7th Annual Theory and Simulation of Disruptions Workshop

Latest results on runaway electrons experiments at FTU

FTU Experimental Campaign 2019-C1

M. Baruzzo, W. Bin, F. Bombarda, L. Boncagni, P. Buratti, L. Calacci, D. Carnevale, C. Castaldo, S. Ceccuzzi, C. Centioli, C. Cianfarani, F. Cordella,

O. D’Arcangelo, B. Esposito, L. Gabellieri, S. Galeani, S. Garavaglia, G. Granucci, M. Iafrati, A. Grosso, S. Magagnino, F. Martinelli, C. Mazzotta, F. Napoli, L. Panaccione, M. Passeri, V. Piergotti, G. Pucella, G. Ramogida,

G. Rocchi, A. Romano, M. Sassano, A. Sibio, B. Tilia, O. Tudiscoand the FTU team

G. Papp, G. Pautasso, J. Mlynar, C. Reux

7th Annual Theory and Simulation of Disruptions WorkshopPrinceton Plasma Physics Laboratory, Princeton, New Jersey August 5-7, 2019

https://pppltsdw.princeton.edu/

SUMMARY

D. Carnevale and L. Boncagni | FTU experimental results | TSDW 2019

Post-disruption RE beam scenarios: natural disruptions

• Diagnostics, additional equipment and control tools active on FTU

• Temperature increase of the background plasma: possible mechanisms

• REs instabilities observed at FTU

• Heavy metal interaction with quiescent REs and post-disruption beams

• Initial results on the use of ECRH on REs (quiescent and beam)

• REs beams trapped into MHD modes

• Long discharges

• Conclusions2

RE studies: diagnostics

Plasma current: up to 1.5MAToroidal field: up to 8TMajor radius: 0.96m, minor radius 0.3mPulse duration: up to 4s

3

FTU experimental results| TSDW 2019

New releasein September

New diamonddetectors

RE studies: additional equipment

4


DEUTERIUM PELLET INJECTOR

Small D2 pellet: 1x1020 ≈ 1200 m/sLarge D2 pellet: 2x1020 ≈ 1000 m/s -> time to reach the plasma core ≈ 0.3msInjection on a single discharge (horizontal): 2 small + 2 large

Used to rise density (fueling) up to 8x1020 with Ip=1.2MA (8T) [2001]

Diagnostics: Halpha, CO2 scanning interferometer (65 μs), Mirnov coils (MHD)Only horizontal pellet injector is available.

Fig. A: schematic of pellet system

Laser Blow Off (LBO) InjectorUsed in the past to trigger disruptions (E. Esposito studies on disruption mitigations)

Metal impurities injector by laser ablation of deposited metals on thin layers:Molybdenum, Iron, Tungsten, Zirconium, Yttrium.

Measurements: HXR, X-VUV spectrometer Schwob, Cherenkov.

RE studies: Control tools

• Current Quench and Onset RE plateau detectors (sensing Ip)• Current ramp-down reference or pre-programmed generic waveform (with

electrical field limitations to reduce MHD instabilities and radial drift – FTU/TCV)• Multiple switching integrators• Enhanced hybrid (switching) PIDs to control the beam position (radial and vertical)• Fast ramp-like controller to counteract fast events• Newly tested slow ramp-down controller to improve RE beam position tracking

during the non constant currents• Current allocation scheme among poloidal field coils• Vloop controller (used to impose loop voltage oscillations)

5


0

1

2

3

4

I p, Pel

lets

×105

0

50

100

ne(1

E18)

0

50100150

Hα

0 0.1 0.2 0.3 0.4 0.5 0.6time (s)

00.20.40.60.8

HXR

Pellets and LBO: early plateau phase (1/2)

Early plateau phase: approximatively within 300ms the plateau onsetDeuterium pellets: they do ablates (Halpha) but they do not ionize, on the contrary, temperature is decreased down-to recombination and density goes low. Cleaning effect.LBO: any sensible effect with all metallic species after the plateau onset.

6


Small pelleton flat-top

LBOs

Pellets

#41898

#41899Pellets: ne drops

Pellets: ne increases

Instabilities (fan-like)

Pellets and LBO: early plateau phase (2/2)

Early plateau phase: approximatively within 300ms the plateau onsetDeuterium pellets: they do ablates (Halpha) but they do not ionize, on the contrary, temperature is decreased down-to recombination and density goes to zero. Cleaning effect.LBO: any sensible effect with all metallic species after the plateau onset.

7


#41899 ECE Spectrum t= 0.35

t= 0.43

t= 0.35: RE dominated

t= 0.43: valley@300GHz due to thermal plasma absorption

0.51

1.52

2.5 ×105IpIprefp0p1p2p3

-0.0200.020.040.060.08 MHD

soft-XHXR

1.1 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.180

0.01

0.02

Che

renk

ov 58keV157 keV359 keV

0

0.5E20

1E20

n e

Pellet and LBO: late plateau phase (1/2)

The first (small) pellet is ablated and ionized as the second (large) pellet, while the third (large) cool down to recombination the background plasma leading almost to zero the density.

ECE spectral absorption and pellet ionization support plasma background temperature increase (and/or runaway energy decrease)

8


ZOOM

0.8 sPellets injected much later in the plateau phase do increase density.

157 keV Cherenkov probe got damaged

0.8

11.2

×105 42648IpIprefp0p1p2p3

-0.05

00.050.1 HXR

0

0.050.1

FC

0

500

HXR

x>40kex>100kex>180ke

0.7 0.8 0.9 1 1.1 1.2024

Che

renk

ov ×10-3

58keV157 keV359 keV

1010

neu213BF3

Pellet and LBO: late plateau phase (2/2)

Fe LBO has been used on a post-disruption RE beam undergone a current ramped-down. The OH coils does not sustain the current and instabilities have increased the background plasma temperature enough to ionize iron yielding increased Ipdecaying rate due to increased collisionality by high-Z particles.

9


0

1

2

3

4 ×105 42648 IpIprefp0p1p2p3

012

FC

0

500

HXR


0 0.2 0.4 0.6 0.8 1 1.2 1.40

0.5

Che

renk


1010neu213BF3

0

0.51

HXR

0.75 s

LBO injection

The pellets, cleaning the discharge, restore a flat current profile

0

1

2

I p, IpR

E:

×105 41902

-10123

V loop

0.35 0.4 0.45 0.5 0.550

0.05

0.1

0.15

0.2

0.25

Interferom

.[102

0m

−2]

r=0.90012r=0.92875r=0.95738r=0.98601r=1.0146r=1.0433r=1.0719r=1.1005r=1.1292r=1.1578r=1.1864

-0.5

00.5 MHD

HXRsoft-X

Plasma background heaters: instabilities (1/2)

Each time the Ohmic coil starts inverting current derivative (almost null flux or slightly negative provided to REs) anomalous doppler appears (at different frequency)

10


Anomalous doppler (fan-like)

There are different instabilities that are capable of heating the surrounding plasma in different ways.

Slower frequency instabilities (involving higher energy REs - FC) do not seem to be fan-like, no clear MHD signs and inverted ECE signal wrt fan-like.

Not fan-like

Note that the ’’not fan-like’’ induces vertical line density changes only on the middle (radial) part. The fan-like involves all of them.

Interferometer: vertical line density of mainly “cold” background plasma.

Plasma background heaters: instabilities (2/2)

11


Check differences between fan/not fan-like instabilities (ECE raw signal is E-C-PC.CH06)

REs current decay induced by fan-like inst.

Continuous deuterium puff, similar to D2 pellets, wipes out heavy particles and given the extremely low temperature during the RE plateau phase, recombination takes place yielding to negligible densities and almost null drag: the REs current evolves almost freely (T coil current constant before 0.24s).

0123

I p

×105 42661 IpIREneD2 puff

-0.2

00.20.40.6

HXR

MHDHXR

0

0.2

0.4

0.6

FC(1

E14)

0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.40

1020

V loop

Being the voltage non negative, the REs do not loose their energy and when a quick current shut down is imposed at 0.37 s (pulse program error), the beam energy is released against the vessel.

FC signal is practically zero during the plateau phase (except when fan instability sets in) since the beam is well confined. 12


0123

I p, Pel

lets

×105 42661Ippellet 0pellet 1

-0.1

0

0.1

0.2

HXR

MHDHXRVsaddle

0

500

HXR

E>40keE>100keE>180ke

0.1 0.15 0.2 0.25 0.3time (s)

0

0.2

0.4

Che

renk


0.2 0.25 0.30

0.0050.01

120kA130kA

Pellets usually induce prompt REs loss.

Instability sets in at about 0.22 s increasing the REs losses as marked by Cherenkov and HXR detectors.

13


REs current decay induced by instability

Total current Ip shows a mild decrease only due the instability ().

A second pellet destroys the instability and losses disappear (not unique). Do pellets affects RE beams instabilities similarly to MHD instabilities on plasma discharges?

Remark:There are a series of peculiar marks that seems to reveal the start of fast energy release/conversion (RE thermalization):• HXR, NEU213 begin to decrease• Radial flick.• Spike in the Vloop• ECE emission shows a non-cold

background plasma• HXR spectra peaks

Interestingly, similar marks have been found on two unique discharges at TCV where full conversion has been obtained (Electrical Field oscillations).

14


0

1

2

3

4

I p, IpR

E:

×105 42747

IpIREDensD puffNe

0

0.5

1

HXR

MHDsoft-XHXR

0 0.5 10

1

2

FC(1

E14)

,ete

rd:

-5

0

5

10

15 100

105

1010

gammaneu213BF3

0 0.5 11.05

1.11.15

1.2

Long lasting plateau: RE energy suppression

On long lasting RE beams the energy is: • Dissipated (drag, instabilities, error fields)• Transferred to the background plasma (instabilities)• Reduced (via central solenoid)

Harmless final loss(position control has limited capability on ITER, backup and/or simultaneous to others)

Vloo

pR

ext

LBO: continuous expulsions

In the shot #42655 a LBO injection of Fe at 0.9 s induced MHD instabilities that seem to continuously expel REs.

3

3.5

I p

×105 42655

0

200400

HXR


0.6 0.7 0.8 0.9 1 1.1 1.2 1.30

0.558keV157 keV359 keV

100

1010

n, γ gamma

neu213BF3

0

0.20.4

HXR MHD

soft-XHXR

0.8 0.9 1 1.1 1.2 1.3 1.4time (s)

1010

15


LBO on quiescent runaways at ramp-down

123

I p

×105 42645

-0.1

00.10.2 MHD

soft-XHXR

00.20.4

FC

0

50100

HXR


0.8 0.9 1 1.1 1.2 1.3 1.40

0.1

0.2

Che

renk


100

1010

n, γ gamma

neu213BF3

Fe induces large MHD instability burst leading to RE expulsion.

Low density discharge with REs (quiescent) in current flat-top: LBO injection at 1.24 s of Fe.

16


LBO on quiescent runaways at ramp-down

Low density discharge with REs (quiescent) in current flat-top: LBO injection at 1.2 s of Fe.

2

3

I p, Pel

lets ×105 42647

-0.1

00.10.2

HXR

MHDsoft-XHXR

0

200

400

HXR


0.95 1 1.05 1.1 1.15 1.2 1.250

0.050.1

0.1558keV157 keV359 keV

100

1010

n,γ gamma

neu213BF3

Fe induces large MHD instability burst leading to RE expulsion.

17


LBO: considerations

LBO seems to be able to induce “small” disruptions expelling REs in flat-top (quiescent) as well as ramp-down and post-disruption: abruptly or continuously.

The main key parameter seems to be the level of (background) plasma temperature that allows for ionization.

ITER: current quenches of hundred of milliseconds would possibly allow continuous LBO injections to limit REs formation (particularly second mechanisms) - provided enough temperature for ionization.

ECRH: • Increase the temperature of the background plasma • Provide extra free electrons ionizing puffed deuterium during CQ to

absorb flux and to possibly limit the maximum RE energy • Do heated electrons by ECRH possibly reduce the CQ drop?

18


ECRH on REs: initial results

ECRH is configured as for plasma current assisted break-down experiments:

• Continuous ECRH on early RE beam phase: no clear effects on density.

• On RE Soft-stop (ramp-down triggered due to high HXR) with D2 puff, the RE expulsion synchronizes with ECRH pulse modulation (20Hz).

Note the multiple small CQs (0.5s, 0.75s, 1.05s): reduced CQ drop?

19


0

2

4

6

8

I p, IpR

E:

×105 43211IpIrefneD2 puff

0

0.10.2

FC

06

V loop

γ

0 0.2 0.4 0.6 0.8 1 1.20

0.10.2 58keV

157 keV359 keV

-0.2

00.20.40.6 HXR

MHD

ECRH on REs: initial results

• On quite low energetic post-disruption REs with continuous D2 puffing, the RE expulsion synchronizes with ECRH pulse modulation (50Hz): with higher frequency no exact correspondence with MHD spikes, no multiple plateaus…

Note that current reference profile is not a simple ramp-down:

20


0

2

4

6

8

I p, IpR

E:

×105 43214IpIrefneD2 puff

0

0.050.1

FC

0V lo

op γ

0 0.2 0.4 0.6 0.8 1 1.2 1.40

0.05

0.1 58keV157 keV359 keV

-0.2

00.20.40.6 HXR

MHD

AIM: Increase the current to measure the fraction of flux absorbed by REs and heated background plasma (energy limiting).

REs trapped by MHD in quiescent REs

0

1020

Cher

enko

v

E>40keE>100keE>180ke

0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5time (s)

00.050.1

HXR

-0.2

00.2

Vsad

0.20.30.4

Soft-

x 45

I p

×105 42540Ipne

On flat-top low density discharges large rotating MHD modes marked by voltage oscillations of the saddle coil (Vsad) trap REs.

Cherenkov probes level confirm REs confinement.

21


Images of the visible camera, sampled at 50Hz, shows peculiar REs marks trapped by MHD modes.

Long discharges: density cycles

0

2

4

6

I p [A]

×105 #42666

IpneD puff

-0.1

0

0.1

MH

D [A

.U.]

0 0.5 1 1.5 2 2.5 3 3.5time [s]

100

1010

γ [c

ount

s] gammaneu213BF3

starts withREs

Long discharges density cycles obtained by feedback action resulted not smooth and reached to low values triggering mode growth.

22


In the next experiments we will provide a pre-programmed reference (dashed black) to smoother the density evolution and avoid MHD growth.

Density cycles with constant current on quiescent plasma and the REIS should provide good data for RE energy models validation.

Conclusions

- Long lasting RE beams with almost no residual energy (more than 1.1s, extendable)- Temperature increases again: RE phase transitions and/or energy transfer to

background plasma.- Dissipation effects of RE losses due to fan-like/MHD instabilities on

quiescent and post-disruption RE beams. Plasma waves emissions by REs are studied, waves in the 400-800 MHz range have been detected.

- LBO can increase/induce RE dissipation/losses- Pellets seem to stabilize instabilities on post-disruption RE beams- Quantifiable effects of RE losses due to fan-like instability on a RE beam- REs orbits trapped by MHD large modes on quiescent REs- ECRH initial results on quiescent REs, post-disruption and Soft-Stop (w/o

modulation)Control dream/obsession: inject the minimum amount of materials for thermo-mechanical loads, reduce the current drop (more current curriers provided by ECRH and less drag thanks to D2 pellets/puff cleaning effect) and the REs peaking energy (flux absorption by many electrons and larger interaction with higher temperature background plasma)…improved controllability for ITER!!!

FTU (almost certainly) conclusive campaign: are you interesting to participate?! 23


24

ADDIONAL SLIDES

Final Loss: new findings

25

FTU: approximately 150 ms after the CQ the beam exhibits (80%) a sudden radial inner movement (HXR drop below saturation) .

TCV: two ramp-down with Vloop oscillations shown sudden loss of all REs, radial inner movement (T increase, Li decrease) and Ohmic plasma since then: never seen before.

Radial shift approx. FTU: lost/conversion of high energetic REs

TCV: lost/conversion of all REs

RE dynamics are affected by hysteresis: oscillations of Vloop might enhance the RE conversion (overcrossing the hysteresis threshold) into thermal electrons.A new possible strategy to limit magnetic to kinetic RE energy conversion at final loss.

WRE is the RE energy

LBO: W on low level of quiescent REs

Runaway control | WIP 25 Mar 2019 | Page 2

A small effect of LBO injection of W at 0.9 s has been observed on the shot #42653. Probably, given the low level of REs, the temperature was high enough to ionize the W and its effect is slightly visible.

3.4

3.5

3.6

I p

×105 42653

0

200400

HXR


0.6 0.7 0.8 0.9 1 1.1 1.20

0.1

0.258keV157 keV359 keV

-0.1

00.10.2

MHDsoft-XHXR

1010

1012

n, γ

gammaneu213BF3

26

RE losses: Fan instability

27

Study of correlations with Electron Cyclotron Emission and HXR from NE213 scintillator.

Latency after disruption(large photo-neutron signal)

ECE spikesECE spikes are correlated with MHD and NE213

RE losses: Fan instability

28

l ECE: not the usual thermal emission, it is the low-frequency tail of synchrotron emission by RE.

l ECE increase at microsecond time scale can only be due to anomalous pitch angle scattering.

l Anomalous pitch angle scattering due to the “fan instability” is well known (Vlasenkov 1973, Parail 1976, Coppi 1976).

l NOT MHD but kinetic instability, driven by momentum space anisotropy of RE.

l HXR increase due to larger diffusion at larger pitch angle

l MHD spike due to increase of perpendicular beta.

Importance:l Anomalous pitch angle scattering can play an important role in RE

beam dynamics. In fact, with an increase of the pitch angle synchrotronlosses increase and the maximum RE energy decreases.

Documents

7th Annual Theory and Simulation of Disruptions Workshop