Microwave probe diagnostic for the lower hybrid four-way-splitter antenna on Alcator C-Mod O. Meneghini, I. Faust, D. Johnson, R. Parker, S. Shiraiwa, D. Terry, R. Vieira, G. Wallace, S. Wukitch MIT-PSFC June 24, 2010 * Work supported by USDOE awards DE-FC02-99ER54512 and DE-AC02-76CH03073 O. Meneghini (MIT-PSFC) EPS 37 th Plasma Conference 2010 - Dublin June 24, 2010 1 / 18
18
Embed
Microwave probe diagnostic for the lower hybrid four-way ...€¦ · l=0.001 m l=0.005 m l=0.010 m l=0.015 m l=0.020 m l=0.025 m l=0.050 m l=0.100 m Quickly scan the antenna phasing.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Microwave probe diagnostic for the lower hybridfour-way-splitter antenna on Alcator C-Mod
O. Meneghini, I. Faust, D. Johnson, R. Parker, S. Shiraiwa,D. Terry, R. Vieira, G. Wallace, S. Wukitch
MIT-PSFC
June 24, 2010
* Work supported by USDOE awards DE-FC02-99ER54512 and DE-AC02-76CH03073
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 1 / 18
Abstract
The new lower hybrid launcher (LH2) of the Alcator C-Mod tokamak is based on anovel 4-way-splitter concept. A diagnostic based on the microwave probes concept[Jacquet et al. 1997] has been installed to verify the LH2 design and study thephysics of LH wave coupling. A total of 32 dedicated probes measure the forwardand reflected power in a carefully selected set of the active and passive waveguidesof the LH2 grill. A new technique which relies only on the microwave probes formeasuring the edge density profile in front of the launcher is proposed.
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 2 / 18
Lower Hybrid current drive (LHCD)
• Slow wave branch with n‖ > 1 forωci << ω << ωce
• Highly efficient non-inductivecurrent drive and current profilecontrol
• LH waves are well described by:• Cold plasma wave theory
(governs wave propagation)• Electron Landau Damping
(governs wave damping)
• n‖ is the control nob• Accessibility ↑ as n‖ ↑• ηELD ↑ as n‖ ↑• ηCD ↑ as n‖ ↓
Slow waveFast wave
ncutoff
Density [1E18 m-3]
N||=1.4
N||=1.5
N||=1.3
N2
Dispersion relation of fast (dashed) and slow (solid)waves in a cold plasma for Alcator C-Mod plasmaparameters
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 3 / 18
C-Mod LH2
4.6 GHz, 16x4 grill (60x7mm WGs), based on 4-way-splitter concept
RF probes
ReflectometerLangmuir probes
• Design goals:• Increase transmission
efficiency• Improve wave coupling• Lower breakdown probability
• Several diagnostics installed
0 5 10 15 20 25 300
1
2
3
4
5
6
7
8
9
10
Ver
PLT
Asdex
Alcator-A
FT
PBX-M
Alcator-C
TSJET
C-Mod LH1 C-Mod LH2(10 200kW Klystrons)
C-Mod LH2(16 200kW Klystrons)
f2b (GHz cm)
Power Flux (kW/cm2)
Hard conditioning limit
Soft conditioning limit
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 4 / 18
C-Mod LH2
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 5 / 18
4-way-splitter concept
• Power splitting at front end of thelauncher
• Simplified low loss feeding network(WR187, copper)
• Input impedance is resilient touneven poloidal plasma load
• Flexible N‖ spectrum
• Directivity is almost independentof plasma load
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 6 / 18
4-way-splitter simulated performances
0 0.5 1 1.5 2 2.50
5
10
15
20
25
30
Density at grill [1E18 m-3]
% average reflection coefficient
N||=1.6
N||=1.9
N||=2.2
-6 -4 -2 0 2 40
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1Normalized power spectrum
N||
N||=1.6 , ne=0.5E18 [m-3]
N||=1.6 , ne=1.25E18 [m-3]
N||=1.9 , ne=0.5E18 [m-3]
N||=1.6 , ne=1.25E18 [m-3]
N||=2.2 , ne=1.25E18 [m-3]
||=2.2 , ne=0.5E18 [m-3]N
More power is coupled through the rowswhich have more favorable density.
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 7 / 18
Microwave probe diagnostic
• Two probes displaced by ∆ = λg/4sense the total wave field in thewaveguide
Probes mounted on modules fed directly by a klystronProbes mounted on modules fed by a split klystronProbes mounted on dummy columnsUnused probes housings
e-, current drive direction
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 8 / 18
Microwave probe diagnostic
• Coupling of ≈ −65 ± 2 dB:• Microwaves coupled through
small circular hole• Central conductor is welded to
the opposite side of the housing
• Silicon-Dioxide (SiO2) cables areused in-vessel to ensure:
• Stability of phase with respectto temperature variations
• Low losses
• Phase and amplitude are measuredby off-the-shelf homodyne IQdetectors
• 250 KHz digitization rate
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 9 / 18
Measuring the edge density profile
Common practice to use the ambiguity of the edge density profile to makecoupling simulations fit to the experimental measurements
0 2 4 6 8 10
x 1018
0
0.2
0.4
0.6
0.8
1
Probe Density [m−3
]
Γ2
λ = 40000 [m−1
]
n||=1.5
n||=2.3
n||=3.1
Simulation vs experimental measurements for LH1, G. Wallace
Detail of langmuir probes and reflectometer horns on LH2
Necessity to eliminate free parameters:
• Self-consistent validation of LHcoupling codes
• Optimize design of future antennas forspecific plasma profiles
Measurement of density profile is on LH2:
• SOL X-mode reflectometer at threepoloidal locations
• Langmuir probes of different length(1 mm and 2 mm) estimate edgedensity and gradient
• Phase scan method, a new techniqueto fit a model density profile using linearcoupling theory
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 10 / 18
Phase scan method - General idea
Reflection coefficient strongly dependson the N‖ of the launched wave (i.e.antenna phasing)
0 20 40 60 80 100 120 140 160 1800
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
phase
Γ
Γ as a function of phasing (ne0
=4E17[m−3
])
λ=0.001 m
λ=0.005 m
λ=0.010 m
λ=0.015 m
λ=0.020 m
λ=0.025 m
λ=0.050 m
λ=0.100 m
• Quickly scan the antenna phasing.Assuming that linear theory holdstrue, the density profile can beinferred by the least square fittingof the measured reflectioncoefficients to the ones predictedby linear coupling theory
• LH grill antenna as a edge densityprofile diagnostic
• Average antenna reflectioncoefficient Average density profile
• On LH2, microwave probes allowdensity measurement at each row
• On LH2, X-mode reflectometer willbe used to validate this method
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 11 / 18
Phase scan method - Example
• E.g. Assume linear densityprofile in the experimentne0 = 9E17 [m−3] andλ = 15E−3 [mm] was to bemeasured
• Least square with fit twovariables linear density model:ne0 and λ
• ALOHA1D code, in the range1.2× 1017 < ne0 < 30× 1017
m−3 and 0.1 < λ < 100 mm• RMS error of the density
profile fitting procedure, for a50 to 145 degrees phase scan(to keep experimental Γ2
within acceptable limits)
5 10 15 20 25 30
10
20
30
40
50
60
70
80
90
100
RMS error for ne0
and λ fit
ne0
[10−17
m]
λ [10
−3 m
]
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 12 / 18
Relative: Obsorber loads in each of thefingers of a 4-way-splitter andmeasure coupling of incidentwave from each of the probes.
Absolute: Measure true amplitude andphase of the incident wave withOdaptor
• IQ detector assembly: IQ detectors,
amplifiers, attenuators, filters, DC breaks
• Back-mapping procedure(compensates for IQnon-linearities). Proven to beresilient to temperaturevariations. Assumes constantamplitude LO signal.
Obsorber: ECCOSORB 117 piramids comb (-30 dB
simulated, -27 dB measured) 1.5cm insertion depth
Odaptor: reduced waveguide to WR187 adaptor with
low insertion loss (-30 dB simulated, -20 dB
measured) at 4.6 GHz
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 13 / 18
1D simulation of grill antenna
The LH grill antenna theory gives analytic equations for the electric field and theplasma surface impedance for the linear density profile [1,2].Simulation results were tested against those Airy function based expressions,showing very good agreement.
N ||
-14 -12 -10 -8 -6 -4 -2 0
1.5
1.0
0.5
0
Z
Gamma
3000
2000
1000
0
Anal
ytic
eq.
is in
valid
[1]M. Brambilla, Nulear Fusion,16, 47 (1976)
[2]S. F. Knowlton and M. Porkolab, Nulear Fusion, 29, 1543 (1989)O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 14 / 18
2D model of Grill antenna
The modeling of cold plasma + collisions can be done in a straightforward manner.The following grill antenna simulation shows the well-known resonant conepropagation [P. Bellan and M. Porkolab, PRL 34 (1975) 124 ].
• Port boundary conditions (specify the forward power at the wave guide)
• Radiative boundary implemented by collisions
• Cold plasma dielectric εr
• Collisions are introduced by imaginary part of mass
Phasing
WG ports
Pla
sma
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 15 / 18
S matrix shows good agreement with TOPLHA
Comparison with TOPLHA (a LH antenna code based on the boundary elementmethod) shows good agreement with FEM model.
ne
ne0 =0.5×1018m-3
dne/dx=1.0×20m-4
ne0 =0.5×1018m-3
dne/dx=2.5×20m-4
ne0 =2.0×1018m-3
dne/dx=1.0×20m-4
WG# WG# WG#
• Thick lines : TOPLHA
• Thin lines : 2D FEMdne/dr
ne0
r
LH grill
Vgap
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 16 / 18
3D cold plasma modeling with FEM
The remaining difference between COMSOL and TOPLHA motivated the 3Dsimulation of the grill antenna.
3D antenna model of the traditional grill
COMSOLTOPLHA
Comparison of the waveguide reflectivitycalculated by COMSOL and TOPLHA
Coupling shows good agreement between between COMSOL, TOPLHA andALOHA-2D (not shown here).
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 17 / 18
3D FEM simulation shows better agreement
ne0 =0.5×1018m-3, dne/dx=1.0×20m-4
ne0 =0.5×1018m-3, dne/dx=2.5×20m-4
2D COMSOL and TOPLHA 3D COMSOL and TOPLHA
O. Meneghini (MIT-PSFC) EPS 37th Plasma Conference 2010 - Dublin June 24, 2010 18 / 18