advancing tokamak physics with the iter hybrid … tokamak physics with the iter hybrid scenario on...
TRANSCRIPT
Advancing Tokamak Physics with theITER Hybrid Scenario on DIII–D
byP.A. Politzer
Presented atForty-Ninth APS Meeting ofthe Division of Plasma PhysicsOrlando, Florida
November 12–16, 2007
with C.C. Petty*, R.J. Jayakumar† , T.C. Luce*, J.C. DeBoo*,E. J. Doyle#, J.R. Ferron*, P. Gohil* , R. Groebner*,W. Heidbrink§, C.T. Holcomb†, A.W. Hyatt *, J. Kinsey*,R.J. La Haye*, M.A. Makowski† , R. Nazikian‡ , T.W. Petrie*,Q. Ren¶, E. Ruskov§ , G. Staebler*, and M.R.Wade*
*General Atomics, San Diego, California.†Lawrence Livermore National Laboratory, Livermore, California.#University of California-Los Angeles, Los Angeles, California.§University of Irvine-Irvine, California.‡Princeton Plasma Physics Laboratory, Princeton, New Jersey¶academy of Sciences Institute of Plasma Physics, China
APS/DPP Orlando – 15Nov07 – Politzer – 2
The Hybrid Scenario is an High PerformanceInductive Operating Regime for ITER
Conventional H-mode:βN ≈ 1.8, q95 ~ 3, fBS << 1baseline
Advanced inductive:βN ≤ βno-wall, q95 ~ 3, fBS ~ 0.3high gain
Hybrid:βN ≈ βno-wall, q95 ≥ 4, fBS ~ 0.5high fluence
Advanced tokamak:βN ≤ βideal-wall, q95 ≥ 5, fBS ~ 0.8steady-state
inc
rea
sing
co
mp
lexi
ty &
be
nefit
“hyb
rids”
• The major tokamak programs have been developing the hybrid scenario for a number of years.
• It provides high gain and high neutron fluence options for ITER operation.
• This talk will cover two areas where we’ve made significant advances, leading to better capability to forecast performance:
– MHD & the current profile,
– rotation & confinement,
– and will give a brief sampling of other areas of tokamak physics being addressed in hybrids.
APS/DPP Orlando – 15Nov07 – Politzer – 3
Stationary, High Performance Hybridsare Studied in DIII-D
• Stationary conditions are maintained for many τE and τR.➜ limited only by hardware.
0.00.51.01.5
024
012
05
10
time (s)
0.02.55.0
0 2 4 6 8 10 12
120675
Ip (MA)
PNB (MW/10)
βT (%)βN
H89PH98y2
ne (1019m–3)ZeffDα
Bn=1 (G; rms)~Bn=2 (G; rms)~
βT ≈ 4%βN ≈ 2.7
H89P ≈ 2.3H98y2 ≈ 1.3
Zeff ≈ 2
10 τE τR
ITER scenarios#2(base)
2.5%1.8 1.9
1.0 1.0
2.8%
#3(hybrid)
≈ 0.58 0.43 0.32βNH89q952
2.4 2.1
APS/DPP Orlando – 15Nov07 – Politzer – 5
MHD Activity is an Integral Part of Hybrid Operation– Usually a m/n = 3/2 NTM in DIII-D
• The effect of the ~stationary NTM
– is to broaden the current profile ➜ raises q(0) sawteeth are reduced (for q95 ≤ 4) or eliminated (for q95 > 4) › better confinement › removes one trigger for the 2/1 NTM
– increases stability of 2/1 mode
– with only a modest confinement reduction est ΔτE/τE ≈ – 6-15% depending on q95, rotation
– leading to high β operation;βN ~ 4li (~ no-wall limit)
0 1 2 3 4 5 6time (s)
H89P0
2
4
0.40.81.2
0.0
li
012302460
20
40125470125469
βN
Bn=2,rms (G)~
Bn=1,rms (G)~
➤ Without a 3/2 mode, the discharge evolves to an unstable 2/1 tearing mode
➜ controlled shut-down
APS/DPP Orlando – 15Nov07 – Politzer – 6
The Current Near the Axis is Reduced, Increasing q(0)
10 ρ
0
1
2
–1
117740.5000
q=3/2
Johm+Jbs+JNB
Jtotal
J (M
A/m
2 )
ΔJ
DIII-D current profile simulation
• Measured current profile shows deficit at center compared to sum of calculated currents.
• Only ~5% of total current, but strongly affects q(0).
• TRANSP simulation switches to neoclassical resistivity and current transport at 3.5 s.
– current profile peaks and q(0) drops.
2.5 3.0 3.5 4.0 4.5
time (s)
1.1
1.0
0.9
0.8TRANSP
plasma
q(0
)
●
APS/DPP Orlando – 15Nov07 – Politzer – 7
The NTM is Responsible for Modifying the q Profile:Changing the NTM Amplitude with ECCD Affects Sawteeth
• co-ECCD at q=1.5• suppresses 3/2 NTM• sawteeth appear
• counter-ECCD at q=1.5• enhances mode• sawteeth suppressed
n=23/2 NTM
n=1sawteeth
• Decreasing NTM amplitude increases sawtooth size, indicating peakingof the central current and reduction of q(0), and vice versa.
δB (
G; r
ms)
0246 119729
0123
ECCD on
3 4 5 6t (s)
δB (
G; r
ms)
119736642
0
012
3
ECCD on
3 4 5 6t (s)
• Raises the question: how does the NTM act on the current profile?
APS/DPP Orlando – 15Nov07 – Politzer – 8
Several Mechanisms Suggested to Explain the Effectof the 3/2 Mode on q & J
• Making progress, but no definitive conclusion yet.
– Direct current drive by the 3/2 mode – as seen from the magnetic axis, the NTM island looks like an Alfvén wave antenna
• interesting physics; calculated magnitude too small
– Broadening of the fast ion spatial profile by the 3/2 mode• change in fast ion profile is observed;
modeling indicates small effect on current profile
– Modulation of the NTM amplitude by ELMs – asymmetry in time ⇒ flux pumping (analogous to the effect of sawteeth on q & J)
• evidence for effect is seen
– Dynamo – conversion of kinetic to magnetic energy via 〈v×B〉
• no data; need nonlinear resistive MHD modeling
~~
➜
➜
APS/DPP Orlando – 15Nov07 – Politzer – 9
The Central Fast Ion Density is Reduced by the 3/2 NTM
0.4
0.3
0.2
0.1
0.0
be
am
cur
rent
de
nsity
(M
A/m
2 )
0.0 0.2 0.4 0.6 0.8 1.0ρ
Dfi = 0.01 m2/s
0.1
1.0
4.0
• Preliminary TRANSP analysis using uniform fast ion diffusion shows a large drop in central NBCD.
• TRANSP indicates that most of this is replaced by increased ohmic current, yielding a very small change in q(0).
• The FIDA (fast ion Dα) diagnostic indicates that the density of fast ions drops in the inner region of the plasma when the 3/2 NTM appears.
129257
n=2
no MHD
1.7 1.8 1.9 2.0 2.1R (m)
400
300
200
100
0
FID
A fa
st io
n d
ens
ity (
arb
)
ρ = 0 ρ = 0.5
APS/DPP Orlando – 15Nov07 – Politzer – 10
Modulation of the 3/2 NTM by ELMs Leads to Current Profile Broadening
1042761
0
–1
–2
–3
–4
ΔBz (mT)
ρ0.0 1.0 0.0 1.0ρ
0.04
0.02
0.00
–0.02
–0.04
Δq
Profile change at an ELM(from direct MSE analysis)
• Averaging over many ELMs, analysis of MSE data shows that,at an ELM, q increases inside the q=3/2 surface.
100
0
–100
dB/
dt (
T/s)
118444
0
4
8
Dα
(a
rb)
4.0 4.1t (s)
• This modulation can lead to a poloidal magnetic flux pumping effect(similar to the process whereby sawteeth maintain q0 ~ 1).
[Petty, JP8.00084]
APS/DPP Orlando – 15Nov07 – Politzer – 12
Hybrid Performance Depends on Toroidal Rotation
• Most of the tokamak experience base has been limited to plasmas with strong toroidal rotation (thanks to NB heating).
• There is concern that ITER (& DEMO & reactors) will have low rotation.
➜ To study this issue, we’ve used the recently modified NB configuration in DIII-D to study the effect of rotation on the performance of hybrid plasmas. (5 sources co-NBI; 2 sources counter-NBI.)
– We did systematic scans of rotation for both hybrid (q95 ~ 4.2 & 4.6) and advanced inductive (q95 ~ 3.2) plasmas.
– The central Mach number has been reduced by up to a factor of 5, to M(0) ≈ 0.1, maintaining stationary conditions.
➜ Both the confinement and the MHD properties are affected.
APS/DPP Orlando – 15Nov07 – Politzer – 13
Reducing Rotation Strongly Affects Plasma Characteristics
Pnb(MW)
0.0
0.2
0.4
0.6
0.8
M(0)
125499
0 1 2 3 4 5 60
1
2
3
4
t (s)
Bn=2(G; rms)
~
12
4
8
0
co
ctr
total
• Density and βN held constant, under feedback control
• High vs. low rotation: – add counter-NBI – reduce torque; rotation decreases – confinement decreases; total power increases – 3/2 NTM amplitude increases
• Experiments and modeling are sorting out what’s happening.
0.20
0.00
0.10τE (s)
M(0.5)
APS/DPP Orlando – 15Nov07 – Politzer – 14
Density, Temperature, and Current Profilesare Unaffected by Changing Torque
1.5
1.0
0.5
0.0
-0.5
torque density(N-m/m3)
0.0 0.2 0.4 0.6 0.8 1.0ρ
Changing torque and power:
Results in a large change inthe rotation profile:
But the density and temperatureprofiles vary little(βN and density are controlled):
1254990.50.40.30.20.10.0
–0.10.0 0.2 0.4 0.6 0.8 1.0
ρ
M (=vφ/cs)
4
3
2
1
0
Te (keV)
Ti (keV)8
6
4
2
0
8
6
4
2
0
ne (1019 m–3)
0.0 0.2 0.4 0.6 0.8 1.0ρ
0.0 0.2 0.4 0.6 0.8 1.0ρ
0.0 0.2 0.4 0.6 0.8 1.0ρ
Also, JNBCD is halved(0.12→0.06 MA), butthe effect on q and Jtotalprofiles is negligible.
APS/DPP Orlando – 15Nov07 – Politzer – 15
Angular Momentum is Roughly Proportional To Torque
• Vary the applied torque (T) over the range 0.4-4.6 N-m.
• Characterize rotation by either L = total angular momentum M(0.5) = Mach number at ρ = 0.5
› these are strongly correlated
• L ∝ T⇒ τΩ = L/T ~ independent of L,
except for a possible indication of ‘inherent’ rotation at low torque.
• Very low (and zero) rotation was inaccessible because of error field penetration and locking of the 3/2 NTM.
0 1 2 3 4 5torque (N-m)
0.8
0.6
0.4
0.2
0.0ang
ula
r mo
me
ntum
(kg
-m2 /
s)
?
uuu
u
uu
u
▲▲▲▲▲▲
▲
▲▲
▲
▲▲nnn
n
n n
n
nnnnnnnnu
u
q95
4.5-4.9
3.1-3.44.1-4.2
APS/DPP Orlando – 15Nov07 – Politzer – 16
Global Confinement Improves As Toroidal Rotation Increases
0.20
0.15
0.10
0.05
0.00
M(0.5)0.50.40.30.20.10.0
nnnnnnnuuunun
nnnn
nn
u
uuuu
▲▲
▲▲▲
▲▲
▲▲▲
▲
▲
τΩ(s)
τE(s)
• Weak dependence of τE and τΩ on M(0.5) at q95 > 4; stronger for low q95.
• Medium & high q95 indistinguishable.
• τE and τΩ close to numerically equal, except at low NB torque.
0.200.150.100.050.00
0.20
0.15
0.10
0.05
0.00
nnnnnunnuu▲uu
uuunnn
n
▲
n
▲▲
▲▲▲
▲▲▲▲▲
q95
4.5-4.9
3.1-3.44.1-4.2
low q95
low q95
med q95high q95
med q95high q95
τE (s)
τΩ(s)
0.20
0.15
0.10
0.05
0.00
M(0.5)0.50.40.30.20.10.0
nnnnnuuuuu
uuu u
nnnnn
n
n
n
▲▲▲
▲▲▲▲
▲▲▲ ▲
▲
APS/DPP Orlando – 15Nov07 – Politzer – 17
ExB Flow Shearing Rate Increases With Rotation
M(0.5) incr.M(0.5)incr.
high q95
0.0 0.2 0.4 0.6 0.8 1.0ρ
200
150
100
50
0
x103 s–1 q=1.5100
80
60
40
20
00.0 0.2 0.4 0.6 0.8 1.0
x103 s–1
ρ
low q95M(0.5)0.120.140.200.25
M(0.5)0.130.160.240.36
q=1.5
• ExB flow shear at high q95 is is ~2x value at low q95.
APS/DPP Orlando – 15Nov07 – Politzer – 18
Turbulent Transport Coefficients Decrease As Rotationand Rotation Shear Increase (Tglf Modeling)
• At high q95: χe & χi are comparable, and well above χi,neo.
• At low q95: χi is comparable to χi,neo, but χe is much larger.
➜ Using χi+χe as a measure of overall turbulent transport, going from high to low rotation the turbulent transport increases by ~25% at low q95 and by ~40% at high q95.
4
3
2
1
0
m2/s
0.1 0.2 0.3 0.4M(0.5)
q95 = 4.5-4.9
χi,turb
χe,turb
χi,neo
●
● ●● ●●
◆
◆
◆
◆ ◆◆
▲ ▲ ▲ ▲ ▲
dashed blue lines give χe due to kθρi > 1 modes
avg ρ=0.5-0.75
0.1 0.2 0.3 0.4
4
3
2
1
0
m2/s
M(0.5)
q95 = 3.1-3.4
●
◆
▲
χi,turb
χe,turb
χi,neo
◆ ◆ ◆ ◆
●●
● ●
▲▲ ▲ ▲
avg ρ=0.5-0.75
APS/DPP Orlando – 15Nov07 – Politzer – 19
Turbulent Transport Coefficients From TGLF ReproduceTrends Seen In Transp Analysis of Profiles
◆●
●
●●●
●
●
●
● ●●
◆
◆
◆
◆
◆◆◆◆ ◆ ◆
(χi–χi,neo)TRANSP
χe,TRANSP
(χi–χi,neo)TRANSP
χe,TRANSP
4
3
2
1
0
m2/s
0.1 0.2 0.3 0.4M(0.5)
q95 = 4.5-4.9
χi,turb
χe,turb
●
● ●● ●●
◆
◆
◆ ◆◆
avg ρ=0.5-0.75
◆
0.1 0.2 0.3 0.4
4
3
2
1
0
m2/s
M(0.5)
q95 = 3.1-3.4
●
◆
χi,turb
χe,turb
◆ ◆ ◆ ◆
●●
● ●
avg ρ=0.5-0.75
APS/DPP Orlando – 15Nov07 – Politzer – 20
Change In 3/2 Island Width Is Calculated To HaveA Moderate Effect On Confinement
– Island width is larger at low q95 at all rotations.
– Change in width with rotation is larger at low q95.
– Largest island is 9.5 cm wide, ~16% of minor radius.
• Assess effect on τE using Chang-Callen ‘belt model’: – At high rotation, island reduces τE to 90-96% of est’d unperturbed value. – For higher q95, increasing width has a small effect. – For low q95, τE is reduced to ~85% of unperturbed value.
M(0.5)0.0 0.1 0.2 0.3 0.4 0.5
τE (with island)/τE (without island)
1.00
0.95
0.90
0.85
0.80
nnnnnnn
n n
nnn
n
n
uuuuu
uu
u
n▲▲
▲▲▲▲
▲▲
▲▲▲▲
NTM
isla
nd w
idth
(m
)
M(0.5)
0.10
0.08
0.06
0.04
0.02
0.000.0 0.1 0.2 0.3 0.4 0.5
n
n
nn▲
nnnnnnn
nnnn uuuu
uuu
u
▲
▲▲▲▲
▲
▲
▲▲▲
q95
4.5-4.9
3.1-3.44.1-4.2
low q95
high q95
low q95
high q95
APS/DPP Orlando – 15Nov07 – Politzer – 22
Many Other Physics Studies And Results Using Hybrids
– The role of Te/Ti; changes in turbulence levels & confinementof energy and momentum. [Doyle, GO3.00006]
– The effect of shaping on confinement (triangularity, squareness,upper/lower null, double null). [Groebner, GO3.00012, Leonard BI1.00003]
– The effect of triangularity and squareness on the pedestal.[Leonard, BI1.00003, Groebner GO3.00012]
– The effect of rotation on the pedestal pressure. [Leonard, BI1.00003]
– Demonstration of radiative divertor operationand divertor heat flux reduction. [Petrie, UP8.00035].
– ELM suppression with RMP. [Fenstermacher, BI1.00002]
– The effect of wall conditioning on hybrid performance. [West, GO3.00005]
APS/DPP Orlando – 15Nov07 – Politzer – 23
Increasing Te/Ti Leads to Increased Fluctuationsand Reduced Confinement
• Increase Te/Ti by adding 2.3 MW of ECH.
• Te/Ti at center increases by 22% (0.67-0.82).
• Rotation is reduced.
• Low-k & medium-k density fluctuations increase.
• To separate Te/Ti and rotation effects, compare with NBI-only plasma, matched for the same rotation and β; – by adding ~0.6 MW of counter-NBI instead of ECH.
• H89 is ~15% lower with ECH.
1.41.21.00.80.60.40.20.0Fl
uctu
atio
n le
vel (
a.u.
)
5004003002001000Frequency (kHz)
ECH (129282)
No ECH (128284)
ALT-CP, Ch1:4,17,18(acc),NP:1024/950, 3000-4000ms
r/a ~ 0.7
BES (0-3.5 cm–1)
FIR scattering (7 cm–1)
APS/DPP Orlando – 15Nov07 – Politzer – 24
Confinement, Pedestal, And ELMs All Depend On Shape
• With increased triangularity:– confinement improves
at fixed βN– pedestal β increases
●
●●
●
■■
■■
1.6
1.4
1.2
1.0
0.81.0 1.5 2.0 2.5 3.0 3.5
βN
H98
y2
high δ
low δ
3.3 3.4 3.5 3.6 3.70
2
4
time (s)Dα
(a
rb)
• With increased squareness:– pedestal β decreases– smaller, faster ELMs
(consistent with theory)– reduce triggering of
2/1 NTM119748119750
APS/DPP Orlando – 15Nov07 – Politzer – 25
Conclusions
• The presence of a stationary NTM is an inherent feature of ITER hybrid scenario plasmas in DIII-D – the NTM contributes to the beneficial modification of the current profile – the confinement penalty due to the NTM is modest
(5-15% lower than no-NTM estimate)
• Rotation and rotation shear have strong effects on confinement:
– Scanned the central Mach number 0.1-0.5
– At low rotation, the fusion performance parameter G (= βNH89/q952) is reduced by 10-30%, but remains above the ITER level for Q fus = 10 operation at low q95.
– 3/2 NTM island width increases as rotation is reduced, with a moderate effect on confinement.
– Primary confinement effect of changing rotation is via changes in ExB flow shear and turbulent transport.
• In addition to being a demonstrated scenario for improved ITER operation, the hybrid scenario is a good place for studying tokamak physics.