Transcript Corsica free-boundary scenario and controller studies for ITER

Corsica free-boundary scenario and
controller studies for ITER
L. L. LoDestro, T. A. Casper, W. H. Meyer, and L. D. Pearlstein
12th ITPA Confinement Database and
Modelling Topical Group Meeting
EPFL, Lausanne
May 7--10, 2007
Ackn: A. Portone, EFDA-CSU Barcelona
Work supported by the U.S. DOE Contract W-7405-48.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p1
Recent Corsica ITER modelling is focussed on
free-boundary predictive simulation:
ITER scenarios:
Assess VS consumption, PF-coil currents, strike- point
locations.
Verify the equilibria pass through the fiducial shapes.
Test the plasma shape-and-position controller.
We are developing code capability. Physics results
here are illustrative.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p2
Outline
• Start-up simulations for ITER and DIII-D
• ITER scenario studies
• ITER controller studies
• Corsica and Corsica/Simulink
• Summary/Conclusions
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p3
ITER start-up experiments at DIII-D will
be used for code validation
EFIT01 shapes
Corsica computed
equilibria
Ip
•EFIT boundaries at selected times provide “fiducial” shapes
•Corsica evolves internal flux/current densities using N,T measurements
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p4
Corsica is also being used to design and
model the DIII-D experiments simulating ITER
Ip
DIII-D
ITER
1st day of expt’s.
Ref. 2 with
faster Ip ramp.
Without trying
very hard, q0 & li
come close.
L-mode .
q0
sawteeth
• DD
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
Ip
More expt.’s to
follow.
li
sawteeth
p5
Simulations of ITER scenarios (following)
employ a special free-boundary technique
Prescribed shape scenario technique:
– Developed for Snowmass
– Plasma boundary is linearly interpolated between
fiducials
– Coil currents and Volt-seconds (VS) are backed out
– Result is similar to that with a perfect controller
Faster and simpler than using a controller.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p6
Simulation of ITER reference scenario 2 finds
PF-coil currents and VS are within limits
• Simulation includes neo, neo, GLF23,
analytic Paux,  heating & diffusion.
• n is prescribed.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p7
Ref. 2 simulation indicates JBS has a significant impact
on the plasma boundary (preliminary result).
With JBS
Without JBS
•Divertor strike-points were not prescribed here.
When they are prescribed along with the
boundary and JBS is included, the plasma shape
can not be held “close” to the fiducials--suggesting control might be problematic.
•Note increased range of divertor strike-points and
limiter point with JBS.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p8
ITER reference scenario 2, no JBS, cont.
• Simulation includes coil ramp-down through
return of VS.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p9
Two implementations for simulating
control with Corsica
• Corsica solves circuit equations
– Can run with implicit time-step  speed
• Corsica is coupled to Mathworks Matlab/Simulink ($$$)
– Circuit eqn’s solved by Simulink -- state-space controllers, explicit
– Modular
• Rapid installation of new controllers
• Entire coil description can reside outside the 1&1/2D code
– Corsica can serve as the plasma model in D. Humphrey’s (GA)
PCS, continue collaboration on devel. of Simulink environment.
• Can split the control: fast (VS) with Corsica, slow (shape)
with Simulink
p 10
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
Corisca/Simulink simulation is
distributed over network
Corsica on beowulf cluster
MDSPlus server
mdsip
Xwindows
and jScope
Rpc &
mdsip
X
Matlab node
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
We have successfully controlled VDEs & MD1,2 in
Corsica/Simulink with the JCT Feb 2001 controller
• Ohm’s law transport only.
• Inputs to controller:
Vertical Stability: d ZCS/dt
Shape: gaps, Ip, IPFC
• Outputs:
VS: voltage to PFC2--5
Shape: CS and PF voltages
• Delays, etc., included
Tested 0 ok (nec for implicit)
• Linear plasma model tests
were successful
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p 12
Corsica – JCT/Feb 2001 controller during
Vertical Displacement Event
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
Corsica models the large-scale plasma
disturbances
Minor Disruptions:
MD1
Instantaneous li drop of
0.2(li -0.5) and p
drop of 0.2p followed
by a 3s exponential
recovery.
Corsica p drop
Prescribed p drop
MD2
Instantaneous p drop
of 0.2p followed by
a 3s exponential
recovery.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
Corsica li drop
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
Graphics produced with MDSPlus jScope
Corsica – JCT/Feb 2001 Controller
during MD1
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
Graphics produced with MDSPlus jScope
Corsica – JCT/Feb 2001 Controller
during MD2
Controlled simulations with Te,i transport are being
carried out with Corsica performing the control
• Capability was developed for
original ITER-EDA
– Exercised with Ohm’s law only
– ITER-FEAT results agree with
Corsica/Simulink
• Here we include JBS. Other
transport and sources as before.
• Plasma bulges toward the upper
x-point.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p 17
Summary and conclusions
• Physics conclusions:
– Reasonable VS & PF currents found for ITER
reference scenario 2 (Ohm’s law + Te,i transport).
– Large edge bootstrap current is an issue
• affects shape & divertor strike-points
– The ITER-FEAT controller successfully controls 1cm
VDE’s and MD1, MD2 minor disruptions
• tested with Ohm’s law only, without large edge JBS
– Location of upper control point cf. upper X-point with
large edge JBS is an issue
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p 18
Summary and conclusions, cont.
• Code status:
– Two types of predictive free-boundary simulation are available:
backing out PF currents (ready); and solving circuit equations
with controllers.
– Two environments for plasma control simulations are available
(same physics): Corsica/Simulink code-coupling; or all
Corsica.
• Ohm’s law: working well.
• Ohm’s law plus Te,i transport: minor warts.
• Not in this talk: alternative start-up and performance scenario
simulation; profile control; synthetic diagnostics.
LoDestro, ITPA CDBM, Lausanne, May 7--10, 2007
p 19