IAEA, AMD Unit Page - mon crp neutral, …mon 19 jun 2017, j ko, iaea‐crp‐neutral, vienna, austria 19 Reliable inputs possible from the existing KSTAR MSE KSTAR#7331F6C4 656 658

1mon 19 jun 2017, j ko, iaea‐crp‐neutral, vienna, austria

J. Ko1,2, J. Chung1

Experimental validation of atomic data for motional Stark effect diagnostics

The 1st IAEA Research Coordination Meeting onData for Atomic Processes of Neutral Beams in Fusion Plasmas

19 – 21 June 2017, IAEA Headquarters, Vienna, Austria

1National Fusion Research Institute, Daejeon, Korea2University of Science and Technology, Daejeon, Korea

mon 19 jun 2017, j ko, iaea‐crp‐neutral, vienna, austria 2

Scope of the project

•High‐precision measurements of beam‐emission spectra from KSTAR discharges

•Development of a spectra analysis tool with a modulated interface for atomic data

Experimental validation of atomic data for motional Stark effect diagnostics


Issues to address

Calculated spectra are not always reproduce experimental observations.

… especially, when we are dealing with polarized light.


Issues to address #1Reflections on mirrors distort polarization properties.

Vacuum window

PEM

LensesMirror

Linearpolarizer

Dichroicbeam splitter

MSE focal plane CES focal plane

Incident light

Plane‐of‐incidence

Normal reflective surface

P‐pol

S‐pol

Lee et al, Fusion Eng. Des. (In press)

Plane‐of‐incidence

(Multi‐layer) dielectric‐coated surface


Issues to address #1Specialized dielectric coating works for present tokamaks…

Vacuum window

PEM

LensesMirror

Linearpolarizer

Dichroicbeam splitter

MSE focal plane CES focal plane

Incident light

P‐pol

S‐pol

Lee et al, Fusion Eng. Des. (In press) But not for ITER …


Issues to address #2Faraday effect rotates the polarization angle.

ch16, 008, 20151103, ic=stef3

0.0 0.5 1.0 1.5 2.0Bt (T)

89.089.590.090.591.0

pol (

deg)

-0.50.00.51.01.52.0 δpol (deg)δpol (from Bt = 0) = -0.007 + 0.828 x Bt

Ko et al, Rev. Sci. Instrum. 88, 063505 (2017)


Issues to address #3Background polarized light does exist in a tokamak.

0.10

1.00

10.00

0.1 1 10 100signal-to-background ratio

nl_04 (1e20)0.2 ~ 0.40.6 ~ 0.81.0 ~ 1.2

0.1 1 10 100Ch1 (87.21 cm) Ch9 (81.75 cm)

Statisti cal errorin pitch angle (deg)Alcator C‐Mod

Ko, PhD Dissertation, MIT (2006)


Issues to address #3Background polarized light does exist in a tokamak.

Statistical error in pitch angle (deg)

1 10 100Signal to background ratio

0.1

1.0

10.0nebar (1e20/m3)

0.20.5

0, /home/m

sea/mse_ps/m

se_study_bkg_20160818.ps, Mon Apr 18 17:57:29 2016

0.10

1.00

10.00

0.1 1 10 100signal-to-background ratio

nl_04 (1e20)0.2 ~ 0.40.6 ~ 0.81.0 ~ 1.2

0.1 1 10 100Ch1 (87.21 cm) Ch9 (81.75 cm)

Statisti cal errorin pitch angle (deg)Alcator C‐Mod

Ko, PhD Dissertation, MIT (2006)

KSTAR

0.2

0.6

2060

100

0 2 4 6 8 10 sec0.01.02.03.0

659 660 661 662 nm

0, mse_spec_show

_10000.ps, Fri Oct 17 10:28:27 2014

KSTAR#10000

ne (1e19/m3)

Ip (MA)

σππ

σππ

σ ππ

V2 V3Beam energy (keV)


Issues to address #4Practically, multiple ion sources are used in NBI heating

Ko et al, J. Inst. 8, P1009 (2015)

12

3


Issues to address #4Practically, multiple ion sources are used in NBI heating

0, mse_spec_show

_10515.ps, Fri Oct 17 10:59:15 2014

0.2

0.6

2060

100

0 1 2 sec0.01.02.03.0

659 660 661 662 nm

KSTAR#10515

ne (1e19/m3)

V1V2

Ip (MA)

V3Beam energy (keV)

Ko et al, J. Inst. 8, P1009 (2015)

12

3


Outline of the activities

Spectral analysis

High‐resolution spectrummeasurements ( 0.05 nm)

Atomic models(NOMAD, ADAS etc)

Mirror reflections

Faraday rotation

Polarized background

Multi‐ion‐source injection

Existing (conventional) MSE‐ Input for model validation‐ Comparison with spectral analyses



Spectral analysis



Mirror reflections

Faraday rotation




Code validation



Spectral analysis



Mirror reflections

Faraday rotation




Code validation

ITER MSEprototype


How the known pitch angles help the code validation?

x

z (v)

y

B

B T

E = v BX Tαθ

φ

θe

ey

Voslamber, Rev. Sci. Instrum. 66, 2892 (1995)Ko et al, Rev. Sci. Instrum. 81, 10D702 (2010)

I: Intensity of pure sigma & pi when viewed perpendicular to the Lorentz electric field E.

,


Using ADAS for I’s can reduce the systematic error

1100713 (45 shots)

0.34 0.36 0.38 0.40 0.42 0.44 0.46On-axis |B| from ADAS (T)

0.34

0.36

0.38

0.40

0.42

0.44

0.46

On-

axis

|B| f

rom

theo

retic

mod

el (T

)

1, 1100713a_a4_adas.ps, Wed Jul 28 03:14:59 2010

1100714 (97 shots)

0.34 0.36 0.38 0.40 0.42 0.44 0.46On-axis |B| from ADAS (T)

0.34

0.36

0.38

0.40

0.42

0.44

0.46

On-

axis

|B| f

rom

theo

retic

mod

el (T

)

1, 1100714a_a4_adas.ps, Wed Jul 28 03:15:04 2010

2 ~ 3 % of systematic error can be reduced over the range of typical MST operating conditions, when compared with the analytic model* which tends to overestimate |B|.

*W. Mandl et al. Plasma Phys. Controlled Fusion 35, 1373 (1993)Ko et al, Rev. Sci. Instrum. 83, 10D513 (2012)


Reliable inputs possible from the existing KSTAR MSE

Front optics

Filter & detector module

VW

PEM (M

DB

P

NBI

tan(2P ) I(2I(2

Fibers( 40 m)

Ko et al, Rev. Sci. Instrum. 88, 063505 (2017)Ko et al, Fusion Eng. Des. 109‐111 (2016) 742‐746Chung et al, Rev. Sci. Instrum. 87, 11E503 (2016)Chung et al, Rev. Sci. Instrum. 85, 11D827 (2014)

• Operational since 2015• Conventional polarimetric (photo‐elastic modulation) method

• 25 channels with radial resolutions < 2 cm with ~ 10 msec interval

• Most of the systematic errors (Faraday, mirror reflections, geometric projections) have been calibrated.



Front optics

Filter & detector module

VW

PEM (M

DB

P

NBI

tan(2P ) I(2I(2

Fibers( 40 m)

KSTAR#7331F6C4

656 658 660 662 nm

0

Zeeman-split Dα

Zeeman-split Hα

Two Zeeman-split CII(657.81, 658.29)

MSE multiplets

> 20 times

a.u.

0.5e4

1.0e4

1.5e4



KSTAR#7331F6C4

656 658 660 662 nm

0

Zeeman-split Dα

Zeeman-split Hα


MSE multiplets

> 20 times

a.u.

0.5e4

1.0e4

1.5e4



KSTAR#7331F6C4

656 658 660 662 nm

0

Zeeman-split Dα

Zeeman-split Hα


MSE multiplets

> 20 times

a.u.

0.5e4

1.0e4

1.5e4

Spectral region for photo‐elastic modulation‐ Narrow bandpass filter is used


Very simple spectral analysis already helping bandpass filter calibrations

ch12, 2000.5mm, 2.9T, 100/00/00keV, , fwhm = 0.18nm, w_qr = 0.2/0.2, +0.30 nm, 3.44 deg, 12640a

658 659 660 661 662 6630.0

0.2

0.4

0.6

0.8

1.0

Filter function

Convolutedtransmission onlyaccepts components.

0 1 2 3 4

MeasurementModel

•Filter tuning: 0.2 ‐ 0.4 nm red‐shift from +3pi


Very simple spectral analysis already helping bandpass filter calibrations

90919293949596

pitch angle (deg)

0.140.160.180.200.220.240.26

lpf

0.0 0.2 0.4 0.6cwl offset from +3pi (nm)

0.0005

0.0010

0.0015

0.0020

0.0025

sqrt(a12^2+a22^2)

2.2T, ch03

•Quantitative analysis on lpf, cpf etcenables precise determination of the optimized amount of filter offset.

•Recently, this calibration has been performed from beam‐into‐gas injection tests (independent of the regular plasma run time).

0.91.01.1

-1.0 -0.5 0.0 0.5 1.0temporal frac�on around a crash

0.9

1.0

1.1

(a) Te (keV)

(b) q0


Direct derivation of q from pitch angle –q0 correlated with sawtooth events

Ko, Rev. Sci. Instrum. 87 (2016) 11E541

• q0 evolving around 1 – expected not to be well below 1 after the offset correction.

• q0 < 1 occurs before the sawtooth crash (confirming the internal kink) and the current build‐up (q0 > 1) is slow after the crash.

R. Giannella et al, Rev. Sci. Instrum. 75 (2004) 4247‐4250C. Petty et al, Plasma Phys. Control. Fusion 47 (2005) 1077‐1100

0.91.01.1

-1.0 -0.5 0.0 0.5 1.0temporal frac�on around a crash

0.9

1.0

1.1

(a) Te (keV)

(b) q0


Direct derivation of q from pitch angle –q0 correlated with sawtooth events

Ko, Rev. Sci. Instrum. 87 (2016) 11E541

• q0 evolving around 1 – expected not to be well below 1 after the offset correction.

• q0 < 1 occurs before the sawtooth crash (confirming the internal kink) and the current build‐up (q0 > 1) is slow after the crash.

Shot #18186 (Tue 23 May 2017)

1.0

1.5

2.0

2.5

10.0 10.1 10.2 10.3 10.4 10.50.90

0.95

1.00

1.05

1.10

0, ./mse_ps/m

se_display_qprofile_18186a.ps, Fri Jun 9 10:05:43 2017


Flat (or near‐hollow) q profiles with q0 1.5 during steady ITB formation.

0.0 0.2 0.4 0.6 0.8 1.0r/a:0

2

4

6

8

sa�e

y fa

ctor

, q

16498 (ITB/H/L) 2.8 sec (ITB) 4.5 sec (H-mode) 6.7 sec (L-mode)

1, ./mse_ps/m

se_display_qprofile_itp_160928d.ps, Fri Oct 28 09:38:36 2016

Inferred directly from MSE data using analytic models

ITB H‐mode L‐mode





0.0 0.2 0.4 0.6 0.8 1.0r/a:0

2

4

6

8

sa�e

y fa

ctor

, q


1, ./mse_ps/m


ITB




0.0 0.2 0.4 0.6 0.8 1.0r/a:0

2

4

6

8

sa�e

y fa

ctor

, q


1, ./mse_ps/m


17250 (steady ITB) 2.8 sec (ITB) 5.0 sec (ITB) 6.8 sec (ITB)


ITB



Time line for work plan • Construction of self‐calibrated high‐precision ( 0.05 nm) spectral diagnostic

• Development of a single‐ion‐source‐injection fitting routine and interface for existing atomic data and modeling packages (such as NOMAD, ADAS etc)

• Introduction of the polarization‐distortion effects to the spectral analysis suite• Systematic comparison with the PEM‐base MSE

• Evaluation and assessment of the atomic data used in the spectral analyses• Optimization of the spectral analysis suite for ITER application

Year 1 (2017)

Year 2 (2018)

Year 3 (2019)

Documents

IAEA, AMD Unit Page - mon crp neutral, …mon 19 jun 2017, j ko, iaea‐crp‐neutral, vienna, austria 19 Reliable inputs possible from the existing KSTAR MSE KSTAR#7331F6C4 656 658