Effect of resonant magnetic perturbations on low collisionality - - PowerPoint PPT Presentation

▶
effect of resonant magnetic
SMART_READER_LITE
LIVE PREVIEW

Effect of resonant magnetic perturbations on low collisionality - - PowerPoint PPT Presentation

Oct 27, 2023 183 likes •701 views

Effect of resonant magnetic perturbations on low collisionality discharges in MAST and a comparison with ASDEX Upgrade Andrew Kirk on behalf of I. Chapman, Yueqiang Liu, C. Ham, J.R. Harrison, S. Pamela, D. Ryan, S.Saarelma, R. Scannell,

slide-1
SLIDE 1
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

1

Andrew Kirk

  • n behalf of
  • I. Chapman, Yueqiang Liu, C. Ham, J.R. Harrison, S. Pamela, D. Ryan, S.Saarelma, R. Scannell,

A.J.Thornton, M. Valovic CCFE

  • W. Suttrop, T. Eich, M. Dunne, C. Fuchs, B. Kurzan, R. Fischer, R McDermott , B. Sieglin, E. Viezzer

Max-Planck Institut für Plasmaphysik Garching

  • M. Jakubowski

Max-Planck Institut für Plasmaphysik, Greifswald Yunfeng Liang FZ Julich

  • P. Cahyna, M. Paterka

EURATOM/IPP.CR, Prague

CCFE is the fusion research arm of the United Kingdom Atomic Energy Authority

Effect of resonant magnetic perturbations on low collisionality discharges in MAST and a comparison with ASDEX Upgrade

slide-2
SLIDE 2
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

2

Motivation

  • The natural type-I ELMs frequency in ITER is predicted to be too low to avoid either

W accumulation at low IP or damage to PFCs at high IP Required increase in ELM frequency

slide-3
SLIDE 3
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

3

  • The natural type-I ELMs frequency in ITER is predicted to be too low to avoid either

W accumulation at low IP or damage to PFCs at high IP

  • One technique that has been shown to reduce the size of ELMs is the application of

Resonant Magnetic Perturbations (RMPs)

  • Need to understand how RMPs control ELMs to make predictions for ITER – a good

way of doing this is by making in depth comparison across devices Situation at IAEA 2012

Motivation

slide-4
SLIDE 4
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

4

  • The natural type-I ELMs frequency in ITER is predicted to be too low to avoid either

W accumulation at low IP or damage to PFCs at high IP

  • One technique that has been shown to reduce the size of ELMs is the application of

Resonant Magnetic Perturbations (RMPs)

  • Need to understand how RMPs control ELMs to make predictions for ITER – a good

way of doing this is by making in depth comparison across devices New low collisionality results from AUG and MAST

Motivation

slide-5
SLIDE 5
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

5

Non-axisymmetric coil system

Can produce configurations n=1,2 or 4 MAST is equipped with

  • 6 coils in the upper row
  • 12 coils in the lower row

AUG is equipped with

  • 2 rows of 8 coils each

Can produce configurations n=1,2,3,4 or 6

slide-6
SLIDE 6
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

6

Examples of ELM mitigation - MAST

All nRMP can mitigate ELMs Very small window for n=2 between ELM mitigation and H-L transition RMPs cause a density pump out and braking of toroidal rotation

slide-7
SLIDE 7
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

7

Examples of ELM mitigation - AUG

ELM mitigation has been achieved with nRMP=1, 2 and n=4 magnetic perturbations Sustained ELM mitigation demonstrated with nRMP=2 and 4 RMPs cause a density pump out and braking of toroidal rotation W Suttrop EX/P1-23

slide-8
SLIDE 8
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

8

Examples of ELM mitigation - AUG

ELM mitigation has been achieved with n=2 and n=4 magnetic perturbations

ELM mitigation not suppression as very small high frequency (800 Hz) ELMs remain

  • But are they type I?
slide-9
SLIDE 9
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

9 9

Pros of mitigation

target heat load ELM mitigation decreases: DWELM

slide-10
SLIDE 10
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

10 10

The problem is that a density pump out occurs across the entire plasma while T

e ~ constant

  • leading to a large drop in

confinement

Cons of mitigation

slide-11
SLIDE 11
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

11

Minimising the effect of the RMPs

  • n confinement
slide-12
SLIDE 12
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

12

Minimising the density pump out - MAST

The application of RMPs to a shot that is not fuelled in the H-mode period leads to an increase in ELM frequency and reduction in the plasma density

slide-13
SLIDE 13
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

13

Minimising the density pump out - MAST

The application of RMPs to a shot that is not fuelled in the H-mode period leads to an increase in ELM frequency and reduction in the plasma density The density pump out is often large enough to lead to a back transition to L-mode

slide-14
SLIDE 14
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

14

Using a feed forward waveform and slow IRMP ramp can keep at constant density Application of n=6 RMPs to LSND

Minimising the density pump out - MAST

slide-15
SLIDE 15
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

15

Using a feed forward waveform and slow IRMP ramp can keep at constant density Application of n=6 RMPs to LSND

Minimising the density pump out - MAST

M Valovic EX/P4-36 Also possible using pellets

slide-16
SLIDE 16
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

16

Restoring the density - MAST

The density and temperature profiles show that not only has the core density been recovered but also the edge density The ELM averaged line average density and stored energy are similar So mitigation achieved with little effect on stored energy Application of n=6 RMPs to LSND

slide-17
SLIDE 17
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

17 17

Pressure pedestal evolution - MAST

Natural ELM cycle – pressure pedestal evolves to a maximum value determined by the Peeling Ballooning modes stability boundary just before ELM crash P-B boundary

slide-18
SLIDE 18
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

18 18

Application of RMPs leads to 3D distortions of plasma shape

  • > produces regions of enhanced

ballooning mode instability – reducing the PB boundary and hence triggering type I ELMs at lower Pped

Pressure pedestal evolution - MAST

Infinite n ballooning stability calculated using COBRA from a VMEC equilibrium

C Ham et al., ‘Tokamak equilibria and edge stability when non-axisymmetric fields are applied ‘ submitted to PPCF

P-B boundary

slide-19
SLIDE 19
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

19 19

Pressure pedestal evolution - MAST

Previously observed

  • n MAST

Application of RMPs leads to 3D distortions of plasma shape

  • > produces regions of enhanced

ballooning mode instability – reducing the PB boundary and hence triggering type I ELMs at lower Pped

slide-20
SLIDE 20
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

20 20

Pressure pedestal evolution - MAST

So how can Pped stay the same and yet fELM increases? Application of RMPs leads to 3D distortions of plasma shape

  • > produces regions of enhanced

ballooning mode instability – reducing the PB boundary and hence triggering type I ELMs at lower Pped

slide-21
SLIDE 21
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

21 21

Pressure pedestal evolution - MAST

If the pedestal evolved to a saturated value early in the ELM cycle

slide-22
SLIDE 22
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

22 22

Pressure pedestal evolution - MAST

Then could increase fELM at almost constant Pped

slide-23
SLIDE 23
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

23 23

Pressure pedestal evolution - MAST

Then could increase fELM at almost constant Pped For these shots on MAST Pe

ped spends a

large amount of times near to a saturated value during the ELM cycle and the mitigated ELMs are triggered near to the point at which the maximum is obtained

slide-24
SLIDE 24
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

24 24

Pressure pedestal evolution - MAST

Then could increase fELM at almost constant Pped It is likely that if the frequency was increased further then the peak Pe

ped

reached would be reduced

slide-25
SLIDE 25
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

25 25

Pressure pedestal evolution - MAST

Then could increase fELM at almost constant Pped Note: Max Pe

ped prior to ELM

AND Min Pe

ped after ELM are similar in

natural and mitigated ELMs So why is DWELM so different?

slide-26
SLIDE 26
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

26 26

Pedestal affected area - MAST

The ELM affected area is much smaller for the mitigated ELMs Dne(R) = ne

before ELM(R) – ne after ELM(R)

slide-27
SLIDE 27
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

27

Parameters determining the

  • nset of ELM mitigation
slide-28
SLIDE 28
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

28

Vacuum resonant field - MAST

Normalised resonant radial field component (br

res) in the vacuum approximation

IP = 400 kA IP = 600 kA On MAST ELM mitigation scales ~ linearly with br

res above a threshold value

This threshold is scenario and nRMP dependent

slide-29
SLIDE 29
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

29

Vacuum resonant field - AUG

On AUG ELM mitigation scales ~ linearly with br

res above a threshold value

which is scenario and nRMP dependent 2013 2014

slide-30
SLIDE 30
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

30

Vacuum resonant field - AUG

2013 2014 On AUG ELM mitigation scales ~ linearly with br

res above a threshold value

which is scenario and nRMP dependent BUT there are some clear outliers

slide-31
SLIDE 31
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

31

Vacuum resonant field - AUG

Differential phase scan between the currents in the upper and lower coils W Suttrop EX/P1-23

  • > a pitch angle/equilibrium

field alignment scan Similar increase in fELM

  • bserved at Df = 90 and

180⁰ 90⁰ 180⁰=Odd

slide-32
SLIDE 32
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

32

Vacuum resonant field - AUG

But br

res (vacuum) very different

Similar increase in fELM observed at Df = 90 and 180⁰

slide-33
SLIDE 33
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

33

Plasma response - AUG

Included plasma effects using MARS-F, which is a single fluid linear MHD code that solves the full resistive MHD equations in full toroidal geometry – the code allows for plasma response and screening due to toroidal rotation to be taken into account Clear screening of resonant components br

res now similar for

90 and 180⁰

slide-34
SLIDE 34
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

34

Plasma response - AUG

Included plasma effects using MARS-F, which is a single fluid linear MHD code that solves the full resistive MHD equations in full toroidal geometry – the code allows for plasma response and screening due to toroidal rotation to be taken into account Preliminary results indicate that maximum ELM mitigation is obtained near to where the peeling response of the plasma is maximum Plasma response composed of kink (core) and peeling (edge) eigenfunctions

slide-35
SLIDE 35
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

35

ELM type during mitigation

slide-36
SLIDE 36
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

36

RMP off RMP on Natural and Mitigated ELMs look very similar

Effect of RMPs on ELM filaments - MAST

slide-37
SLIDE 37
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

37

RMP off RMP on An analysis of the mode number of the filaments suggests that:

  • the mitigated ELMs are still

type I ELMs

  • they are just smaller and

more frequent

Effect of RMPs on ELM filaments - MAST

slide-38
SLIDE 38
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

38

RMP off RMP on An analysis of the mode number of the filaments suggests that:

  • the mitigated ELMs are still

type I ELMs

  • they are just smaller and

more frequent

Effect of RMPs on ELM filaments - MAST

slide-39
SLIDE 39
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

39

RMPs producing small ELMs on MAST

Application of n=3 RMPs to a particular discharge in MAST caused a density pump out which resulted in the establishment of a small ELM regime

slide-40
SLIDE 40
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

40

RMPs producing small ELMs on MAST

Application of n=3 RMPs to a particular discharge in MAST caused a density pump out which resulted in the establishment of a small ELM regime – which had a different mode number

slide-41
SLIDE 41
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

41

RMPs producing small ELMs on MAST

Effect of RMPs on pedestal characteristic Pedestal characteristics compatible with type IV ELMs Type IV = low ne-high Te branch of type III

slide-42
SLIDE 42
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

42

Natural type IV ELMs on MAST

Without RMPs the naturally occurring type IV ELMs frequency increases with decreasing pedestal density

slide-43
SLIDE 43
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

43

Effect of RMPs on pedestal - AUG

The mitigated ELMs move to the region of the Pedestal operation space typically associated with type IV ELMs

slide-44
SLIDE 44
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

44

Effect of RMPs on pedestal - AUG

ELM mitigation increases as pedestal density is decreased Similar to the trend observed on MAST suggesting it may be a transition to type IV ELMs

slide-45
SLIDE 45
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

45

ELM mitigation or type I ELM suppression?

slide-46
SLIDE 46
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

46

However in at least some of the cases it appears there is a suppression of type I ELMs and a transition to different ELM type

ELM mitigation or type I ELM suppression?

slide-47
SLIDE 47
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

47

Regimes with tolerable ELMs can be established over a wide operating space in a range of devices

ELM mitigation or type I ELM suppression?

slide-48
SLIDE 48
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

48

  • Sustained ELM mitigation has been obtained at mid to low collisionality on

MAST and AUG using RMPs with a range of toroidal mode numbers resulting in

  • smaller ELMs (DW) and reduced peak heat loads (qpeak)
  • reduction in density and stored energy
  • On MAST in one type of discharge the drop in density has been eliminated resulting

in reduced peak divertor heat flux with minimal drop in confinement – the smaller ELMs being a result of a change in the region of the plasma affected by the ELM.

  • While the size of the resonant magnetic field component plays some role in

determining the onset of ELM mitigation – this cross machine comparison has clearly indicated the need for studying the effects of the plasma response.

  • There appears to be several mechanisms by which ELMs can be mitigated –

increasing the frequency of type I ELMs or a transition to a different ELM regime

Summary

slide-49
SLIDE 49
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

49

Backup slides

slide-50
SLIDE 50
  • A. Kirk 25th IAEA FEC, St Petersburg, Russia, October 2014

50 50

Fuelling with pellets – MAST

Possible to refuel to densities higher than original using pellets while still keeping mitigation 10 % drop in stored energy M Valovic EX/P4-36