Runaway Electron Beam Stability and Decay in COMPASS O

Ficker DOI:10.1088/1741-4326/ab210f EX/P1-26

Runaway Electron Beam Stability and Decay in COMPASS O. Ficker1,2, J. Mlynář1, E. Macusova1, J. Cerovsky1,2, A. Casolari1, M. Farnik1,2, A. Havranek1,3, H. Ales1,3, M. Hron1, M. Imrisek1,4, P. Kulhanek1,3, T. Markovic1,2, D. Naydenkova1, R. Panek1, M. Tomes1,2, J. Urban1, M. Vlainic5, P. Vondracek1,2, V. Weinzettl1, J. Decker6, G. Papp7, Y. Peysson8, M. Gobbin9, M. Gospodarczyk10, V. V. Plyusnin11, M. Rabinski12, and C. Reux8

The COMPASS Team and EUROfusion MST1 Team 1Institute of Plasma Physics AS CR v.v.i., Prague, Czech Republic 2Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University, Prague, Czech Republic 3Faculty of Electrical Engineering, Czech Technical University, Prague, Czech Republic 4Faculty of Mathematics and Physics, Czech Technical University, Prague, Czech Republic 5Institute of Physics, University of Belgrade, Belgrade, Serbia 6Swiss Plasma Center (SPC), École polytechnique fédérale de Lausanne (EPFL), 1015 Lausanne, Switzerland 7Max-Planck-Institut für Plasmaphysik, Garching, Germany 8Institut de Recherche sur la Fusion par confinement Magnétique (IRFM), Commissariat à l’énergie atomique (CEA/Cadarache), 13108 St. Paul lez Durance, France 9Consorzio RFX, Associazione EURATOM-ENEA sulla Fusione, Padova, Italy 10Università di Tor Vergata, 00173 Rome, Italy 11Institute of Plasmas and Nuclear Fusion (IPNF), Association EURATOM/IST, Lisbon, Portugal 12National Centre for Nuclear Research (NCBJ), Świerk, Poland Corresponding Author: O. Ficker, fi[email protected] Runaway electrons (REs) as one of the yet unsolved threats for ITER and future tokamaks are a topic of intensive research at most tokamaks. The experiments performed on COMPASS are complementary to the experiments at JET and MST (Medium-Size Tokamaks), building on the flexibility of the diagnostics set-up and low safety constraints at this smaller (R0 “ 0.56 m, a “ 0.23 m) device. During the past couple of years two different scenarios with the RE beam generation triggered by gas injection have been developed and investigated. The first one is based on Ar or Ne massive gas injection (MGI) into the current ramp-up phase leading to a disruption accompanied by runaway plateau generation, while the second uses smaller amounts of gas in order to get runaway current dominated plasmas. The successful generation of the beam in the first scenario depends on various parameters, including the toroidal magnetic field. The generated beam is often radially unstable, and the stability seems to be a function of various parameters, including the value of current lost during the CQ. Surprisingly, the current decay rate of the stable beams is rather similar in most discharges. The second scenario is much more quiescent, with no observable fast current quench. This allows to better diagnose the beam phase and also to apply secondary injections or resonant magnetic perturbations (RMP) to assist the decay of the beam. In this regard, interesting results have been achieved using secondary deuterium injection into a runaway electron beam triggered by Ar or Ne injection and also using n “ 1 error field generated by top and bottom RMP coils. While D dilution is clearly able to almost stop the beam decay, RMPs help to accelerate the beam decay. The effect of RMPs seems to be very different when acting on Ar and Ne background plasmas. Very interesting effects have been observed also by the high-speed cameras, including filamentation during the application of the RMPs and a slow local variation of the light intensity similar to turbulence during the beam decay.

Published as a journal article in Nuclear Fusion http://iopscience.iop.org/article/10.1088/1741-4326/ab210f Ficker et al.

RUNAWAY ELECTRON BEAM STABILITY AND DECAY IN COMPASS

O. Ficker1,2, E. Macusova1, J. Mlynar1, J. Cerovsky1,2, A. Casolari1, M. Farnik1,2, A. Havranek1,4, M. Hron1, M. Imrisek1,3, P. Kulhanek1, T. Markovic1,3, D. Naydenkova1, R. Panek1, M. Tomes1,3, J. Urban1, M. Vlainic5, P. Vondracek1,3, V. Weinzettl1, D. Carnevale6, J. Decker7, M. Gobbin8, M. Gospodarczyk6, G. Papp9, Y. Peysson10, V.V. Plyusnin11, M. Rabinski12, C. Reux10, COMPASS team1 and the EUROfusion MST1 Team*

1 Institute of Plasma Physics of the CAS, CZ-18200 Prague 8, Czech Republic 2 FNSPE, Czech Technical University in Prague, CZ-11519 Prague 1, Czech Republic 3 FMP, Charles University, CZ-12116 Prague 2, Czech Republic 4 FEE, Czech Technical University in Prague, CZ-166 27 Prague 6, Czech Republic 5 Institute of Physics, University of Belgrade, 11080 Belgrade, Serbia 6 Universita' di Roma Tor Vergata, 18-00133 Rome, Italy 7 Swiss Plasma Centre, EPFL, CH-1015 Lausanne, Switzerland 8 Consorzio RFX, 35127 Padova, Italy 9 Max-Planck-Institut für Plasmaphysik, Garching D-85748, Germany 10 CEA, IRFM, F-13108 Saint-Paul-lez-Durance, France 11 Instituto de Plasmas e Fusão Nuclear, IST, Lisbon, Portugal 12 National Centre for Nuclear Research (NCBJ), Otwock-Swierk, Poland

* See the author list "H. Meyer et al 2017 Nucl. Fusion 57 102014"

Email: [email protected]

Abstract

Runaway electrons (REs) as one of the yet unsolved threats for ITER and future tokamaks are a topic of intensive research at most of the European tokamaks. The experiments performed on COMPASS are complementary to the experiments at JET and MST (Medium-Size Tokamaks), building on the flexibility of the diagnostics set-up and low safety constraints at this smaller device. During the past couple of years two different scenarios with the RE beam generation triggered by gas injection have been developed and investigated. The first one is based on Ar or Ne massive gas injection (MGI) into the current ramp-up phase leading to a disruption accompanied by runaway plateau generation [1], while the second uses smaller amounts of gas in order to get runaway current dominated plasmas [2]. The successful generation of the beam in the first scenario depends on various parameters, including the toroidal magnetic field. The generated beam is often radially unstable, and the stability seems to be a function of various parameters, including the value of current lost during the CQ. The second scenario is much more quiescent, with no observable fast current quench and it is highly reproducible. This allows to reasonably diagnose the beam phase and also to apply secondary injections or resonant magnetic perturbations (RMP) to assist the decay of the beam. In this regard, interesting results have been achieved using secondary deuterium injection into a runaway electron beam triggered by Ar or Ne. The current of the RE beam can be controlled at a fixed value, however only using relatively high loop voltage. The radial position is not fully controlled and the request for the stabilising field is changing independently on plasma current which implies that the RE energy plays a key role in correct approach to beam position control. The experiments with elongated beams were also carried out. Last but not least it seems that Ar and Ne behave differently in terms of radiated power and HXR intensity during the beam decay.

1. INTRODUCTION

Runaway electrons (REs) have been extensively studied at COMPASS within the framework of EUROfusion MST1 work package as they still present an issue with respect to the safe operation of ITER [3]. Without securing mitigated disruptions with no RE being generated and/or without developing a fully reliable runaway beam mitigation technique, ITER can never succeed. It seems that that the typical timescales of disruptions in ITER will be crucial [4] for the beam generation and that the position stability in the post- disruptive phase remains the critical issue. Enough time to mitigate the RE beam can be only secured in case that the beam position is stabilized which may be extremely difficult. The shaping field will cause the vertical instability of the beam just after the disruption while correct stabilizing vertical field can be hardly optimized early enough to secure the radial position stability. Once the stability of the beam is ascertained, various mitigation techniques can be employed, including injection of large amount of noble gases, preferably based on shattered pellet injection that is selected as a mitigation technique for ITER. However, the amount of gas already present in the chamber from the disruption mitigation injection may be sufficient to slowly mitigate the beam,

1 IAEA- EX/P1-26

again in case that the beam position is reasonably stable. This challenging task has to be tackled at the currently operated machines and using complex modelling approach.

COMPASS and RE experiments

The COMPASS tokamak is one of the Europe’s flexible smaller machines that has been operated at IPP

Prague since 2008. The vacuum chamber is D-shaped with open divertor and major radius R0 is 0.56 m and minor radius a is 0.21 m. The toroidal magnetic field Bt ranges from 0.8 to 1.6 T while plasma current up to 350 kA can be driven in the machine. The toroidal magnetic field and plasma current directions can be both clockwise and counter-clockwise. The machine is equipped with two 40-keV NBIs with heating power up to 350 kW each which allow to reach H-mode, however COMPASS is able to reach even an ohmic-heating-driven H-mode as well [5]. For overview of the recent ITER-related results ranging from detachment studies to disruption analysis, see poster by Pánek et al. at the Conference [6]. The RE research on COMPASS gains from the machine flexibility and cooperation with groups working on relevant diagnostics and models. The RE generation in quiescent scenarios and losses related to various MHD phenomena and external field errors were studied as well as massive gas injection (MGI) triggered disruptions in the ramp-up scenario [1,7] – see the description in the next section. In the flattop discharges without gas injection the losses of RE seem to be modified to a large extent by sawtooth instability, magnetic island rotation and poloidal field power supply oscillations (introduced by flywheel rotation and set of 12-pulse convertors) [8]. The reaction of REs to the perturbed magnetic fields was recognized as an important topic and thanks to the rich variety of resonant magnetic field perturbation (RMP) coil system at COMPASS it is possible to run very detailed scans of the RMP effect on RE beam. So far, n=1 and n=2 LFS off-midplane coil configurations were tested with promising results [9] and good agreement with the results achieved at ASDEX-U [10]. Despite the fact that ITER ELM coils are probably too weak to affect the RE orbits in the confined beam as they are too far from the plasma, these experiments will contribute to the understanding of the beam behavior in the perturbed fields and to validation of theoretical predictions and models. These results as well as other experiments related to the two scenarios described below are published in [2] that is currently in press. The most urgent question is what situation is considered to be a successful beam mitigation and how to reliably achieve it. From the ITER point of view, the worst danger is that the beam can be lost in very localized manner and concentrated flux of high energy particles may damage first wall including the cooling elements or other important structures exposed directly to the plasma. The unmitigated RE beam termination may occur on much shorter timescales than a typical plasma disruption, at COMPASS, terminations of extremely low-density plasmas in the slide-away regime on microsecond timescales were recorded [2] due to radial position instability. These terminations also caused very localized hotspots. On the other hand, even with minor amount of high Z impurity gas added, no such events were observed and in the case of good position control, the current decay has been rather moderate. Although most of the relativistic electrons were possibly lost to the limiters, it did not happen in a focused manner. It is a question whether a similar mitigation strategy is sufficient in ITER and whether the position control can be maintained. The control of the RE beam is one of the key topics in Europe, see [11,12].

Diagnostics and gas injection valves

COMPASS is equipped by a rich set of magnetic diagnostics (multiple poloidal Mirnov coil arrays, flux loops, saddle loops, full and partial Rogowski coils, etc.) which contributes not only to reconstruction of dominant field configuration but also to study the magnetic fluctuations related to various instabilities. The density is measured and controlled using a 2 mm microwave interferometer, the density and temperature profiles are measured using Thomson Scattering (TS) system. The detection of REs is carried out using multiple methods, the low energy RE (about 100 keV) confined in the vessel volume are detected using a vertical ECE system [13] and the lost RE can be detected by Cherenkov detector [14,15] but namely using several HXR detectors measuring the secondary radiation. The photo-neutrons may be also detected by multiple neutron detectors, however the contribution of photo-neutrons is mixed with HXR in case of large fluxes that affect even the shielded detectors. The wide-angle view and detailed observation of the beam core in the visible range is provided by fast cameras Photron MINI UX-100 and Photron SA-X2, respectivelly. Spectral data in visible,

Ficker et al.

near IR and near UV regions are acquired using various mini-spectrometers. To some degree, SXR cameras and AXUV cameras may also contribute to the analysis of RE experiments, despite the fact that they can be affected by HXR radiation from the walls. The gas injection experiments use two valves: the ex-vessel piezo valve injecting smaller amount of gas from the HFS divertor (also for seeding experiments), can inject gas atoms at rates roughly 2.1020 s-1 when used with Ar and pressures around 1 bar, the ex-vessel MGI solenoid valves (3 installed at COMPASS) can injected at rates 1.1022 s-1 when used with Ar and at pressure 2 bars, while much higher pressures can be used.

2. RUNAWAY ELECTRON BEAM SCENARIOS

2.1. Ramp-up scenario

COMPASS is characterised by relatively low toroidal magnetic field which seems not to be beneficial for post-disruptive RE beam generation as was shown, e.g. in TEXTOR [16]. Despite this, RE beams in the ramp- up phase following a classical disruption triggered by massive gas injection (fast TQ, often current spike, CQ and beam plateau) were achieved irregularly, see an example discharge in FIG. 1 where Bt=1.15 T. The scenario includes carefully tuned fuelling waveform, injection at early times (low currents), optimised position reference and argon MGI with at pressure 1-3 bars and short valve opening time. Despite the relatively low reproducibility, a systematic analysis brought very interesting results [1, 7], more recently the crucial role of the toroidal magnetic field has been confirment in a dedicated scan [2]. No RE beams were created in the very same scenario at fields lower that 1.1 T despite the fact that the beam was reliably produced at higher fields. These beams are often characterised by extremely interesting first stage, where bright filaments in the camera correlate with the short spikes in the ECE and Mirnov coil data and later also in the HXR and Cherenkov detector data. The RE beams are typically radially unstable – the instability is more severe in the case of a smaller beam current, i.e. larger drop of current during the current quench. While cases with no or only small beam generated terminate at the HFS, well confined beams expand in the major radius and are lost to the LFS, or partially lost and then stabilised as in the case of the discharge below.

FIG. 1: The ramp-up scenario for RE beam generation: Plasma current ramp-up is interrupted by Ar MGI which leads to a current quench and RE generation. The density increases, and the early beam phase is accompanied by spikes in the ECE data and increased level of HXR. Later the signals of higher-energy Cherenkov detector channels and shielded HXR detector increase as well.

3 IAEA- EX/P1-26

Flat top scenario

Due to the low reproducibility of ramp-up scenario, which would be problematic especially in the RE beam decay experiments, an alternative, more quiescent scenario has been developed. In these discharges, the current flattop is reached and the fueling is turned off which leads to the decay of the thermal plasma density and a rise of the RE current in several tens of ms. Ar or Ne injection is then introduced using piezo valve puff or MGI (which leads to much faster decay, see [2]). With a short delay (5 ms) after the injection, the derivative of the current in the primary windings (which creates the external loop voltage) is set to zero, see FIG. 2 for the detailed overview. During this stage, additional puffs or RMPs [2,9] may be applied to investigate the influence on the natural decay of the beam. The puff causes the quench of the thermal plasma while the REs are practically unaffected, the thermal quench (TQ) is very slow in the case of piezo gas-puff lasting roughly 5 ms (see the profile evolution in FIG.2) while in the case of MGI, it seems to last much less than 1 ms. The amount and time of the gas injection plays a crucial role. While this scenario reliably works with the piezo valve (slow TQ) injection at almost any time during the discharge, including late ramp-up, the MGI typically causes immediate current quench when injected too early, while gradual beam decay is the result of the later injection, see FIG. 3. This indicates that the slower TQ allows to reach sufficient RE energy and/or current while fast MGI may only preserve the beam in the case that the RE population is already well evolved.

FIG. 2: The flat-top scenario with RE beam generation – left: time evolution of plasma current in blue, electron desity measured by TS , Ul in orange, current in the primary windings in red, Ar gas puff in green and HXR detector signals in brown (>50 keV) and pink (>500 keV), right: the profiles of temperature and density from the TS scattering

3. RUNAWAY ELECTRON BEAM STABILITY

3.1. Current control

The current in the RE beam can be controlled in the case that the amount of injected gas is low. However, the coordinated control of radial position and beam current proved to be difficult. When the external loop voltage is removed, the beam current usually steadily decreases with a rate related to the type and amount of injected gas(es) [2], while the application of a feedback on the plasma current at a large setpoint requires loop voltage up to 2-4 V and the radial position is driven unstable. However, the beam current was successfully sustained when FIG. 3:. The timing scan with MGI the current setpoint was decreased, see FIG. 4, blue case of Ar injection shows that the sufficient RE discharge #14592. The current is typically related to the seed or energy is needed for creation and number of runaway electrons as the change in velocity is small slow termination of a RE beam for relativistic electrons. Therefore, the task for a current

Ficker et al.

feedback in this scenario is to compensate the loss of particles by creating new RE, which might be rather difficult in the situation with increased density and rather cold thermal electron population. However, this has also another effect – due to a high loop voltage, the energy of confined runaway electrons is further increased.

This can be clearly seen in FIG. 4, where the vertical magnetic field Bv is indicated – despite decreasing beam current (see the next subsection), the necessary vertical field to sustain radial position is increasing up to very large values. Moreover, the effect of ongoing increase of kinetic energy of runaway electrons occurs also in the case of spontaneously decaying current – the induced loop voltage might not be sufficient for primary runaway electron generation however it is sufficient for further acceleration of REs.

FIG. 4: Different attempts to control the current of runaway electron beam, current measurement and reference

(dashed), total external vertical field - approximate value at R0 and loop voltage measured on the HFS; blue (#14592): Ar injection and control of current at 60 kA, orange (#14598): Ar and D injection with current feedback at the original setpoint (130 kA), green (#14599): Ar + D puff and zero loop voltage after injection

Radial stability of the relativistic electron beam and role of RE energy

The runaway orbits in equilibrium magnetic field and even the contribution of the runaway current to the total equilibrium were studied in many publications [17,18]. The radial stability of the RE beam seems to be incompatible with the standard feedback scheme applied on COMPASS. The radial position on COMPASS is actuated by two different systems – equilibrium field power supply – EFPS and fast vertical magnetic field power supply – BV. The controller of the first one contains a part proportional to plasma current (as a result of the Grad-Shafranov equilibrium) and proportional and integral terms depending on the radial position error of the current centroid with respect to the reference. The fast BV is dependent only on the position error and is also weaker. The system performs excellently in case of high temperature plasma without RE, however the performance is poor with a RE beam in the low temperature plasma background as can be seen in FIG 5. In the figure, various scenarios are compared in terms of plasma current evolution, normalised value of current in the dominant radial feedback windings (IEFPS/(7000+Ip) – the constant is added so the function is stable when approaching zero), radial position and loop voltage which drives the current in plasma or runaway electrons. It is obvious that while in the green case (standard shot) the value of the function in the second frame is constant, in case of various RE scenarios (RE beams triggered by MGI or piezo-valve Ar gas injection and plasma current feedback or zero loop voltage), the quantity is quickly increasing. The faster is the current decay of the RE dominated plasma or loop voltage, the higher is the request for the vertical field normalised by the plasma current value. Notice that despite large values of current in the stabilizing windings the beam still drifts to LFS. From this observation it is obvious that the radial feedback dependence on plasma current is too strong. The standard tokamak request for vertical field that results from Grad-Shafranov equation can be expressed as [19]:

2 ∂ 2 2 퐹퐹′ 2π푅0퐼푝퐵푣 = 1/2푅 0 퐼푝 (퐿푒 + 퐿푖) − 2π ∫ 푑푟 푟 (푝′ − 2 ) ∂푅0 푅0

5 IAEA- EX/P1-26

where the first term on the right side is related to external and internal part of "hoop force" (force between the current elements within the plasma ring), the second to the tire "tube force" expansion due to kinetic pressure gradient and the last term changes the direction with the value of 훽푝. The first three terms are always outward and typically depending on plasma current squared (the pressure term via the 훽푁 value). This means that vertical magnetic field should be also dependent on 퐼푝. If we consider the case of the runaway electron beam – the hoop force terms should be still valid but the current density and therefore the internal inductance of the beam might be very different compared to thermal plasma. However, the gradient of pressure is practically negligible for low temperature plasma background that has often even a large neutral fraction. Therefore, the dependence of vertical field on current in the beam ring should be weaker in the case of RE.

3.2.1. Additional component of vertical field required for radial stabilization

Comparison with the betatrons, namely the modified ones (including toroidal stabilizing field) [20] turns to be appropriate – the beams in these high current accelerators are very similar to RE beams in tokamak. The average energy of beam electrons plays a crucial role in the 퐵푣 value necesary to keep the beam particles on radially stable orbits. If Larmor radii in the vertical magnetic field is considered (which applies in the case of low beam current), the vertical field should be proportional to the flux swing or kinetic energy 퐸푘 of the accelerated electrons – this is how betatrons typically work. A necessary value of the vertical field to confine a high current RE beam in the tokamak would be based on the mix of this contribution and the hoop force compensation:

1 ∂ 퐸푘 퐵푣~ 퐼푝 (퐿푒 + 퐿푖) + in the ultrarelativistic case 4휋 ∂푅0 (푒푐푅0) However, in practice it would be easier to simply decrease the relative contribution of Ip-proportional part with respect to the contribution of radial position error term in the controller. Most of the tokamaks are close to the situation where 퐵푣 is optimal for given major radius and energies of RE in the range of tens of MeV. Therefore, the standard feedback may perform reasonably, but the equation above can offer an increase in efficiency. On the other hand, the disturbed feedback algorithm may be a source of information about the average energy of RE in the beam.

FIG. 5: Comparison of radial control performance in standard discharge (green) and various RE beam scenarios – Ar injection by MGI (16636) or piezo valve (16625) or Ar injection followed by D injection (14598, 14599). Second plot

shows that necessary 퐵푣 is not a function of Ip only; large loop voltage values are measured during the decay.

Vertical stability

To investigate the vertical stability of RE beam created using the flattop recipe with Ar, a scan in the amplitude of the elongating field was carried out. It seems that the generation of RE beam using destruction of thermal

Ficker et al.

component with piezo gas injection is not much affected by the elongation of the plasma. As expected, the beam is suspect to a vertical instability during the decay of its current and the instability occurs the earlier, the higher is the elongating field – this is reported in FIG. 6 below. The highest value of elongation requested in the scan leads to early loss of current and slightly higher loop voltage must be applied to sustain the value. The VDEs last several ms.. In this case the high loop voltage at the disruption (beam termination) does not lead to generation of further RE population as the beam is suddenly and completely lost.

FIG. 6: Elongation scan with piezo gas-puff - Ar injection – generation of beam is not much affected, the beam allows shaping, however when critically low value of current is reached, VDE occurs.

4. DECAY RATE AND ENERGY LOSS CHANNELS

The beam energy can be lost through several channels – direct particle loss, energy radiated through excited atomic transitions, induced currents in the vessel structures due to loss of the beam current, etc. Based on the gas amount scans with Ar and Ne, it seems that the two gases behave very differently in terms of the first two beam energy loss channels – while RE beam in Ne is radiating with larger power in the AXUV spectral region, Ar causes larger averaged signal of the least sensitive HXR/Photoneutron detector available (that still performs well in these discharges with extreme HXR fluxes) during the beam decay, see FIG 7. As the current decay rate is comparable for the two gases, these two channels of the loss of the RE beam energy and current should be indeed complementary to each other, however the large difference may present another argument to prefer Ne over Ar in ITER as more energy is radiated and perhaps less energetic RE are lost to the wall if Ne is used.

5. CONCLUSIONS

Experiments using two different scenarios with gas injection triggered RE beam generation are under investigation at COMPASS. The ramp-up scenario, which includes typical disruption features like current quench is relevant to larger machines. However, the reproducibility and control possibilities were not sufficient so far. Therefore, a flattop scenario with high current RE beam was developed using various amounts of Ar or Ne to isolate the beam from the thermal component. The current of a RE beam created in this way may be kept at the desired value in the Ar or Ne background plasma at the cost of a relatively high loop voltage, however it is even more interesting to switch off the external drive and observe the self-consistent decay of the beam. The radial position feedback is not performing very well, perhaps due to the absence of pressure gradient that would require vertical magnetic field proportional to beam current and due to the role of RE energy in the correct request on the vertical magnetic field that would secure stable orbits. This hypothesis will be tested in the next campaign. Regarding, the vertical position, the beam seems to be stable even at relatively high values of elongation, unless the current drops too much. During the beam decay, some of the total energy is lost in the interaction with the gas (excitation, ionization and subsequent radiation) or directly through RE loss to the wall

7 IAEA- EX/P1-26

– Ar and Ne seems to behave oppositely. While Ar causes more high energy HXR signal indicating larger RE losses, Ne causes stronger radiation in the AXUV spectral region. The work on this scenario will further continue with feedback optimization, puffing of gas mixtures and analysis of instabilities in order to provide the largest amount of information possible to medium sized and large machines, where safety constraints are much larger compared to COMPASS.

FIG. 7: Gas amount scans for Ar and Ne – left: the radiated power calculated using AXUV tomography increases with the amount of gas injected (and current decay rate) and it is higher for Ne; right: the average value of HXR and photoneutron intensity during the decay as measured by 3He detector is much higher for Ar, though it also scales well with the amount of injected gas (and current decay rate).

ACKNOWLEDGEMENTS

This work has been carried out within the framework of the EUROfusion Consortium and has received funding from the Euratom research and training programme 2014-2018 under grant agreement No 633053 with the Co-fund by MEYS project number 8D15001. The views and opinions expressed herein do not necessarily reflect those of the European Commission. This work has been also supported by the Czech Science Foundation research grant No GA18-02482S and the MEYS projects LM2015045. This work was supported by the Grant Agency of the Czech Technical University in Prague, grant No. SGS17/138/OHK4/2T/14.

REFERENCES

[1] Vlainic, M. et al. Journal of Plasma Physics 81, (2015). [2] J. Mlynar et al. Runaway Electron experiments at COMPASS in support of the EUROfusion ITER physics research. Accepted in Plasma Phys. Control. Fusion (2018) [3] Hender, T.C. et al. Nuclear Fusion 47 (6), S128 [4] Boozer, A. H. Nucl. Fusion 57, 56018 (2017). [5] Pánek, R. et al. Plasma Phys. Control. Fusion 58, 14015 (2016). [6] Panek, R. et al. 29th IAEA FEC, Gandhinagar, India OV/P-3 (2018). [7] Plyusnin, V. V. et al. Nucl. Fusion 58, 16014 (2018). [8] Ficker, O. et al. Nucl. Fusion 57, 76002 (2017). [9] Macusova, E. et al. 28th Symposium on Plasma Physics and Technology, Prague (2018). [10] Gobbin, M. et al. Plasma Phys. Control. Fusion 60, 14036 (2018). [11] Esposito, B. et al. Plasma Phys. Control. Fusion 59, 14044 (2017). [12] Carnevale, D. et al. Runaway Electron Beam Control, submitted to Plasma Phys. Control. Fusion (2018). [13] M Farnik et al. 45th EPS Conference on Plasma Physics, Prague, P1.1010 (2018) [14] Rabinski, M. et al. J. Inst. 12, C10014 (2017). [15] J Cerovsky et al. 45th EPS Conference on Plasma Physics, Prague, P1.1006 (2018) [16] Zeng L. et al 2013 Phys. Rev. Lett. 110 235003 [17] Knoepfel, H. & Spong, Nucl. Fusion 19, 785 (1979). [18] Yoshida, Z. Numerical analysis of runaway tokamak equilibrium. Nucl. Fusion 30, 317 (1990). [19] Freidberg, J.P. Ideal Magnetohydrodynamics, Plenum Press, New York (1987). [20] Manheimer, W. M. The Plasma Assisted Modified Betatron, Naval Research Lab, WASHINGTON DC (1984).