Transcript 1.09_Martin

RFX/RFP mode control issues
Piero Martin & Sergio Ortolani
Consorzio RFX
Associazione Euratom-ENEA sulla fusione
Padova, Italy
Presented by P. Martin
at the 2003 workshop on “active control of MHD stability:
extension to the burning plasma regime”
University of Texas-Austin
Nov. 3-5, 2003
CONTRIBUTORS

L. Marrelli, G. Spizzo, P. Franz, P. Piovesan, I.
Predebon, T.Bolzonella, S. Cappello, A. Cravotta, D.F.
Escande, L. Frassinetti, S. Martini, R. Paccagnella,
D. Terranova and the RFX team
And fruitful collaborations with the



MST team: B.E. Chapman, D. Craig, S.C. Prager, J.S.
Sarff
EXTRAP T2R team: P. Brunsell, J.-A. Malmberg, J.
Drake
TPE-RX team: Y. Yagi, H. Koguchi, Y. Hirano
Hopefully this is the last workshop…



…. without RFX !
…. RFX reconstruction is in the final phase
and first plasmas are expected in Sept. 2004
The new RFX will be a “state of the art” MHD
MODE CONTROL FACILITY:

192 ACTIVE COILS, INDEPENDENTLY DRIVEN,
COVERING THE WHOLE PLASMA SURFACE
Outline of the talk
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
The new RFX device
Main new components:
1. 192 saddle coils, covering the whole plasma
boundary, each independently powered and
feedback controlled
2. a smoother and thinner shell
3. the first wall with higher power handling
capabilities
4. an in-vessel system of magnetic and electrostatic
probes
5. the toroidal field power supply
R /a = 2.0 / 0.46
(m) - I up to 2 MA
Overview of the magnetic boundary
Active
saddle
coils
vessel
shell
- 4 coils in the poloidal direction: 90° spaced
– 48 coils in the toroidal direction: 7.5° spaced
Saddle coil performance
each independently powered
24 kAt: 400 A x 60 turns
Wide spectrum of Fourier
components can be produced:
•m=1,2
•n ≤ 24
•DC < f < 100 Hz
Significant amplitude available.
For example: edge br for (1,8)
mode:
•
•
20 mT@10 Hz
1.3 mT @100 Hz
The new RFX shell
• 3 mm Cu layer
• 50 ms time constant
(450 ms before)
• High reduction of gap
field error
• 1 overlapped poloidal
gap
• 1 toroidal gap on high
field side.
Integrated System of the Internal Sensors (ISIS)
•97 Electrostatic (Langmuir) probes
•139 Magnetic pick up probes
•8 Calorimetric probes
Langmuir probe
Magnetic probe
Test of RFX mode control equipment in T2R
16 coils
Partial coverage


Some RFX digital controllers have been moved to Stockholm to
be tested in an “intelligent shell” experiment, which will be
done in the EXTRAP T2R device as a joint T2R-RFX
collaboration.
Preliminary results with analog controllers developed by KTH
positive!
MORE
IN JIM DRAKE’S TALK WEDNESDAY!
Outline of the talk
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
What shall we use the new RFX for ?


To explore the RFP physics and to optimize the RFP
confinement performance in a steady fashion in the
MA current range
To contribute to the worldwide program on MHD
modes control in fusion devices
What do we use the new RFX for ?

To explore the RFP physics and to optimize the RFP
confinement performance in a steady fashion in the
MA current range

Control of MHD modes:


m=1 “dynamo” modes (resonant inside the Bt reversal surface)
m=0 non-linearly generated and/or linearly unstable
The standard RFP has many of these modes with 1% B amplitude
simultaneously present, and they spoil confinement!

RWM when the shell is resistive (reactor relevance)
The modes we have to deal with
m=0

various n: resonant at the Bt
reversal surface
m=1



|n| ≥ 2R/a, resonant inside the Bt
reversal surface (resistive kink,
“dynamo modes”)
|n| ≤ 2R/a, internally non resonant
from above (RWM, with the same
helicity as the “dynamo” modes and
the same handedness as the core B)
|n| ≤ R/a, externally non resonant
(RWM, with opposite helicity)
The RFP dynamo issue


The electrical currents flowing in a RFP can not be directly
driven by the inductive electric field Eo
….but RFP plasmas last for times much longer than the resistive
diffusion time ! (actually, as long as Eo is applied)
The RFP dynamo: E + vxB = hJ



An additional electric field, besides that externally
applied, is necessary to sustain and amplify the
toroidal magnetic flux.
A Lorentz contribution vxB is necessary, which
implies the existence of a self-organized velocity
field in the plasma.
The origin of this contribution is the classical RFP
dynamo problem
Turbulent dynamo: remarkable self-organization

A wide experimental and numerical database supports the MHD
turbulent dynamo theory:

~
~
Ed  v  b
2
-2
log b
1n
-4
the dynamo electric field is
produced by the coherent (and
non-linear) interaction of a large
number of MHD modes:
-6
-8
1000
3000
5000
t/
7000
A
Multiple Helicity (MH) dynamo
The standard Multiple Helicity (MH) RFP
… and severe plasma-wall
interaction if the modes lock
in phase and to the wall !


m=1 “dynamo” modes (resonant inside the Bt reversal surface)
m=0 non-linearly generated and/or linearly unstable
Magnetic stochasticity
allover the plasma !
The strategy towards dynamo modes

Keep them rotating in the lab frame



Make their amplitude lower



Reduces amplitude br
Optimizes the basic standard target plasma
…but you must provide dynamo electric field from outside
Run the plasma in a regime where resort to dynamo is reduced
Work in a regime where their spatial spectrum is
monochromatic, i.e. dynamo is driven only by ONE INDIVIDUAL
SATURATED MODE
ALL THESE TOPICS MIGHT BE
INFLUENCED BY ACTIVE CONTROL !
Outline of the talk
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
Mode dynamics in RFPs

Previous experimental evidence in several different
devices shows that the evolution of MHD modes,
including the dynamo modes, depends on the magnetic
boundary, and in particular on the shell:



thickness
proximity
geometry
The RFP synopsis
Experiment
R/a (m)
b/a
RFX92
2/0.457
1.24
450
150
1/3
RFX new
2/0.459
1.11
50
~ 150
≥3
MST
1.5/0.51
1.07
400
60-90
¼
TPE RX
1.72/0.45
1.08
1.16
10
330
60
6
1/5
T2R
1.24/0.183
1.08
6
20
>3
shell (ms) pulse (ms)
pulse/shel
TPE-RX spontaneous mode rotation
b1/a = 1.08 thin shell, b2/a = 1.16 thick shell
tshell =10 ms
tshell =330 ms
Spontaneous Rotation in EXTRAP T2R
RWM’s
RWM’s
•Tearing modes rotate
From Malmberg Brunsell PoP 2002
MST modes spontaneous rotation

….listen Brett Chapman’s invited talk !
Conclusions on mode rotations





Mode rotation is beneficial and depends on magnetic boundary
Modes were locked to the wall in the old RFX and this lead to a
serious deterioration of performance
Slow rotation of modes was actively driven in RFX ( Bartiromo et
al, PRL 99)
Dynamo modes are spontaneously rotating in RFP devices with
boundary conditions similar to the new RFX.
There is a reasonable basis to hope for spontaneous mode
rotation in the new RFX (even if a threshold in current might
exist)
Outline of the talk
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
Magnetic chaos is not intrinsic to RFP
2
-2
log b
1n

~
~
Ed  v  b
-4
-6
DYNAMO CAN BE PRODUCED BY A
SINGLE MHD MODE
-8
1000
3000
5000
t/
7000
A
The Single Helicity (SH) dynamo


a theoretically predicted state with a unique m = 1 saturated resistive
kink (a pure helix wound on a torus),
Stationary LAMINAR dynamo mechanism with good helical flux
surfaces
Escande et al., PRL 85 (2000)
Magnetic order with SH dynamo
Good magnetic flux surfaces in SH
Overlapping of
many modes !
SH
Turbulent (MH)
Helical states in the experiment

Quasi Single Helicity (QSH) spectra have been observed in all
RFP devices, under a variety of boundary conditions (Martin, NF
2003).

The mode spectrum is dominated by one geometrical helicity

The other modes have still non-zero amplitude
Stationary Quasi-Single Helicity

Stationary QSH spectra have been observed in the RFX
device

with a helical coherent structure emerging from magnetic chaos
in the plasma core.
Toroidal mode number n spectrum vs. time
RFX pulse length
Flow velocities measurements in QSH plasmas
Remember:

~
~
Ed  v  b
Plasma flow velocity fluctuations measured in MST with Doppler
spectroscopy (Den Hartog et al., Phys Plasmas 99)

In QSH not only the spectrum of magnetic
fluctuation spectra become narrower in
comparison with MH, but also that of flow
velocity fluctuations
Work due to D. Craig, L. Marrelli, P. Piovesan in MST
Magnetic and Flow velocity fluctuations toroidal spectra
~
b
~
v
Dynamo electric field in QSH

Dynamo in QSH becomes more concentrated in one
mode than in standard MHD plasmas!
~
~
(v  b )
MH vs. QSH vs. SH
QSH and mode wall locking


The access to QSH regime is beneficial for the
problem of modes wall locking.
The dominant (big) mode might be more prone to lock
to the wall (see Brett Chapman’s talk), but…



Non-linear interaction between modes decreases
The “strength” of mode locking decreases
Easier rotation for secondary modes
Spontaneous mode rotation in RFX during QSH
Dominant and Rotating
Mode amplitude
Phase of Rotating Mode
From Bolzonella,Terranova, PPCF 2002
This is consistent with theoretical calculations (Guo-Chu, Fitzpartick,Chapman),
which predict that modes rotate more easily if they are smaller
PWI in QSH is milder anyway !


Vertical
displacement
of the plasma
column in
RFX
MH
QSH
Edge
fluctuating
magnetic
field
Toroidal angle
Toroidal angle
Favorable conditions for QSH active control
1.
There are already plasma regimes where
monochromatic spectra are more easily obtained
spontaneously
a.
b.
2.
At higher plasma current
With shallower reversal
We can also “select” the toroidal mode number n we
wish to be the dominant one.
Mode selection

Note that the pre-programming the magnetic equilibrium
allows to select efficiently the mode that will dominate the
spectrum!
Shallow reversal, m=0 modes and QSH





Shallow reversal brings outwards the
reversal surface, where q=0.
Narrower stochastic region produced
by m =0 when the reversal surface is
closer to the plasma edge.
This provides a smoother plasma
boundary, which helps the onset of
QSH
In the new RFX m=0 modes are not
any more LINEARLY unstable, as
they were in the OLD device.
Positive feedback, since in QSH nonlinear generation of m=0 modes is
strongly reduced.
Numerical studies on active control: successful drive and
sustainment of m=1 n=7 SH state starting from MH conditions
Wr( 0, 1)
Wr( 0, 2)
Wr( 1, 1)
Wr( 1, 2)
Wr( 1, 3)
Wr( 1, 4)
Wr( 1, 5)
Wr( 1, -1)
Wr( 1, -2)
Wr( 1, -3)
Wr( 1, -4)
Wr( 1, -5)
Wr( 1, -6)
Wr( 1, -7)
Wr( 1, -8)
Wr( 1, -9)
Wr( 1,-10)
Wr( 1,-11)
Wr( 1,-12)
Wr( 1,-13)
eps42wf3_4/f5_10
0.001
0.0001
10-5
10-6
•Low dissipation
conditions
•Thin shell

•Simoultaneous active
control of RWMs !
10-7
Remember we can apply with the
saddle coils up to ~20 mT on a single
m=1 mode
10-8
10-9
0.5
1
1.5
Time
2
2.5
Paccagnella, MHD workshop 2002
Conclusions on Single Helicity



There is enough theoretical understanding and
experimental evidence which support the idea that
this regime might lead to significant improvement of
RFP performance
We know how to produce a target plasma, which more
easily could achieve a QSH spectrum.
From an active control point of view, QSH is a robust
state:


One big mode, well identified, selectable in advance
RFX has more than enough power to deal with QSH
Outline of the talk
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
Dynamo modes active reduction

Pulsed Poloidal Current Drive
(PPCD):


the induction of a poloidal
current at the plasma edge
causes a dramatic reduction
of the magnetic turbulence
(MST + RFX PRLs) and
STRONG PLASMA
HEATING
ROBUST TECHNIQUE
(recently performed in T2R
at high aspect ratio-many
modes to suppress!Cecconello, PPCF 2003)
It is TRANSIENT, but in RFX a quasi-stationary version has been implemented
Oscillating Poloidal Current Drive (OPCD)
Periodic
(oscillating) applied inductive variations of the poloidal
electric field allows to extend the PPCD benefits in a stationary
fashion
450
T
e0
(eV)
350
stationary average
improvement of
confinement obtained
with OPCD in RFX
(Bolzonella et al., PRL
2001)
250
150
60
W
1,sec
(a.u.)
40
20
0
60
40
20
W
1,7
(a.u.)
0
20 30 40 50 60 70 80 t (ms)
PPCD action strongly affects MHD in the RFP

Though transient (but quasi stationary version is feasible),
PPCD is an efficient tool to interact actively with MHD modes
in the RFP.
Why might be useful for future operation in the new
RFX ?
1.
“per se”: the new RFX coils system allow optimized, high power,
quasi stationary PCD. This is a technique for improving
confinement.
2.
It can set-up an improved collisionless target plasma on which to
work with feedback


Record reduction of magnetic stochasticity
Change the properties of broadband magnetic turbulence
Active control issue: PPCD triggers QSH spectra

Evolution of m=1 modes in MST following the
application of PPCD
Heat pulse propagation observed
Significant transport barrier present in the plasma
SXR brightness (W/m2)

time (ms)
Modes decreased to very low amplitude!

Beating in the SXR signals of
the frequencies corresponding
to TWO helical structures:


the (1,6) and the (1,7) mode
Beating recorded:

at the dominant frequency


envelope @ Df ~ f1,6-f1,7
and at the first harmonic

envelope @2*Df
Two rotating islands


Modes are so small that magnetic chaos is strongly reduced
We observe individual TINY rotating magnetic islands
associated with the (1,6) and (1,7) modes: a marker of
stochasticity suppression.
t1 t2 t3
1
2
Franz, Marrelli et al, submitted to PRL
3
t4
4
Numerical simulation confirms

ORBIT used for Poincare’
maps of the magnetic field
lines (with experimental
mode amplitudes as inputs)
confirms the presence of
two islands
Outline of the talk
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
RFP & RWMs (see J. Drake’s talk)


EXTRAP T2R
experience (thin
shell device)
demonstrates that
RFP discharges with
duration ≥2 shell
times can be
produced without
significant effects
due to RWMs
If EXTRAP T2R experience is confirmed in RFX, time scales of dynamo
and RWMs are separated and we should have a reasonable amount of
time (≤ 100 ms) to cope with dynamo modes before dealing with RWMs.
Outline of the talk
1.
Status of the RFX reconstruction
2.
RFP mode control issues
3.
a.
Mode rotation
b.
The monochromatic dynamo
c.
Suppression of dynamo modes
d.
RWMs
RFX operational scenarios
Control strategies and plans - 1
i.
Benchmark and improve old RFX performance
(passive or m=0 perturbations)
ii.
Explore new regimes with active actions on
the MHD through the coils
RFX operational scenarios - 1
Benchmark and improve old RFX performance

Actions through an applied m=0 mode (TF
coils):
a.
Synchronous driving torque for mode rotation
also in closed loop mode
b.
PPCD
c.
OPCD
RFX operational scenarios - 2
Active actions through 192 saddle coils:

Apply m=1 magnetic perturbations


Work on individual modes: one at the time or several simultaneously
Realize an intelligent shell


Zeroing of radial field at the edge to maintain an effective
close fitting shell.
Might be interesting for QSH studies, since we have
evidence that a smooth magnetic boundary facilitates their
onset.
RFX operational scenarios - 3
Drive of m=1 magnetic perturbations

Apply a monochromatic perturbation to affect one individual
mode:
i.
ii.
iii.

“pumping” the mode to drive QSH states through helical fields at
the plasma boundary
Feedback stabilization of individual modes
inducing rotation of a single mode
Apply several simultaneous geometrical helicities (various n’s):
a.
b.
c.
damping of main “dynamo modes”
feedback stabilization of RWM
breaking phase locking among “dynamo modes” with induction of
modes differential rotations
RFX operational scenarios - 4

Low current scenario
1.
2.

Theoretical work (Guo, Fitzpatrick et al) and experimental data
(TPE-RX, EXTRAP T2R, MST) suggest that low current operation
could allow for spontaneous dynamo mode rotation.
If dynamo modes rotate, this scenario is more suitable to
concentrate efforts on RWM control
High current scenario (> 1 MA)


Better for confinement improvement techniques (OPCD)
and for interaction with “dynamo” modes (but higher
wall-locking probability).
Passive shell (and EXTRAP T2R experience) might
postpone RWM issue up to ≈50-100 ms
RFX IS THE ONLY EXPERIMENT DESIGNED TO EXPLORE
RFP PHYSICS IN THE MA REGIME !
Thanks to the flexibility of the coils system, different
actions can coexist in the same pulse
Active drive of (1,ndom)
> 1 MA
PPCD
QSH
Self-similar decay
+ QSH
T
~50-70 ms
~50 ms
feedback on m=0
Feedback on RWM
… comments, and
collaborations, welcome !