Demonstration of sawtooth period control with EC waves in KSTAR - - PowerPoint PPT Presentation

demonstration of sawtooth period control with ec waves in
SMART_READER_LITE
LIVE PREVIEW

Demonstration of sawtooth period control with EC waves in KSTAR - - PowerPoint PPT Presentation

Demonstration of sawtooth period control with EC waves in KSTAR plasma J. H. Jeong , M. Joung, M. H. Woo, S. H. Hahn, W. S. Han, Y. S. Bae and J. G. Kwak National Fusion Research Institute, Daejeon, Korea D. H. Kim, T. Goodman and O Sauter CRPP


slide-1
SLIDE 1

Demonstration of sawtooth period control with EC waves in KSTAR plasma

  • J. H. Jeong, M. Joung, M. H. Woo, S. H. Hahn, W. S. Han, Y. S. Bae and J. G. Kwak

National Fusion Research Institute, Daejeon, Korea

  • D. H. Kim, T. Goodman and O Sauter

CRPP EPFL La sanne S it erland CRPP - EPFL, Lausanne, Switzerland

  • K. Sakamoto, K. Kajiwara and Y. Oda

Japan Atomic Energy Agency, Naka, Japan

  • W. Namkung and M. H. Cho

Department of Physics, POSTECH, Pohang, Korea

  • H. Park

School of Electrical and Computer Engineering, UNIST, Ulsan, Korea

  • J. Hosea and R. Ellis

Princeton Plasma Physics Laboratory, Princeton, USA

KSTAR Conference 2014

February 24~26, 2014, Gangwon-do, Korea Email address: jhjeong@nfri.re.kr

slide-2
SLIDE 2

Sawtooth instability

Destructive periodic magnetic instability (divided into three phases):

  • Ramp-up phase precursor phase Collapse

τsawtooth

  • Ramp-up phase precursor phase Collapse

ne & Te increase in the core MHD instability grows with internal kink structure ne & Te crashes toward

  • utside of q=1 surface

Sawtooth crash criterion [F. Porcelli et al, PPCF (1996)] : I. S1 > S1,crit

) (r q r s

1 1 1

where, ′ ⋅ = ˆ

* A

W τ ω δ

Id l MHD t bl h δW <0

II.

2

* 1 A i

S W τ ω δ π − <

Ideal-MHD unstable when δW <0

FAST KO MHD

ˆ ˆ ˆ ˆ W W W W δ δ δ δ + + =

2

slide-3
SLIDE 3

Sawtooth control actuator: localized ECCD

Change the magnetic shear near q=1 surface (S1=r1·q′ (r1)) using localized CD as follows;

  • Deposition location of ECCD power influences period
  • Direction of current change the period

F db k t l th i d th h th d iti

  • Feedback control: measure the period, then, change the deposition

( slow and determined by the performance of antenna mechanics)

  • Another method: sawtooth locking (or pacing)
  • Active control using modulated power

locking to the requested period

  • Possibility of open-loop control

S (q=1) (q ) ECCD

3

slide-4
SLIDE 4

Motivation: Link between sawteeth and TM/NTMs

Sawtooth lengthening by the ICRF resonance q ~ 1

(JET shot number of 51794)

Purpose of sawtooth lengthening Increase the n & T in the core by long inter crash time Increase the ne & Te in the core by long inter-crash time Result Increased sawtooth period big crash triggering

  • f n=2 MHD activity

Sauter O et al 2002 Phys. Rev. Lett. 88 105001

4

slide-5
SLIDE 5

Motivation: KSTAR & ITER

ITER will have large fusion born alpha particle population in the core

KSTAR #7950

population in the core Long sawtooth periods are expected

NB1 NB2 ECRH

Trigger NTMs are expected

ECRH ICRH

Sawtooth crash

☞ Sawtooth period has to be controlled with small & frequent (to avoid NTM triggering)

ECE @1.855 m

Triggering of n = 1 tearing modes in H-mode plasma

5

slide-6
SLIDE 6

Sawtooth behavior by the EC beam deposition position

Actuator:

  • 110 GHz, 0.4 MW X2 mode
  • co-CD 15 deg oblique injection

co CD 15 deg oblique injection βN=0.45, Ip= 0.4 MA

110 GHz ECH

us

#7072: Off-axis ECCD

#7072: Off-axis ECCD

nversion radiu

#7072: Off-axis ECCD

In rsion radius

#7070 O i ECCD

Ray tracing using

#7070:On-axis ECCD

Invers

#7070: On-axis ECCD

y g g Toray-GA code

6

slide-7
SLIDE 7

Sawtooth behavior according to the direction of CD

Actuator:

  • 110 GHz, 0.4 MW X2 mode
  • co-CD 20 deg oblique injection
  • Off-axis injection with z=10 cm in Rres

co CD 20 deg oblique injection βN=0.45, Ip= 0.4 MA q=1 surface q 1 surface r1=2.07, ρ=0.37 20 deg. cnt-CD injection 20 deg. co-CD injection

Stabilization: cnt-CD injection inside the q=1 surface Destabilization: co-CD injection inside the q=1 surface

7

slide-8
SLIDE 8

Preliminary results of sawtooth locking by the periodic

  • Preliminary results of sawtooth locking by the periodic

forcing of EC beam in KSTAR (year 2013)

– System upgrade: feed-forward control of antenna mirror system in poloidal direction

Feed-forward control of antenna mirror

counts] Requested position r position [c q p Measured position Launcher

latency ~ 100 ms

8

slide-9
SLIDE 9

Background & Motivation of sawtooth locking

ECH/CD is very efficient to modify the sawtooth period [1]

1 By tailoring the profiles of jCD which is associated by the magnetic shear on the q=1 surface (= s1) 1. By tailoring the profiles of jCD which is associated by the magnetic shear on the q=1 surface (= s1)

  • Experimental results are well corresponded with the other tokamaks (year 2012, using 110 GHz ECH)

2. Possible of injection locking by the modulated EC beam [2,3] (year 2013, using 170 GHz ECH)

[1] C. Angioni, T.P. Goodman, M.A. Henderson and O. Sauter, Nucl. Fus. 43 455 (2003) [2] T. P. Goodman et al., Phys. Rev. Lett. 106 245002 (2011) [3] G. Witvoet et al., Nucl. Fusion 51 103043 (2011).

slide-10
SLIDE 10

Activities on Real-time sawtooth period control in 2013 KSTAR

  • Implementation of Launcher Control System in PCS

PCS wave server

RECR Scan of vertical position To find optimized EC target position PCS wave server Vertical target position q=1 PCSrt1

RFM

cal Time

PCSrt5 (HICS)

RFM

Optic Antenna pivot: (2.8m, 0, -0.3m)

(HICS)

(launcher status, encoder) RS232, 115200 bps (100Hz/50Hz) TCP/IP

All PCSs have a real time processor and RFM. Preset time evolution of the vertical target position But calculate the angle with position in real time

Micro-PLC

DC motor motor ETOS ECH OPI (EPICS)

RS232 (19200 bps, 150 msec)

Driver

But calculate the angle with position in real time Transfer the angle to ECH local launcher controller based in PLC Mirror angle is changed by motor. Returned angle from encoder is also calculated in EC b t t iti t PCS i l ti

Micro PLC

Encoder T/L PLC

Antenna

Upgraded by FPGA (after campaign)

10

beam target position at PCS in real time. Block diagram for controlling antenna mirror by PCS

slide-11
SLIDE 11

EC beam poloidal scan (#9145)

Target position: just outside q=1 surface for stabilization of sawtooth

  • Poloidal mirror scan: z=10 cm ~ 30 cm (from mid-plane)
  • τsaw maximized at z~26 cm (= EC-beam deposit position for the sawtooth locking experiments)

KSTAR #9145 (EC sweeping)

Te [keV] EC [MW] τsaw [ms] Poloidal mirror scan EC [MW] τ Z [cm]

τsaw maximized with 100 ms at z=26 cm

Time [sec] Z

EC beam

EC start

11

EC start

slide-12
SLIDE 12

Sawtooth locking experiments (#9215)

EC beam Target position fixed at z=26 cm

170 GHz X2 at BT=2.9 T IP = 700 kA

g p

<ne> = 2.7 ×1019m-2 βN = 0.5 q95 = 5.43

ECE

locking locking

τsaw

Sawteeth are regulated at τsaw=43 msec

Sawtooth period is fully extended with τsaw=100 C C f

12

by EC modulation freq. of 23 Hz & 70 % duty

msec by CW EC beam injection outside q=1 surface

slide-13
SLIDE 13

Injection-locking range in KSTAR 2013

EC modulation period

BT=2.9 T, IP = 700 kA, <ne> ≈ 2.7 ×1019m-2 βN ≈ 0.5

  • 95 5 43

#9147 Non-locking Sawtooth locking range in KSTAR during 2013 campaign #9146

q95 ≈ 5.43

non- locking partially-locking (#9146) sawtooth period alternates two values #9146 #9147 #9146 partially locking locking two values #9215 p y g

τsaw

EC beam

Sawtooth locking range during 2014 campaign?

13

Modulation enabled at t=3.5 sec to 6.5 sec

g g g p g

slide-14
SLIDE 14

Future work: Injection-locking range in KSTAR ?

Wh t d d? What do we need?

We can find two X points: min τsaw, min w/o EC and max τsaw max with full EC power τsaw, max with full EC power We have to choose proper EC period and duty cycle for both locking and non-locking cases possibility of l ki i KSTAR b d locking in KSTAR can be proved We can obtain the condition of τsaw NTM triggering

We can control the sawtooth period in real-time to avoid NTM triggering

  • D. H. Kim, sawtooth locking range in TCV

gg g

g g

14

slide-15
SLIDE 15

Future plan:

Conceptual design & requirements for the real-time sawtooth control

Gyrotron RF modulation

Timing & synchronization for RF modulation RF modulation for sawtooth locking and pacing

reference τsaw signal PI controller

(d t i ti f

Feedback

KSTAR PCS

+/-

(determination of antenna angle & EC freq.)

Detect crash

(real-time τsaw determination)

ECHres

(ray-tracing)

Antenna

T fil

Corr- elation

rt- equili brium RINVERSION

Mag. & Flux Ne profile ECE

(T )

controller

Te profile Soft X-ray (detect crash)

Plasma

Flux

(B, Ψ)

profile

(Te)

(detect crash)

15

slide-16
SLIDE 16

Future plan (system upgrades)

Goal: real-time feedback control using ECCD to investigate relation between the sawtooth i d d t i i NTM

  • Feed-forward control of the ECCD mirror

(2013) period and triggering NTM (2013)

  • Conceptual design of real-time sawtooth

TCV #42357

p g control (2014)

  • Development of real-time sawtooth control

system by PCS

  • Upgrade of antenna for fast movements and
  • Upgrade of antenna for fast movements and

feed-back control (2015)

  • Validation & application of the real-time

feedback sawtooth control in KSTAR (2016)

16

Goodman T. P. et al 2011 Phys Rev. Lett., 106 245002

slide-17
SLIDE 17

Upgrade plan of ECH system

2014 2015 2016 2017 EC Power

170GHz, 1MW, 30s 170GHz, 1MW, 50s 105/140GHz, 1MW 170GHz, 1MW 105/140GHz, 2MW 170GHz, 1MW 105/140GHz, 3MW

Modulation

1 kHz Up to 5 kHz Up to 5 kHz Up to 5 kHz

U d ld l h

Launcher

Test steady state launcher (2* 1MW)

Upgrade old launcher Design of two beam launcher (2*1MW)

Fabricate two beam launc her (2*1MW, 2MW) 2*1MW, 2MW

Speed

Motor design New motor 10 degree/sec > 50 degree/sec > 50 degree/sec > 50 degree/sec

Beam width

r ~ 40mm at core Analysis of valuable beam size ? ?

St d t t

W t l d fi d d U d t l d All t l d St d t t

Steady state

Water cooled fixed and steering mirrors Upgrade water cooled All water cooled Steady state

Deposition position

Compare experiment with ray tracing code Make data table Te, ne profile, rtEFIT Implement ray tracing code in PCS

position

y g

e e p

Controller

Upgrade motor driver Power timing control < 100 msec New motor driver <10 msec <10 msec

17

slide-18
SLIDE 18

Summary and future work

Sawtooth experiments using ECCD has been demonstrated in KSTAR

b diff t i j ti d iti d di ti f CD

  • by different injection depositions and direction of CD

Sawtooth locking experiments has been demonstrated in KSTAR Sawtooth period can be controlled in accurate way by EC beam injection conditions Sawtooth period can be controlled in accurate way by EC beam injection conditions Future work

  • Investigation of locking-range in KSTAR by the scan of τEC and duty ratio (to apply sawtooth control)
  • Relations between the sawtooth period and triggering of TM/NTM will be investigated (to avoid

TM/NTM triggering) TM/NTM triggering)

  • Development of real-time feedback control of sawtooth period to avoid triggering TM/NTM instability

18