Assessment and mitigation of wall light reflection in ITER ... Documents/Fusion... · Assessment and mitigation of wall light reflection in ITER by ray tracing Shin Kajita Nagoya

Assessment and mitigation of

wall light reflection in ITER

by ray tracing

Shin Kajita

Nagoya University

Evgeny Veshchev, Maarten De Bock,

Robin Barnsley, Manfred von Hellemann, Michael Walsh

ITER Organization

The views and opinions expressed herein do not necessarily reflect those of the ITER Organization.

Acknowledgement ITER Organization

Steve Lisgo, Michael Walsh

RSC Kurchatov Institute

Andrei Kukushkin

QST

Eiichi Yatsuka, Hiroaki Ogawa, Tatsuo Sugie,

Kiyoshi Itami

Wall reflection in fusion devices

Da During an ELM

• reflections is small ~typically<10% for divertor view channels, but can be significant ~50% for other view chords.

E. Hollmann, RSI (2003).

DIII-D (2003) JET (2013), ILW

The ratio of divertor stray light (DSL) to SOL light, DSL/SOLL vary from ~1 to 5

A.B. Kukushkin, EFDA-JET-CP(13)04/05 (2013).

-Reflectance is ~50% for W and Be. -Reflectance is higher than carbon based machines (~10%) (In IR range, it is much higher (~90%).)

-Reflectance should increase by roughly one order of magnitude when changing from C to W.

ITER: full metallic walls

Contrast of emission/signal can be high in ITER

• Emission in the divertor region, which is

several (2-3) orders of magnitudes greater

than those from SOL.

• CXRS, beam attenuation can be three orders

of magnitude. (Core signal is very very weak

compared with the edge signal)

Typical emission

profile in ITER (calc.

by S. Lisgo)

Model

Assessment of stray light in ITER

• Ha monitor (SOL spectroscopy)

• DIM (divertor impurity monitor)

• CXRS (Charge exchange recombination

spectroscopy)

• Edge laser Thomson scattering

Mitigation of stray light

Contents

LightTools (Illumination Design Software) is a 3D optical

engineering and design software product that supports

virtual prototyping, simulation, optimization, and

photorealistic renderings of illumination

applications. (synopsys home page)

Ray tracing software LightTools

https://optics.synopsys.com/lighttools/

Stray light modeling in SOL

• First walls and divertor tiles

were installed.

• To reduce the time of the

calculation and the file size,

only eighteenth of the

whole toroidal section (20

degree) was installed to the

model

• Two perfect mirrors were

installed for the boundaries.

• Hundreds of toroidal shaped sources are installed.

• Ray tracing has been done using LightTool for the assessment of

the influence of the optical reflection at the wall.

Stray light modeling using LightTools

• First walls and divertor tiles

were installed.

• To reduce the time of the

calculation and the file size,

only eighteenth of the

whole toroidal section (20

degree) was installed to the

model

• Two perfect mirrors were

installed for the boundaries.

• Ray tracing has been done using LightTool for the assessment of

the influence of the optical reflection at the wall.

• Source was simplified. The number of the source was reduced to

~500.

+y -y

+y

-y +y

-y

Ray tracing using LightTools

• Since the source profile is toroidally asymmetric, 3D full vessel model is required for (i) active CXRS.

• 1/18th toroidal section model is used for (ii) cold components and (iii) brems, in which the sources are toroidally symmetric.

Assessment of wall reflection for

Edge Thomson Scattering

Laser beam

Intermediate image

FOV of collection optics

DSM aperture

M3

M4 M2 M1

Entrance pupil

This part was modeled precisely.

• 1:1 relationship between the positions on FOV and the intermediate image. • Real size entrance pupil.

Specular and diffuse reflectance

• Optical reflectance property is one of the important parameters.

• Diffuse reflectance, Rd, and specular reflectance, Rs.

• Specular reflectance, we usually assume Gaussan with 12 degree in full 1/e width.

Optical reflectance property (DIM)

• Reflectance is ~50% for Be

and W in visible range.

• For divertor impurity monitor

system, measured BRDF was

used.

• The profile can be well fitted

with double Gaussian and

Lambertian profiles.

Model





spectroscopy)



Contents

Case Pedestal/

Edge

Far-SOL

v_perp (m/s)

Far-SOL

Te (eV)

d L-mode 30 10

e H-mode 30 10

f L 30 20

g H 30 20

h L 65 10

i H 65 10

j L 65 20

k H 65 20

l L 100 10

m H 100 10

n L 100 20

o H 100 20

1000

800

600

400

200

0

Em

iss

ion

po

we

r [W

]

d e f g h i j k l m n o

From divertor

From SOL

Emission in different discharge scenarios

•Emission profiles of Ha and Be I

(457.4 nm) calculated by S. Lisgo are

used in the model.

•Scenarios l-o are quite different from

scenarios d-k.

Ha

Be I (457.4 nm)

6

103

2

4

6

104

2

4

6

105

2

4

Em

iss

ion

po

we

r [W

]

d e f g h i j k l m n o

From divertor

From SOL

Typical radiance profiles (2D)

300

200

100

0

-100

-200

-300

y [

mm

]

200 100 0 -100 -200

x [mm]

2.5 Mrays

-13

-12

-11

-10

-9

-8

Lo

g(

Illu

min

an

ce [

W/m

m2]

)

300

200

100

0

-100

-200

-300

y [

mm

]

200 100 0 -100 -200

x [mm]

2.5 Mrays

-13

-12

-11

-10

-9

-8

Lo

g (

Ill

um

inan

ce [

W/m

m2]

Without

reflection

With reflection

Reflection 50%

(25% diffusive reflection, 25% specular

reflection (6 degree Gausian))

Comparisons in case d and case o (Ha)

• The ratio of diffusive to

specular reflection was

changed from 98% to 2%.

• Increase in the diffusive

components flatten the

emission profile. However,

the influence to the global

picture is not significant.

Total reflection is important.

• Stray light is estimated to be

one to two orders of

magnitude greater than the

real signal.

S. Kajita et al., PPCF(2013).

Model





spectroscopy)



Contents

• For DP, one peak around

strike point.

• For UPP and EPP, two

peaks exist around the

inner and outer regions.

• The stray light level is not

so significant compared

with the SOL cases.

• Typically the stray light is

comparable or less than

of the real signal.

]

Ha cases DP-outer DP-inner

UPP EPP

S. Kajita et al., JNM (2015).

]

Be cases

• Different from H-alpha

cases, stray light could be

much greater in some

parts.

• E.g., for EPP and UPP, the

stray light above divertor

dome could be one to two

orders of magnitude

greater than the real

signal.

• Almost the same situation

for SOL measurements.

DP-outer

DP-inner

UPP EPP

S. Kajita et al., JNM (2015).

Model





spectroscopy)



Contents

Diagnostic beam D, H

H(NB) + Az+ H+ + A(z-1)+ (n=n1) H+ + A(z-1)+ (n=n2)+hn

(i) Reflection of active CXR signals

(ii) Cold components

(iii) Bremsstrahlung

Diagnostic beam

Active cxr signal can be identified in the

high density case

Typical power profile at the receiver (upper port)

Difference is clearly identified in

UPP receiver in the high density

scenario. Other cases, the

influence is not significant.

Fraction of diffusive reflection component is important


• He emission radiance for three FOV.

Fraction of divertor reflection component is important

The peak of the cold component from the divertor and SOL is more than 105 times greater than the CXRS signal, indicating that the detector can be easily saturated even if it has a 16 bit resolution.

It is necessary to take into account this effect.

Brems. can be twice

higher by the reflection

In high density case, the SNR is lower than 10 in r<0.4. In those region, the increase in SNR may have some impact even by ~50%.





spectroscopy)



Contents

Brems light reflection from div. region is OK, but line

emission from div. region could be crutial

28

450-600 nm

Core bremsstrahlung

Edge/SOL bremsstrahlung

Line emissions Solid lines: Reflective wall Dotted lines: Absorptive wall

Emission profile in plasma was referred from one of the attachment (#2436: qpeak=10 MW/m2 on divertor target and Prad=43 MW) of [ITER_D_LFL2FN] which was obtained thorough a density scan for the carbon-free ITER divertor with Ne seeding computed by SOLPS 4.3 code.

Intensities of line emissions especially for visible lines are drastically increased due to wall reflection because the source of line emission is dominantly near the X-point.

700-800 nm 900-1000 nm

https://user.iter.org/?uid=LFL2FN&version=v2.3



Then, how to mitigate?

• Viewing dump

• Detailed spectrum analysis

• Ray tracing may help to mitigate it.

…

Then, how to mitigate?

• Viewing dump

• Detailed spectrum analysis

• Ray tracing may help to mitigate it.

…

Viewing dump may decrease the stray

light level by an order of magnitude

Depending on the position, the viewing dump can decrease the stray light level to less than 10%. S. Kajita, PPCF (2013).

Usage of transfer matrix

ai is the coefficient corresponding the intensity

of the source. The intensity profile for source k

at position x is denoted as Ik(x).

Ji(x): ray tracing transfer matrix.

If the set of ai is obtained, radiance profile can be reconstructed.

Simplified 60 sources were used for the model.

Radiance is the summation of

the contribution of sources.

w/o reflection w/ reflection

• Obtain the set of ai by comparison to the case with reflection.

• Reconstruct the radiance without reflection using the transfer matrix w/o reflection.

• Reflection was well compensated.

• No further ray tracing run is necessary once developing the matrix.

Ha case

S. Kajita et al., Contrib. Plasma Phys. (2016)

Stray light level could be decreased significantly

by usage of ray tracing transfer matrix

Total stray light level could be decreased significantly typically less than 10 % if the original stray light level is 50-100%.

• Using the reconstructed

a and another transfer

matrix, stray light in

SOL FOV was assessed.

• SOL stray light was well

assessed.

Using DIV-IMP data,

SOL stray light can be

assessed with the help of

ray tracing simulation

Prediction of stray light

After the transfer matrix was constructed, no simulation run

required; this can be implemented in the analysis procedure.

• Influence of the stray light is modeled using LightTool

for Ha monitor, DIM, and CXRS, edge TS in ITER.

• In Ha monitor, stray light will dominate the signal.

Viewing dumps are inevitable.

• In DIM, since the real signal itself is high, the stray light

seemed to be minor for Ha. For Be, it could be

significant in some parts where the real intensity is weak.

• In CXRS, the influence of reflection should be taken into

account in the core region. (CXRS signal from edge, cold

component, and brems)

• For TS, reflected line emission from divertor can be

significant contribution of photon noise.

• Ray tracing can also be used to mitigate the stray light

using ray transfer matrix method with no additional run.

Conclusions

• Emission of active signal profile was calculated using SOS code.

• In low density case, the signal decreases by two orders of magnitude, while it decreases by three orders of magnitude in the high density scenario.

Emission source profile

Be cases

• In Be cases, the influence

was better than H alpha

cases.

• The stray light level can be

still more than one order

of magnitude greater than

the real signal.

• Diffusive reflectance may

increase the stray light in

some parts; the influence

is minor.


Stray light can be increased in the

range of Rd/Rs < 20 %

0.01

0.1

1

10

100

Str

ay

lig

ht

/ s

ign

al [%

]

100806040200

Rd/Rs [%]

Low density scenario

r

r

r

r 0.01

0.1

1

10

100

Str

ay

lig

ht

/ s

ign

al [%

]

100806040200

Rd/Rs [%]

High density scenario

r

r

r

r

Rd: diffuse reflectance, Rs: specular reflectance

Stray light significantly increased decreased for 0<Rd/Rs<20%.

When Rd/Rs~20%, the stray light can be ~20% in the core.

BRDF (bidirectional reflectance distribution function) is important to know

• Emission of active signal profile was calculated using SOS code.

• In low density case, the signal decreases by two orders of magnitude, while it decreases by three orders of magnitude in the high density scenario.

Discharge scenarios / emission profile

Three different wall reflections in CXRS

(i) Reflection of active CXR signals

Strong active CXR signal at the edge can be harmful to the core.

(ii) Cold components

Emission from the cold edge/divertor emission line will complicate the analysis.

(iii) Bremsstrahlung

Bremsstrahlung intensity can be increased by the wall reflection.

Transfer matrix

Simplified 60 sources were used

for the model.

W/o reflection, the profiles have

a simple shape with one peak.

When considering the reflection,

the profiles become complicated.



Usage of ray tracing for mitigation

• Ray tracing calculation can

be implemented in the

analysis procedure to

consider the effects of reflection.

Radiance is the summation of the

contribution of sources.

In matrix expression

Jk is transfer matrix for kth receiver a is radiance profile.

If radiance profile (a) can be

reconstructed using transfer

matrix, it can be used for

mitigation.

Transfer matrix example

Simplified 60 sources were used for the model.

W/o reflection, the profiles have a simple shape with one

peak. When considering the reflection, the profiles become

complicated.


Two discharge scenarios

• Two different scenarios, i.e. low density and high density scenarios, are chosen.