Transport and turbulence reduction with negative triangularity : Correlation ECE measurements in TCV A. Pochelon, M.Rancic, V.S.Udintsev 1, T.P.Goodman, E.Fable 2, B.Labit, O.Sauter, R.Behn, A.Bottino 2, S.Brunner, Y.Camenen 3, A.Casati, P.K.Chattopadhyay 4, S.Coda, B.P.Duval, L.Federspiel, S.Jolliet 5, A.Karpushov, N.Kirneva 6, A.Marinoni, Ph.Marmillod, B.McMillan, S.Y.Medvedev 7, J-M.Moret, A.Pitzschke, L.Porte, L.Villard, and the TCV team Ecole Polytechnique de Lausanne (EPFL) Centre de Recherches en Physique des Plasmas Association EURATOM-Confédération Suisse Lausanne, Switzerland 1 ITER-IO, 2 IPP-Garching, 3 CFSA-Warwick, 4 IPR-Bhat-India, 5 JAEA-Tokyo, 6 RRC-Kurchatov-Moscow, 7 Keldysh-Inst-Moscow 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 1
MOTIVATIONS Why to study also plasma shapes different from ITER? A tool for test and validation of transport modeling δ=+0.4 δ=-0.4 Confinement in the core improves towards negative triangularity (at least in L-mode!) 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 2
Energy confinement improves towards δ < 0 (at low ν eff ~ 0.2-1) [ms] 15 I p [ka] 260 P tot MW τ Ee improvement is not explained by gradient geometrical (=shape) factor, -> χ e is thus varying with triangularity τ 19 (1.8 10 /ne_av) 0.46 10 5 Ee 260 & 400 260 0 0.6 0.4 0.2 0.0 0.2 0.4 δ 0.3 0.9 1.5 OH EC Plasma conditions: low density ECH plasmas, high R/L Te >7 and T e /T i, low collisionality ν eff -> TEM dominated regime (no ETG due to high Z eff & T e /T i, in range 0.2<ρ<0.7) Coda 98, Pochelon NF99 & EPS99, Weisen NF98 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 3
Energy transport reduces by a factor 2 towards δ < 0 LFS δ=+0.4 δ=+0.4 δ= -0.4 δ= 0.4 Two shapes, δ = ± 0.4 : Expt. 1: adapting the power at δ = - 0.4 to keep the same plasma energy, T e, n e and q-profiles than at δ = + 0.4 : Only half add. power needed, resulting in χ e halfed at mid-radius, Expt. 2: same powers at δ = ± 0.4: resulting in higher energy & lower χ e Camenen NF07 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 4
OUTLINE FROM GLOBAL TO LOCAL quantities... confinement, transport, turbulence 1. Intro: correlation ECE set-up, effect of collisinality on turbulence 2. Turbulence characteristics with triangularity 3. Global linear gyrokinetic simulations FROM LOCAL TO GLOBAL GK simulations 4. Conclusions and Outlook 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 5
TCV facility TCV: R = 0.88 m, a = 0.25 m, R/a ~ 3.5 B < 1.5T, I p 1MA elongation 0.9 < κ < 2.8 triangularity - 0.7 < δ < 1 squareness SN & Snowflake divertor ECRH: 4.5 MW at 2 nd and 3 rd harmonic TCV programmatic brainstorming: 15-16 Sept. 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 6
Turbulence measurements using ECE ECE views 4 normal antennas, equat. LFS with focusing lens 1 sweepable oblique LFS antenna (L7) 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 7
Correlation ECE setup ρ q=1 ρ~-0.5 to 0.7 ρ=1 single Gaussian-beam sight-line measurement from LFS, from 2 narrowband (0.1GHz) tunable (61-89GHz) YIG filters, defining two sampling volumes focussed beam resolves to k θ 1.4 cm -1 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 8
Fluctuation amplitudes versus collisionality cross-spectral density (CSD) decrease with density Ohmic, q~10, κ~1.4, δ~0.3 fluctuation amplitude CSD integrated over<30-130khz> to avoid MHD-modes ρ=0.55 on LFS equator n e increasing f [khz] ν eff T e -fluctuation amplitudes decrease with collisionality ν eff Udintsev & Fable US-TTF09 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 9
T e and n e fluctuations behaviour with ν eff (from GS2 simulations) (δt e /<T e >) 2 (δn e /<n e >) 2 Motivation: outer plasma not totally blackbody in for ECE: since optical depth at ρ v =0.55 is < 3, thus measured fluct. attributable to both T e and n e ν eff from GS2: - both (T e,n e )-fluctuations show same trend with ν eff - direction of propagation characteristic of TEM The amplitude reduction with ν eff is also consistent with TEM drive reduction due to TE collisional detrapping Udintsev, Fable US-TTF09 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 10
Triangularity scan measuring location profiles δ=+0.4 δ=-0.4 ρ v =0.6 on LFS equator (taking into account varied Shafranov shift in δ=±0.4; significative since CSD doubles each ΔR/a~10%) matching δ=-0.4, 250 kw Ohmic profiles by adding 580 kw ECH in δ=+0.4 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 11
Cross spectral density : δ = ± 0.4, OH/EC δ =+0.4 EC δ = ± 0.4 Ohmic profiles δ =+0.4 OH δ = -0.4 OH δ = ± 0.4 same profiles CSD reduced towards δ < 0 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 12
Varying separation of detection volumes Norm. Cross Spectral Density [a.u.] 2.5 x 10 9 2 1.5 1 0.5 0 δ = + 0.4 Δf = 0.24 GHz (#39304) Δf = 0.50 GHz (#39301) Δf = 1.00 GHz (#39302) Δf = 1.50 GHz (#39303) ρ = 0.6 Δf [GHz] - scan, ρ v ~0.6 CSD amplitude decreasing with radial separation 0.5 1 1.5 2 2.5 Frequency [Hz] x 10 5 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 13
Correlation lengths: δ = ± 0.4, OH/EC Radial correlation length λ c : δ = -0.4 OH δ =+0.4 EC δ =+0.4 OH - shorter at δ <0 (factor ~ 2) - larger with EC to reach same T e and n e profiles, but thus with doubled heat flux 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 14
Global linear GK simulation with ORB 30 δ = +0.5 γ [10 4 s 1 ] 20 10 0 0.6 δ=+0.5 δ=+0.2 δ= 0.3 k ρ s 0.4 δ = -0.3 γ/k 2 [m 2 /s] 0.2 30 15 δ=+0.5 δ = -0.3 mixing length transport 0 5 10 15 20 25 n n=10 LORB GK simul. linear, global [Camenen, Bottino05] 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 15
Comparing correlation length: expt. λ c / k lin. global GK simul. 10 3 δ = 0.4, P ECRH = 580 kw δ = 0.4, P ECRH = 0 kw R 12 (0) (δt e /T e ) 2 [a.u.] 10 4 δ = + 0.4 δ = + 0.4 EC n=10 δ = +0.5 10 5 ρ 0.6 0.5 1 1.5 2 2.5 3 Frequency separation Δf [GHz] δ = -0.3 Radial turb. correlation length λ c and λ ~1/k, both reduced by factor ~ 2 linear global GK simulations appear to suggest correct trend! 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 16
Comparing spectra (expt. turbulence / GK TEM transport) 30 Norm. Cross Spectra Density [a.u.] 2.5 x 10 9 2 1.5 1 0.5 0 δ = + 0.4 δ = - 0.4 δ = 0.39, Δf = 0.24 [GHz] (#39304) δ = 0.45, Δf = 0.24 [GHz] (#38230) 0.5 1 1.5 2 2.5 Frequency [Hz] x 10 5 CSD reduced at δ <0, predominantly at the low frequencies ρ = 0.60 γ [10 4 s 1 ] k ρ s γ/k 2 [m 2 /s] 20 10 0 0.6 0.4 0.2 30 15 δ=+0.5 δ = -0.3 mixing length transport δ=+0.5 δ=+0.2 δ= 0.3 0 5 10 15 20 25 χ e_mixing length reduced at δ <0, predominantly at the low n (role of the large, global structures) n 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 17
Correlation length λ c and χ e versus collisionality from corr - ECE from power balance 10 8 δ=+0.4 δ=+0.2 δ= 0.2 δ= 0.4 δ=+0.4 χ e [m 2 /s] 6 4 δ= 0.4 2 0 ρ=0.55 0 1 2 3 4 5 6 2 1/ν T /(ne Z ) eff e eff Camenen 07 Correlation length λ c and χ e show both similar reduction with collisionality and negative triangularity 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 18
Conclusions At negative triangularity, compared to positive, - τ Ee is improved and χ e reduced (both factor ~2) - both radial correlation length λ c, and turbulence amplitudes reduce (by factor ~2) The (so far linear) comparison with global GK code ORB shows similar trends: λ ~1/k is reduced with negative triangularity (similar trend with δ as the measured correlation length λ c, apparently a relevant (linear!) hint for TEM developped turb....) - role of low n, low frequency - large radial structures - in the variation of transport with δ. plasma shape: a tool to investigatate transport properties for model validation - in particular their dependence on equilibrium geometry! 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 19
Outlook next step aims at a more detailed comparision beween expt and GK-modeling, using e.g. non-lin global ORB5 code with an artificial diagnostics, mimicking correce measurements wishful next experimental step: evolve to multi-channel correce (from present 2-point-correlation diag.) : - opens to new physics domain: non-local transport, avalanches, meso-scale structures (as in present NL global GK codes) - allows for a more efficient tokamak use, 2-point-only-correlations limiting investigations further explorations with oblique line-of-sight angles to resolve turbulence structures (e.g.: link between λ c and potential cell orientation) 3 rd EFDA TTG & 15 th EU-US TTF 7-10 Sept 2010 20