New schemes for MHD stability control New schemes for MHD stability control in RFX-mod using the real-time MARTe in RFX-mod using the real-time MARTe framework framework Chiara Piron 3 rd PhD Event in Fusion Science and Engineering York, June the 23 rd 2013 boration with: G.Manduchi, L.Marrelli, P.Piovesan, P.Zanca, A.Garofalo, M. O
New schemes for MHD stability control in RFX-mod using the real-time MARTe framework Chiara Piron 3 rd PhD Event in Fusion Science and Engineering York,
Mar 27, 2015
Welcome message from author
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
New schemes for MHD stability control in RFX-mod New schemes for MHD stability control in RFX-mod using the real-time MARTe frameworkusing the real-time MARTe framework
Chiara Piron
3rd PhD Event in Fusion Science and EngineeringYork, June the 23rd 2013
In collaboration with: G.Manduchi, L.Marrelli, P.Piovesan, P.Zanca, A.Garofalo, M. Okabayashi
York, June the 23rd 2013 Chiara Piron
The RFX-mod experimentThe RFX-mod experiment
10-0.1
1
r/a
ar
Padova RFX-mod: a Reversed Field Pinch (RFP) that can operate also as a
Tokamak
Bt directionreversed
Comparable Bt, Bpamplitude
Bt
Bp
p
t
RB
rBq
1/10
York, June the 23rd 2013 Chiara Piron
The MHD stability problemThe MHD stability problem
Ip[kA]
q(a)
2/1 RWM Br [mT]
# 29774 ctrl OFF # 29797 ctrl ON
Ip[kA]
q(a)
1/6 RWM Br [mT]
# 18542 ctrl OFF # 18541 ctrl ONRFPTOKAMAK
AIM OF MY PhD ACTIVITY: develop and implement in realtime new feedback control strategies to counteract MHD instabilities to optimize the performance both in tokamak and RFP plasmas. Such control strategies can be used as control schemes in reactor-relevant scenarios, such as advanced
tokamak regimes
2/10
York, June the 23rd 2013 Chiara Piron
The RFX-mod feedback controlThe RFX-mod feedback control
How is it possible to control MHD instabilities?
3/10
York, June the 23rd 2013 Chiara Piron
The RFX-mod feedback controlThe RFX-mod feedback control
192 MAGNETIC SENSORS (Br, Bt, Bp)How is it possible to control MHD instabilities?
3/10
York, June the 23rd 2013 Chiara Piron
The RFX-mod feedback controlThe RFX-mod feedback control
192 MAGNETIC SENSORS (Br, Bt, Bp)How is it possible to control MHD instabilities?
3/10
York, June the 23rd 2013 Chiara Piron
The RFX-mod feedback controlThe RFX-mod feedback control
192 MAGNETIC SENSORS (Br, Bt, Bp) + 192 ACTIVE COILSHow is it possible to control MHD instabilities?
3/10
York, June the 23rd 2013 Chiara Piron
The RFX-mod feedback controlThe RFX-mod feedback control
192 MAGNETIC SENSORS (Br, Bt, Bp) + 192 ACTIVE COILS
MARTe fCYCLE = 5kHz
It is like opening and closing the shower turncock…
5000 times per second!
How is it possible to control MHD instabilities?
5kHz
3/10
York, June the 23rd 2013 Chiara Piron
Par == 1 ? CMC
Par == 2 ?
Par == 4 ?
Par == 3 ?
PLASMA RESPONSE
DOUBLE SHELL CMC
POLOIDAL CTRL
Extrapolation?
EXTRAPOLATOR
PID
FEEDFORWARDCURRENT
CURRENTFREEZING
PID
Cleaning?
Par == 0?
ReadingParameters
All Pars = 0
RAW CTRL
The MARTe frameworkThe MARTe framework
Multi-threaded Application Real-Time Executor [1]: a C++ development enviroment for real-time applications. It has been introduced in RFX-mod, where both software and hardware upgrades in the feedback control have been recently carried out
T F
T
T
T
T
T
T
F
F
F
F
F
Br, Bt spectrum Bp spectrum
active coils
4/10
[1] A. Netò et al., IEEE Trans. Nucl. Sci. 57 (2010)
York, June the 23rd 2013 Chiara Piron
Par == 1 ? CMC
Par == 2 ?
Par == 4 ?
Par == 3 ?
PLASMA RESPONSE
DOUBLE SHELL CMC
POLOIDAL CTRL
Par == 0? RAW CTRL
The MARTe frameworkThe MARTe framework
T
T
T
T
T
F
F
F
F
[2] Strait E.J. et al. 2003 Nuclear Fusion 43 430 [3] Liu et al. Plasma Phys. Control. Fusion 51 (2009)
Why so many control schemes?To contribute to the scientific debat on the optimal sensors for feedback:Poloidal sensor control is successful in tokamak experiments[2, 3] BUT the sensor are located inside the vessel, thus they can not be used in a fusion reactor. Radial sensors can be used instead BUT they are less effective because of sideband aliasing.
challenge: find the optimal algorithm to clean the radial measurements and make the radial magnetic field sensor control as much effective as the poloidal one
The Fourier harmonics of the signals
An analytic cylindrical model is applied to clean the sideband aliasing
A refinement of CMC where both the vessel and the shell are considered in the model
A control scheme that uses a 8X4 array of poloidal sensor
The field produced by the active coils is subtracted from the measurements
5/10
York, June the 23rd 2013 Chiara Piron
Par == 1 ? CMC
Par == 2 ?
Par == 4 ?
Par == 3 ?
PLASMA RESPONSE
DOUBLE SHELL CMC
POLOIDAL CTRL
Extrapolation?
EXTRAPOLATOR
PID
FEEDFORWARDCURRENT
CURRENTFREEZING
PID
Cleaning?
Par == 0?
ReadingParameters
All Pars = 0
RAW CTRL
The MARTe frameworkThe MARTe framework
Multi-threaded Application Real-Time Executor [1]: a C++ development enviroment for real-time applications. It has been introduced in RFX-mod, where both software and hardware upgrades in the feedback control have been recently carried out
T F
T
T
T
T
T
T
F
F
F
F
F
Br, Bt spectrum Bp spectrum
active coils
6/10
[1] A. Netò et al., IEEE Trans. Nucl. Sci. 57 (2010)
York, June the 23rd 2013 Chiara Piron
• The Plasma Response control scheme fails in suppressing the instability, regardless the applied proportional gain.Experiments with derivative gains are planned in the next future.
• Preliminary results seem to suggest that a proportional gain threshold exists for the growth rate saturation, as predicted by [1]
The Plasma Response control schemeThe Plasma Response control schemeFeedback variable = the component of the radial magnetic field that is obtained by subtracting the field produced by the active coils from the measurements
[4] G.Marchiori, et al., Nucl. Fus, 52 (2012)
RMC: unstablePlasma Response: unstableCMC: stable
RMC: unstablePlasma Response: unstableCMC: stable
GainGain
[4]
7/10
York, June the 23rd 2013 Chiara Piron
• Two similar discharges: #33576 reference shot#33634 frozen currents at t=[0.5, ..] s
• When q(a)<2 the 2/1 RWM always grows despite error fields are compensated
Error Field Correction (EFC)Error Field Correction (EFC)
Ip[kA]
q(a)
2/1 RWM Icoil [A]
2/1 RWM Br [mT]
# 33576 # 33634 freezing ON
Why does EFC seem not to work in RFX-mod?Probably because we are using the wrong
feedback variable Try to use the poloidal sensors instead
8/10
EFC control scheme was developed at the DIII-D tokamak to operate at high beta values and to investigate the RWM physics [5]
[5] A. Garofalo et al., PRL, 23 (2002)
Can the 2/1 RWM at q(a)<2 be destabilized by error fields?
York, June the 23rd 2013 Chiara Piron
Poloidal magnetic sensor controlPoloidal magnetic sensor control
#33808
Θ=342°
Θ=297°
Θ=252°
Θ=207°
#33808Bp signals [T] array @ φ=322.5°
2,1 RWM Bp [mT]#33808
9/10
2D FFT
beforeafter algorithm processing
before before (realtime) after algorithm processing
York, June the 23rd 2013 Chiara Piron
Conclusions and future plans..Conclusions and future plans..
• Active feedback control is an essential and effective tool to counteract the growth of MHD instabilities both in RFP and tokamak plasmas
• The development and of new MHD feedback algorithms contributes to the scientific debat on the optimal feedback control strategies (ITER relevance for advanced tokamak scenarios)
• The Plasma Response and the EFC are just two of the available RFX-mod control schemes that put light on the inner physics of RWM dynamics and act as benchmarks for theoretical models
To do..
Plasma Response, EFC and other control schemes are planned to be further tested in real experiments both in Tokamak and RFP campaigns
Data analyses on the collected data will be carried out to characterize the nature of the MHD instabilities and to extract theoretical model to describe their dynamics
10/10
Related Documents
