2005, M.Maraschek, IPP-Garching. ASDEX Upgrade

Transcription

1 ASDEX Upgrade Max-Planck-Institut für Plasmaphysik Control of core MHD Instabilities by ECCD in ASDEX Upgrade M. Maraschek (), G. Gantenbein (), S. Günter (), F. Leuterer (), A. Mück (), A. Manini (), H. Zohm (), and the ASDEX Upgrade-Team () () Max-Planck-Institut für Plasmaphysik, EURATOM Association, Boltzmannstr., D-8748 Garching, Germany () Institut für Plasmaforschung, Pfaffenwaldring, D-769, Stuttgart, Germany Introduction and motivation Sawtooth tailoring / avoidance with co / counter ECCD NTM stabilisation with co-eccd Deposition width scan Modulated ECCD Early application of ECCD FIR-NTMs and their triggering with ECCD Summary and future plans, M.Maraschek, IPP-Garching

2 Resonant surfaces and MHD modified with ECRH / ECCD #9,.99 s deposition q=, (flat or reversed in the centre in adv. scen.): sawteeth, fast particle driven fishbones, (/)-modes sawtooth tailoring avoidance of NTM trigger burn control q=4/: (4/) NTM, ideal (4/) modes during FIR-NTM artificially trigger / avoid (4/) FIR-NTM transition q=/: (/) NTM stabilisation and suppression of (/)-NTM ECRH beam ECRH deposition layer, controlled by Bt q=: (/) NTM, (/) classical current driven tearing modes stabilisation and suppression of (/)-NTM (j, ne, Te) profile tailoring for advanced scenarios control of current drive and depostion by Bt and toroidal and poloidal launching angle, M.Maraschek, IPP-Garching

3 typical "natural" time-development of an NTM PNI # W.E+7 E+7.4 E+6 βn βonset βsat βcrit W/ t [a.u.]. βp/lp FIR-NTM MRM:MPR() ρp=.6, m= 4 4 time (s) W [cm]. 4 βp,crit : onset at βp,onset > βp,crit, + a seed-island, such as sawteeth : growth to saturated size, proportional to βp,sat FIR-NTM phases at βn >. stabilisation experiments with ECCD : reduction of βp by heating until βp,crit is reached 4: βp = βp,crit, mode decouples from βp : βp < βp,crit for all times mode decays away again βp,sat βp,onset W [a.u] Wsat βp Wsat(FIR,ECCD) < βp, M.Maraschek, IPP-Garching

4 Sawtooth tailoring with ECRH / ECCD collaboration with T.P.Goodman, O.Sauter (CRPP) normalised sawtooth period τ τo HFS PNBI = MW PECCD.4MW co-eccd inversion radius complete stabilization LFS ECR deposition in ρpol co-eccd: - stabilisation / full suppression outside inversion radius - destabilisation for on-axis explainable with critical shear criterium: dq/dr r/q > (dq/dr r/q)crit pure heating (= % co and counter-eccd): - similar behaviour as for co-eccd, but less pronounced counter-eccd: - stabilisation for on-axis - inverse behaviour - effect on (/) mode plays an additional role impurity and He removal burn control in the core, M.Maraschek, IPP-Garching A.Mück, EPS, St.Petersburg A.Mück, PPCF,

5 Optimization of sawtooth tailoring by varying the deposition width. Wide ECCD ρ inv MSE ρ q= P ECH =7kW co-eccd cnt-eccd Bt-ramps PNBI = MW. Narrow ECCD ρ inv MSE ρ q= P ECH =7kW co-eccd cnt-eccd NBI τ / < τ saw saw >.. τ / < τ NBI saw saw > ρ.. ECCD.. ρ.. ECCD critical shear criterium can explain general behaviour clearest impact on τst for narrow deposition, especially for counter-eccd heating effect similar to co-eccd heating always present and stronger for broad deposition (> counter-eccd) jeccd or IECCD/d more important than total IECCD co-eccd more efficient for sawtooth tailoring, M.Maraschek, IPP-Garching A.Mück, PPCF, A.Manini, nd EPS, Tarragona

6 High power NBI experiments: NTM avoidance at high βn =.8 collaboration with T.P.Goodman, O.Sauter (CRPP) #78, counter-eccd NBI *ECR βn (/)-NTM amplitude #7, co-eccd W e6 ECE time (s) 8e6 4e ev 4 higher PNBI = MW to reach NTM-threshold βn.8 fixed by βp feedback sawtooth tailoring less clear with higher PNBI MW off-axis co-eccd no NTM during ECCD no sawteeth, first large sawtooth triggers on-axis counter-eccd fishbone triggered NTM during ECCD (/) mode further outside, no big seed-island NTM avoidance achieved, M.Maraschek, IPP-Garching A.Mück, EPS, St.Petersburg

7 (/)-NTM stabilisation with narrow co-eccd at reduced q9 at βn =.7 PNI = MW * PECCD = MW ( gyrotr.) #97. MW.. BT (/) ampl.. βn βn, Beta Normalized = q time (s) complete stabilisation at βn =.7 with PECCD = MW and PNBI = MW, ITER relev. q9 =.8 βn / PECCD =.7/MW, M.Maraschek, IPP-Garching Bt = -.6T q9 = T Bt =.T Ip =.MA q9 =.8 higher βn βn increase with more PNBI even higher βn achievable (re-excitation, Shafranov-shift)

8 Influence of the deposition width on the (/)-NTM stabilisation Jeccd [MA/m^] φ = -. Ieccd = 7.9 ka, d =.4 cm I/d = 878 ka/m jeccd =.66 MA/m^ φ = -. Ieccd =. ka, d =.9 cm I/d = ka/m jeccd =.66 MA/m^ φ = -. Ieccd =. ka, d =.6 cm I/d = 9 ka/m jeccd =.86 MA/m^ #9,./.9/.66s Peccd =. MW 4 GHz,. MW, θ = -. ρp I/d/PECCD/(Te/ne) [ka/m/mw/kev*e9m^-] peaking of driven current q tor. angle [deg] Wmin / Wsat narrow depostion: I / d = current density maximal for - (TORBEAM) full stabilisation with reduced PECCD / PNBI possible higher βn achievable at stabilisation ( βn / PECCD, βn / (PECCD/PNBI) ) W < d reduces the stabilisation efficiency ECCD modulated by mode (only O-point) might be required for ITER (further modulation experiments will be performed in 6), M.Maraschek, IPP-Garching

9 Stabilisation efficiency as function of deposition width and driven current. βn/peccd [/MW] , M.Maraschek, IPP-Garching I/d/PECCD/(Te/ne) [ka/m/mw/kev*e9m^-] peaking of driven current q9 variation reason for some scatter PECCD might be larger than required for stabilisation (PECCD > PECCD,marginal) the figure of merrit βn / PECCD gives the achievable βn at the stabilisation βn / PECCD "maximum" at maximal current peaking I/d similar improvement achieved for (/)-NTM

10 (/)-NTM stabilisation with broad ECCD (9 ) at early and late application collaboration with A.Lazaros, E.Westerhof (FOM) PECRH PNBI #8 #8 MW.. PNBI = MW PECCD =.7MW βn? βn comparable Bt = -.9T from Bt-scan W % time (s).6s delay of mode onset, with identical fishbone activity (B/f)... mode amplitude comparable due to β-drop power requirement for subsequent stabilisation not resolved more peaked deposition I/d gave no clear answers, M.Maraschek, IPP-Garching

11 Stabilisation of NTMs by non-modulated ECCD (intermediate width) P ECRH modulated: 4kW non-mod.: 8 kw Experiment ECRH Power (not calibrated) #8 MW.. I /I ECCD p odd n db dt even n n=, amplitude n= amplitude non-modulated β N modulated time (s) H. Zohm et al. NF 9 (999) d / W stabilisation also effective for non-modulated current drive for d < W, but: for W d the required current increases significantly for non-modulated ECCD, compared to modulated ECCD, M.Maraschek, IPP-Garching Q.Yu, PoP 4

12 NTM stabilisation with modulated broad ECCD PECRH PNBI #44 W.E+7 E+7 E+6. βn, Beta Normalised. PECRH E+6 (/) NTM amplitude time (s), M.Maraschek, IPP-Garching (/) NTM width reduction of island size, βn recovery phasing correct, but DC gives similar behaviour FIR-NTM phase at βn =.8 yet no clear answer possible, PECCD too small

13 FIR-NTMs - a general NTM behaviour for βn >. collaboration with D.F.Howell (UKAEA) W W. smooth open = ASDEX Upgrade full = JET even n -.. #87 n= amplitude. f (khz). 4 (4,) odd n (,). FIR 7. (,) amplitude (,)... β N,onset t (s) common behaviour of FIR-NTM for βn >. for JET and ASDEX Upgrade stabilility of required coupled ideal (4/)-mode (high p, low q infernal mode) ELMs have a similar effect at JET for βn >.9 for low Bt, low q9 presence of q= surface modifies behaviour in improved H-mode, M.Maraschek, IPP-Garching

14 Triggering / suppressing of FIR-NTMs with ECCD PNBI [MW] (4/) burst w [a.u.] w [a.u.] 4 4 PECRH [MW] βn - feedback controlled ctr-eccd (,) mode amplitude co-eccd (,) mode amplitude time (s) f [khz] #79, SXR:C:... time [s] triggering of ideal pressure driven (4/) mode by q-flattening with ECCD (ideal: growth time, duration; p, q - dependence as for infernal modes) FIR behaviour of NTM at lower / higher βn can be triggered / suppressed control of FIR-NTM feasible complete stabilisation main target, M.Maraschek, IPP-Garching

15 Present status and plans for the future feed-forward Bt - scan feedback stabilisation: () realtime detection of (m/n) mode, its localisation and deposition of the ECCD () feedback loop for the resonant surface (ρeccd = ρntm) () steerable ECCD launchers and tunable gyrotrons immediate reaction at still small island efficiency? PNBI increase to raise βn keep ECCD on q-surface without an NTM ultimate goal is not only removal, but avoidance of NTM feedback loop on ρeccd = ρ(q) with equilibrium q-profile seed-island avoidance (such as sawteeth and/or fishbones) co-eccd to "prevent" bootstrap hole at the resonant surface global tailoring of the j-profile ( effect) or the ne-profile (bootstrap is driving term via ne) to reduce drive for MHD mode, M.Maraschek, IPP-Garching

16 Detection of mode and ECCD on ECE (SENSOR) Radius [m]. cut-off Shot #74, NTM and ECRH position determination by correlation analysis Noise: ELMs NTM position detection ECRH position detection ECRH position detection ECRH position simulation NTMs can be directly measured from high time and radial resolution ECE ECCD modulation (9%) mode can be detected at the same time on ECE input quantities for NTM feedback stabilisation available.89 NBI Power (MW) ECRH Power (MW) Magnetic Perturbation (a.u.) #74 W W.. - high time resolution - realtime capabilities - Toroidal Field T Normalised Beta time (s), M.Maraschek, IPP-Garching. A.Keller, EPS, St. Petersburg

17 The new ECRH system on ASDEX Upgrade (ACTOR) power: 4 MW, provided by 4 gyrotrons pulse length: sec frequency: / 4 GHz as a -f-gyrotron / 7 / 7 / 4 GHz as a step tunable gyrotron change of frequency between pulses launcher: feedback controled deposition via poloidal launching angle toroidal angle can be set between pulses heating and current drive, in particular for advanced tokamak regime suppression of tearing modes control of transport and pressure profile, M.Maraschek, IPP-Garching

18 Summary and outlook local co / counter-eccd has been shown to be a powerful tool to control core MHD narrow deposition layer, well controlable deposition and width sawtooth tailoring at intermediate PNBI, deposition width, NTM avoidance at high PNBI deposition width scan I/d, βn/peccd narrow deposition reduces required power (NTM,sawteeth) early ECCD delays NTM onset modulation of ECCD, no clear answer yet trigger and suppress FIR-NTM physical understanding Outlook: application of feedforward technique: deposition width and modulation experiments extension for more general scenarios (ITER hybrid scenario) realtime feedback control with increased ECCD power and control capabilities will be applied in 6 for stabilisation and avoidance, M.Maraschek, IPP-Garching

19 possible routes to influence NTMs by ECCD sawtooth tailoring / avoidance of seed-island / NTM avoidance reduced size, trigger q= early ECCD q = m/n [A.Mück, NF, ] triggering transition to FIR-NTM, M.Maraschek, IPP-Garching βn βp/lp βonset PNI FIR-NTM # MRM:MPR() ρp=.6, m= 4 4 time (s) W.E+7 reduce magn. shear for (4/)-mode q=4/ active stabilization [S.Günter, PPCF, 4] of saturated island deposition width / island width modulation required for ITER? reduced requirements for early ECCD? (W < Wsat) q = m/n [G.Gantenbein, PRL, ] βsat βcrit E+7 E+6 W [cm]

20 Dependence of the sawtooth frequency on the NBI selection SXR C #667 W/m** variation of tangency radius governs the fast particle distribution from NBI fast particle stabilisation sawtooth period [ms] S S7 S4 S8 S S6 S S time (s) SXR B # time (s) variation in the particle energy between kev and 6 kev has an additional impact significantly different deposition profiles for different sources correction for sawtooth frequency required!, M.Maraschek, IPP-Garching

21 Sawtooth tailoring with ECRH / ECCD collaboration with T.P.Goodman, O.Sauter (CRPP) normalised sawtooth period τ τo normalised sawtooth period τ τo HFS HFS co-eccd inversion radius complete stabilization ECR deposition in ρpol inversion radius complete ST prevention ctr-eccd ECR deposition in ρpol PNBI = MW PECCD.4MW, M.Maraschek, IPP-Garching LFS LFS co-eccd: - stabilisation / full suppression outside inversion radius - destabilisation for on-axis explainable with critical shear criterium: dq/dr r/q > (dq/dr r/q)crit pure heating (= % co and counter-eccd): - similar behaviour as for co-eccd, but less pronounced - addtional stabilisation when (/) mode is directly hit counter-eccd: - stabilisation for on-axis - effect on (/) mode plays an additional role impurity and He removal burn control in the core A.Mück, EPS, St.Petersburg A.Mück, PPCF,

22 Modelling the effect of co-eccd / conter-eccd / pure ECRH on sawteeth COUNTER CD ECRH CO CD. τ ST [ms] ρ dep max ρ(q = ).8.6 ρ.4 modelling of the sawtooth period for pure (!) ECCD and ECRH sweep of the deposition layer over the q= surface ρ dep sweep time [s] heating effect similar to co-eccd with comparable size and broader effect small counter-eccd (braod dep.) effect can be understood in the experiment, M.Maraschek, IPP-Garching C.Angioni, NF 4 (), p.4

23 Power dependence of the sawtooth behaviour MW.6 MW.4 MW.8 SXR B. SXR B co-eccd. SXR B. co-eccd co-eccd ρpol ~ time (s) ρpol ~ time (s) time (s), M.Maraschek, IPP-Garching #67 #67 #679 W/m** W/m** W/m** Frequency [Hz] W/m variation of (/) ampl. with constant sawteeth (/) mode survives role of the (/) mode SXR B ECR time (s) ρpol~. #847 signal: B, index: A.Mück, PPCF,

24 (/)-NTM stabilisation with narrow co-eccd at βn =.6 PNI =.MW * PECCD = MW ( gyrotr.) MW #9 BT complete (/)-stabilisation T βn, Beta Normalized B[a.u.] (/) NTM amplitude time (s) complete stabilisation at βn =.6 with PECCD = MW and PNBI =.MW βn / PECCD =.6/MW βn increase with more PNBI not considered even higher βn achievable (re-excitation), M.Maraschek, IPP-Garching

25 (/)-NTM stabilisation at low q9 =.4 PNBI PECCD =.MW * PECRH #97 MW... βn =.. q9 = W = MOD () EvenN Bt time (s) T - -. Bt =.T, Ip =.MA, higher βn (narrow ECCD deposition) at more ITER relev. q9, but no complete stabilisation achieved, due to lower Te and less driven current, M.Maraschek, IPP-Garching

26 (/)-NTM stabilisation with narrow co-eccd at βn =. MW PNI = MW * PECCD =.4MW (4 gyrotr.) #944 BT complete (/)-stabilisation T - -. (/) NTM amplitude (n=) βn, Beta Normalized time (s). B [a.u.] - V. Locked Mode signal -. stabilisation at βn =. [.9] with PECCD =.4MW [.9MW], PNBI = MW [6.MW] βn / PECCD =.64/MW [./MW] stabilisation of the (/) NTM requires more power (βp,marg, less current drive) faster unlocking of (/)-NTM injection in the O-point of the locked mode works, M.Maraschek, IPP-Garching

27 stabilisation efficiency as function of deposition width, driven current,... I/d/PECCD/(Te/ne) [ka/m/mw/kev*e9m^-] βn/peccd [/MW] q tor. angle [deg] Wmin / Wsat q9 variation reason for some scatter for complete stabilisation PECCD might be to large (PECCD > PECCD,marginal, IECCD > IECCD,marginal) the figure of merrit βn / PECCD gives the achievable βn at the stabilisation βn / PECCD "maximum" at maximal I/d, M.Maraschek, IPP-Garching

28 Nonlinear modelling allows separation of different terms relative island width w/a.. shift of deposition radius during Bt-scan at cm/s (,), (6,4) and (,) component together (,) comp. (,) comp. time dependent Fokker-Planck code to calculate jeccd(r,t) [G. Giruzzi et al.,nf,9(),7(999)] nonlinear D-MHD (circular cylinder) [Q. Yu and S. Günter, POP,7, ()].... time [s] typically -% of plasma current driven at resonant surface (Fokker-Planck-Code) Modelling of DC co-eccd with scan of deposition and Fourier analysis : helical current ((,)-comp.) and -effect ((,)-comp.) are of similar importance complete stabilisation only due to synergy of both effects, M.Maraschek, IPP-Garching

29 Stabilization of neoclassical modes by non-modulated current drive non-modulated co-eccd (AUG) w /a Numerical modelling ECCD #7 ECRH Power x (MW) NBI Power (MW) 8... counter-cd Mirnov even n Mirnov even n n= Amplitude (a.u.) Shift of EC Resonance (cm).6 Stored energy time (s) non-modulated counter-eccd (AUG) n= Amplitude (a.u.).6 Stored energy time (s), M.Maraschek, IPP-Garching I /I =. ECCD P co-cd.... relative island width w/a stabilisation due to synergy of helical current and change in.. shift of deposition radius at cm/s (,), (6,4) and (,) component only (,) component.... time [s] t / τ R

30 Stabilisation of NTMs by non-modulated ECCD (intermediate width) P ECRH odd n db dt even n β N Experiment ECRH Power (not calibrated) #8 n=, amplitude n= amplitude ac: 4kW dc: 8 kw MW time (s) H. Zohm et al. NF 9 (999) w /a Numerical modelling modulated CD ECCD non-modulated CD I /I =.4 ECCD P t= w ECCD /a = (including heat transport and current diffusion) Q. Yu, S. Günter, PPCF 4 (998) 977 t / τ R modulated ECCD in O-point: PECCD/PNI 4-8%, 4% βn recovery with mode reduction stabilisation also effective for non-modulated current drive for d < W, M.Maraschek, IPP-Garching

31 Stabilisation of NTMs by non-modulated ECCD (intermediate width) P ECRH odd n db dt even n β N modulated: 4kW non-mod.: 8 kw Experiment ECRH Power (not calibrated) #8 n=, amplitude n= amplitude MW time (s) H. Zohm et al. NF 9 (999) PECCD [kw] W B [a.u.].. dw/dt [a.u.] #8 modulated ECCD in O-point: PECCD/PNI 4-8%, 4% βn recovery with mode reduction stabilisation also effective for non-modulated current drive for d < W clearly highest reduction rate at early modulated phase, M.Maraschek, IPP-Garching time [s]

32 phase adjustment of the modulated ECCD B = MHD GYRTRIGi #44 V -. D = MHD GYRTRIG V I = E+F+G+H; E time (s) correct phase adjustment for modulated deposition from the equilibrium ECCD is capable for modulation with f = khz requirement for ITER for W > d?, M.Maraschek, IPP-Garching

33 FIR-NTMs by nonlinear mode coupling with(m+,n+) modes and (,) mode even n f (khz) #87 n= amplitude. (,) amplitude odd n (4,) (,) (,) t (s) presence of both (m+/n+) mode and (/) mode required (,) amplitude (a.u.) f (khz) (4,) (,) t (s) # 68. f(,) f(,) phase locked resonance required, M.Maraschek, IPP-Garching A.Gude, Nucl. Fusion 4 () 8

34 Triggering / suppressing of FIR-NTMs with ECCD NBI heating power [MW] ECRH [MW] NBI heating power [MW] w [a.u.] w [a.u.] toroidal magnetic field β N ctr-eccd co-eccd ctr-eccd co-eccd (,) amplitude (,) amplitude.... Time (s) w [a.u.] w [a.u.] 4 4 β N ECRH power [MW] ctr-eccd (,) mode amplitude co-eccd (,) mode amplitude time (s) triggering of ideal pressure driven (4/) mode by q-flattening with ECCD FIR behaviour of NTM at lower / higher βn can be triggered / suppressed, M.Maraschek, IPP-Garching

35 General idea of a feedback loop for NTM stabilisation NTM detection mode numbers ECCD localisation mode localisation localisation from calibrated / improved equilibrium mode phase ECCD and loop off ECCD and loop on first guess launch angle of ECCD ρntm = ρeccd ECCD modulation NTM amplitude =, βn - max, WMHD = max,... launch angle Rplasma, Bt at DIII-D ECCD deposition meas. by ECE TORBEAM calculations for "first guess" for new scenarios ECCD modulation available trigger, M.Maraschek, IPP-Garching

36 Newly developed tools for the stabilisation (SENSOR) PNI MOD:EvenN PECCD #9 W.E+7 E+7 E amplitude from FFT tracing RVE:MHD low pass from ECCD trigger time (s).8.4 V 8 4 EvenN Trigger detection of odd n ((/)-NTM, but (/) also) and even n ((/)-NTM) diagnostic upgrade provides realtime n=, n=, n= detection detection of localisation of the mode and ECCD via realtime ECE / SXR, M.Maraschek, IPP-Garching