Effects of stochastic magnetic boundaries on divertor optimizations - PowerPoint PPT Presentation

1st IAEA Technical Meeting on Divertor Concepts 29 Sep. – 2 Oct. 2015, IAEA Headquarters, Vienna, Austria Effects of stochastic magnetic boundaries on divertor optimizations M. Kobayashi National Institute for Fusion Science, Oroshi-cho 322-6, Toki-city 509-5292, Japan 1

Introduction Stochastic magnetic boundaries appear in Stellarators due to the intrinsic non-axisymmetirc magnetic configuration  zero toroidal current favourable for steady state operation Tokamaks when symmetry breaking perturbation is applied  Aiming at edge plasma control, ELM mitigation/suppression Stochastic field as a tool for Intrinsic edge stochastization in Stellarators RMP for ELMs mitigation controlling edge plasma What do we benefit from the stochastic boundary for divertor optimization? 2

Summary: Effects of stochastic magnetic boundary Benefit Disadvantage Cost of divertor volume Issues to be assessed (10~20% of a) Confinement region Laminar region Ergodic region PFC Enhanced radiation Challenge for Density pump-out Control of radiation & detachment engineering Fueling efficiency ↓ (RMP coils, 3D ELM mitigation/suppression shape) PL-H ↑ Change of Er & turbulent transport Strike line splitting (non-uniform power/particle load) Peak power load? Impurity screening Decontamination Change of divertor Core performance? density regime Pumping efficiency? Energy transport barrier 3

Onset of stochastic instability by island overlapping: σ Chir > 1 ~ With increasing , island B r becomes large. ι = ι = + n / m n /( m 1 ) ~ w ∝ ι B / ' r m , n ∆ + ( m , m 1 ) n X-point + 0 . 5 ( w w ) σ = > ( n , m ) 1 2 1 Chir ∆ + ( m , m 1 ) n ? ? increases O-point Island “overlaps” Field lines in overlap region share B r ~ B field with neighboring island, and separatrix “overlap” “forget” from which island come from. w ( n , m ) 2 Stochastic trajectories + w ( n , m 1 ) 1 4 K.H. Finken et al., “The structure of magnetic field in the TEXTOR-DED”

Field line structure in stochastic magnetic boundary Schematics of field line structure Perturbation Short B lines field Long B lines PFCs Perturbation (edge surface layers) coil current Laminar region Connection length (LC) distribution in TEXTOR- DED Stochastic region r Poloidal turns θ φ Remnant islands Laminar region Short & long LC coexist radial (edge surface layers) No clear separatrix, Stochastic region locally ergodic, remnant islands 5 poloidal K.H. Finken et al., PRL 98 (2007) 065001.

Effects on divertor density regime 6

Absence of high recycling regime prior to detachment in the 3D configurations In helical devices as well as tokamaks with RMP, the modest density dependence is often observed. High recycling in 2D tokamaks The modest dependence in 3D configuration ~ − 3 1 ∝ − ~ 1 2 n n ∝ ∝ ∝ T n T n n n div up div up div up div up ~nup-2 ~nup3 ~nup1 Te div (eV) ne div (10^19 m-3) Te_div (eV) ndiv (10^19 m-3) Tdiv (eV) 2D ndiv (10^19 m-3) Tdiv (eV) 3D ndiv (10^19 m-3) 2D ~nup1.5 TEXTOR-DED probe_masuzaki_29265 Tdiv (eV) 3D probe_masuzaki_29265 ndiv (10^19 m-3) 3D nup^1 ndiv (10^19 m-3) 3D nup^1.5 W7-AS (m/n=6/2) 10 2 ~nup-1 Detach TEXTOR-DED 10 0 (m/n=6/2) ~nup-0.3 Detach LHD LHD 10 1 10 -1 1 10 1 10 n up (10 19 m -3 ) n up (10 19 m -3 ) Y. Feng et al., PPCF 44 (2002) 611. 7 M. Clever et al., Nucl. Fusion 52 (2012) 054005. S. Masuzaki et al., JNM 313-316 (2003) 852.

Effects of enhanced ⊥ interaction of momentum transport on divertor regime 2D axi-symmetric divertor ndown is suppressed 1.0 Pressure conservation along flux tube With (a) Divertor plate 0.8 m -3) -V// +V// 0.6 (- φ ) (+ φ ) 0 2 0 1 ( 0.4 radial w n d o n 0.2 poloidal 0 0 0.2 0.4 0.6 0.8 1.0 3D configuration (e.g. stochastic layer, ID) n up (10 20 m -3 ) //-Momentum loss due to counter flows //-Temperature drop Tup/Tdown becomes Divertor plate Divertor plate small 1000 (b) Open field lines f m0 =5 radial T up +V//-V// V e 100 ( ) w T d o n poloidal p , 10 T down Tu τ D L = ⊥ ∝ m // // f ⊥ Ø loss of //-Momentum m 0 τ λ 2 V ⊥ m // m 10 18 10 19 10 20 8 n up (m -3 ) Y. Feng et al., Nucl. Fusion 46 (2006) 807.

LHD m/n=7~2/10 W7-AS HSX m/n=7/8 Simulation (EMC3-EIRENE) (a) LHD m/n=9/5 1.0 nV // (10 23 /m 2 /s) 5.0 4.0 0.5 3.0 2.0 m 1.0 0 Z( ) 0 -1.0 -2.0 -0.5 -3.0 -4.0 -5.0 -1.0 Mach probe scanning path A. Bader et al., Nucl. Fusion vol.53 (2013) 113036. Y . Feng et al., Plasma Phys. Control. Fusion 3.0 3.5 4.0 4.5 R (m) vol.53 (2011) 024009. DIII-D m/n~10/3 TEXTOR-DED Tore Supra m/n=6/2 m/n=18/6 Mach number Mach number Mach number 9 Y . Corre et al., Nucl. Fusion vol.47 (2007) 119. H. Frerichs et al., Nucl. Fusion vol.52 (2012) 023001.

Flow alternation is detected in experiments Simulation (EMC3-EIRENE) (a) LHD 1.0 nV// (1023 /m2/s) 5.0 4.0 0.5 3.0 2.0 Tore Supra Z (m) 1.0 0 0 -1.0 -2.0 -0.5 -3.0 -4.0 -5.0 -1.0 Mach probe (b) scanning path Experiments 3.0 3.5 4.0 4.5 R (m) Mach Number Simulation (EMC3-EIRENE) J.P. Gunn et al., JNM 290-293 (2001) 877. 10 Z (m)

High recycling regime can be recovered in 3D configuration for some cases Tore Supra (m/n=18/6) TEXTOR-DED (m/n=3/1) W7-X (m/n=5/5) numerical simulation ∝ 3 n n EMC3-EIRENE down up ned 1019 m-3 ∝ 3 n n down up ∝ 4 n n down up δ m~8 cm nu 1019 m-3 B. Meslin et al., J. Nucl. Mater. 266-269 (1999) Y. Feng et al., Nucl. Fusion 49 (2009) 095002. 318. M. Lehnen et al., J. Nucl. Mater. 337-339 (2005) 171. In these cases, the following effects are small, τ D L ⊥ m // // = ∝ f m τ 2 λ V ⊥ m // m χ ⊥ q n ⊥ = e e κ 2 2 . 5 q ( B / B ) T // e r t 0 e e due either to Large separation of counter flow ( λ m ~2 π a/m ) 1. Relatively high Te in SOL 11 2.

Multi-machine comparison for divertor density regime netic α ≈ Ø ⊥ loss of //-Momentum shea ( 3 ) Larg mag α High recycling ∝ La e n n rg m e > p p down up α < r ( 3)  //-pressure drop No High recycling LCFS div tau_m// / tau_m perp mid tau_m// / tau_m perp mid LHD τ D L tau_m// / tau_m perp mid = ⊥ m // // (m/n~5/10, strong shear) parameters_devices_density_regime_mod4_high-recycling τ 2 λ V 10 1 ⊥ m // m W7-AS λ m: ⊥ characteristic scale length for momentum (m/n=9/5) loss (e.g. ~2 π a/m) 10 0 ⊥ HSX m τ TEXTOR-DED Ø Replacement of //-energy flux with ⊥ -flux (m/n=4/4) No high recycling 10 -1 Tore Supra (m/n=12/4) (m/n=7/8) /  reduction of //-conduction energy // (m/n=18/6) m TEXTOR-DED τ χ ⊥ q n ⊥ = (m/n=6/2) e e 10 -2 W7-X κ 2 2 . 5 q ( B / B ) T DIII-D (m/n=5/5) // e r t 0 e e (m/n≈10/3) 10 -3 TEXTOR-DED EAST (m/n=3/1) ITER Operation domain for (m/n~3/1) 10 -4 a (m/n~10/3) high recycling regime 10 -2 10 -1 10 0 10 1 Large 2 q_perp_e / q_//_e mid     τ q − ⊥ 5  m //   e  < × q e q / 3 . 6 10 collisionless     ⊥ // e τ q     collisional ⊥ m // e B / r B large t Possible Impacts on divertor functions due to the absence of high recycling regime: Pumping efficiency ↓, physical sputtering ↑, detach. onset density ↑ (!?) 12 Can be avoided in detached Preferable for core performance (?) phase

Effects on impurity screening 13

Impurity screening has been observed in many devices with edge stochastic layer Experiments with density scan shows better screening at high density (low Te) C6+ (10^17 m-3) Tore Supra Carbon_Tore_Supra 15 10 TEXTOR-DED 5 C concentration (%) M. Lehnen et al., PPCF 47 (2005) B237. C concentration (%) m/n=3/1 Icoil=2kA C-concentration_fig8b_DED 0 0 0.5 1 n e edge (10 19 m -3 ) 1 Y. Corre et al., Nucl. Fusion 47 (2007) 119. 0.5 (m/n=3/1) 0 2 2.5 3 3.5 n e (10 19 m -3 ) LHD ne (1019 m-3) CVI/ne normalized C_all_Nebar#81896to81933_080407_375_woLID Enhanced outward particle flux due to braiding B field, Ø CVI/ne (a.u.) 1 (~ n C5+ ) density pump-out  Effective friction force 0.5  Drive impurity towards divertor 0 High edge density Ø 0 2 4 6 8 n e (10 19 m -3 )  screening of CX flux, shallow penetration of neutral impurity 14 ne (1019 m-3)

SOL thickness dependence of impurity screening: thicker stochastic SOL  better screening already at low density λ Ø Thicker stochastic layer/SOL ( ) relative to neutral st − SOL λ impurity penetration ( ) Thin stochastic layer imp λ λ ↑  better screening / − st SOL imp LHD ratio 3.60m ratio 4.00m 10 VI C in core IV ∝ Thick stochastic layer C source 1 Better 0.1 screening 0 2 4 6 8 10 ne (1019 m-3) 15 M.B. Chowdhuri et al., Phys. Plasmas 16 (2009) 062502.

Impurity screening: species dependence Better screening for non-recycling impurity Tore Supra Ratio of boundary to core impurity concentration Better screening for O & N (non-recycling) than Ne (recycling) is due to the wall pumping (!?). 16 Ph. Ghendrih et al., NF 42 (2002) 1221.