe). A delay between the two channels can be seen,
indicating a radial propagation from the inner to the outer
channel. This outward propagation is found at line densities
between 2–2.5 1019 m3. At densities above
3 1019 m3, the propagation direction in some cases
reverses after a short time without oscillations. Analysis of
the delay yields propagation velocities within the range
50–200 m/s with a radial trend as shown in Fig. 3. In this
figure the vertical bars represent the error in the estimation
of the propagation velocity and the horizontal ones represent
the radial separation between the two radial channels.
The radial propagation velocity decreases as the turbulence-
flow front approaches the Er shear position. The
inward propagation velocity has similar values but no
clear radial dependence can be inferred. The extreme values
of the Er oscillations are comparable in both cases but
the time evolution shows differences. As the turbulenceflow
front propagates outward, the increase in the turbulence
level produces an increase in the inner shear; however,
a turbulence-flow front propagating inward produces
an increase in the outer shear.
Fig. 1. Er profiles in L-mode (black) and during the intermediate
oscillatory phase: Er maxima (red) and minima (blue)
measured at ne ~ 2.0–2.5  1019 m3.
Fig. 2. The time evolution of (a) rms(ñe) and (b) Er measured
at = 0.8 (blue) and = 0.75 (red). (c) The time evolution
of Er shear and (d) relation between Er shear and
rms(ñe).
-400
0
400
-30 -25 -20
174.5 - 175.5 ms
dE
r
/dr (kV/m
2
)
S (dB)
d)
-30
-20
-10
ch1
ch2
S (dB)
#27137 a)
-12
-9
-6
-3
E
r
(kV/m)
b)
-500
-250
0
250
500
172 174 176 178
dE
r
/dr (kV/m
2
)
time (ms)
c)
Stellarator News -3- February 2012
Summary and discussion
The spatiotemporal evolution of the oscillation pattern
could be linked to the radial spreading of plasma turbulence
from the plasma core to the edge barrier. Previous
experiments performed in TJ-II have shown signatures of
radial spreading of the turbulence as the plasma
approaches the H-L back transition condition [7]. Similarly,
in the present experiments, as the turbulence propagates
towards the barrier, the associated turbulence-driven
flow generates the inner shear layer, which in turn regulates
the turbulence level. The present observation could
be also understood in terms of turbulent bursts propagating
toward the plasma edge. These turbulent bursts could be
generated in the plasma interior by instabilities linked, for
instance, to the magnetic topology. To explain the present
experimental observations, each turbulent burst should be
accompanied on its way to the plasma edge by a shearedflow
layer. The deceleration in the oscillation-pattern
propagation as it approaches the edge shear layer together
with its absence at outer radial positions suggest an
absorption process at the Er shear layer. In this process, the
turbulence-flow events generate a dual shear layer and
thus enhance the formation of the Er well.
A reversal in the front propagation velocity is observed in
some cases at the final stage of the discharge after a quiet
period without oscillations. In those cases the oscillation
pattern arises at the outer shear layer position and propagates
towards the plasma interior.
The present experimental results indicate that the edge
shear flow linked to the L-H transition can behave either
as a slowing-down, damping mechanism of outward propagating
turbulent-flow oscillating structures, or as a source
of inward propagating turbulence-flow events.
The reported results show the need to approach L-H transition
studies within a 1-D spatiotemporal framework.
T. Estrada and the TJ-II team
E-mail: teresa.estrada@ciemat.es
Phone: 34 91 346 0845
Laboratorio Nacional de Fusión. As. Euratom-CIEMAT, 28040
Madrid, Spain
References
[1] F. Wagner et al., Phys. Rev. Lett. 49, 1408 (1982).
[2] E.-J. Kim and P.H. Diamond. Phys. Plasmas 10, 1698
(2003).
[3] T. Estrada et al., Europhys.. Lett. 92, 35001 (2010).
[4] P.H. Diamond et al., Plasma Phys. Control. Fusion 53,
124001 (2011).
[5] T. Estrada et al., Phys. Rev. Lett. 107, 245004 (2011).
[6] T. Happel et al., Rev. Sci. Instrum. 80, 073502 (2009).
[7] T. Estrada et al., Nucl. Fusion 51, 032001 (2011).
Fig. 3. Radial propagation velocity of the turbulence-flow
front. The striped area indicates the Er-shear position.
Stellarator News -4- February 2012
W7-X status report
W7-X torus completed
On 16 November 2011, the last of the five field-period
modules that comprise the Wendelstein 7-X (W7-X) stellarator
in Greifswald, Germany, was placed on its foundation
with millimeter precision, as shown in Fig. 1. The
entire procedure required only 3 hours although the assembly
team had expected a considerably longer process
because for the first time, it was necessary to avoid collision
at both ends of the module. As little as 8 mm of clearance
was available—often at several points
simultaneously—in maneuvering the 120 tonne module
into position.
In the coming months, each module will be connected to
its two neighbors. The separate cryopiping, instrumentation,
and bus systems of the individual modules will then
be joined using superconducting joints where necessary.
The sections of the central support ring will be bolted
together, the thermal insulation joined at the seams, and
the plasma and external vessels joined by welding.
At the beginning of 2012, assembly of the in-vessel components
will commence by cladding the plasma vessel
with stainless steel cooling panels. Regions that will be
subjected to high thermal loads must be protected with
carbon tiles.
Ports provide access to the plasma
Before assembly of the numerous in-vessel components
can begin, however, it is necessary to install the ports that
will provide the links between the interior and exterior of
the plasma vessel.
There is a total of 254 vacuum-tight ports, some as long as
3 m. Roughly half of the ports are devoted to diagnostic
purposes; the remainder provide access for plasma heating
systems and for the vacuum pumps. Also accommodated
by the ports is the water-filled piping, which is used for
cooling the vacuum vessel. See Fig. 2.
The shapes of the port openings vary from circular with a
diameter of 150 mm to approximately rectangular with
dimensions of 1000  4000 mm2. The largest ports are
reserved for microwave and neutral-beam heating systems
and to provide physical access to the torus for maintenance.
Placement of the ports is done from outside the torus. This
involves insertion through the opening in the external vessel
and maneuvering through the cryo chamber to the target
position on the plasma vessel, after which both ends
are welded into place.
To accommodate movement of the plasma vessel during
experimental operation of W7-X, connection of the ports
to the external vessel makes use of bellows. Because the
ports must pass through the extreme cold of the cryo
chamber, it is mandatory that they be carefully insulated.
Fig. 1. Wendelstein 7-X on 16 November 2011.
Fig. 2. Real and virtual perspectives: the photo on the left shows a module without ports in the experimental hall. On the
right is a CAD view of the same region (with external vessel removed) with the ports highlighted in red.
Stellarator News -5- February 2012
In accordance with the project timetable, port assembly
has been completed for three of the five modules. This was
achieved in spite of the technical challenges faced during
the process. Indeed, ports weighing as much as a tonne
have been maneuvered and welded in place with the
require millimeter precision.
Determination of correct lengths and the complicated
curves along which the ports must be cut to shape was also
technically demanding. The initial approach, which
involved provisional placement of each port and subsequent
cutting to achieve a satisfactory fit, has now been
replaced with a three-dimensional (3-D) construction technique
(Fig. 3) which accounts not only for the plasma vessel
shape and the port orientation, but also distortion that
can be expected during the welding process. Once the port
has been cut to the correct shape it must be thermally insulated.
To prevent subsequent collisions in construction and
in the operation of W7-X, the insulation must also be
applied with millimeter accuracy.
Zero tolerance—engineering pushed to its limits
A handful of ports pose special challenges as they require
accuracy at the level of technical feasibility. Although in
most cases tolerances of a few millimeters are allowed,
these special ports have tolerances that are essentially
zero. The reasons for this include maximization of the port
cross sections and the necessity of avoiding collisions with
other components.
An example is the port for neutral particle heating. Reduction
of the port cross section by as little as 1 mm would
significantly increase the number of fast particles that
never reach the plasma and instead deposit their energy
onto the port. An increase in port dimensions is impossible
in this case due to the surrounding components, including
the magnetic field coils. To improve such situations an
iterative design and test procedure is used in which the orientation
and welding properties of prototype ports are
investigated with current results incorporated into each
subsequent design.
This approach requires teamwork of the highest level as it
represents an interplay of scientific calculations, modern
CAD modeling, accurate 3-D measurements, cutting-edge
engineering and a fine human touch.
Prof. Dr. Robert Wolf and Dr. Andreas Dinklage for the W7-X
Team
E-mail: w7xnewsletter@ipp.mpg.de
Max-Planck-Institut für Plasmaphysik
Association Euratom – IPP
Wendelsteinstraße 1
Germany - 17491 Greifswald
Fig. 3. For the accurate placement of the ports it was necessary
to develop a carriage system encompassing 5
degrees of freedom at heights of up to 12 m above the floor
of the experimental hall.