Toroidally resolved measurements of ELMs in RMP and non-RMP H-mode discharges on DIII-D M.W. Jakubowski 1 , T.E. Evans 3 , C.J. Lasnier 4 , O. Schmitz 2 , M.E. Fenstermacher 4 , R. Laengner 2 , R.C. Wolf 1 , L.B. Baylor 3 , J.A. Boedo 5 , K.H. Burrell 3 , J.S. deGrassie 3 , P. Gohil 3 , R.A. Moyer 5 , A.W. Leonard 3 , C.C. Petty 3 , R.I. Pinsker 3 , T.L. Rhodes 5 , M.J. Schaffer 3 , P.B. Snyder 3 , H. Stoschus 2 , T. Osborne 3 , D. Orlov 5 , E. Unterberg 3 , J.G. Watkins 6 1 Max-Planck-Institut für Plasmaphysik, IPP-EURATOM Association, Greifswald, Germany 2 Forschungszentrum Jülich, IEF-4, Association FZJ-EURATOM, TEC, Jülich, Germany 3 General Atomics, P.O. Box 85608, San Diego, California, 92186-5608 U.S.A. 4 Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, CA 94550, U.S.A. 5 University of California, San Diego, La Jolla, CA 92093, U.S.A. 6 Sandia National Laboratory, Albuquerque, New Mexico, U.S.A. M.W. Jakubowski, et al., PSI San Diego (2010)
Transcript
Toroidally resolved measurements of ELMs in RMP and non-RMP H-mode discharges on DIII-DM.W. Jakubowski1, T.E. Evans3, C.J. Lasnier4, O. Schmitz2, M.E. Fenstermacher4, R. Laengner2, R.C. Wolf1, L.B. Baylor3, J.A. Boedo5, K.H. Burrell3, J.S. deGrassie3, P. Gohil3, R.A. Moyer5, A.W. Leonard3, C.C. Petty3, R.I. Pinsker3, T.L. Rhodes5,M.J. Schaffer3, P.B. Snyder3, H. Stoschus2, T. Osborne3, D. Orlov5, E. Unterberg3, J.G. Watkins6
1 Max-Planck-Institut für Plasmaphysik, IPP-EURATOM Association, Greifswald, Germany2 Forschungszentrum Jülich, IEF-4, Association FZJ-EURATOM, TEC, Jülich, Germany3 General Atomics, P.O. Box 85608, San Diego, California, 92186-5608 U.S.A.4 Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, CA 94550, U.S.A.5 University of California, San Diego, La Jolla, CA 92093, U.S.A.6 Sandia National Laboratory, Albuquerque, New Mexico, U.S.A.
M.W. Jakubowski, et al., PSI San Diego (2010)
Outlook
• Motivation
• Experimental set-up with two toroidally separated
infra-red cameras
• Evolution of ELM behavior from non-RMP to mitigated
RMP scenario.
Motivation• The common picture of ELMs is that they are filamentary
structures creating toroidally symmetric heat loads, when averaged over time. How does it change with RMP?
• On DIII-D application of n=3 resonant magnetic perturbation fields (RMP) allows to achieve H-mode with ELMs suppressed or mitigated. What can we say about heat loads due to mitigated ELMs?
• As ELMs prevent accumulation of impurities inside the plasma volume a scenario with very small, well controlled ELMs would be beneficial for ITER.
3.64.0
q95
510
20
579 Ptot [MW]
1000 1500 2000 2500 3000 3500 4000 45000
2
4Icoilu [kA]
time [ms]
0.61
1.2
Ptot [MW]
PIR [MW]
H98
In RMP: once H98 comes back to pre-RMP value very small ELMs appear
1 2 3 4
• We have realized four discharges with different q95: 3.5, 3.9, 4.1, 4.3.• Four different phases of the discharge.
Wetted area increases linearly with ELM size in H-mode discharges
Wetted area is defined as:
M.W. Jakubowski et al., Nuclear Fusion 49 (2009) 095013
Slope in non-RMP H-mode is a function of plasma current
Wetted area is defined as:
M.W. Jakubowski et al., Nuclear Fusion 49 (2009) 095013
Slope of wf = f(Edep) changes with plasma current
A = tan a
1.1 1.15 1.2 1.25 1.31
2
3
4
5
6
plasma current [MA]A 1
0-3 [m
/kJ]
3.53.94.14.3
q95
non-RMPRMP
M.W. Jakubowski, et al., PSI San Diego (2010)
Without RMP evolution of ELM structures shows 3D dynamics
• Evolution and structure of heat flux density distribution on the surface of lower divertor can be very different (bottom graphs), but there are cases, where the evolution is rather similar (top example).
• Energy deposited per ELM is defined as:
• With toroidal symmetries defined as:
j = 160° j = 55°
j = 160° j = 55°
smaller ELM
larger ELM
M.W. Jakubowski, et al., PSI San Diego (2010)
With RMP heat deposition patterns follow structure of stochastic boundary
• Introducing RMP (5 kA, q95 = 3.5) changes evolution of type-I ELMs to small events following topology of the stochastic boundary.
• Smaller ELMs “fill” two lobes of striated footprints
• Larger ELMs expel enough energy to “fill” the third lobe (bottom graph).
j = 160° j = 55°
j = 160° j = 55°
smaller ELM
larger ELM04080120160200240280320360-1
0
1
2
3
4
5
6
7Magnetic footprint ISP #139745@2800ms
toroidal angle [deg]
s
wal
l from
ISP
outw
ard
[cm
]
L c [m]
200
400
600
800
1000
1200
1400
1600
1800
0
4
8
12
= 160° = 55°
1000 1500 2000 2500 3000 3500 4000 45000
2
4
time [ms]
0,6
1
1,4
1
2
6
9
Without RMP larger variability of deposited energy and wetted area between toroidal locations
• Application of RMP reduces significantly ELM energies. Higher heating power (9 MW) results in stronger ELM mitigation.
• Without RMP some ELMs show toroidal asymmetries up to 50% on average there is no toroidal asymmetry (RE) between energy deposited on two toroidal locations
• Without RMP there is also rather strong variability of wetted area (Rw) between two locations.
• Introducing RMP reduces variability of deposited energy and wetted area, but creates small asymmetries in deposited.
M.W. Jakubowski, et al., PSI San Diego (2010)
On average no toroidal asymmetry without RMP
• Without RMP (blue curves):• average ELM energy almost does
not change with q95 with variability of deposited energy of about 50%.
• on average no toroidal asymmetry in deposited energy and slight asymmetry in wetted area.
0
10
20
30
579P tot [MW]
1600 1800 2000 2200024
Icoilu [kA]
time [ms]
P IR[MW]
3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.3
0,81
1,21,4
2468
1012
3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.30.6
0.8
1
1.2
1.4
q95
toroidal symmetry of energy - RE
no RMP, Ptot
= 6 MW
toroidal symmetry of wetted area - Rw
deposited energy Edep [kJ]deposited energy [kJ]
symmetry of wf - Rw
symmetry of Edep - RE
M.W. Jakubowski, et al., PSI San Diego (2010)
RMP at q95 closer to resonant window reduces ELM energies and variability.
• Without RMP (blue curves):• average ELM energy almost does not
change with q95 with variability of deposited energy of about 50%.
• wetted area change rate increases linearly with q95
• on average no toroidal asymmetry.• RMP at Ptot = 6 MW (red curves):
• reduces average energy by factor of 2• toroidal asymmetry depends on q95
0
10
20
30
579P tot [MW]
2400 2600 2800 3000024Icoilu [kA]
time [ms]
P IR[MW]
3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.3
0,81
1,21,4
2468
1012
3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.30.6
0.8
1
1.2
1.4
q95
toroidal symmetry of energy - RE
no RMP, Ptot
= 6 MW
RMP, Ptot
= 6 MW
toroidal symmetry of wetted area - Rw
deposited energy Edep [kJ]deposited energy [kJ]
symmetry of wf - Rw
symmetry of Edep - RE
M.W. Jakubowski, et al., PSI San Diego (2010)
Higher Ptot enhances coupling of RMP with plasma: smaller ELMs and toroidal asymmetries.
• Without RMP (blue curves):• average ELM energy almost does not
change with q95 with variability of deposited energy of about 50%.
• wetted area change rate increases linearly with q95
• on average no toroidal asymmetry.• RMP at Ptot = 5 MW (red curves):
• reduces average energy by factor of 2• toroidal asymmetry depends on q95
• RMP at Ptot = 9 MW (green curves):• reduces ELM energy even better (8 kJ 3 kJ) • toroidal asymmetries are slightly higher 0
10
20
30
579
P tot [MW]
3800 4000 4200 4400024
Icoilu [kA]
time [ms]
P IR[MW]
3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.3
0,81
1,21,4
2468
1012
3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.30.6
0.8
1
1.2
1.4
q95
toroidal symmetry of energy - RE
no RMP, Ptot
= 6 MW
RMP, Ptot
= 6 MW
RMP, Ptot
= 9 MW
toroidal symmetry of wetted area - Rw
deposited energy Edep [kJ]deposited energy [kJ]
symmetry of wf - Rw
symmetry of Edep - RE
M.W. Jakubowski, et al., PSI San Diego (2010)
Without RMP ELMs spanned over wide spectrum of energies.
0 2 4 6 8 10 12 14 16 18 200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
prob
abili
ty o
f the
eve
nt
1500ms < t < 2200 ms
deposited energy [kJ]
q95 = 3.5q95 = 3.9q95 = 4.1q95 = 4.3
• Without RMP one observes rather wide spectrum of ELMs (up to 20 kJ).
• Shape of the population distribution does not depend on q95
0
10
20
30
579P tot [MW]
1600 1800 2000 2200024
Icoilu [kA]
time [ms]
P IR[MW]
M.W. Jakubowski, et al., PSI San Diego (2010)
Population curve of ELMs has two subgroups different ELMs?
0 2 4 6 8 10 12 14 16 18 200
0.2
0.4
0.6
0.8
1
prob
abili
ty o
f the
eve
nt
2300ms < t < 3000 ms
deposited energy [kJ]
q95 = 3.5q95 = 3.9q95 = 4.0q95 = 4.3
• Introducing RMP at 6 MW heating power suppresses large ELMs.
• In case of q95 = 3.5 almost 90% of ELMs are below 3 kJ.
0
10
20
30
579P tot [MW]
2400 2600 2800 3000024Icoilu [kA]
time [ms]
P IR[MW]
M.W. Jakubowski, et al., PSI San Diego (2010)
At Ptot = 9 MW and q95 = 3.5 virtually all ELMs below ITER limit.
• At higher power coupling of RMP with plasma is better
• Strong dependence of population distribution on q95
• In the case of q95 = 3.9 virtually all ELMs deposit less than 3 kJ).
0 2 4 6 8 10 12 14 16 18 200
0.2
0.4
0.6
0.8
1
deposited energy [kJ]
prob
abili
ty o
f the
eve
nt
ELM distribution, inner leg, SBFP, 3700ms < t < 4500 ms
q95 = 3.9q95 = 4.2q95 = 4.4
0
10
20
30
579
P tot [MW]
3800 4000 4200 4400024
Icoilu [kA]
time [ms]
P IR[MW]
M.W. Jakubowski, et al., PSI San Diego (2010)
Amplitude and frequency is very sensitive to q95
T. Evans, et al., IAEA, Geneve (2008)
M.W. Jakubowski, et al., PSI San Diego (2010)
Summary• On average in DIII-D H-mode plasmas without RMP type-I ELMs do not
introduce toroidal asymmetries in energy deposition. However, individual events show up to 50% toroidal asymmetries in deposited energy to the divertor and rather different evolution of heat flux density patterns.
• The wetted area increases linear with ELM size. Slope is a function of plasma current.
• Applying RMP at proper q95 significantly reduces energy deposited per ELM keeping virtually all events below certain level, which is compatible with ITER guidelines.
• Their structure of deposition patterns follows 3D topology of stochastic boundary, which also results in small toroidal asymmetries.
• RMP allowed to realize a scenario with very small, well controlled ELMs.
M.W. Jakubowski, et al., PSI San Diego (2010)
ELMs in RMP phase follow stochastic boundary
M.W. Jakubowski et al., NF 49 (2009) 095013
M.W. Jakubowski, et al., PSI San Diego (2010)
Weak effect of RMP in the initial phase
0 2 4 6 8 10 12 14 16 18 200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
prob
abili
ty o
f the
eve
nt
1000ms < t < 1400 ms
deposited energy [kJ]
q95 = 3.7q95 = 4.0q95 = 4.2q95 = 4.4
• In the initial RMP phase most of the ELMs deposit energy between 2 and 6 kJ
• Rather weak effect of RMP on ELM behavior
• Shape of the population curve does not depend on q95
0
10
20
30
579P tot [MW]
1100 1200 1300 1400024
Icoilu [kA]
time [ms]
P IR[MW]
M.W. Jakubowski, et al., PSI San Diego (2010)
Change of heat flux density
0 5 10 15 200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
maximum of the heat flux density [MW/m2]pr
obab
ility
of t
he e
vent
ELM distribution, inner leg, SBFP, 3700ms < t < 4500 ms
q95 = 3.9q95 = 4.2q95 = 4.4
0 5 10 15 200
0.05
0.1
0.15
0.2
0.25
maximum of the heat flux density [MW/m2]
prob
abili
ty o
f the
eve
nt
ELM distribution, inner leg, SBFP, 1500ms < t < 2200 ms
