Overview of transport and MHD stability study and impact of magnetic field topology in the

Large Helical Device

Katsumi Ida on behalf of LHD experiment group

National Institute for Fusion Science

25th IAEA Fusion Energy Conference 13-18 October 2014

St Petersburg

Russia

OV/2-3 2014/10/13

OUTLINE 1 Extension of operation regime in LHD

48 minute long pulse discharge Simultaneous achievement of high Te and Ti plasmas

2 Transport analysis beyond power-balance

Non-diffusive term of momentum and impurity transport Heat and momentum transport in modulation ECH experiment

3 MHD–Transport interaction

Impact of stochastic magnetic field on MHD Pellet injection into the magnetic island The effect of energetic particle driven MHD to transport

4 Detachment and impurity control

stabilization of detached plasma by RMP impurity behavior in LHD

5 Summary

48 minute long pulse discharge

Progress of high-performance steady-state plasma and study of PWI in LHD

H. Kasahara, Y. Yoshimura, K. Nagasaki*, M. Tokitani, N. Ashikawa, Y. Ueda**, G. Motojima, M Shoji, T. Seki, K. Saito, R. Seki, S. Kamio, G. Nomura, S. Kubo, T. Shimozuma, H. Igami, H. Takahashi, S. Ito, S. Masuzaki, H. Tanaka, J. Miyazawa, H. Tsuchiya, N. Tamura, K. Tanaka, T. Tokuzawa, T. Mutoh and LHD Experiment Group National Institute for Fusion Science, 322-6, Oroshi-cho, Toki-city, 509-5292, Japan *Institute of Advanced Energy, Kyoto University, Gokasho, Uji, Kyoto, 611-0011, Japan **Graduate School of Engineering, Osaka University, 2-1, Yamadaoka, Suita, Osaka, 565-0871, Japan e-mail: kasahara.hiroshi@LHD.nifs.ac.jp An ultra-long-pulse plasma with a duration time (τd) of 47 min, a line-averaged electron density (ne) of 1x1019 m-3, and electron and ion temperatures (Te, Ti) of 2 keV has been achieved by the averaged heating power (PICH+ECH) of 1.16 MW and a proportional-integral-derivative (PID) control of particle fueling in LHD. The heating energy injected into plasma (Wheat) reached 3.36 GJ, which is a new world record in toroidal plasmas. The ultra-long-pulse plasmas are often terminated by a sudden increase in carbon impurity, which might be caused by ejection of mixed-material deposition layers on the plasma facing surface (PFS). A large amount of mixed-material layers, consisting mainly of carbon (> 90%) and iron impurities, are formed over a wide surface area of the PFS. Carbon impurity originally comes from the divertor region (domes and divertor plates) by physical sputtering, while iron impurity comes from the first wall. They are deposited on the PFS during the steady-state operation (SSO). On the PFS, helium radiation damage is also created just under the mixed-material deposition layers, and increases in retention of helium. Such a long deposition process plays an important role of sudden impurity source to the plasma termination and particle source in the SSO.

In LHD, high-performance plasma experiments with ne > 1x1019 m-3 and Te > a few keV have been carried out in order to understand PWI effect on a long time scale (> a few thousand seconds). Figure 1 shows an ultra-long-pulse plasma discharge of a helium plasma with ne ~ 1 x1019 m-3, Te ~ Ti ~ 2keV, PRF ~ 1.2 MW (PICH ~ 0.9MW, PECH ~ 0.3MW), and τd ~ 2,859 sec. The newly achieved heating energy of 3.36GJ is twice as the previous record, 1.6GJ, with PRF ~ 0.5MW and ne ~ 0.4x1019 m-3 in LHD [1], which is a world record in toroidal plasmas. Three types of ICRF antennas were installed at different toroidal sections in order to avoid local hot spots around each ICRF antenna. Real-time feedback control for impedance matching, antenna phasing and heating power boosting for unexpected impurity influx were also developed. These heating techniques enabled a stable plasma heating as shown in Fig. 1(d). Wall pumping was gradually reduced during the long-pulse discharge with τd > 1000 sec. The gas-fueling by a proportional-integral-derivative (PID) control was adopted to maintain a constant level of the plasma density as shown in Fig. 1(a). Radiation power (Prad) was less than approximately 17 % of PRF, where the carbon and iron impurity accumulation was negligible in the discharge. The spike frequency of the line intensity for the carbon spectrum began to increase after 600 sec. At that time, the divertor temperature

EX-D

Fig. 1. Ultra long-pulse plasma sustained by RF heating for hydrogen minority helium plasma. Typical time evolutions are as follows: (a) line-averaged electron density (ne), (b) ion (Ti) and central electron temperature (Te), (c) total radiation power and line intensity of carbon spectrum, (d) RF heating powers of ICH and ECH.

The ultra-long pulse (τd ~ 48 min): neτETi ~ 3.5x1018 keVm-3 s, ne ~ 1.2x1019 m-3 meausred with FIR Te ~ Ti ~ 2 keV,

Dusts (M. Shoji EX/P6-33) induce frequent spikes of emission of iron and carbon impurities and trigger the termination of discharge

H. Kasahara EX/7-3

Carbon spikes

Simultaneous injection of ICH (T. Seki FIP/P5-3) +ECH PRF~ 1.2 MW Pinj ~ 3.4 GJ.

(T. Akiyama FIP/P8-31)

Mixed-material (C-Fe) deposition layers increase retention of helium and contributes to the wall pumping

Simultaneous Achievement of High Te and Ti plasmas High Te and Ti plasmas are realized with the simultaneous achievement of electron and ion ITB by applying central focused ECH (T. Shimozuma EX/P6-34) and by reducing neutral density, which is confirmed by Hα measurements (K.Fujii EX/P6-31)

K. Nagaoka PPC/2-1

.

The reduction of turbulence transport is evaluated to be factor more than three by using heat deposition code GNET-TD (S. Murakami TH/P6-38) and transport code TASK-3D (M. Yokoyama EX/P6-27).

e-ITB i-ITB

Transport analysis beyond power balance

Power balance analysis (database of χi, χe, D)

Dynamic transport analysis Flux-gradient relation

Radial propagation of plasma response to perturbation

Te

ECH

gradient grad-T grad-V grad-n/n

Flux Q/n P/(nm) Γ/n

S. Inagaki EX/2-1 M. Yoshinuma EX/P6-30

time

M. Yokoyama EX/P6-27

Non-diffusive term of momentum transport

ITG mode M. Nunami TH/P7-9. Analysis tool M. Emoto FIP/P8-28

The reversal of intrinsic torque from counter-direction to co-direction is observed to be associated with the formation of ion ITB, where the ITG mode is expected to be unstable due to the increase of ion temperature gradient and flattening of electron density profile.

Flux-gradient relation Radial profile of intrinsic torque

K. Ida K., et al 2013 Phys. Rev. Lett. 111 055001

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

Nor

mal

ized

Rad

ial F

lux � c

/nc

(m/s

)

3.9 3.925 3.95 3.975 4time (sec)

0

0.2

0.4

0.6

0.8

1

�

(b)

Out

war

dIn

war

d

Non-diffusive term of impurity transport

M. Yoshinuma EX/P6-30

The reversal of radial flux starts at reff /a99 = 0.65, where the ITB also starts, and propagates toward the edge and center in the time scale of 15 ms which is much shorter than the time scale of change in the mean plasma parameter.

Normalized radial particle flux of carbon impurity Flux-gradient relation

inward outward

15 ms

Ion-ITB foot

center

edge

L-mode

ITB

Dynamic method to study turbulence and turbulence transport S. Inagaki, K. Ida1, S.-I. Itoh, K. Itoh1, T. Tokuzawa1, N. Tamura1, S. Kubo1, T. Shimozuma1, K. Tanaka1, H. Tsuchiya1, Y. Nagayama1, T. Kobayashi2, N. Kasuya, S. Sasaki, A. Fujisawa, Y. Kosuga3, K. Kamiya4, H. Yamada1, A. Komori1 and LHD experiment group1 Research Institute for Applied Mechanics, Kyushu University, 6-1 Kasuga-Koen, Kasuga-city, 816-8580, Japan 1National Institute for Fusion Science, 322-6, Oroshi-cho, Toki-city, 509-5292, Japan 2Interdisciplinary Graduate School of Engineering Sciences, Kyushu University, 6-1 Kasuga-Koen, Kasuga-city, 816-8580, Japan 3Institute for Advanced Study, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka-city 812-8581, Japan 4Japan Atomic Energy Agency, 801-1 Mukouyama, Naka-city 311-0193, Japan e-mail: inagaki@riam.kyushu-u.ac.jp Here we developed research methods of plasma turbulence transport associated with the non-local features. The ECH modulation experiment and the higher harmonic analysis of the heat wave indicated: (i) propagation of the change of Te at the time of switch-off/on of ECH power is about 5 times faster than that of perturbation itself, (ii) propagation of the higher (7th) harmonic of the Te perturbation is 5 times faster than prediction by the diffusive model. New bi-spectral analysis of fluctuations demonstrated a non-linear coupling of micro-fluctuations at different radial locations. These results are beneficial for control of plasma dynamics in future fusion reactors. For more than 20 years, dynamic behavior of magnetized plasma has indicated limitations of a local diffusive picture of transport in which the transport flux is expressed in terms of mean parameters and their spatial derivatives at the same location and achieved wide recognition during recent years [1]. More recently, a new spatiotemporal analysis of a dynamic response of plasma has been developed and new global hysteresis in the gradient-flux relation was discovered [2]. The heat flux is a multiple-valued function of gradient, so that the dynamics in the temperature perturbation is far from a simple diffusive response. This issue has a critical impact on the predictive capability of temporal evolution of future burning plasmas and thus should be clarified. We applied this analysis technique extensively to LHD experiments and the results made establish a research method of plasma turbulence transport. Furthermore, we experimentally discovered a basic process concerning to the break of a local transport model. The ECH modulation (fmod = 25 Hz) is imposed on a low-density (n0 = 1.35×1019 m−3) NBI heated plasma (balanced injection of 2 MW) confined in LHD (Rax = 3.6 m and Bax = 2.75 T). The ECH power of 2 MW is absorbed in the central core reff/a99 ~ 0.2, here reff is the effective minor radius and a99 is the minor radius in which 99% of the total stored

EX-C

Fig. 1 Time evolution of ECH power (a), the normalized temperature perturbations at reff/a99 = 0.37 and 0.81 (b) and spatiotemporal evolution of δTe (c)

Heat pulse propagation

S. Inagaki (Kyushu Univ.) EX/2-1

The propagation length of higher harmonic frequency was found to be larger than that predicted by the diffusive model because of the fast propagation.

Bulk propagation (fundamental frequency) à slow Front propagation (higher harmonic frequency)àfast

EX2-1 4

The conventional methods estimates χhp from the radial propagation of the fundamental harmonics of the perturbation. However, it is found to be powerless because the conventional method does not consider “the two time scales” (i.e. hysteresis) as shown in Fig. 2. The nonlinear feature associated with the hysteresis in gradient-flux relation should appear in the response of extremely-higher harmonics [13]. The conditional averaging technique decreases the noise level significantly and allows us to observe higher harmonics up to tenth harmonics. Figure 3 shows the Fourier spectrum of the conditional averaged Te perturbation: more than ten harmonics are evident in the spectrum [14]. The power spectrum shows that the m-th harmonic decays as p(mf0) ∝ m-a

where α = 3.5-4. The square wave-like ECH power modulation does not have even numbers of harmonics. However, even number of harmonic of the Te perturbation is observed and thus the spectrum has some informations concerning non-linear processes of heat transport. We can capture the non-linear features of turbulent transport by observing the radial propagation of higher harmonics of Te perturbation. The Te perturbation is fitted to δTe ≈ Σexp(-imω0t+ikmreff), here m is the number of harmonics and km = kmr+ikmi. The real and imaginary parts of radial wavelength (kmr, kmi) of heat wave are estimated from radial dependence of the phase (θ) and amplitude (A), respectively (kmr=∂θ/∂reff, kim=-∂lnA/∂reff). In the diffisive transport model, the following dispersion relation kmr = kmi = (mω0/2χ)1/2 is obtained [14]. The radial decay length, i.e., the e-folding length of the m-th harmonic is shorter than that of the fundamental mode by a factor of m1/2 in the diffusive transport. Figure 4 shows the radial dependence of the amplitude of harmonics [14]. It is clearly shown that the decay of amplitude of higher harmonics is much slower than the

Fig.4: Radial profile of the intensity of the higher harmonics.

Fig.3: Typical power spectrum for an electron temperature perturbation. Noise due to the power supply (60 Hz and its higher harmonics) is not compensated in this figure.

Fig.5: Radial wave number of the heat wave.

Turbulent Transport is far from diffusive model

Fourier transform analysis

Intrinsic rotation at LCFS

The modulation of toroidal rotation Vφ at the LCFS is clearly observed. The driven rotation is in the counter-direction for positive Er and it propagates inward from LCFS.

This experiment suggests the existence of strong intrinsic toque at the boundary of plasma.

K. Kamiya (JAEA)

10

Figure 2. K. Kamiya et al.

10

Figure 2. K. Kamiya et al.

Modulation ECH is applied near the plasma boundary of reff ~ 0.88 to make a perturbation of temperature gradient at LCFS in LHD

LCFS

LCFS

counter

co

Effect of stochastic field on MHD and transport

S. Ohdachi EX/P6-29

When the stochastization of the magnetic field is enhanced by the RMP, the pressure driven mode is suppressed even without a change in the pressure gradient itself.

Amplitude of m/n = 2/3 mode are reduced by the RMP field when the normalized RMP coil current IRMP/B exceeds 0.7.

Related Simulations K. Ichiguchi TH/6-2 and A. Ishizawa TH/P6-40

Pellet injections into magnetic island

T. Evans(GA) EX/1-3

Hydrogen pellets are injected into the O-point of the magnetic island. The magnetic island prevent the inward shift of density. The significant peaked pressure profile inside the magnetic island is observed.

pellet

Magnetic island (with RMP) at reff = 0.5 – 0.6 m

Electron density

No magnetic island (without RMP)

Electron density

Effective Ion heating by energetic particle driven MHD

M. Osakabe EX/10-3

The increases of effective ion temperature evaluated from the neutral particle spectra is observed associated with EGAM excitation.

This is a clear evidence for the interaction between MHD instability and heating/transport in the plasma.

-3

0

3a) �B

�

�B� [x

10-6

T]

�t

0

0.5

1

4.9 4.95 5 5.05

c)

Tieff.T

ieff. [k

eV]

time[sec]

~7.5ms Decay time = ~79.3ms

0.43keV

0.26keV

1018

1020

1022 b) 0.9keV1.3keV

1.9keV2.4keV

Neu

tral

Flu

x [a

.u.]

MHD burst excited by energetic particle driven GAM is observed in the very low density plasma of 5x1017 m-3.

With RMP

X-point of island

Carbon radiation distribution by EMC3-EIRENE

Without RMP

Inboard side

LOS

of

mea

sure

men

ts

Ch1

Ch16

AX

UV

_P7.

5_D

1.0_

3.0_

C1_

NU

P7.

0_D

P1D

T3_t

emp1

6_2m

m_N

PIN

2 11

:15:

07 2

012/

10/0

2

AX

UV

(a.u

.) t=

3.0-

3.5s

Inte

nsi

ty(m

W/m

**2)

Cha

nnel

Experiments

Simulation

Intensity (mW/m2

)

Cha

nnel

AX

UV

_P08

_D1.

0_3.

0_C

1_N

UP

7.0_

DP

1DT

3_te

mp

18_2

mm

_NP

IN2

11:0

0:57

201

2/10

/02

AX

UV

(a.u

.) t=

4.0-

6.0

sIn

ten

sity

(mW

/m**

2)

Cha

nnel

Intensity (a.u.) C

hann

el

Stabilization of Detachment Plasma by RMP

M. Kobayashi OV/4-4

K. Mukai EX/P6-25 N. Ohno EX/P6-26 G. Kawamura TH/P6-39

Detached plasma without RMP

Detached plasma with RMP

The detached plasma can be stabilized by localizing the radiation spot at the X-point of magnetic island produced by the RMP.

The chord-integrated intensity has a reasonable agreement with the experimental results.

unstable

stable

��

��

��

��

�

�� �� � � �� ���

� �����������

�����������������

i-root�e-root�

Accumulation�

w/o Accumulation�

w/o Accumulation�

����

�� �

impurity hole with high �Ti�

Ti(0)= 6keV�

Short confinement of tracers�

Long confinement of tracers�

���

PS regime�

Plateau regime�

Impurity behavior in LHD

S. Sudo EX/P6-32

Tungsten impurity transport : D. Kato MPT/P7-36 I. Murakami EX/P6-28.

Impurity accumulation takes place only in the narrow parameter space in LHD.

Impurity hole (very hollow carbon profile) is observed in the high performance plasma with ion ITB

Simultaneous achievement of high temperature and low impurity concentration is expected in the future heliotron/stellarator devices.

Er > 0

Er < 0

C6+ profile

Impurity hole

Er ~ 0

Summary 1 Transport analysis beyond power-balance provides new understanding of transport ●The time scale of sign flip of momentum/particle non-diffusive term is shorter than the time scale of profile change. à non-diffusive term is directly connected to turbulence states rather than gradients. ●Heat pulse driven by MECH shows “fast” and “long” front propagation à Diffusive transport model is too simple to explain the front propagations. 2 MHD–Transport interaction ●Stochastic magnetic field ---- stabilizes the MHD instability ●Magnetic island ---- change the transport (not just flattening but playing a role of transport barrier) 3 Detachment and impurity control ●RMP --- stabilizing the detached plasma by controlling radiation spot ●Stochastic magnetic field at the plasma edge --- reducing impurity influx

Posters 10/14 (Tue.)

H. Wang TH/P1-12 10/15 (Wed.)

M. Jakubowski (IPP) EX/P3-4 10/16 (Thur.)

T. Seki FIP/P5-3 K. Mukai EX/P6-2 N. Ohno EX/P6-2 M. Yokoyama EX/P6-27 I. Murakami EX/P6-28 S. Ohdachi EX/P6-2 M. Yoshinuma EX/P6-30 K. Fujii EX/P6-31 S. Sudo EX/P6-32 M. Shoji EX/P6-33 T. Shimozuma EX/P6-34 Y. Narushima EX/P6-35 X. Du EX/P6-36 S. Sakakibara EX/P6-37 S. Murakami TH/P6-38 G. Kawamura TH/P6-39 A. Ishizawa TH/P6-40

10/17 (Fri.) M. Nunami TH/P7-9 D. Kato MPT/P7-36 T. Goto FIP/P7-16 M. Emoto FIP/P8-28 T. Akiyama FIP/P8-31

0rals 10/14 (Tue.)

T. Evans (GA) EX/1-3 DIII - LHD joint experiment M. Kobayashi OV/4-4 Topical Overview

10/15 (Wed.) S. Inagaki (Kyushu Univ.) EX/2-1 MECH experiment

10/16 (Thur.) A. Dinklage (IPP) FIP/3-1 W7X – LHD – TJ-II joint paper

10/17 (Fri.) H. Kasahara EX/7-3 Long pulse operation K. Nagaoka PPC/2- 1 High Te and Ti K. Ichiguchi TH/6-2 MHD with RMP M. Osakabe EX/10-3 EGAM

List of LHD presentation