Docstoc

EFDA confinement

Document Sample
EFDA confinement Powered By Docstoc
					Recent progress in the analysis in
the JET hybrid scenario discharges




Jörg Hobirk
25th September 2007 JET Seminar
Jörg Hobirk     1 (45)   JET seminar, HOW   25th September 2007
                             Thanks to …


 E. Joffrin, I. Voitsekhovitch, O. Zimmermann, M. Brix,
 C. Giroud, B. Alper, P. Beaumont, A. Boboc, J.
 Brzozowski, J. Bucalossi, R. Buttery, C. Challis, V.
 Cocilovo, S. Conroy, Y. Corre, E. de la Luna, P. de
 Vries, A. Ekedahl, N. Hawkes, S. Jachmich, K. Kirov,
 H.R. Koslowski, H. Leggate, T. Loarer, P. Lomas, G.
 Maddison, D. McDonald, A. Meigs, I. Nunes, V.
 Parail, K. Rantamaki, C. Sozzi, M. Valisa, F. Villone,
 K.D. Zastrow, M. Zerbini, V. Zoita



Jörg Hobirk   2 (45)   JET seminar, HOW    25th September 2007
                                 OUTLINE


         • Definitions of a hybrid scenario and
           examples
         • Confinement improvement or not?
         • The q-profile and its consequences
         • How to proceed?
         • Summary




Jörg Hobirk    3 (45)   JET seminar, HOW   25th September 2007
                       q profile in Hybrid is formed,
                            in H-mode relaxed
                          Ip=1.4MA, BT=1.7T




Jörg Hobirk   4 (45)     JET seminar, HOW      25th September 2007
              Hybrid mode with q-profile shaping
                             •Hybrid pulse in Mode B with LH
                             preheat
                             •Target q0 after LH preheat and
                             current ramp between 1 and 1.5
                             and a profile which might be
                             slightly reversed
                             •Low power NBI phase for slow
                             relaxation of q-profile without
                             significant MHD
                             •High power NBI phase to reach
                             target beta and to keep q
                             stationary by adding off-axis
                             bootstrap current, timing depends
                             on MHD stability and q-profile
                             shape
Jörg Hobirk   5 (45)   JET seminar, HOW          25th September 2007
                                     Definitions…
      From ITER (Green, 2003, PPCF):
      A hybrid mode of operation, in which a substantial fraction of the plasma
      current is driven non-inductively by current-drive power and the bootstrap
      effect, is a promising route towards the establishment of true steady-state
      modes of operation.
      A Candidate (ITPA): Improved H-mode from AUG because it combines high
      q95 with improved confinement properties to that the performance reduction is
      limited
      • q0 above (around, slightly below,… ) 1
      • No (or mild, or small) sawteeth
      • Does have mild (no) NTMs
      • Broad q-profile, low shear
      • Has improved core (or pedestal, or none)
        confinement
      • Operates at high beta (above 1 or 2 or 2.5,….)
Jörg Hobirk         6 (45)    JET seminar, HOW                  25th September 2007
                                                Confinement at high normalised pressure

    No advanced tokamak data in scaling                                                                         hybrid scenario data different
                               4 From H-mode database                               0.5


                                           ITER
Normalised pressure: bN (TH)




                                                                    Advanced                                        DIII-D




                                                                                     Measured confinement [s]
                                         advanced
                               3                                    Scenario                                      (bNTH<2.7)
                                         scenario
                                                                                                                                                      DIII-D
                                                                                                                                        JET
                               2     ITER                                                                                            (bNTH<2.2)
                                    H-mode
                                                                                                                                                       AUG

                               1                                                                                          AUG
                                                                                                                        (bNTH<3.2)
                                            JT-60   JET    DIII-D    AUG                                                                                JET
                               0                                                   0.03
                               0.001                          0.01                     0.03                                                   0.5
                                                                                                                 Confinement scaling [s]
                                       Normalised ion Larmor radius                                                                           Wade Nuc Fus 2005
                                                                                                                                              & Sips Nuc Fus 2007
                                   Confinement dependence at high pressure is an issue

                                       E. Joffrin, EPS 2007
                   Jörg Hobirk                            7 (45)           JET seminar, HOW                                           25th September 2007
                 On ASDEX Upgrade significant
              confinement improvement is observed
                              •Target q0 after current ramp
                              between 1 and 1.5
                              •Low power NBI phase for slow
                              relaxation of q-profile without
                              significant MHD
                              •High power NBI phase to reach
                              target beta
                              •q0 stationary above 1, backed up
                              by non 1/1 MHD activity in plasma
                              •Confinement significantly
                              enhanced without ITB




Jörg Hobirk    8 (45)   JET seminar, HOW         25th September 2007
                 Improved confinement in the pedestal ?


              ASDEX Upgrade                                             DIII-D



                      plasma pressure


                                            core
                                                                ELMs


                                            pedestal

C. Maggi, EPS 2005
                                        0                         a
                                             plasma radius
                     No clear signs of improved pedestal confinement


Jörg Hobirk                9 (45)            JET seminar, HOW          25th September 2007
              Summary of JET hybrid experiments


   • Most experiments carried out in HT3 
     high natural density and collisionality
   • ICRH based experiments done at high
     field and current  low beta
   • Current varied between 1.1MA and 2.3MA
   • TF varied between 1.4 and 3.35T
   • q95 between 2.7 and 5
    Large variation of plasma parameters in
     data base

Jörg Hobirk    10 (45)   JET seminar, HOW   25th September 2007
                                     Hybrid database extended to low q95


               2006-2007 data

                                        H-mode
G= H.bN/q952




                                       (mode D)
               q95=2.7
                                                       • Hybrid performance similar to H-mode
                                 q95=3.2
                                                       • Wide range in q95
                                                       •Upper and right end of plot power limited
                                      q95=4.0




                          2003 data


                            ~ bp




         Jörg Hobirk            11 (45)         JET seminar, HOW               25th September 2007
                     Density range expanded for the hybrid regime



      Filled symbols: fuelled discharges
H98
                                           q95=4
                                       q95=3.2     • Only HT3 pulses included here
                                       q95=2.7     • Greenwald fraction increased to
                                                   above unity
                                                   • Also confinement slightly better
                                                   than in 2003
                                                   • Confinement stays good up to
                                                   high densities

                    2002/3 data

                n/nG



 Jörg Hobirk             12 (45)       JET seminar, HOW              25th September 2007
              Hybrid modes, similar performance to H-modes

                               •Hybrid pulse in Mode B with LH
                               preheat, H-mode pulse in Mode D
                               •Even at higher current li becomes
                               very similar in late phase (q95=2.7)
                               •NBI power slightly higher in H-
                               mode
                               •Energy the same
                               •ELMs the same
                               •Time evolution slightly different



                                Within measurement
                               accuracy H-mode pulse and
                               hybrid pulse are the same
Jörg Hobirk     13 (45)   JET seminar, HOW          25th September 2007
              Profiles of H-mode and Hybrid almost
                             identical




Jörg Hobirk   14 (45)   JET seminar, HOW   25th September 2007
              Profiles of H-mode and Hybrid almost
                             identical




                                 •All profiles identical (some
                                 problems with Ti)
                                 •All global values identical
                                 •Transport identical?




Jörg Hobirk   15 (45)   JET seminar, HOW            25th September 2007
              Confinement does (or not) depend on beta?

                                           •Database includes L-modes
                                           and type III H-modes
                                           •ICRH FP not included in H
                                           factor calculation




•Lower boundary has a trend, but upper boundary not
•Maximum H factors around 1
Jörg Hobirk   16 (45)   JET seminar, HOW            25th September 2007
                                Summary

• An overview over the hybrid experiments in
  C15-C18 has been given
• Hybrid scenario has been extended to low
  q95=2.7 and to type III ELMs
• Going down in q95 increases the normalised
  performance H89*bN/q952
• Differences in performance between hybrid and
  H-mode small in a one to one comparison



Jörg Hobirk   17 (45)   JET seminar, HOW   25th September 2007
                    At least one question left…




              What’s about q?



Jörg Hobirk    18 (45)   JET seminar, HOW   25th September 2007
                        q < 1 in early phase of hybrid




   •q below 1 even for 46s
   •Different q-profile diagnostics agree within reason
Jörg Hobirk   19 (45)      JET seminar, HOW     25th September 2007
                        q < 1 in early phase of hybrid


                                              •What is q=1 radius, n=1
                                              or FB ?
                                              •Outer n=3 radius does
                                              not fit
                                              •Centre of n=3 activity
                                              does not agree with
                                              plasma centre




   •q below 1 even for 46s (1.5s after NBI)
   •q profile does agree with MHD mode locations
   •Different q-profile diagnostics agree within reason
Jörg Hobirk   20 (45)      JET seminar, HOW            25th September 2007
              Hybrid and H-mode q-profiles are identical

                                             •Comparison at maximum
                                             beta
                                             •Comparison at identical
                                             timing relative to start of NBI
                                             •Sawtooth inversion radius for
                                             H-mode not well defined,
                                             wrong time (2s early), pivot
                                             point is a region




    •In high heating phase q profiles are identical

Jörg Hobirk     21 (45)   JET seminar, HOW            25th September 2007
              Hybrid and H-mode q-profiles are identical

                                             •Comparison at maximum
                                             beta
                                             •Comparison at identical
                                             timing relative to start of NBI
                                             •Sawtooth inversion radius for
                                             H-mode not well defined,
                                             wrong time (2s early), pivot
                                             point is a region
                                             •MSE and FR agree but
                                             scatter allows for variation
                                             between the shots




    •In high heating phase q-profiles are similar

Jörg Hobirk     22 (45)   JET seminar, HOW            25th September 2007
                        q reaches final value early


                                            •Different diagnostics
                                            and EFIT runs do agree
                                            within reason
                                            •MHD constrains do not
                                            allow to discriminate
                                            between EFIT runs




   •q is constant from 47s on
   •Occurrence of MHD modes do agree with time
   dependence
Jörg Hobirk   23 (45)    JET seminar, HOW           25th September 2007
              Current diffusion converges q to same state


                                            •Similarity in q more
                                            pronounced in MSE q
                                            •At high beta all q-profile
                                            reconstructions agree




   •Initial q very different, LH preheat vs. long diffusion
   •Current diffusion quickly results in same q0
Jörg Hobirk    24 (45)   JET seminar, HOW             25th September 2007
                MHD similar in hybrids and H-modes




   •n=1 dominating the spectrum
   •n=1 consists of sawteeth, fishbones and precursors
   •n=3 only present in hybrid
   •Sawteeth stabilised at high beta, likely due to fast particles

Jörg Hobirk     25 (45)   JET seminar, HOW           25th September 2007
              q-profile agreement not always good




   •Here q from MSE and FR do disagree over large
   part of the profile
Jörg Hobirk   26 (45)   JET seminar, HOW   25th September 2007
              FR q-profile depends on density reconstruction
                                             •Faraday data good reproduced
                                             •Interferometer line integrals
                                             good reproduced (ch 2 missing)
                                             •Major disagreement between
                                             LIDAR and inversion




 •Density reconstruction plays key role for Faraday Rotation

Jörg Hobirk     27 (45)   JET seminar, HOW              25th September 2007
              Only few pulses with q0 larger than 0.9

                                           •Consistency between FR
                                           and MSE not always good
                                           •Underestimation of q in FR
                                           likely because of density
                                           reconstruction
                                           •3He and electron heated
                                           hybrids have bad data




•Most pulses have a more H-mode like q-profile
•At higher q0 more disagreement between diagnostics
Jörg Hobirk   28 (45)   JET seminar, HOW            25th September 2007
                        Confinement drops with q0?

                                             •Large scatter of H factors
                                             •q’s at low values doubtful
                                             •ICRH FP not included in H
                                             factor calculation
                                             • most high q points have low
                                             beta and low triangularity




                                                     FR


•Confinement data has no trend up to 1
•Confinement degrades above 1?
Jörg Hobirk   29 (45)     JET seminar, HOW            25th September 2007
                        Confinement drops with q0?

                                             •Large scatter of H factors
                                             •ICRH dominated and most
                                             3He pulses have bad MSE
                                             data
                                             •Very few points left above 1




                                                     MSE


•Confinement data has no trend up to 1
•Confinement degrades above 1?
Jörg Hobirk   30 (45)     JET seminar, HOW            25th September 2007
              q information before NBI not reliable




    •Tendency with increasing LH power not clear
    •Very large scatter, quality of q-profile?
Jörg Hobirk   31 (45)   JET seminar, HOW   25th September 2007
                   MHD onset and q dependence?

                                           •Database includes L-modes
                                           and type III H-modes




•Threshold is lower with higher starting q?
•Threshold for n=3 is lower than n=2
Jörg Hobirk   32 (45)   JET seminar, HOW            25th September 2007
              Influence of preheat on final q is small

                                           •Data is not sufficiently
                                           validated (FR only) for start of
                                           the NBI heating
                                           •Data might be spoiled by
                                           density profile changes




                                                    FR


•Small tendency for q in main heating to increase
•Slope very small
Jörg Hobirk   33 (45)   JET seminar, HOW             25th September 2007
              Influence of preheat on final q is small

                                           •Data is not sufficiently
                                           validated (FR only) for start of
                                           the NBI heating
                                           •Data might be spoiled by
                                           density profile changes




                                                    MSE


•Small tendency for q in main heating to increase
•Slope very small (y=0.752+0.054*x)
Jörg Hobirk   34 (45)   JET seminar, HOW             25th September 2007
              TRANSP qualitatively agrees for late phase



                                           •Run starts late (45s)
                                           •Central part of profile
                                           “frozen” for 1s (Wrong
                                           history?)
                                           •Time before 45s cannot
                                           be reproduced
                                           •Data before 45s
                                           questionable (density
                                           profile evolution)


   •Drop in central q similar
   •Agrees for having an early and low final q

Jörg Hobirk   35 (45)   JET seminar, HOW            25th September 2007
                 Bootstrap current low in first phase




         •Bootstrap current only 25%
         •Ohmic current dominates current profile
         •Total current very peaked
Jörg Hobirk     36 (45)   JET seminar, HOW   25th September 2007
              Bootstrap current low in second phase




    •Bootstrap current only 31%,FP not resulting in BS
    •Ohmic current still dominates current profile
    •Total current peaked
Jörg Hobirk   37 (45)   JET seminar, HOW   25th September 2007
               On ASDEX Upgrade the q0 is stabilised




•Initial drop similar in calculation and measurement
•After brief small q phase the experimental q increases
•ASTRA simulation does not reproduce time dependence!
 Jörg Hobirk   38 (45)   JET seminar, HOW   25th September 2007
                  S. Günter et al, NF 39, 1999
Jörg Hobirk   39 (45)   JET seminar, HOW         25th September 2007
                  S. Günter et al, NF 39, 1999

Jörg Hobirk   40 (45)   JET seminar, HOW         25th September 2007
Jörg Hobirk   41 (45)   JET seminar, HOW   25th September 2007
                 Higher q95 raises q0 but H98 lower




                           •At q95=5 more easily q>1 obtained
                           •No NTMs/FB or Sawteeth in #69369
                           • H98=0.8 compared to 0.95 at q95=4
                           •Same current


Jörg Hobirk   42 (45)   JET seminar, HOW        25th September 2007
                              Where to go?
   • q drops very fast  more heating power in
     first phase ?
   • BS small  change of density or current?
   • Additional current drive necessary ?
       – LH in main heating
       – Optimise NBI injection (Off-axis) but NBCD in
         question and duct limits might prevent this
   • Confinement improvement? q-profile
     dependence?
   • Similarity experiment with AUG or DIII-D at
     high q0 ?
Jörg Hobirk    43 (45)   JET seminar, HOW    25th September 2007
               simulations of Hybrid shot 67976 (F. Imbeaux)

              Experiments by A Ekedahl, K Rantamäki
  Current diffusion calculated with
  CRONOS (t=45s to t=53s)                     Typical LHCD during main heating
  LHCD calculated by DELPHINE
  for: n// = 1.8, n// = 2.3, and no LH
  in main heating phase.

  •100% accessibility for n//=2.3.
  Similar power deposition as                      n// = 2.3
  n//=1.8 + peak on-axis.
  ILH = 160 kA (IP = 1.4MA)                                    n// = 1.8

 (However, simulations give poor
 agreement on VLoop and li
 Need for more non-inductive
 current and broader current profile,       LH power deposition at  ~ 0.6-0.8
 or problem with Zeff ?)                    ILH / IP ~ 10%

Jörg Hobirk         44 (45)     JET seminar, HOW                25th September 2007
                 Do we have a hybrid at JET?
 • Operationally? Yes, with LH preheat and different
   beta phases all criteria are fulfilled
 • MHD? Yes, most pulses have no or very small
   sawteeth, NTM activity is there but confinement can
   still be okey.
 • Beta? Yes, up to bN=3.6 has been reached,
   discharges are stable against 2/1
 • q-profile? No, almost all pulses have a relaxed
   current profile
 • Confinement? Indefinite, all pulses have more or
   less H-mode confinement but connection to q-profile
   unclear
Jörg Hobirk   45 (45)   JET seminar, HOW   25th September 2007
                        Thanks to everyone!


 • Thanks to everyone in the JET team
 • It was a pleasant time with you
 • I wish you success with the next campaigns

                        Good bye.



Jörg Hobirk   46 (45)   JET seminar, HOW   25th September 2007
                Hybrid modes at low q95~2.7 reach bN~3



                              •Very good MHD stability at q95=2.7
                              •Standard type I ELMs
                              •Good confinement
                              •High normalised density
                              •H89*bN/q952 =0.72 i.e. almost 2 times
                              ITER target
                              •But sawteeth present, is it a real
                              hybrid?
                              •TF M experiments show that
                              operation up 20% above 4*li is
                              possible without RWM



Jörg Hobirk   47 (45)   JET seminar, HOW           25th September 2007
                   Hybrid type III ELM scenario
#68515 Reference High D fuelling (D=5×1022)
#68532 Radiative feedback (D=5×1022+N=3 ×1022)      Radiative feedback: frad=0.7


                Total: D+N


                                                         Strong gas injection is
                                                        required to get Type III
                 GIM9: D                                     ELM ~1023e.s



                                  Type III ELMs




                                                    By Y. Corre TF S1 meeting 7/12/06
Jörg Hobirk          48 (45)     JET seminar, HOW                 25th September 2007
                     How much does it costs?
 #68507 Reference low D fuelling
 #68532 N-seeded
                     bN
                                            2.5


Price to pay
                   H98Y



                               Zeff (vertical)


         +2MW
                                                  Ptot




                                                  By Y. Corre TF S1 meeting 7/12/06
Jörg Hobirk        49 (45)   JET seminar, HOW                  25th September 2007
                      What is the benefit?
 Type III:                     f=500hertz
 Type I compounds :            f~50hertz


               H (divertor)


                                                     #68507 Reference
                                                     #68532 N-seeded




                                             By Y. Corre TF S1 meeting 7/12/06
Jörg Hobirk    50 (45)    JET seminar, HOW                 25th September 2007
              Hybrid modes, similar performance to H-modes

                               •Hybrid pulse in Mode B with LH
                               preheat, H-mode pulse in Mode D
                               •Even at higher current li becomes
                               very similar in late phase
                               Need higher current and power
                               to decouple bN from bpol

                               •NBI power slightly higher in H-
                               mode
                               •Energy the same
                               •ELMs the same
                               •Time evolution slightly different
                                Within measurement
                               accuracy H-mode pulse and
                               hybrid pulse are the same
Jörg Hobirk     51 (45)   JET seminar, HOW           25th September 2007
                        Density in hybrid higher




Jörg Hobirk   52 (45)     JET seminar, HOW   25th September 2007
                    MHD very similar at high beta




              Hybrid                          H-mode
 •Small 4/3 NTM                           •Small 4/3 NTM
 •Fishbones                               •Fishbones
 •Sawteeth                                •Sawteeth
 •5/4 NTM developing                      •5/4 NTM developing
Jörg Hobirk      53 (45)   JET seminar, HOW          25th September 2007
              Confinement - affect of beta normalised?




D.C. McDonald, TF S1 meeting, 15th June 2006, only 2003 data
Jörg Hobirk    54 (45)   JET seminar, HOW     25th September 2007
                 Hybrid scenario: improved confinement ?
Most hybrid discharges have higher b than normal H-modes!
  In AUG, DIII-D and JET the confinement of                 Wth exp [MJ]
 the Hybrid scenario is improved by 20-40% with         2.8 (for BT=1.7T & bN>2)
 respect the standard scaling (H=1.2 up to 1.4):        2.6
                                                                           IPB98y2
                                                        2.4
     tstandard ~ *-2.70 b-0.90 n*-0.01 q-3.0          2.2
                                                                                          Pure
                                                         2                             gyro-Bohm
               Strong b degradation !                   1.8
                                                        1.6
  scaling better suited to high bN discharges:         1.4
 (electrostatic scaling):
                                                        1.2
   tnew ~ *-3 b0 n*-0.1 q-1.7                          1
                                                              1      1.5         2         2.5     3
                                                                           Wth scaling [MJ]

     At high bN, the higher confinement observed in hybrid regime could
      be related to strong the b-0.90 dependence of the standard scaling
                                                                  From: E. Joffrin, ICPP 2006
Jörg Hobirk           55 (45)        JET seminar, HOW                        25th September 2007
                    Degradation of confinement with bN ?

              BtE  1/bN
                                              • Three point scan in Hybrid
                                              achieved with good matches in
         H98=0.99                             *, n*, q, Ti/Te, ….

B tE
                                              • Normalised tE falls by 35% as
                                              b rises by 40%
                    H98=0.98


                                H98=0.94      • Fall in tE with b in line with
                                              IPB98(y,2) and in contradiction
                                              to H-mode results
                        bN th

                                     From L. Laborde, TF S1 meeting 30/11/06

Jörg Hobirk         56 (45)     JET seminar, HOW              25th September 2007
                    Definition of hybrid unclear
   • All different ITPA groups dealing with
     hybrids use different definitions
   • DIII-D, AUG and JET have hybrid pulses
     with sawteeth
   • Confinement improvement in JET not
     obvious in data, b dependence of ITER
     extrapolation not clear
   A Definition:
   The hybrid scenario is an H_mode regime operated at
   high normalised pressure (bN>2.5) minimising the
   impact of sawteeth on stability by using a broad q profile
   above unity in the plasma core (ITPA SSO group).
Jörg Hobirk    57 (45)   JET seminar, HOW       25th September 2007

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:8
posted:3/16/2012
language:
pages:57
jennyyingdi jennyyingdi http://
About