fast ion losses and redistribution induced by low...

NSTXNSTX 12th IAEA TM EP – Fast Ion Redistribution by LF MHD in NSTX, A. Bortolon (09/08/2011) 1

Fast Ion Losses and Redistribution Induced by Fast Ion Losses and Redistribution Induced by Low Frequency MHD in NSTX PlasmasLow Frequency MHD in NSTX Plasmas

A. Bortolon, G.J. Kramer, J. Manickam, M. Podestà N.A. Crocker, D.S. Darrow, E.D. Fredrickson,

W.W. Heidbrink, S.Medley and the NSTX Research Team

12th IAEA Technical Meeting on“Energetic Particles in Magnetic Confinement Systems”

Austin, Texas, USA7-10 September 2011

Culham Sci CtrU St. Andrews

York UChubu UFukui U

Hiroshima UHyogo UKyoto U

Kyushu UKyushu Tokai U

NIFSNiigata UU Tokyo

JAEAHebrew UIoffe Inst

RRC Kurchatov InstTRINITI

NFRIKAIST

POSTECHASIPP

ENEA, FrascatiCEA, Cadarache

IPP, JülichIPP, Garching

ASCR, Czech Rep

Columbia UCompXGeneral AtomicsFIUINLJohns Hopkins ULANLLLNLLodestarMITNova PhotonicsNew York UORNLPPPLPrinceton UPurdue USNLThink Tank, Inc.UC DavisUC IrvineUCLAUCSDU ColoradoU IllinoisU MarylandU RochesterU WashingtonU Wisconsin

NSTXNSTX Supported by


Introduction and motivation

• Fast particle transport and losses in presence of Low Frequency MHD modes has been long studied (TFTR, DIII-D, ASDEX, NSTX, ...)

• Often core modes have been addressed, well described by single helicity radial perturbation (Tearing Mode, Neoclassical TM, internal kink)

• Former studies on NSTX focused on (m=2,n=1) internal kink– Depletion at particle energies below the injection energy (NPA)– Passing particles (E<Einj) are preferentially affected and lost

• Here we address early low frequency MHD activity on NSTX– Extends to the plasma periphery– Strongly affects fast ion population– Appears to be an important element for the destabilization of High

Frequency Alfvénic modes


Outline of the talk

• Experimental scenario– Typical discharge evolution, MHD activity– FIDA observations

• Mode characterization– Experimental signature on Soft X-Rays and Reflectometry– Mode structure from ideal stability computations

• Full-orbit simulations results– Beam ion losses – Redistribution of energetic ions in phase space


• “Fiducial” configuration, t<300ms– PNBI=4-6 MW

– βN~2.5-3.5

• MHD activity at different frequencies:– Toroidal AE (bursting)– Reversed Shear AE– Global/Compressional AE

(bursting/continous)• Onset of LF mode at t=0.22s

– Rotation collapse– βN ramp stops

• Mode vanishes after 100 ms, as density increases

Experimental scenario: H-mode, Bt=0.4T, Ip=900kA

TAE

rsAE CAE/GAE


High Frequency MHD associated with LF mode

• Low Frequency mode enters with multiple toroidal harmonics

• n=1 and n=2 pertsist• Initial chirp follows the toroidal

rotation drop (-15kHz)

• Compressional AE cluster in1-2.5 MHz frequency range

• Co-propagating modes, n=9-13• Appear after onset of LF MHD• Associated with bump-on-tail

beam ion distribution function

• Small effect on neutron rate• FILD measured losses <5%


FIDA spectra show strong effect on fast ions

• Spectral signal decreases in a broad range of wavelength/energies• Vertical view → sensitive to low pitch (p=|v||/v| <0.6, E~30-60 keV)• Low Frequency MHD activity affects the trapped population

• Fast Ion D-Alpha diagnostic observes the hot tail of Deuterium Balmer-α spectral line (656.1nm) emitted by recombined fast ions


• FIDA density:

• nFIDA provides local information about fast ion density

Strong depletion of FIDA density at mode onset

• Depletion nFIDA consistently observed after mode onset: – up to 30% reduction– 10 ms time scale– outboard plasma is affected

first and more

Redistribution in real/velocity space ?


Low Frequency Mode extends to plasma boundary

• Radial displacement of plasma pedestal from reflectometer

• Difficult to determine internal mode structure (high pedestal ne)

• Mirnov array indicates n=1 (Bz~15G at probe location), weaker n=2 • No clear evidence of magnetic island (e.g. in Te profile)

• Edge Toroidal SXR array (MESXR) captures peripheral dynamic: – expansion-compression– 8 kHz, r/a>0.6


• Consider t=250ms, saturated phase, 30 ms after onset

• Plasma configuration from LRDFIT equilibrium reconstruction– constraints on measured profiles of

pressure and pitch angle• Only n=1 component considered• |m|<40 poloidal harmonics included• Computation up to 99.98% of ψe

• Configuration is linearly unstable to n=1 kink under these conditions:– Free boundary– High pressure gradient at pedestal– Reversed shear in plasma core

PEST code predicts instability to n=1 kink


Kink amplitude maximal at outboard plasma periphery

• High order poloidal harmonics contribute in the peripheral region• Mode amplitude is larger in the LFS (m=3-4 effective structure)• Fine structure in the HFS, but smaller amplitude

m=6

m=5

m=4

m=3

ξr (φ=0)


Kink structure validation and normalization (1)

• 3D ne and Te perturbation from radial displacement:

• SXR emissivity assuming carbon impurity only:

• Rigid toroidal rotation at mode frequency (8 kHz)

Mode structure checked against measurements assuming saturated structure is similar to linear computation


Kink structure validation and normalization (2)

Reasonable agreement with data if ξ scaled to 2% of PEST output

• δBz at Mirnov coil ~ 15G

• Fluctuating MESXR profile• Pedestal displacement ±9mm


Full-Orbit Monte-Carlo code SPIRAL used to predict the perturbed ion distribution function and particle losses

• Adequate description for NSTX (90keV ion gyroradius ~19cm)• Beam ions orbits solved in 2 magnetic configurations

– Perturbed fields (PEST n=1 kink, scaled) + unperturbed ref. case – Random selection of ionizing neutrals introduced at uniform rate

along 25 ms simulated time window (birth profile from NUBEAM)– Energy slowing down time ~15ms for 90keV ion → final distribution

assumed representative of the steady state – Particles hitting the realistic wall model are considered lost

G.J.Kramer Poster 1.18


Full-Orbit Monte-Carlo code SPIRAL used to predict the perturbed ion distribution function and particle losses

• Adequate description for NSTX (90keV ion gyroradius ~19cm)• Beam ions orbits solved in 2 magnetic configurations

– Perturbed fields (PEST n=1 kink, scaled) + unperturbed ref. case – Random selection of ionizing neutrals introduced at uniform rate

along 25 ms simulated time window (birth profile from NUBEAM)– Energy slowing down time ~15ms for 90keV ion → final distribution

assumed representative of the steady state – Particles hitting the realistic wall model are considered lost

Without kink With kink Increment

Total Beam Ion Losses 17.4 % 20.6 % + 3.3 %

Predicted losses consistent with FILD estimate (<5%)

G.J.Kramer Poster 1.18


Redistribution in real and velocity space (preliminary)

• Redistribution of core confined particles

• Kink effect from differential distribution ∆F = Fkink- Fequi

↑Increaseswith kink

Decreases with kink

↓-20%

• Slowing down distribution shifts towards v|| / v = 1


Redistribution in real and velocity space (preliminary)

• Redistribution of core confined particles

• Slowing down distribution shifts towards v|| / v = 1

• Net decrease in the region of maximal FIDA response

↑Increaseswith kink

Decreases with kink

↓

FIDAresponse

• Kink effect from differential distribution ∆F = Fkink- Fequi

-20%


Conclusions

• Early LF MHD is observed to affect strongly the fast ion population– FIDA density reduced as much as 30%– Decremental effect on neutron emission rate

• Mode nature and structure inferred by coupling ideal MHD stability calculations to experimental observations– global kink nature, finite edge amplitude, associated to a residual reversed shear– a kink perturbed equilibrium has been constructed consistent with SXR emission

and reflectometry observations

• Full-Orbit simulations with SPIRAL code allow to study the kink effect on fast ion confinement– Fast ion losses increase by ~3%– Redistribution in both real and velocity space is found as result of kink– Shift of fast ion d.f. towards p=1 may account for the depletion of nFIDA

– Increase of particle statistics will permit to address HF modes destabilization


Back up slides


SPIRAL fast ion density (preliminary)


SPIRAL fast ion density 30<E<60keV, |p|<0.5 (preliminary)


CAE response to NBI heating scheme

• Constant input power• Switch NBI sources

• Source B provide deeper energetic ion deposition

• More passing ion are generated with source B

• Condition more favorable to CAE destabilization

Fredrickson et al. Phys. Rev.Lett. 2001Gorelenkov et al. Nucl. Fus. 2002


Rotation braking at kink onset

← Kink frequency 8kHz

• Kink locked at the plasma rotation of qmin location


P=4MW, no concurring CAE activity


FIDA measurement concept

• Active Charge eXchange– Measures hot tails of Balmer alpha– Large Doppler shift of recombining fast ions– Background subtraction is crucial

• Effective average over velocity space– Viewing angle– NBI geometry– Effective CX cross section

• Weighting Wλ(E,p) function gives the sensitivity to different velocity space regions (pitch parameter p=v||/v)

•+

An approximate Fast Ions Density nFIDA can be obtained from

Fast Ion

CX Rec.

Balmer αEmission

FIDAspectrum

Weightfunction FI distribution

function


Example of FIDA spectrum (NSTX vertical view)

CII

BE

ColdDα

OV

CVI

OV+CII BV


NSTX FIDA response

• Wλ(E,p) evaluated at – R=1.2 m,– Eλ=35 keV (652.1 nm)


Reflectometer cutoff 142296 250ms

fast ion losses and redistribution induced by low...

Documents