Page 1 - Effects of Rotation on Turbulent Transport in Field Reversed Configuration
P. 1
The C-2W real time inference and plasma control system
J. A. Romero, C. Finucane, K. Phung and the TAE Team
TAE Technologies, Inc., 19631 Pauling, Foothill Ranch, CA 92610
Motivation
C-2W is a high beta Field Reversed Configuration (FRC) sustained by neutral beam injection.
Efficient neutral beam capture and retention requires maintenance of plasma axisymmetry to avoid stochastic diffusion losses.
Magnetic field lines are very elastic at high beta, so FRCs require stabilization mechanisms to approximately maintain plasma axisymmetry.
Plasma must be kept well positioned inside confinement vessel.
A conductive wall provides stabilization for discharge duration of
the order of the wall time.
As C-2W are lasting longer than the wall time, active feedback
mechanisms are required to maintain plasma axisymmetry and control plasma shape and position as pressure builds up.
Digital control system infrastructure
Data acquisition @ 2.5 MHz
Feedback control@ 100 kHz
SpeedGoat system (matlab/simulink)
FPGA IO Modules
FPGA /CPU control
Control system functions
Sensors and actuators
Chip inductor 3 axis probe
Planned expansion
Magnetic actuation mechanisms
Non axisymmetric control
Useful for correction of static or slowly rotating low n order modes.
Measurement loops embedded in saddle magnets.
Flux conserver emulation when Vloop controlled to zero.
Superconductor Resistive
Loop voltage
Plasma position control strategy
External anti-mirror field (transversally/ axially stable/unstable) Axial instability is then feedback stabilized ( bang-bang control) Stabilization requires low latency ( ~25 μs)
Flux control experiments
Acts on magnet currents to control flux at the wall.
Plasma starts at t=0
Flux swing control t <0 Flux conservation t >0
Preprogrammed Active control
Preprogrammed Active control
• 384 magnetic sensors and 36 magnet current measurements.
• Flux loops, diamagnetic loops and 3 axis magnetic probes
• 20 Axisymmetric magnets.
• 16 Saddle magnets
• Optional in vessel magnets.
L dIcoil −
coil dt dt
EQ 1–4 coils
dφ
disp =0
coil disp L dI +R I =V +dφ
Mirror 1&2 coils
coil coil coil coil dt
V ! α ⎛ dφdisp − L loop coil
dt dIcoil ⎞
• Possibility to add additional sensors for kinetic control etc. ( 60 free optical links)
• NBI acceleration voltage
• Divertor electrodes.
• Puff valves
⎜⎝ dt
dt ⎟⎠
Plasma beta.
Plasma control. System
control is required for FRC steady state operation at high control project at C-2W is focusing initially on magnetic
Conclusions
Bayesian observer for shape /position
Gaussian process tomography Tested off line with C-2W data Tested to run in FPGA @ 100kHz
Plasma shape control options
Prescribed flux at the wall. Flux conserver emulation ( Tested) . Prescribed excluded flux profile ( Planned) .
Current centroid, second current moment ( Planned) .
Zero flux at prescribed separatrix contour ( Planned) .
Prescribed separatrix position and size / vessel gaps ( Planned) . Requires transverse stability and axial position control.
methods.
C-2W plasma control system is in its initial commissioning phase. C-2W plasma control is already assisting operations using flux
control.
This will be followed by flux control using saddle magnets, plasma
shape and position control, and NBI acceleration voltage tuning with B field intensity.
References
J. Romero. Systems and methods for FRC plasma stability. PCT/ US2016/061730
J. Romero et al. Inference of field reversed configuration topology and dynamics during Alfvenic transients. Nature communications 9, 691 (2018)
J Svensson (2011). Nonparametric tomography using Gaussian Processes, JET. Internal report, EFDA-JET-PR(11)24, 2011.
T. Roche. Magnetic diagnostic suite of the C-2W field-reversed configuration experiment. Review of Scientific Instruments 89(10): 10J107.
expansion for kinetic control has been provided.
Long term strategy for plasma control based on Bayesian
Magnetic control
Kinetic control ( planned)
Axis-symmetry control
Adjusts NBI acceleration voltage to field intensity.
Plasma position control
Electron density control with puff valves.
Plasma shape control
Plasma rotation control using electrode biasing
Forward model schematics.
Resistive decay inference