Stabilization of Neoclassical Tearing Mode by Electron Cyclotron Current Drive and Its Evolution

Simulation on JT-60U Tokamak

K. Nagasaki 1), A. Isayama 2), N. Hayashi 2), T. Ozeki 2), M. Takechi 2), N. Oyama 2), S. Ide 2),

S. Yamamoto 1) and JT-60 Team

20th IAEA Fusion Energy Conference, Nov. 1-6, 2004, Vilamoura, Portugal

1) Institute of Advanced Energy, Kyoto University2) Japan Atomic Energy Research Institute

• Electron cyclotron current drive (ECCD) is considered to be one of attractive methods to stabilize the neoclassical tearing mode (NTM), which compensates the missing bootstrap current in the magnetic island.

• The necessary EC power for complete stabilization is estimated as high as 30MW in ITER. The reduction of EC power and the determination of EC conditions will help more reliable operation of high plasmas without NTM.

• The nonlinear NTM physics is still yet understood well. It is necessary to understand the NTM physics such as the excitation and the disappearance.

Introduction

Contents

1. Introduction

2. NTM stabilization experiment by ECCD in JT-60U

Dependence of 3/2 NTM on EC injection timing

Stabilization at high N plasma using 2nd harmonic ECCD

3. Simulation of 3/2 NTM stabilization

Modified Rutherford equation (MRE) coupled with 1.5D transport code and EC code.

Determination of coefficients in MRE

Comparison with early and late ECCD experiment

4. Conclusion

Early ECCD is more effective to suppress 3/2 NTM.

4000 5000 6000 7000 8000 9000 10000 110000

1

2

3

4

0

10

20

30

0.0

0.5

1.0

1.5

2.00

1

2

3

4

NBI

EC

Time (msec)

Power (MW)

0.000

0.005

0.010

|B| (a.u.).

N

E41693

ne (1019m-3)

(a)Early ECCD

4000 5000 6000 7000 8000 9000 10000 110000

1

2

3

4

0

10

20

30

0.0

0.5

1.0

1.5

2.00

1

2

3

4

Time (msec)

Power (MW)

NBI

EC

0.000

0.005

0.010

|B| (a.u.).

N

E41650

ne (1019m-3)

(b) Late ECCD

K. Nagasaki, et al., Nucl. Fusion 43 (2003) L7-L10

• High p ELMy H-mode plasma

• Ip=1.5MA, Bt=3.66T, q95=3.8, PNB=20MW

• Only the m/n=3/2 mode is excited.

• PEC=0-2.7MW

• IEC: 25kA/MW, 0.17MA/m2 at q=3/2

• Rotation frequency: 10-12kHz 0

5

10

15

0

1

2

3

4

5

6

0 0.2 0.4 0.6 0.8 1

E41693(5.5s for Ti, Te; 5.3s for q)

Ti,

Te

[keV

]

q

r/a

Ti

Teq

2 3 4 5 R [m]

0

-1

1

E41693(5.3s)

EC

fEC=110GHz

The EC power for 3/2 NTM suppression is lower in early ECCD.

• The EC power for the complete suppression of the m/n=3/2 mode is lower than in the late ECCD.

• The stabilization effect becomes weak when the current location is deviated by about half of the maximum magnetic island width.

0.0 0.2 0.4 0.6 0.8 1.00.0

0.1

0.2

0.3

0.4

0.5

19deg

21deg

J (M

A/m

2 )

23deg

Calculated EC current profile

PEC=4MW

0 1 2 30.0

0.2

0.4

0.6

0.8 Early ECCD Late ECCD

|B|/

f (a

.u.)

PECH

(MW)

.

0.45 0.50 0.55 0.60 0.650.0

0.1

0.2

0.3

0.4

0.5

PEC

=2.9MW

PEC

=2.3MW

PEC

=1.7MW

|B|/f

(a.

u.)

res

.

Maximum island width

Dependence on EC power Dependence on EC position

• The Te perturbation measured with ECE is more suppressed as the EC power is injected earlier.

• The Te perturbation decreases asymmetrically in the radial direction. The magnetic island looks to decay from the smaller radius position.

Dependence on injection timing

5 6 7 8

3.7

3.6

3.5

3.4

E42833

Time (sec)

R (

m)

5 6 7 8 905

10152025

0.0

0.5

1.0

1.5

2.0

0

1

2

3

40

5

10

15

20

PN

B (M

W)

Time (sec)

N

PNB

(MW)

E42833

PEC

(MW)

3/2 modefrequency (kHz)

/e eT T

5 6 7 8

3.7

3.6

3.5

3.4

Time (sec)

R (

m)

E42832

5 6 7 805

10152025

0.0

0.5

1.0

1.5

2.0

0

1

2

3

40

5

10

15

20

PN

B (M

W)

Time (sec)

N

PNB

(MW)

E42832

PEC

(MW)

3/2 modefrequency (kHz)

/e eT T

0 1 2

0

2

4

6 EC 2 Units EC 3 Units

rT

e (c

m)

EC Injection Timing (sec)

Inner side

NTM onset

Dependence on injection timing

0 1 23

4

5

6

7

8NTM onset

EC 2 unitsEC 3 units

rT

e (cm

)

EC Injection Timing (sec)

Outer side

• When the ECCD is applied before the onset of NTM, the Te perturbation is more suppressed.

• The critical EC timing for effective ECCD is associated with the NTM onset phase.

3.6 3.7 3.8-0.04

-0.02

0.00

0.02

0.04

0.06 #42832

dT e

/Te

R (m)

Time(msec) 5600 5620 5640 5660 5680 5700

Island center

5 6 7 8

3.7

3.6

3.5

3.4

E42833

Time (sec)

R (

m) /e eT T

Inner side

Outer side

2nd harmonic ECCD at high N ELMy H-mode plasma

• The early ECCD is also effective in a high performance plasma of p1.7, H89PL1.8, N3.0 and H89PLN5.

• The NB power for keeping N is lower when the 3/2 NTM is suppressed.

• The stabilization effect is sensitive to the EC current position as well as the fundamental early ECCD.

Ip = 0.85 MA,

Bt =1.7 T, q95 =3.5

PEC=2MW

JEC=0.3-0.4MW/m2

EC=0.42-0.50

45º

47º

49º

Numerical model

Modified Rutherford equation

EC code

1.5D tokamak simulation code ( TOPICS )1D transport equations for density and temperatures1D current diffusion equation2D MHD equilibrium: Grad-Shafranov equation ( Fixed boundary )

Physical values at rational surface

Additional current source

Geometry &Plasma profilesCurrent profile

ECCD term

N. Hayashi, et al., Nucl. Fusion 44 (2004) 477

GGJ

Polarization

Classical Bootstrap

ECCD

0

dWdt

kc W 2

kBS0Lq jBSBp

WW 2 Wd

2 kGGJs2p

Lq2

sLp1

1q2

2 1W

k pol s1.5 p

piLqLp

2

2 1W 3 kEC 0

Lq s

B p

EC

I ECa2

1W 2

Parameters of kBS, kGGJ, kpol, kEC are constant.

Value of Wd depends on theoretical models to limit the parallel heat transport

).

(kBS, kGGJ, kpol, kEC, Wd) should be estimated by fitting to experiments.

Parameter fitting at NTM growing and stabilizing phases

Growth phase

kc =1.2, kBS ~ 5,

kGGJ < 10 ( not important ),

kpol < 3,

Wd = 0.008

( ~ flux-limit model )

8 9

0.05

0.10

0.15

EC

=0.40

EC

=0.39

EC=0.42

EC

=0.34

EC=0.46

EC=0.50

W/a

Time (sec)

Real time control

ECCDE41666

6 7 8

0.05

0.10

0.15

W/a

Time (sec)

E36705

Estimated coefficients are consistent between the growing and stabilizing phases.

0 0.05W-0.1

0

0.1

(s-1

)

ECCD

-0.1

0

0.1

(s-1

)

BS

GGJ

Polarization

dW/dt

dW/dt

Classical

without ECCD

with ECCD

Stabilization phasekc =1.2, kBS =4.5,

kGGJ=1, kpol=1,

Wd =0.02

( ~ flux-limit model )

kEC ~ 3-4, EC=0.025

• Higher EC current induces larger movement of the 3/2 surface.

The rational surface is moved by EC current, resulting that the stabilizing effect is reduced.

IEC

/ Ip = 0.02

IEC

/ Ip = 0.04

10 11time (s)

3

10 11time (s)

W

s

WEC

2

EC current just on island center

0.4

0.5

0.6

10 11

time (s)

1

EC current location

10 11time (s)

W

s

WEC

20.4

0.5

0.6

10 11

time (s)

1

10 11time (s)

3

W larger than that for low EC current

Comparison with Experiment

0 1 2 30.00

0.05

0.10

0.15

Late

Early

Sim

Threshold

Wsa

t / a

PEC (MW)

Background noise level

Exp

Bootstrap current term

(Destabilizing term)

Classical tearing term(Stabilizing

term)PEC=0.8 MW

(IEC=20 kA)-37% +117%

PEC=1.7 MW

(IEC=43 kA)-23% +71%

• The power dependence in late ECCD agrees with the simulation.

• The EC power reduction for complete stabilization can be explained by the simulation. The NTM starts to grow with larger seed island width.

• The decrease of the bootstrap current term and/or the increase of classical tearing parameter term may contribute to the smaller island width in early ECCD.

Conclusion

• The stabilization of m=3/n=2 NTM has been studied experimentally and theoretically in high p ELMy H-mode plasmas on the JT-60U tokamak.

• The early ECCD is more effective than the late ECCD, and the critical EC timing is associated with the mode onset phase.

• The 3/2 NTM has been suppressed in a high p ELMy H-mode plasma of N=3.0 by the second harmonic early ECCD. The dependence on the EC current position is similar to the fundamental ECCD stabilization.

• The simulation on the basis of the modified Rutherford equation with 1.5D transport code and EC code well reproduce the time evolution of the 3/2 NTM experimentally observed both at the growing and stabilizing phases.

• The mode evolution can be explained by the simulation in late ECCD case. The EC power reduction for complete stabilization in early ECCD agrees with the simulation result.