1 st iaea meeting on fusion data processing, validation and analysis, june 2015, nice, france peter...
TRANSCRIPT
![Page 1: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/1.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 1IDM UID: QRGTY3
Requirements for triggering the ITER Disruption Mitigation System
P.C. de Vries1, G. Pautasso2, D. Humphreys3, M. Lehnen1,
S. Maruyama1, J.A. Snipes1, Vergara1, L. Zabeo1.
1 ITER organization, Route de Vinon sur Verdon, 13115 St Paul Lez Durance, France.2 Max-Planck-Institut für Plasmaphysik, 85748 Garching, Germany.
3 General Atomics P.O. Box 85608, San Diego, California 92186-5608, USA.
Disclaimer: This presentation includes new directions for management of disruptions that are not yet introduced into the ITER technical baseline. These results don’t commit the nuclear operator. The views and opinions expressed in this paper
do not necessarily reflect those of the ITER Organization.
![Page 2: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/2.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 2IDM UID: QRGTY3
IntroductionA disruption of the tokamak discharge is an unfortunate phenomenon, in which the control and confinement of the plasma is lost within a very short duration.
The fast release of plasma energy could result in large thermal and electromagnetic loads that may affect the life-time of its components.
Therefore, such events should be avoided and otherwise mitigated.
![Page 3: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/3.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 3IDM UID: QRGTY3
Disruption prediction
Disruption avoidance, prediction and mitigation
The following aspects are relevant to disruption avoidance
Detailed preparation and validation of tokamak operationsbefore a discharge
Continuous analysis to resolve problems and optimize further operationsafter a discharge
Real-time event detection and control/avoidance/termination by PCSduring a discharge
However if a disruption is inevitable, the last line of defense at ITER is to trigger the disruption mitigation system (DMS).
![Page 4: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/4.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 4IDM UID: QRGTY3
Scope of this presentationThe term ‘disruption prediction’ can have a too broad interpretation.
This presentation will discuss the requirements specifically to trigger the ITER disruption mitigation system (DMS)
To develop these requirements the following questions need to be answered: What is a disruption and how can they be detected? What mitigating techniques will be applied? What is their impact and what are the device design limits? How many disruptions are to be expected and can be tolerated? How will ITER be operated?
Information relevant to the requirements for the DMS trigger will be indicated in red on various slides.
![Page 5: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/5.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 5IDM UID: QRGTY3
What is a disruption?A tokamak disruption is made up out of several different facets, and the order in which they appear may differ:
Vertical displacement event (VDE) loss of VS Thermal quench (TQ) loss of confinement Current quench (CQ) too high a resistivity Runaway electrons (REs) too fast a CQ
The following variants can occur …
![Page 6: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/6.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 6IDM UID: QRGTY3
What is a disruption? Minor disruption starts with TQ but is not followed by CQ.
t = -10ms t = 0ms t = +10ms
![Page 7: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/7.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 7IDM UID: QRGTY3
What is a disruption? Major disruption starts with a TQ followed by CQ (and possibly VDE,RE)
t = -10ms t = 0ms t = +10ms
![Page 8: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/8.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 8IDM UID: QRGTY3
What is a disruption? Hot VDE starts with a VDE which triggers a TQ and CQ (and possibly RE)
t = -10ms t = 0ms t = +10ms
Note that, besides these main three, further variants exist
![Page 9: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/9.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 9IDM UID: QRGTY3
How can these events be detected?
For a predictor one may first think of disruption precursors, but for a trigger to the DMS, the most obvious might be the detection of the disruption itself:
VDE Detect maximum vertical excursion DzMAX[1]
CQ Detect large |dIp/dt|[2]
TQ Too fast and prediction required.
Variants of the first two schemes are used on several devices and it has been shown that such triggers could be sufficient to still properly mitigate forces and some of the heat loads[2].
[1] Y. Zhang, et al., Nucl. Fusion 51 (2011) 063039.[2] C. Reux, et.al., Fus. Eng. Des (2012).
A disruption is either initiated by a TQ or a VDE, hencethe detection/prediction of both a VDE and TQ are essential for the trigger
![Page 10: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/10.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 10IDM UID: QRGTY3
Detection/prediction of a VDE
[1] D. Humphreys, et al., Nucl. Fusion 49 (2009) 115003.
DIII-D: increasing elongation in a single discharge to increase growth rate Z until VDE
Uncontrollable VDE occurs at[1]:
Zmax/<Z>noise ~ 2–3
Consistent with controllability threshold DIII-D: VS control lost at Zmax/a~2% (red dashed line) Typical <Z>noise/a ~ 0.7% Zmax/<Z>noise ~ 3
Projecting evolution of Zmax via simulations can predict impending loss of controllability!
VDE
![Page 11: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/11.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 11IDM UID: QRGTY3
Requires a highly reliable and timely prediction that a TQ is inevitable and no other control options, but DMS are possible.
There are a number known precursors to precede a TQ, such as the growth of large magnetic islands, ideal kink modes, etc.
Prediction of the TQ
[1] P.C. de Vries, Nucl. Fusion 49 (2009) 055011
For Locked Mode Threshold
![Page 12: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/12.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 12IDM UID: QRGTY3
Requires a highly reliable and timely prediction that a TQ is inevitable and no other control options, but DMS are possible.
Improving the physics basis for these individual TQ precursors allows transfer of thresholds to future devices, such as ITER.
Prediction of the TQ
[1] P.C. de Vries, Proc. of 41st EPS conference (2014, Berlin)
![Page 13: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/13.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 13IDM UID: QRGTY3
How to mitigate their impact?Traditionally the thermal and electromagnetic loads due to disruptions are mitigated by the massive injection of high Z impurities (Neon or Argon)
Done either in the form of gas mixtures (Massive Gas Injection, MGI[1,2,3]) or by shattering pellets (Shattered Pellet Injection, SPI[4]).
The high Z impurities will: increase radiation and reduce the energy that is convected to PFCs. affect the post TQ resistivity and thus affect the CQ duration. affect the generation of runaway electrons.
[1] D G Whyte et al Phys. Rev. Lett. 89 (2002) 055001.[2] G. Pautasso, et al., Nucl. Fusion 47 (2007) 900.[3] M. Lehnen, et al., Nucl. Fusion 51 (2011) 123010.[4] N. Commaux, et al., Nucl. Fusion 50 (2010) 112001.
![Page 14: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/14.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 14IDM UID: QRGTY3
How to mitigate their impact?The ITER Disruption Mitigation System (DMS) is currently being designed (CDR completed in 2012) will have multiple (individual) injectors (a MGI and SPI hybrid) grouped together on different ports[1].
[1] S. Maruyama, et al., Proc. 24th IAEA FEC (2012, San Diego, USA)[2] M. Lehnen, et al., Proc. SOFE conference (2015, Austin, USA)
UPP (Upper Port Plugs) #02, 08 and 12
EPP (Equatorial Port Plug) #08
When triggered, the DMS should be told how to fire, i.e. which individual injector, for example, to avoid radiation asymmetries.
![Page 15: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/15.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 15IDM UID: QRGTY3
How to mitigate their impact?The ITER Disruption Mitigation System (DMS) is currently being designed (CDR completed in 2012) will have multiple (individual) injectors (a MGI and SPI hybrid) grouped together at different ports[1].
Typical reaction times[2]: Delivery time by SPI: tactuator=25-30ms (UPP), tactuator=15-20ms (EPP)
Delivery/pre-TQ time by MGI: tactuator=10-15ms (UPP) or 2-3ms (from inside PP)
[1] S. Maruyama, et al., Proc. FEC (2012, San Diego, USA)[2] M. Lehnen, et al., Proc. SOFE conference (2015, Austin, USA)
UPP (Upper Port Plugs) #02, 08 and 12
EPP (Equatorial Port Plug) #08
Hence, a working assumption for the minimum trigger time is approximately t >30ms.
![Page 16: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/16.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 16IDM UID: QRGTY3
What impact is to be expected at ITER?A disruption has different facets and therefore also can impact the device in multiple ways and there are different tolerances for each of them:
Heat loads due to the fast release of the thermal energy (TQ)
Heat loads due to the release of part of the magnetic energy (CQ)
Heat loads due to the impact of runaway electrons (REs)
Forces due to too fast a CQ and eddy current forces
Forces due to too slow a CQ with respect to the VDE
![Page 17: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/17.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 17IDM UID: QRGTY3
What impact is to be expected on ITER?
Unmitigated MD heat loadsWth=350 MJ (worst case)
Be/W melt limit: ~25/50 MJm-2s-0.5
80 - 320 MJm-2s-0.5
130 - 280 MJm-2s-0.5
up to 770MJm-2s-0.5
Thermal loadsCaused by the loss of magnetic energy (CQ), the fast loss of thermal energy (TQ), and impact due to possible runaway electrons (REs).
[1] M. Lehnen, et al., Proc. PSI conference (2014, Japan)
Even for low current (Ip=5-6MA) a CQ can lead to local melting[1].
At TQ a Wth=25MJ can lead to shallow melting of the inner divertor[1].
Timely prediction of a TQ is needed for Wth>25MJ.
From Ip=5-6MA heat loads need to be mitigated.
![Page 18: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/18.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 18IDM UID: QRGTY3
Electromagnetic forces Fast current quenches (CQ) lead to large eddy current forces
Too slow a CQ with respect to the VDE leads too large halo current forces
What impact is to be expected at ITER?
![Page 19: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/19.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 19IDM UID: QRGTY3
Balancing halo and eddy current forces
Cat III
Cat II
Force due to eddy currents (MN)0 14
0
16
For
ce d
ue t
o h
alo
cur
rent
s (M
N)
downward hot VDE Ip=15MA tCQ = 36ms
Forces on blanket module No. 1
7
8
[1] M. Lehnen, et al., Proc. PSI conference (2014, Japan)
During the start-up of ITER operations, the disruption load models need to be confirmed experimentally.
![Page 20: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/20.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 20IDM UID: QRGTY3
Balancing halo and eddy current forces
Force due to eddy currents (MN)0 14
0
16
For
ce d
ue t
o h
alo
cur
rent
s (M
N)
Forces on blanket module No. 1
7
8
DMS should be given information how to fire, such as to optimize the CQ rate and hence, balance the halo and eddy current forces.
Too slowCQ
Too fastCQ
OK
![Page 21: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/21.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 21IDM UID: QRGTY3
How many can be tolerated over ITER life-time?
Cat III
Cat II
Force due to eddy currents (MN)0 14
0
16
For
ce d
ue t
o h
alo
cur
rent
s (M
N)
Forces on blanket module No. 1
7
8
~3000 events
~1-2 events
Above about Ip = 8.4 MA, it is possible for unmitigated disruptions to enter the Cat. III range[1] .
[1] M. Lehnen, et al., Proc. PSI conference (2014, Japan)
![Page 22: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/22.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 22IDM UID: QRGTY3
Disruption accounting and prediction performance
Ncorrect = Npulses
Nunintentional disr.
Nmitigated disr.
Nnatural disr.
Nnon-disruptive Nfalse Nlate Nintent.
Ndisruptions
* The number of false alarms can only be assessed with non-active predictor
*
![Page 23: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/23.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 23IDM UID: QRGTY3
Disruption accounting and prediction performance
Ncorrect = Npulses
Nunintentional disr.
N*mitigated disr.
Nnatural disr.
Nnon-disruptive Nfalse Nlate
The prediction performance can be defined by: Rprediction Ncorrect / Nunintional disr.
Rlate Nlate / Nunintional disr.
Rfalse Nfalse-alarm / (Npulses- Nintent.)
Runintentional disr. Nunintional disr. / (Npulses- Nintent.)
Nintent.
Ndisruptions
* This assumed that a correct trigger to DMS ensures a proper mitigation
![Page 24: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/24.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 24IDM UID: QRGTY3
Disruption accounting and prediction performance
The prediction performance can be defined by: Rprediction Ncorrect / Nunintional disr.
Rlate Nlate / Nunintional disr.
Rfalse Nfalse-alarm / (Npulses- Nintent.)
Their requirements depend on: How many unmitigated disruptions can be tolerated? Rtol = Ntol/Npulses
How many and what kind of pulses will be scheduled? Npulses
How many disrupt? disruption rate Rnatural or Runintentional disr
The false alarm rate should not dominate the disruption rate: Rfalse < 0.5 Rnatural
![Page 25: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/25.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 25IDM UID: QRGTY3
Disruption accounting and prediction performance
The disruption rate is determined by the success of the avoidance methods and therefore the prediction requirements are too.
For low tolerances (i.e. 1-2 events in 10000 pulses), even for very low Rdisr., the prediction performance should be Rprediction>98-99%
![Page 26: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/26.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 26IDM UID: QRGTY3
How many disruptions are to be expected?
This requires information on how ITER is going to be operated define the operational/research plan.
Below, a rough outline of ITERs progressive start-up, a possible path from low current non-active operation to full performance active operation[1].
[1] D. Campbell, The ITER Research Plan
0 5000 10000 15000 20000 250000
2
4
6
8
10
12
14
16
Pla
sma
cu
rre
nt (
MA
) H He H DD DT
Number of pulses
Q=10
![Page 27: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/27.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 27IDM UID: QRGTY3
How many disruptions are to be expected?
The disruption rate is determined by the success of the avoidance methods.
One can set reasonable targets for ITER operation[1,2] , based on experience from present day devices, such as JET.
[1] M. Sugihara, et al., 24th IAEA FEC (2012, San Diego)[2] P.C. de Vries, Nucl. Fusion 51 (2011) 053018
0 5000 10000 15000 20000 250000
2
4
6
8
10
12
14
16
Pla
sma
cu
rre
nt (
MA
) H He H DD DT
Number of pulses
![Page 28: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/28.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 28IDM UID: QRGTY3
Required performance
Although the design limits are well known, the tolerances to lower disruption impact, such as shallow melting due to a TQ, are not easy to assess.
The detection and prediction performance and false alarm rate have been calculated based on a simple estimate of tolerable number of disruptions[1]
.
Number of pulses
Prediction requirements differ for each operational phase.
[1] M. Sugihara, et al., 24th IAEA FEC (2012, San Diego)
![Page 29: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/29.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 29IDM UID: QRGTY3
Disruption prediction development First operations: medium requirements (Rpredict<80% and Rfalse~10%)
Can be achieved by simple, physics based disruption thresholds 1st level of prediction
For Ip=15MA operations high requirements (Rpred.>98% , Rfalse~2.5%). Currently are only achieved by advanced predictors[1] . 2nd level of prediction
[1] J Vega, et al., Fus. Eng Des. 88 (2013) 1228
At ITER may have time to develop/assess disruption prediction
![Page 30: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/30.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 30IDM UID: QRGTY3
Prediction and detection of disruptions
Different levels of the DMS trigger
Detection of the disruption itself (VDE, CQ)fall back option
Simple prediction of TQ or VDEfirst level of prediction
Balancing input from advanced predictors and forecasting methods by PCSsecond level of prediction
![Page 31: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/31.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 31IDM UID: QRGTY3
Requirements list
Heat loads due to the CQ need to be mitigated from Ip=5-6MA.
Timely prediction of VDEs and TQ (for Wth>25MJ) is needed.
A working assumption for the minimum trigger time is t >30ms.
DMS should be given information how to fire.
False alarms should not dominate the disruption rate: Rfalse < 0.5 Rnatural
For early operations, lenient performance req. Rpredict<80%, Rfalse<15%
But for higher currents (Ip>8.4MA): Rpredict>98% and Rfalse<2.5%
ITERs progressive start-up may allow time to develop disruption prediction.
![Page 32: 1 st IAEA Meeting on Fusion Data Processing, Validation and Analysis, June 2015, Nice, France Peter de Vries – © 2015, ITER Organization Page 1 IDM UID:](https://reader030.vdocuments.us/reader030/viewer/2022032707/56649e305503460f94b21009/html5/thumbnails/32.jpg)
1st IAEA Meeting on Fusion Data Processing, Validation and Analysis , June 2015, Nice, France Peter de Vries – © 2015, ITER Organization
Page 32IDM UID: QRGTY3
Summary
A disruption is made up of different facets (VDE, TQ, CQ, REs) that each develop on different time-scales and create different impacts, to which the device will have different tolerances. Therefore the prediction/trigger requirements may have to be determined per impact type.
Often the scope of ‘disruption prediction’ is too broadly defined. The requirements for event detection/prediction can only be determined clearly, if the event and its related actions are well defined.
This presentation aimed to give basic requirements of the trigger to the ITER DMS. Further details will depend on,
final design of the DMS, development of mitigation physics, improved tolerance assessment detailing of the operation schedule/research plan.