Second Technical Meeting on Plasma Disruptions and their Mitigation
ITER Headquarters
KEY DATES AND DEADLINES
5 June 2022 Deadline for submission of abstracts through IAEA-INDICO for regular contributions
5 June 2022 Deadline for submission of Participation Form (Form A), and Grant Application Form (Form C) (if applicable) through the official channels
10 June 2022 Notification of acceptance of abstracts and of assigned awards
The most important initiative on fusion R&D is currently ITER: the international tokamak reactor scale experiment being assembled in France. In tokamaks, instabilities can develop under certain operational conditions or as a consequence of a loss of plasma control. These instabilities eventually lead to the rapid loss of thermal and magnetic energy, a phenomenon known as plasma disruption.
Plasma disruptions cause thermal and mechanical loading to the tokamak components. Due to the high amounts of stored thermal and magnetic energies in ITER, the in-vessel components, such as the first wall panels and the divertor, will receive significant thermal loads. Furthermore, the in-vessel components, the vacuum vessel and the coils in the tokamak must also bear substantial mechanical loads. Disruption mitigation will be essential to reducing thermal and mechanical loading in order to guarantee the lifetime of these components.
Objectives
The event aims to serve as a forum to help coordinate experimental, theoretical and modelling work in the field of plasma disruptions with special emphasis on developing a solid basis for possible mitigation strategies in ITER and next generation fusion devices.
Target Audience
The event aims to bring together junior and senior scientific fusion project leaders, plasma physicists, including theoreticians and experimentalists, and experts (researchers and engineers) in the field of plasma disruptions.
-
-
Registration
-
Opening Council Room
Council Room
ITER Headquarters
-
1
Welcome (IO)Speaker: Tim Luce (ITER Organization)
-
2
Welcome and Organization (IAEA)Speaker: Matteo Barbarino (International Atomic Energy Agency)
-
3
ITER ProjectSpeaker: Alberto Loarte (ITER Organization)
- 4
-
1
-
10:10
Coffee Council Room Lobby
Council Room Lobby
ITER Headquarters
-
Disruption Consequences Council Room
Council Room
ITER Headquarters
-
5
Present status and challenges of disruption load modelling
One of the high priority research needs for the ITER project is the development of a solid physics basis of plasma disruptions and their mitigation. The thermal and electromagnetic loads taking place during these events pose important constraints on the lifetime of tokamak components [1, 2]. The extrapolation of these loads from experimental data to new machines entails large uncertainties, thus, detailed load modelling and validation is essential for the success of future tokamaks such as ITER and DEMO.
In this talk, we present a summary of previous efforts on disruption load modelling and validation as well as the present status for the different load types. Thermal loads can result from plasma convective and conductive losses, radiated plasma energy or deposition of runaway electron kinetic energy into the plasma facing components. Electromagnetic loads arise on the conducting structures due to eddy and halo currents induced by MHD instabilities. A review of the main codes and frameworks used for the simulation for each of these loads is presented. Special attention is paid to the main modelling assumptions, present capabilities and main physical and numerical limitations. Finally, we review and discuss the main needs, challenges and planned efforts in order to simulate the relevant disruptive loads for future devices.
[1] Hender, T. C., et al. "MHD stability, operational limits and disruptions." Nuclear fusion 47.6 (2007): S128
[2] Lehnen, Michael, et al. "Disruptions in ITER and strategies for their control and mitigation." Journal of Nuclear Materials 463 (2015): 39-48.Speaker: Dr Francisco Javier Artola Such (ITER Organization) -
6
CarMa0NL modelling of halo currents on COMPASS
During vertical displacement events (VDEs) plasma column hits the wall and scrape-off layer (``halo'') currents can reach significant amplitudes [1]. Therefore, the related electromagnetic loads on plasma facing components (PFCs) should be thoroughly evaluated to guarantee their structural integrity. Modelling of halo currents for next generation tokamaks is a challenging task. For example, non-linear (3D plasma + 2D wall) MHD codes, like JOREK [2, 3], M3D [4], M3D-C1 [2, 5] and NIMROD [2, 6] are appropriate for study of the disruption physics, but might be too demanding in case one needs to analyse many scenarios and geometry configurations. Disruption-oriented (2D plasma + 3D wall) numerical tool CarMa0NL [7, 8] is more practical for design purposes, but it requires some physical insight for definition of the halo width $w_h(t)$. Such criteria is found here empirically by comparing magnetic measurements during COMPASS VDEs [3] with results of CarMa0NL modelling for a wide range of parameters.
It is shown that the halo width correlates with the value of safety factor on the last closed flux surface $q_\lambda(t)$. The best fit with experiment is obtained for $q_\lambda(t) \approx 1$, which suggests that m/n=1/1 kink instability might play a role in defining $w_h(t)$. Further, we discuss application of the findings to design of COMPASS-U tokamak [9].
References
[1] Strait E, Lao L, Luxon J and Reis E 1991 Nucl. Fusion 31 527
[2] Artola F J, Sovinec C R, Jardin S C, Hoelzl M, Krebs I and Clauser C 2021 Phys. Plasmas 28 052511
[3] Artola F J et al. 2021 Plasma Phys. Control. Fusion 63 064004
[4] Strauss H R, Paccagnella R and Breslau J 2010 Phys. Plasmas 17 082505
[5] Clauser C F, Jardin S C and Ferraro N M 2019 Nucl. Fusion 59 126037
[6] Bunkers K J and Sovinec C R 2020 Phys. Plasmas 27 112505
[7] Villone F, Barbato L, Mastrostefano S and Ventre S 2013 Plasma Phys. Controlled Fusion 55 095008
[8] Chen S L et al. 2019 Nucl. Fusion 59 106039
[9] Yanovskiy V V, Isernia N, Pustovitov V D, Scalera V, Villone F, Hromadka J, Imrisek M, Havlicek J, Hron M and Panek R 2021 Nucl. Fusion 61 096016Speaker: Vadim Yanovskiy (Institute of Plasma Physics of the Czech Academy of Sciences) -
7
Parallel transport induced tokamak thermal quench and its mitigation
Thermal quench (TQ) marks the point of no return in a tokamak
disruption. It not only brings a thermal load management issue at the
divertor plates and first wall, but also determines the runaway
seeding for the subsequent current quench (CQ). There are two ways to
trigger a TQ, one is the globally stochastic magnetic field lines that
connect the hot core plasma to the cold boundary, while the other is
high-Z impurity injection. In both situations, a nearly collisionless
plasma is made to intercept a radiative cooling mass (RCM), being that
an ablated pellet or a vapor-shielded wall. Previous JET experiments
and more recent DIII-D data have demonstrated a wide range of TQ
durations with and without high-Z pellet injection, which is
concerning for future tokamak reactor operations.With fully kinetic VPIC simulations and analytical theory, we have
uncovered three underlying parallel transport mechanisms that govern
the thermal collapse of a fusion-grade and hence nearly collisionless
plasma. They are: (1) thermal collapse of surrounding plasmas due to a
localized RCM is dominated by convective energy transport as opposed
to conductive energy transport, and as the result, TQ comes in the
form of four propagating fronts with distinct characteristic speeds,
all originated from the RCM, and core thermal collapse is a lot slower
than one would expect based on electron thermal conduction of
Braginskii or free-streaming; (2) cooling of perpendicular electron
temperature closely follows that of parallel electron temperature, and
in a nearly collisionless plasma, is mostly driven by fast
electromagnetic kinetic instabilities of two kinds, sequentially in
time; (3) the overall TQ inevitably has a transition from the
collisionless phase to the collisional phase, the duration of which
have distinct physics scalings, and the two of which are sandwiched by
a transition period of its own unique physics scaling. Altogether, we
can now predict the TQ history of a plasma at given density and
temperature as a function of the magnetic connection length.These physics advances inform the strategies for avoiding and
mitigating the deleterious effects of TQ on both thermal load
management and the subsequent Ohmic-to-runaway current conversion. In
the ITER scenario of high-Z pellet injection for spreading plasma heat
load via radiation, pellet assimilation and spatial homogenization are
both tied to the TQ physics. The staged pellet injection suffers
particularly strong constraints that result in severe performance
degradation. In situations where runaway avoidance is a priority, one
can no longer rely on impurity radiation for thermal load management.
The drastically different TQ durations in the collisionless and
collisional regimes point to the alternative mitigation approach that
relies on dilutional cooling via massive hydrogen injection to place
the entire TQ in the collisional phase, in which case the CQ and TQ
span the same period. If there is not enough lead time for
predisruption pellet injection, the physics insights place stringent
constraints on what an optimal passive mitigation method would entail,
especially when runaway avoidance is also a consideration.Work supported by U.S. DOE OFES and OASCR.
Speaker: Xianzhu Tang (Los Alamos National Laboratory)
-
5
-
12:00
Lunch
-
Posters Entrance Lobby
Entrance Lobby
ITER Headquarters
All posters will be displayed on both days.
Please choose among the following formats for your poster presentation:
a) In person: please upload your e-poster in Indico and display your printed poster at the ITER site during the two posters sessions (19-20 July).
b) Remote: please upload your e-poster in Indico with audio recording as well as a summary slide (which may be presented during the appropriate discussion session).-
8
An interpretable, transferable and real-time disruption predictor in HL-2A based on deep learning
A disruption predictor based on deep learning is developed in HL-2A. It has an accuracy of 96.1% on Shot Nos. 32000-36000. Novel 1.5-D CNN + LSTM structure is used to get such a high accuracy. [1] In recent years, further investigations and updates are carried out on the basis of the original algorithm, which bring it interpretability, transferability and real-time capacity.
For the interpretability, HL-2A’s algorithm give saliency maps indicating the correlation between the algorithm’s input and output. The distribution of correlations shows good coherence with the disruption causes. A disruption recognizer can be realized by using Bayes theorem to inference disruption reasons by correlations distributions. [2]
For the transferability, a preliminary disruption predictor is successfully developed in HL-2M, a newly built tokamak in China. Although only 44 shots are used as the training set of this algorithm, it still gives reasonable outputs with the help of data from HL-2A and J-TEXT.
For the real-time capacity, the algorithm is accelerated to deal with an input slice within 0.3ms with the help of some adjustments on it and TFLite framework. It is implemented into the plasma control system and get an accuracy of 89.0% during online test. Several demo shots are also realized where the algorithm predicted the disruptions and triggered the SMBI to mitigate them. [3]
These three characteristics along with the high accuracy make the deep learning-based disruption predictor in HL-2A a new promising method for the disruption prediction in ITER.References
[1] Zongyu Yang et al, Nuclear Fusion 60, 016017
[2] Zongyu Yang et al, Nuclear Fusion 61, 126042
[3] Zongyu Yang et al, 4th IAEA FDPVA, short talkSpeaker: Zongyu Yang -
9
Analysis of fast camera images for dual shattered pellet injection at KSTAR
In KSTAR disruption mitigation experiments have been progressed since 2019 and four fast visible cameras are currently installed for the dual shattered pellet injection (SPI) system [1] to view pellets before and after shattering. Two fast visible cameras (FVCs) were installed viewing the plasma in toroidal direction at both SPI locations [2]. An operation frame rate of 10 k-fps is capable to measures SPI with velocities of 400 ~ m/s during the penetration into the plasma. For the characterization of the fragment parameters, discharges were selected for which the penetration of the fragments could be clearly identified. During the 2020 campaign, the pellet velocity before shattering was measured using a microwave cavity (MWC) in the flight path and FVC images. Since 2021, an optical pellet diagnostic (OPD) is operational at each SPI system viewing the pellet in the flight path. During the penetration into the plasma, the velocity of the group of shards, the velocity dispersion, and the plume width along the vacuum trajectory were estimated. To observe the shard movement more clearly, each image frame was subtracted with the previous frame. Furthermore, through mapping of the FVC images, it was found that the particle assimilation in the plasma is impacted by the field line pitch angle. Especially the effect of the q=2 surface can be clearly seen. The mapped images are compared to magnetic field lines from EFIT calculation for further analysis of the 3D radiation pattern.
1. S. Park, K. Lee, L. R. Baylor, S. J. Meitner, H. Lee, J. Song, T. E. Gebhart, S. Yun, J. Kim, K. Kim, K. Park, and S. Yoon, "Deployment of multiple shattered pellet injection systems in KSTAR," Fusion Eng. Des. 154, 111535 (2020).
2. J. W. Yoo, J. Kim, M. K. Kim, S. H. Park, B. H. Park, Y. U. Nam, J. W. Kim, H. Wi, M. Lehnen, and W. C. Kim, "Fast visible camera diagnostic for dual shattered pellet injections at KSTAR," Fusion Eng. Des. 174, 112984 (2022).Speaker: Jaungwon Yoo -
10
Analysis of Variability in Pre-Disruption Plasma Parameters and their Effect on Runaway Electron Generation using the JET data-base on RE
Analysis of Variability in Pre-Disruption Plasma Parameters and their Effect on Runaway Electron Generation using the JET data-base on RE
V.V. Plyusnin(1) , C. Reux(2), V.G. Kiptily(3), S. Gerasimov(3), S. Jachmich(4),
M. Lehnen(4), O. Ficker(5), E. Joffrin(2) and JET contributors*EUROfusion Consortium, JET, Culham Science Centre, Abingdon, OX14 3DB, UK
(1) Instituto de Plasmas e Fusão Nuclear, Instituto Superior Tecnico, Universidade de Lisboa, Lisboa, 1049-001 Portugal; (2) CEA, IRFM, F-13108 Saint-Paul-les-Durance, France; (3) CCFE, Culham Science Centre, Abingdon, OX14 3DB, UK; (4) ITER Organization, Route de Vinon-sur-Verdon, CS 90 046 - 13067 St Paul Lez Durance Cedex – France; (5) Institute of Plasma Physics of the CAS, Prague, Czech Republic;
The generation of runaway electrons (RE) during major disruptions in International Thermonuclear Experimental Reactor (ITER) is unacceptable. Disruption Mitigation System (DMS) designed in ITER is based on massive injection of impurities, gaseous (MGI) or solid state pellets (SPI). Such injections should provide an intense radiation of the plasma stored energy in order to mitigate the damaging effect of the heat and mechanical loads and to provide reliable suppression of RE. Despite a significant progress in studies relevant to the ITER DMS design, the set of physical and technology problems remains unsolved. In particular, they concern to understanding of the mechanisms for mixing and assimilation of injected impurities during the rapid shutdown and to the physics of RE, their formation, interaction with surrounding plasma and injected gases (fuel and impurities, frozen and gaseous) and dissipation. Comprehensive analysis of the existing experimental database on RE in JET and other tokamaks, as well as planned new experiments should stimulate further advances in understanding of the physics of RE generated in major disruptions.
This report presents the results of the mapping of RE parameters depending on pre-disruption and post-disruption JET plasma parameters (electron temperature and density, internal plasma inductance, current quench (CQ) rates, etc.). Despite the plasma parameters are poorly known during and after disruptions, this approach enables establishing links between plasma parameters before thermal quench and during CQ, allowing the calculation of accelerating electric fields and RE parameters. Using known models for RE generation: primary mechanism (“Dreicer-Gurevich-Connor/Hastie…”) and Putvinski/Rosenbluth the parameters of RE were calculated and compared to those measured in experiments for a wide range of disrupted JET currents (up to 6.25 MA). Note, that in certain cases the simulations yielded the data, which was in contrary to experimentally observed trends. Study of current quench (CQ) stages revealed different, accelerating and constraining effects of initial plasma configurations (circular (limiter) or X-point) on CQ rates, RE generation and value of current conversion ratio (Ipl/IRE). Analysis of MGI effect from different Disruption Mitigation Valves revealed different effects on disruption dynamics and RE generation.Speaker: Dr Vladislav Plyusnin (Instituto de Plasmas e Fusão Nuclear, Associação EURATOM-IST, Instituto Superior Tecnico) -
11
Assimilation of a Composite Hydrogen/High-Z Plasmoid
Pellet injection is used in tokamaks and stellarators for fuelling, ELM pacing, diagnostics, and disruption mitigation. Injection of shattered pellets is a critical part of the envisaged ITER disruption mitigation system.
A highly localized plasmoid initially expands predominantly along the magnetic field lines. These assimilation dynamics play a critical role in determining the post-pellet plasma energy balance: the energy transferred to the expanding plasmoid is split between the plasmoid electrons, ions, and the radiative losses in the presence of high-Z impurities. If the plasmoid is heated at a constant rate, the ions accelerated by the ambipolar electric field acquire half the total energy transferred to the plasmoid in the absence of radiation losses.
In the present work, we study the plasmoid expansion dynamics for hydrogen-neon mixed pellets. The initially very dense plasmoid is shown to be opaque to line radiation for cases of high impurity content. For low impurity content, the friction force acting on the cold impurity ions transports the impurity together with the hydrogen, which simplifies the analysis. We calculate the radiated energy fraction for ITER-relevant plasmoid parameters.
A distinctive feature of our model is a coupled hydrodynamic description of the cold plasmoid and a kinetic treatment of the ambient electrons and ions. In particular, we describe the reduction of the plasmoid heating rate due to the effect of the ambipolar potential on the hot electrons.
Speaker: Dr Pavel Aleynikov (Max-Planck-Institut für Plasmaphysik) -
12
Bifurcation-driven vertical plasma displacement
The vertical motion of the plasma current during the current quench phase of the uncontrolled major disruption is described analytically. The presented filament-based model interprets the vertical displacement event in the ideal wall limit as an adiabatically slow evolution of the plasma equilibrium. The initial pre-disruption equilibrium becomes unstable in a pitchfork bifurcation. The bifurcation occurs when the decaying plasma current passes a critical value determined by the external magnetic field. Minimizing the threshold plasma current is desirable to reduce the disruption-induced mechanical and heat loads on the first wall.
Speaker: Dmitrii Kiramov (The University of Texas at Austin) -
13
DECAF Code Cross-device Investigation of Disruption Categorization and Timing Indicated by Variations in the Plasma Current and Vertical Position
Timely detection and prevention of plasma disruptions is essential for next-step tokamaks. Plasma disruption is a multi-step process in which the loss of plasma vertical position control and the quenching of the plasma current are typically among the last events that precede complete plasma deconfinement. Given the major role of abnormal behavior of the plasma current and vertical position in the occurrence of disruptions, real-time control of these two plasma parameters has become routine in tokamak operations. Furthermore, deviation from a normal waveform of either parameter has served in various forms as an indicator of the disruption onset n a number of disruption-related studies [1,2]. The definition of the disruption time sets both the time scales and success rates of algorithms that aim to predict disruptions. The present non-uniformity of disruption time definitions across the tokamak community brings ambiguity in cross-comparisons of various predictor performances. Here, we present a systematic study of abnormal plasma current and vertical position waveforms and evaluate their capacity to serve as disruption onset indicators. The study is conducted on a multi-year, multi-device database with a focus on NSTX, KSTAR and MAST/-U tokamak data. The analysis is conducted with the Disruption Event Characterization and Forecasting (DECAF) code [3]. The frequency of occurrence of different types of abnormal waveforms and the disruption categories that they define will be presented in the context of the operational spaces of each device. Interconnection of the abnormal waveforms will be discussed as well. Results obtained hereby might serve to define a reliable indicator of the disruption onset time. *This research was supported by the U.S. Department of Energy under contracts DE-SC0018623 and DE-SC0016614.
[1] V. Klevarova et al., Fusion Engineering and Design 160 (2020) 111945
[2] A. Pau et al., Fusion Engineering and Design 125 (2017) 139-153
[3] S.A. Sabbagh et al., Proc. 27th IAEA Fusion Energy Conference, Ahmedabad, India (2018)Speaker: Veronika Zamkovska (Columbia University) -
14
Disruption Prediction with different wall conditions based on multi-scale deep hybrid neural network on EAST
With the upgrading of the tokamak device, the performance of the plasma disruption prediction model will be affected. In 2019, EAST upgraded the lower divertor of carbon material to tungsten-copper divertor, and EAST upgraded from non-all-metal wall to all-metal wall. In this study, a variety of diagnostic signals (such as radiation array and magnetic probe array, etc.) are selected for the main disruption types of EAST, and a multi-scale deep hybrid neural network model is built according to the physical information of different diagnostic signals. A large amount of non-metal wall experimental data is used to train the multi-scale deep hybrid neural network model. The performance of the model is better under the test of non-metal wall experimental data, and the recall rate is >90%. However, the experimental warning performance of all-metal wall is decreased, and the recall rate is only ~80%. By adding a self-attentional structure into the multi-scale deep hybrid neural network, the network pays more attention to the features closely related to the disruption and reduces the interference of some factors on the model. The test results show that the new network structure model has good warning performance for both non-metal wall experiment and all-metal wall experiment，the recall rate is all >90%. In the future, the network model will provide real-time warning on EAST and verify the model's cross-device warning capability in conjunction with other tokamak devices.
Speaker: bihao guo -
15
Hybrid injection system of gas and metal pellets for disruption mitigation on HL-2A
At present, the massive gas injection (MGI) and shattered pellet injection (SPI) techniques are regarded as the primary injection methods for disruption and RE mitigation. Both of them have their own character and can be used in different applications.
In order to combine the advantages of gas injection and pellet injection for avoidance and mitigation of disruptions, a new hybrid injection system, which including an MGI system and Li pellet injection system, had been developed successfully on HL-2A tokamak. It can realize the global simultaneous cooling of the plasma core and boundary. This impurity injection method is more conducive to enhance the mixing effect of impurities. The hybrid injection system is installed at the midplane port on HL-2A. The MGI injector is a kind of pulse valve, which is activated by eddy current. The piston does not make use of any ferromagnetic materials, so the valve can be connected close to the device through flange. The valve with a short response time (0.25 millisecond), and adjustable throughput (1021~1023) allows to meet the requirement of disruption mitigation. The Li pellet injection system has automatic supplying system and turntable adjustment system to adjust the number of Li pellet. The pipeline is connected with the gas outlet of MGI. The pellets freely fall into the pipe, and then are blew into the device by using the MGI injected gas. The injection speed of pellet can be adjusted by MGI injection volume. The Li pellet can be injected with a speed of 200-400 m/s.
Several different injection scenarios were performed using different gases and Li pellet in various amounts on HL-2A.Speaker: Dr Yunbo DONG (Southwestern Institute of Physics) -
16
Kinetic-MHD simulation of compressional Alfvén eigenmodes excited by runaway electrons in current quench
Kinetic instabilities in the MHz range have been observed during current quench in DIII-D disruption experiments (A. Lvovskiy et al., PPCF 60, 124003 (2018)). In this talk we show the new updates on kinetic-MHD simulation of current quench modes using M3D-C1. It is found that this mode is mainly compressional Alfvén eigenmode (CAE) and has large parallel perturbed magnetic field component. The wave can have resonance with high energy trapped runaway electrons, which have precession frequency close to the mode frequency. Nonlinear simulation shows that multiple eigenmodes can be excited with the same toroidal mode number but different freqencies. The characteristics of the simulated modes, including frequecies and polarization, are consistent with experimental observations.
Speaker: Chang Liu (Princeton Plasma Physics Laboratory) -
17
MHD activity during disruptions and disruption mitigation: Recent insights from JOREK
Mitigation as last line of defense against disruptions needs to simultaneously mitigate electromagnetic forces, heat loads, and runaway electrons (REs) for a safe operation of ITER like tokamaks. Simulations help to prepare robust mitigation scenarios and need to capture various non-linearly interacting physics processes simultaneously in a self-consistent way.
JOREK is an extended non-linear MHD code designed for this purpose in realistic tokamak geometry. It features models for pellet ablation, fluid and kinetic impurities, fluid and kinetic neutrals, fluid and kinetic REs as well as resistive walls and coils. This allows for a comprehensive treatment of disruption physics. A focus lies on validation against experiments; in particular, we aim to quantitatively reproduce bifurcating dynamics, e.g., RE beam formation changing discontinuously as function of plasma parameters. ITER predictions are being continuously refined as more advanced models become available. In this contribution, we review recent developments and results and provide a brief outlook.
Halo currents and wall forces of mitigated and unmitigated disruptions are compared against dedicated experiments; first results show good agreement regarding key parameters. 3D ITER predictions focus on a complete modelling of the current quench including vertical plasma motion, wall forces and the interplay with MHD activity. Direct coupling to the CARIDDI code is progressing for self-consistent modelling of 3D plasma dynamics with accurate 3D conducting structures.
We compare the assimilation of material during deuterium/mixed/impurity shattered pellet injection (SPI) and the resulting radiation patterns to experiments and perform predictions for ITER. For deuterium injection, we introduce an ad-hoc model to capture the drift of small-scale plasmoids that cannot be resolved in simulations otherwise. The effect of pre-existing islands onto the MHD activity triggered by SPI is studied. Radiative collapse and current spike formation during the thermal quench are explored. A fully kinetic impurity model is being established that does not assume a Maxwellian energy distribution.
Using a self-consistently coupled RE fluid model that has been benchmarked against other codes, RE mitigation by multiple material injections is studied predictively for ITER axisymmetrically and the inclusion of 3D MHD activity is progressing. Test particle simulations in stochastic MHD fields provide insights into loss mechanisms for passing and trapped REs. A self-consistent coupling of the kinetic REs to the MHD fluid is planned for the future.Speaker: Dr Matthias Hoelzl (Max Planck Institute for Plasma Physics) -
18
MHD modeling and optimization of a passive helical coil for mitigation of runaway electrons
A helical coil designed to passively generate non-axisymmetric fields during a plasma disruption has been shown, via electromagnetic analysis, linear MHD modeling, and relativistic drift orbit tracing, to be effective at deconfining runaway electrons (REs) on a time scale significantly faster than the plasma current quench. Magnetic equilibria from DIII-D RE experiments are used to calculate the toroidal electric field generated during the current quench phase of a disruption, which in turn drives current in a proposed n=1, m=1 in-vessel helical coil, without the need for any external power supplies or disruption detection diagnostics. Simulations of the plasma evolution using the TokSys GS Evolve code predict the inductive coupling of coil currents up to 12% of the pre-disruption plasma current into the helical coil. The coil geometry is systematically varied to maximize both the non-resonant and resonant components of the 3D magnetic perturbation, resulting in an optimized coil location on the vessel center-post and generating fields as strong as $\delta$B/B~$10^{-2}$ at the plasma edge and a vacuum island overlap width of up to 0.7$\psi_N$. The REORBIT module of the MARS-F code is used to model the full non-axisymmetric magnetic field for toroidal modes up to n=6 and trace RE orbits to determine the effect on RE deconfinement, with up to 70% of the RE orbits lost after 0.2 ms. A two-stage evolution of the RE orbit loss fraction is observed to be caused by resonant trapping between multiple magnetic island chains. Scaling estimates are shown to be favorable for the operation of a passive RE deconfinement coil on a larger ITER-scale device, with only marginal decreases in relative coil current and magnetic perturbation amplitudes. Electromagnetic and thermal stresses on the coil are calculated to be within acceptable limits for both a DIII-D coil as well as a larger ITER-scale coil: worst-case J$\times$B stresses are less than 40% of the stainless steel yield strength, and the total temperature rise caused by joule heating throughout the current quench is less than 50$^{\circ}$C. Electromagnetic modeling of the 3D eddy currents in the vacuum vessel wall suggests that while a conductive wall may slow the coil current rise time by up to a factor of two, the maximum current is not affected. These findings motivate future experimental study of the helical coil concept in DIII-D or other tokamaks.
Work supported by General Atomics Internal funds and US DoE Office of Science under Contract DE-SC0016452 and DE-FC02-04ER54698.
Speaker: David Weisberg (General Atomics) -
19
Numerical simulation of thermal quench triggered by density source in HL-2A
Magnetic flux surfaces are abruptly destroyed and the plasma is no longer confined during thermal quench [1]. For a representative unmitigated disruption occurring in ITER with full deuterium-tritium (DT) performance, about 350 MJ of thermal energy and up to 1 GJ of magnetic energy may be released to the divertor and first-wall (FW) surfaces during several milli-seconds, leading to serious damage of the device [2]. Since occasional disruptions might probably be unavoidable in future fusion reactor, the realization of disruption mitigation is of crucial importance.
The mainstream methods of disruption mitigation system (DMS) include Shattered Pellet Injection (SPI) [3] or Massive Gas Injection (MGI) [4]. The location and uniformity of injected material deposition can affect the mitigation efficiency. To clarify the effect of MHD-modes-induced-transport on the injection penetration, we simulate the thermal quench (TQ) during the pre-disruption phase triggered by pure deuterium (D2) injection at different fixed deposition location employing three-dimensional (3D) non-linear reduced MHD code JOREK. Attention is focused on the characteristics of MHD activity and particle transport responding to the externally injected density source. Results exhibit evidently different n=1 mode dynamics and variation of plasma density profiles, depending on the location of D2 deposition (LoD) relative to the magnetic surface of q=2.
When LoD is outside the q=2 surface, the m/n=2/1 mode tends to be dominant and mainly couple with the m/n=3/1 mode. Magnetic stochasticity firstly happen in the m/n=2/1 island region and then expanding outwards. But when LoD is inside the q=2 surface, the m/n=1/1 mode tends to be dominant, and magnetic stochasticity firstly occur in the m/n=1/1 island region then the core plasma density greatly increases. The LoD is also found to be strongly linked to the growth rate and maximum dominant mode amplitude during the TQ, and consequently affects the TQ duration and the current spike amplitude.[1] T.C. Hender, J.C Wesley, J. Bialek et al., ITER Physics Basis Chapter 3: MHD stability, operational limits and disruptions , Nucl. Fusion 47, S128–S202 (2007).
[2] E. M. Hollmann, P. B. Aleynikov et al., “Status of research toward the ITER disruption mitigation system”, Phys. Plasmas 22, 021802 (2015).
[3]D. Hu, E. Nardon et al., “JOREK simulations of Shattered Pellet Injection with high Z impurities”, 45th EPS Conference on Plasma Physics, P4.1043.
[4] E Nardon, A Fil et al., “Progress in understanding disruptions triggered by massive gas injection via 3D non-linear MHD modelling with JOREK”, Plasma Phys. Control. Fusion 59, 014006(2017).Speaker: Dr shilin hu (Southwestern Institute of Physics) -
20
Pellet fueling technology development in India
Plasma fueling and disruption mitigation studies are two key aspects of the future fusion reactors. To meet these challenges, development of pellet injectors and associated technologies has been initiated in India to cater the domestic fusion program. As a first step towards it, a single pellet injector system (SPINS-IND) is successfully developed. The injector, depending on the barrel and sleeve size, freezes cylindrical hydrogen pellet (size ranging from 1.6 mm - 4 mm) and propels it using helium gas to a velocity in the range of 700 - 1000 m/s (depends on release pressure of propellant gas and pellet size). The system uses a closed loop GM cycle cryocooler, a cryogen free unit, to freeze pellet. The cryocooler provides ease of handling and operational reliability to the pellet freezing process. SPINS-IND is installed on SST-1 tokomak for pellet injection related experiments. System dimension is ~2 m (length) x ~0.5 m (width) x ~2m (height).
Successful development of SPINS-IND gave confidence to initiate development of shattered pellet injector(SPI-IND) to aim at disruption mitigation studies (DMS). The injector cryostat is designed to freeze cylindrical pellets of size up to 10 mm, having equal length and diameter. The integration, commissioning of the system is completed. Preliminary testing of the cryo system and the vacuum system is successfully completed. Formation of pellet is tested.
Diagnostic systems such as light gate and fast imaging camera are used to measure the pellet speed and size, respectively. Additionally, a microwave cavity diagnostic system having resonant frequency of 3.2 GHz is under development for the pellet mass measurement.
In addition to the in-situ pipe gun type injector, development of a twin-screw hydrogen extruder system is ongoing for continuous fueling. It is designed for extrusion throughput in the range of 400 mm^3/s to achieve pellet size of 3 mm (L) x 3 mm (D) with a frequency of 10Hz. In the present design, a counter rotating inter-meshing twin-screw geometry with square thread is selected. Each screw has 28 mm root diameter, 4 mm screw cavity depth and 10 mm pitch with a screw helix angle of 6.5 degree. The main extrusion circuit will have four stage cooling mechanism namely, 80K hydrogen pre-cooler, 40K pre-cooler, liquefier and solidifier. Two numbers of two-stage Gifford-McMahon cycle crycooler will be used for 40K precooling, liquefaction and extrusion application.
To know the hydrogen extrusion process, a non-Newtonian, non-isothermal ANSYS-CFD model was developed. The model is validated with the experimental data of already published literature for single and twin screw extrusion. At present, the assembly and integration process is ongoing.Thus, India is progressing in the field of development of pellet injector technology with the successful demonstration of SPINS-IND, upcoming shattered pellet injector (SPI-IND) and twin-screw extruder system.
Speaker: Dr Ranjana Gangradey (Institute for Plasma Research) -
21
Review of a data-driven adaptive disruption predictor for mitigation based on a nearest centroid approach
Any disruption mitigation system requires a trigger to trigger the corresponding remedial actions. Such trigger is the final step of a chain of events. This chain starts with an alarm that recognises an incoming disruption followed by interlocks protecting particular systems (for example, plasma heating systems). This contribution is a review of a specific disruption predictor that is installed in JET. The predictor uses only one signal, the mode lock normalised to the plasma current (NML), and its feature space, in which the separation frontier between disruptive and non-disruptive behaviour is linear, is two-dimensional. The linear frontier is defined based on two centroids, where each one summarises the disruptive and non-disruptive behaviours of past discharges, respectively. From a conceptual point of view, the predictor recognises a disruptive behaviour when large differences between consecutive samples of the NML appear. The predictor is installed in the JET real-time network from June 2019 (in open loop). The real-time predictions analysed so far confirm the following positive characteristics: fully deterministic (the running time of the algorithm for each prediction is less than 10 s), not based on a simple threshold but on differences of amplitudes, easy physics interpretation (not a black-box), success rates above 96%, false alarm rates about 4%, most of the alarms very close to the disruption (26% of alarms within 10 ms) and average warning times of about 100 ms (can be smaller if assertion times are set-up). Off-line analyses with several databases (JET with C-wall, JET with ILW and JT-60U) have shown full compatibility with an adaptive development from scratch with about 10 re-trainings when tested in more than 1200 discharges. Re-trainings are performed after missed alarms. These properties make the predictor a potential candidate to be used as disruption predictor in ITER for mitigation purposes.
Speaker: Prof. Jesús Vega (CIEMAT) -
22
Runaway electron formation and termination in mitigated ITER disruptions
Uncontrolled termination of post-disruption relativistic runaway electron (RE) current can cause deep localized melting of the plasma facing components and poses a serious challenge to the successful operation of fusion grade tokamaks, including ITER. While RE deconfinement depends on the timescale of flux-surface reformation, the magnetohydrodynamic (MHD) plasma stability itself is affected by the runaway current. Therefore MHD-RE interaction is highly non-linear and determines the eventual impact-profile of REs on the components. This is the motivation of the present work, aimed at studying RE-MHD co-evolution in disruptions using the 3D MHD code JOREK [1,2].
We present numerical predictions for the formation and termination of RE current in vertically unstable plasmas in mitigated ITER disruption scenarios. The REs are modeled as a separate fluid species [3], subjected to field-parallel and drift transport along with an avalanche source including the effect of partially-ionized impurities [4]. The back-reaction of REs on MHD is treated via a current-coupling, while the impurities are assumed to be in coronal-equilibrium. Electrically conducting structures such as the vacuum-vessel and various coils are included via coupling to the STARWALL code. Starting with a 15MA elongated free-boundary X-point plasma equilibrium, a pseudo thermal-quench followed by first injection of Ne+D causes a current-quench and subsequent vertical motion along with a multi-MA RE beam formation.
Through axisymmetric simulations, we investigate the effect of first injection quantities, current-profile flattening, Neon 2nd injection and Neon-flushout via Deuterium 2nd injection. Within the constraint for the current-quench time for ITER (50ms < $t_{CQ}$ < 150ms) and assuming an RE seed current of 0.1A, our simulations predict a multi-MA RE beam for any combination of Ne/D injection quantities, with the RE beam current increasing (upto $I_\mathrm{RE}$ ∼ 9.5MA) with Neon quantity. This is in qualitative agreement with 1D predictions of the GO code [5]. The second injection of Neon causes both a faster decay of RE current, and a correspondingly faster vertical plasma motion towards the wall. However, 2nd injection of Neon is found to be ineffective in reducing the undissipated RE energy deposited onto the wall. Commensurate increase in the poloidal magnetic energy channeled to REs (due to bound-electrons acting as additional avalanche targets), is seen to offset the additional dissipation of RE energy.
In general, vertical plasma motion and scrape-off causes a net decrease in the edge safety factor ($q_\mathrm{95} ∼ a^2 /I_p$ ). Through full 3D simulations, we demonstrate that accessing lower q95 can trigger the fast growth of MHD instabilities leading to magnetic stochastization. This opens up a possible pathway for benign termination of RE beams in ITER via distributed RE losses onto the wall, even without executing a Neon flushout. Dynamics of this process and its sensitivity will be discussed.
[1] G.T.A. Huijsmans et al., Nucl. Fusion 47.7, 659 (2007).
[2] M. Hoelzl et al., Nucl. Fusion 61, 065001 (2021).
[3] V. Bandaru et al., Phys. Rev. E 99, 063317 (2019).
[4] L. Hesslow et al., Nucl. Fusion 59, 084004 (2019).
[5] O. Vallhagen et al., J. Plasma Phys. 86, 475860401 (2020).Speaker: Vinodh Kumar Bandaru (Max-Planck-Institute for Plasma Physics, Garching) -
23
Runaway Electron Related Relaxation Phenomena in EAST Disruptions
Runaway electron (RE) currents of several mega amperes are expected to be generated in ITER disruptions due to avalanche multiplication. An uncontrolled loss of these high-energy electrons to the plasma facing components might cause serious damage. We present here observations of the RE-related relaxation phenomena during different phase of disruptions in the EAST tokamak.
RE formation phase: Electrostatic fluctuations are observed at the beginning of the current quench, which is detected by the HXR signals and related to strong RE losses. Note that the fluctuation is found in the disruptions triggered by amount of argon and neon injection but not helium injection, suggesting that the fluctuation should be driven by REs generated during disruptions. The frequency of the mode, in the range from 20 to 40 kHz, strongly depends on the amount and species of injected impurity and its evolution is suggested to be consistent with the evolution of REs. Mode structure of (m, n=1, 0) has been identified based on SXR arrays. GAM can capture most of the features of the fluctuations and barely-trapped/passing electrons can contribute to excite the mode.
RE plateau phase: Burst-like relaxations during the RE plateau phase cause large RE losses, which is seen by the spikes in the signals of magnetic coils. Two distinct types of RE-related relaxation phenomena are distinguished on the basis of the amplitude of magnetic fluctuations. Large-amplitude magnetohydrodynamic activity with indications of RE loss covering the entire energy range is observed during the RE plateau when the edge safety factor decreases to less than 3, and the external kink mode is discussed to resolve this anomaly. Burst-like relaxations with small-amplitude magnetic fluctuations and ~0.6 kHz frequency are confirmed from the spikes in the hard X-ray array signals under a negative loop voltage, and REs with medium energy are significantly lost at the same time. A possible mechanism for the instability is that due to the negative loop voltage, electric field de-accelerates REs and decrease the energy in the medium energy region, and finally, the modification of RE energy spectrum excites this kinetic instability.
These results will further deepen the understanding of RE losses in EAST and be an important part of RE mitigation or avoidance research in future.
Speaker: Prof. long zeng -
24
Strategy of disruption mitigation using multiple injection of shattered pellets in KSTAR
ITER adopts massive particle injection using shattered pellet injection (SPI) as a basic mitigation method to mitigate three major risk factors that can occur in the process of plasma disruption: heat load, electro-magnetic load, and runaway electrons. The injected particles composed of a combination of hydrogen and neon increase the density of plasma through the assimilation process to prevent runaway electrons and emit stored energy in the form of radiant energy. A safe and effective disruption mitigation strategy in the ITER disruption mitigation system capable of injecting a total of 27 pellets depends on which combination of pellets are injected at what time. Among these strategies, the most basic issue is whether to sequentially or simultaneously inject hydrogen, which increases density, and neon, which emits energy. On the other hand, plasma dilution-cooled by hydrogen becomes plasma with completely different characteristics from typical tokamak plasma due to its high density and low temperature. The pellet assimilation in the dilution-cooled plasma and the radiation of stored energy may be different from those of typical tokamak plasma. However, experiments on dilution-cooled plasma have not been sufficiently conducted. KSTAR with two SPIs that form a symmetry in the toroidal direction can independently inject three different pellets for each SPI. KSTAR has conducted experiments to test the disruption mitigation strategy of ITER using multiple SPIs and diagnostics capable of diagnosing the plasma disruption process.
Speaker: Dr Jayhyun Kim (Korea Institute of Fusion Energy) -
25
Torque balance analysis in real-time of rotating MHD for disruption prediction and avoidance in KSTAR
Advanced tokamak reactors require a low disruptivity ceiling to reach commercial viability. The damaging impact of plasma disruptions on machine components can greatly reduced the lifetime of a device. A precursor to disruptions is the locking dynamic of rotating MHD events that are often neoclassical tearing modes (NTM). The drag of electromagnetic and fluid viscosity torques can cause the slowing down of NTM’s with a saturated island width and lock them to the wall reference frame. A balance of the driving torque from the NBI, and drag from perpendicular viscous diffusion drag and electromagnetic forces on the mode, as well as its inertia is used to model the mode rotation dynamics. Threshold rotation frequencies below which the mode rotation is expected to lead to a locking serve as a disruption forecaster. Mode identification is computed most accurately by Fourier analysis of a toroidal array of magnetic probes, or using simpler approaches generally more amenable to real-time calculation. From the rotation, the torque components are then calculated based on conditions for the expected drag torque ratios at the mode onset, changes in frequency, and Mirnov signal amplitudes. This technique is employed for offline and real-time analysis of KSTAR plasmas with potential to signal use of active mode control or disruption mitigation systems. *This research was supported by the U.S. Department of Energy under contracts DE-SC0018623 and DE-SC0016614.
Speaker: Juan Riquezes (Columbia University) -
26
Update on SPI experiments at JET-ILW in 2019-2020
In 2019, the JETILW was equipped with a Shattered Pellet Injector (SPI) system with a wide capability to allow studies on the efficacy of shattered pellets in reducing the electro-magnetic and the thermal loads during disruptions and the avoidance/suppression of the formation of runaway electrons. This contribution presents various aspects of the SPI experiments, with the three-pellet injector with diameter [4.57, 8.1, 12.5] mm, conducted on JET-ILW in 2019-2020.
The experiment was performed with Ip = (1.1 - 3.1) MA plasma and mainly with Ne + D2 pellet composition, but also with Ar pellets. Experiments were mainly conducted with normal (“healthy”) plasma, not prone to disruption. The current quench (CQ) duration, τ80-20, as well as the radiated energy during the mitigation reflect the effectiveness of mitigation.
A pellet with a high content of Ne or Ar can reduce the CQ duration to below the upper required JET threshold, τ80-20 < 27.5 ms. Plasmas with high internal (thermal + poloidal magnetic) pre-disruptive plasma energy require a high content of Ne or Ar pellet to obtain a short CQ duration, moreover, Ar pellets are more efficient than Ne pellets.
The Ne fraction in the pellet does not affect the mitigation efficiency except for pellets with a small fraction of Ne, Ne/(Ne+D) < 0.5. Moreover, pellets with a very small amount of Ne, and accordingly large amount of D, instead of causing a mitigated CQ, create the conditions for a “cold” VDE, that is the worst-case scenario for plasma termination.
Disruption mitigation is intended to be applied on off-normal or post-disruptive plasmas. SPI plasma mitigation has been shown to be equally effective (in terms of τ80-20) on normal (“healthy”) plasma and post-disruptive plasma. The successful prevention of development of full-scale Vertical Displacement Events (VDE), which could become an Asymmetrical VDE, was also demonstrated.Speaker: Sergei Gerasimov (CCFE) -
27
Evolution of Data-driven Disruption Prediction: from Machine Learning to Deep Learning
Because of the need to obtain a real-time predictive model of the complex disruption behaviour, a great effort has been devoted in the last decades to apply data-driven models to disruption mitigation and avoidance, starting from the first black-box neural network approaches to the more physics-based Machine Learning (ML) models up to the latest models based on Deep Learning techniques. In the present contribution, the authors describe the evolution of the disruption predictors to overcome the inherent limitations of the data-driven approach. In [1], a Generative Topographic Mapping (GTM) maps the high dimensional plasma operational space in a 2D map where different disruption risk can be easily identified. Tracking the discharge evolution on the map, the chain of events leading a disruption is followed. One of the key challenges to obtain a performing ML predictor is to identify, with as much precision as possible, the disrupted phase of the disrupted discharges. A good labelling of the discharges avoids to providing contradictory information to the ML model during its training. To meet this need, a statistical approach has been proposed in [2], framed into the anomaly detection techniques, which detects the off-normal behavior of the plasma in a disruption. This helps to implement a continuous learning system that could be automated overcoming the well-known ageing of whatever data-driven methods. Another critical phase of ML approaches is the need to extract the more informative features from multidimensional diagnostics, such as temperature, density and radiation profiles, which have proved essential for achieving high performance. In the case of ML algorithms, profile information must be synthesized into 0D signals, such as the peaking factors proposed in [1, 2]. In order to avoid the complex hand-engineered feature extraction, in [3] a deep learning prediction model, based on deep Convolutional Neural Network (CNN), has been implemented. The CNN extracts the spatiotemporal information from 1D plasma profiles. The features are automatically produced by a cascade of filtering blocks, interconnected through nonlinear activation functions. A Multilayer perceptron combines them producing the output of the network through a Softmax layer that gives the likelihood of the input to belong to a regularly terminated or a disrupted discharge. The achieved performance is proven to be better than the one obtained with GTM model trained with the 0D peaking factors, reaching, on the test set, about 93% of successful predictions, 4% of false alarms, and alarm times suitable for avoidance actions. The modularity of the deep learning approach eases the introduction of additional 2D and 1D signals from, for example, Fast Visible Cameras or spectrograms from Mirnov coils.
References
[1] Pau A. et al, 2019, Nucl. Fusion 59 106017.
[2] Aymerich E. et al, 2021, Nucl. Fusion 61 036013.
[3] Aymerich E. et al, 2022, Nucl. Fusion 62 066005.Speaker: Prof. Giuliana Sias (University of Cagliari)
-
8
-
Disruption Consequences Council Room
Council Room
ITER Headquarters
-
28
Disruption loads in SPARC
SPARC is a compact, high-field, burning plasma experiment, with the mission to demonstrate net fusion energy and retire risks on a fast-track development path toward an ARC-class pilot plant.
SPARC is designed to operate at 12.2 T at the plasma major radius, 1.85 m, and 8.7 MA in a double null configuration with elongation up to 1.97, for a 10 s flat-top. The device is conservatively engineered assuming disruption frequencies normally observed in tokamaks and prediction accuracy already demonstrated in present day machines. In-vessel components are designed to withstand the fastest current quench of 3.2 ms predicted by the ITPA disruption database (IDDB) scaling [1]. The vessel and its support system are designed to withstand the slowest current quench of 40 ms as scaled from observations in the JET-ILW [2], accounting for a halo current fraction times toroidal peaking factor of 0.7 consistent with the scaling in the IDDB and including magnetic damping.
The SPARC disruption mitigation strategy includes both a massive gas injection system and a runaway mitigation coil. SPARC operational space will expand progressively while its disruptions are characterized and compared with the design load cases to inform the operational limits and preserve the integrity of the SPARC structures. The plasma facing components are optimized for normal operation; they are expected to degrade gently and mostly away from the power handling regions of the divertor.
The material presented will focus on the derivation and application of the disruption electromagnetic loads.
Work supported by Commonwealth Fusion Systems.
Speaker: Valeria Riccardo (CFS) -
29
Thermal Quench in DIII-D locked mode disruptions
The cause of the thermal quench (TQ) in tokamak disruptions has not been well understood.
Recent work identified the TQ in JET locked mode disruptions with
a resistive wall tearing mode (RWTM) [1].
New research finds a similar instability in DIII-D locked mode shot 154576 [2]. The instability is studied with simulations, theory, and comparison to experimental data. Linear theory and simulations show the mode is stable for an ideal wall, and unstable with a resistive wall.
Its growth rate $\gamma$ scales asymptotically as the resistive wall
time $\tau_{wall}$ to a negative fractional power,
$\gamma \propto \tau_{wall}^\alpha,$ which varies between $-4/9 \ge \alpha \ge -1.$
The scaling depends on the tearing stability parameter $\Delta',$ with and without an ideal wall.
The growth rate increases as the edge safety factor approaches $q = 2.$
The growth time is consistent with the experimental thermal quench time.
Nonlinear simulations show that the mode grows to large amplitude, causing a thermal quench.
These results could be important for ITER [3], greatly mitigating the effects of disruptions.
The ITER thermal quench time could be
much slower, because the wall resistive penetration time is
50 times longer than in JET and DIII-D.[1] H. Strauss and JET Contributors,
Effect of Resistive Wall on Thermal Quench in JET Disruptions,
Phys. Plasmas 28, 032501 (2021)[2] R. Sweeney, W. Choi, M. Austin, et al.
Relationship between locked modes and thermal quenches in DIII-D, Nucl. Fusion 58, 056022 (2018)[3] H. Strauss, Thermal quench in ITER disruptions,
Phys. Plasmas 28 072507 (2021)Speaker: Dr Henry Strauss (HRS Fusion) -
30
KORC Modeling of Wall Heating by Avalanche Runaway Electrons During a Final Loss Event in DIII-D
This work extends the recent modeling of runaway electron (RE) mitigation in Ref. {1} by including an avalanche RE source {2} in the Kinetic Orbit Runaway electrons Code (KORC). Our main finding is that REs produced by the avalanche source are the primary contributor to transient high heat loads observed at plasma-facing component (PFC) surfaces as shown in Fig. 1. The magnitude of the calculated heating is comparable to that for DIII-D graphite tile ablation as described in Ref. {3}, and qualitative features of the simulated PFC surface heating agree with infrared imaging of the first wall tiles in DIII-D only when the avalanche source is included.
We model DIII-D discharge 177031, where a RE beam undergoes a rapid final loss event. As the beam advects into the inner wall, a large toroidal electric field is induced in a region where magnetic surfaces connect to the wall. The drift orbit effects of high energy, primary REs allow them to remain confined in this open field region where the large electric fields increase secondary RE generation by the avalanche source. Lower energy, secondary REs have negligible drift orbit effects, causing them to be rapidly deconfined and deposit their energy shallowly into the PFCs because the energy deposition length scales with energy as shown by the green trace in Fig. 2. Furthermore, because the distribution of secondary REs generated by the avalanche source scales inversely with energy, lower energy REs provide the dominant contribution to PFC surface heating as shown by the blue trace in Fig. 2.
The KORC simulations presented in this work evolve a distribution of tracer RE guiding center (GC) orbits using a time series of axisymmetric experimental reconstructions of the electromagnetic fields, ignoring contributions from non-axisymmetric MHD modes. Fokker-Planck and large-angle collisions with the effects of partially-ionized impurities are employed assuming uniform and constant plasma and impurity profiles with magnitudes inferred from experimental observations. To calculate the PFC surface heating due to RE deposition, we have generalized the 1D analytic model from Ref. {4} to include the energy dependence of the deposition length scale. We have approximated the angle of incidence of deposition using the GC angle of incidence, pitch angle, and a randomized gyrophase constrained by the deposition geometry.
{1} Beidler et al., Phys. Plasmas 27, 112507 (2020)
{2} Aleynikov et al., IAEA FEC Paper TH/P3-38 (2014)
{3} Luxon, Nucl. Fusion 42, 614-633 (2002)
{4} Martín-Solís et al., Nucl. Fusion 54, 083027 (2014)*This work is supported by the US DOE under contracts DE-AC05-00OR22725, DE-FC02-04ER54698, and DE-AC02-05CH11231.
Speaker: Matthew Beidler (Oak Ridge National Laboratory)
-
28
-
15:30
Coffee Council Room Lobby
Council Room Lobby
ITER Headquarters
-
Disruption Consequences Council Room
Council Room
ITER Headquarters
-
31
Parametric study of Alfvénic instabilities driven by runaway electrons during the current quench in DIII-D
We report comprehensive investigation of Alfvénic instabilities driven by runaway electrons (REs) during the current quench in the DIII-D tokamak. These instabilities are observed as toroidal magnetic field fluctuations in the frequency range of 0.1–3 MHz and correlate with increased RE loss from the plasma which candidates them to be responsible for non-sustained RE beams and motivates a study to use such instabilities as an alternative or complimentary mean to massive impurity injection to avoid or mitigate RE beams in ITER.
It is found that decreasing the toroidal magnetic field ($B_T$) leads to instabilities shifting to lower frequencies, as expected for Alfvénic instabilities. As $B_T$ decreases, the RE population becomes more energetic, the power of RE-driven instabilities increases, and no RE beam is observed when the maximum energy of REs exceeds 15 MeV (or when $B_T$ is below 1.8 T). This may be caused by worsening conversion of plasma to RE current as $B_T$ decreases and may explain the common empirical observation of high $B_T$ favorable for sustained RE beams. Theoretically, decreasing $B_T$ can decrease the on-axis current density (through $j_0=2B/μ_0qR$) which leads to lower post-thermal electric field and weaker primary RE generation. As a result, small RE population needs to be accelerated to higher energy in order to replace the decaying plasma current, and this more energetic RE population would increase the drive of instabilities.
Analysis of plasma disruptions at different plasma core temperature ($T_e$) shows that the RE population is much less energetic (with maximum energy of only 3 MeV) when $T_e$ reaches about 8 keV, and no RE-driven instabilities are observed in this case. Since reactor-relevant high $T_e$ increases the current conversion, this supports the results of the $B_T$ experiment.
Besides the $B_T$ and $T_e$ experiments performed using Ar impurity injection, disruptions caused by massive gas injection (MGI) of Ne or D$_2$ were also studied. Both Ne and D$_2$ MGI result in no RE beam. Ne MGI leads to highly energetic RE population (with maximum RE energy exceeding 13 MeV even for a very substantial injection of 780 Torr·L) and one of the most clear Alfvénic instabilities. On the other side, no signs of REs nor Alfvénic instabilities are observed after D$_2$ MGI, which can be explained by slow plasma cooling (1.5 ms vs 0.5 ms for Ar MGI).
Measurements of the polarization of Alfvénic instabilities ($\delta B_T/\delta B_P$) indicates that it is of predominantly toroidal (compressional) nature, consistent with estimates and modeling suggesting excitation of Compressional Alfvén Eigenmodes. The toroidal mode number of these instabilities is found to be from −1 to +2, partially supporting the results of modeling presently not predicting n=0 mode.
Speaker: Andrey Lvovskiy (Oak Ridge Associated Universities) -
32
CONS Discussion
-
31
-
-
-
Prediction & Avoidance Council Room
Council Room
ITER Headquarters
-
33
Applications of machine learning for plasma control in fusion reactors
Machine learning techniques have been applied successfully in EAST plasma equilibrium reconstruction and disruption prediction. Regression neural network models are trained to identify the plasma center position and calculate equilibrium plasma parameters including li,β_p,κ,q_0 and a_minor with magnetic diagnostic signals as input features [1][2]. The results on test dataset show good calculation accuracy compared with the tokamak simulation code (TSC) calculated value. Mean absolute errors of the horizontal and vertical positions of current center are within 1 mm and the normalized errors of these equilibrium parameters are within 1%. Alongside that, the Bayes inference method is applied to reconstruct the plasma current density profile and boundary recognition [3]. An accurate boundary can be obtained with only magnetic diagnostic, but to get an accurate current density profile especially in core plasma area, additional diagnostic is needed. Both the neural network and Bayes model can iterate one step at around ~1ms, which is fast enough for real-time equilibrium reconstruction and current profile control. If with appropriate accelerating such as parallel calculating, they can also be used for controller that require faster response, such as vertical displacement control.
Beside the above examples, machine learning also show great potential in area of disruption prediction. A convolutional neural network (CNN) that is trained on a database of EAST disruption discharges is used to recognize disruptive discharges and distinguish them from non-disruptive discharges. The true positive rate of the model increases up to 87.5%, while the false rate decreases to 6.1% [4]. Meanwhile, a real-time disruption predictor using a random forest (DPRF) was developed for high-density disruptions and used in the plasma control system of the EAST tokamak [5]. During the dedicated experiments, when the disruption probability signal increases up to a preset, configurable threshold for more than 10 ms, a trigger signal is sent to the MGI system and neon gas was injected into the plasma to successfully mitigate disruption damage.
With cooperation between EAST, Alcator C-Mod and DIII-D group, disruption databases of three single machine are built and managed with MySQL tablets [6]. Cross-machine disruption prediction is carried out on the database of three machines [7]. A hybrid deep-learning algorithm is trained with a dataset composed with only 20 EAST discharges and with more than a thousand discharges from DIII-D and C-Mod for disruption prediction. Then testing this disruption predictor on EAST and a predictive accuracy of AUC=0.973 is achieved. Machine learning shows great potential to act as the solution for future fusion reactor’s disruption prediction.Ref.
[1] IEEE Transactions on Plasma Science, 2019, 48(1): 54-60.
[2] Chinese Physics B 28.12 (2019): 125204.
[3] Fusion Engineering and Design 172 (2021): 112722.
[4] Plasma Physics and Controlled Fusion 63.2 (2020): 025008.
[5] Nuclear Fusion 61.6 (2021): 066034.
[6] Nuclear Fusion, 2019, 59(9): 096015.
[7] Nuclear Fusion 61.2 (2020): 026007.Speaker: Bingjia Xiao (Institute of Plasma Physics, Chinese Academy of Sciences) -
34
Disruption paths in high performance scenarios at JET with D, T, and DT plasmas
Present contribution aims at comparing the different kinds of disruptions that occurred in the last JET with ITER-like wall (ILW) campaigns with Tritium and Deuterium-Tritium fuels.
Last campaigns performed in JET-ILW with a D fuel showed that the majority (around 80%) of disruptions follow two main paths [1]. The first path (temperature hollowing, TH) is strictly related with the influx of high Z impurities, which can accumulate in the plasma core increasing the radiative losses and deteriorating the electron temperature (Te) profile [2]. The second path (edge cooling, EC) is instead related to the erosion of the edge Te profile; the contraction of Te profiles looks similar to that of a "density limit" disruption [3]. Both paths are found [4] to modify differently the current density profile but always in a way to destabilize a 2/1 mode, which locks before disruption.
In T and in DT campaigns, around the 90% of disruptions can be explained by the occurrence of TH or EC. Furthermore, it is found that in DT the two main scenarios developed at JET [5] are characterized by disruptions following mainly one out of the two paths. In the DT experiments performed in baseline scenario (βN~1.8, q95~3), 12 disruptions out of 13 follow an EC; while in the DT hybrid scenario (βN~2-3, q95~4) 13 disruptions out of 15 occur after a TH in the ramp-down phase.
The two different kinds of disruptions will be compared for the two high performance scenarios and for the different isotopes contents. The comparison will be performed considering the conditions at onset of the 2/1 modes before the disruption, in terms of the power balance and of the density levels.[1] C.Sozzi et al. 2020 28th IAEA Fusion Energy Conference -IAEA CN-978
[2] J. Hobirk et al 2018 Nucl. Fusion 58 076027
[3] J.A. Wesson et al. 1989 Nucl. Fusion 29 641
[4] G.Pucella et al. 2021 Nucl. Fusion 61 2021 046020
[5] L. Garzotti et al 2019 Nucl. Fusion 59 076037Speaker: Edoardo Alessi (Istituto per la Scienza e la Tecnologia dei Plasmi- CNR) -
35
Avoiding sensor induced disruptions with use of machine learning
Control is necessary to keep fusion plasmas stable. This requests a set of real-time diagnostics. These sensors and/or data acquisition systems are prone to failure, especially under the demanding environments of a fusion reactor that has cryogenic and extreme hot conditions, high neutron production and high magnetic fields. Current real-time control algorithm assume the sensors as correct within given error bars. Algorithms that can check and variate each sensor as the plasma evolves (or in between discharges) in minutes/hours is hard to implant and is labor intensive which introducing human error. A promising approach is machine learning based design, where the control/analysis structure is trained using a combination of good and bad data (available or made up). We explain how these ML algorithms can be robust and can avoid the brittleness of hand written code. We present the example of NSTX-U equilibrium reconstruction using robust ML algorithms (NF 2022, in review). The achieved reconstruction makes NSTX-U robust against magnetic sensor failures while at the same time getting rid of the requirement that a human input on which sensors are good. Another benefit of the ML reconstruction is higher quality compared to rt-EFIT. Given the immense sensor fusion task, the high cost of disruptions and the relative low human resource availability at ITER, robust and easy to train ML-based sensor fusion might prove valuable for ITER and future reactors.
Speaker: Azarakhsh Jalalvand (Ghent University) -
36
Insights on disruption physics in MAST using high speed visible camera data
Disruption prediction requires an understanding of the routes that a plasma may take from being in a healthy state to a disruption, such as the analysis carried out on JET [1]. Of particular concern are those routes that give very little warning of an imminent disruption because they give little potential to either take avoiding or mitigating action. We will use the MAST high speed visible camera data to provide insights into disruptions and their precursors. MAST (Mega Amp Spherical Tokamak) operated from 2001 to 2013 at UKAEA. MAST had a very open internal structure without a close fitting first wall which allowed particularly full views of the plasma. A high-speed video camera, taking data at around 100kHz, was deployed for a series of shots in the later MAST campaigns. We have produced spectrograms from these data which show the presence of various MHD instabilities such as ballooning filaments [2], the LLM [3] and NTM [4]. We further use techniques such as EVM (Eulerian Video Magnification) [5,6], which can highlight oscillations in camera data, to look at the structure of these instabilities. We use this and other available data to look at disruption precursors and the disruptions themselves in both LSN and DND MAST plasmas and discuss implications for disruption prediction.
This work has been funded by the EPSRC Energy Programme [grant number EP/W006839/1].
[1] P.C. de Vries et al ‘Survey of disruption causes at JET’ (2011) Nucl. Fusion 51 053018
[2] C. J. Ham, S. C. Cowley, G. Brochard, and H. R. Wilson ‘Nonlinear Stability and Saturation of Ballooning Modes in Tokamaks’ (2016) Phys. Rev. Lett. 116 235001
[3] I.T. Chapman et al ‘Saturated ideal modes in advanced tokamak regimes in MAST’ (2010) Nucl. Fusion 50 045007
[4] H.R. Wilson ‘Neoclassical Tearing Modes’ (2012) Fusion Science and Technology 61 113-121
[5] Hao-Yu Wu, Michael Rubinstein, Eugene Shih, John Guttag, Fredo Durand, and William T. Freeman ‘Eulerian Video Magnification for Revealing Subtle Changes in the World’ (2012) ACM Transactions on Graphics (Proc. SIGGRAPH 2012) 31 (4)
[6] D Ryan ‘Amplification of Plasma Edge Phenomena with Eulerian Video Magnification’ (2014) University of York MSc thesis
Speaker: Christopher Ham (UKAEA-CCFE)
-
33
-
10:55
Coffee Council Room Lobby
Council Room Lobby
ITER Headquarters
-
Prediction & Avoidance Council Room
Council Room
ITER Headquarters
-
37
Mode lock precursors from the electron temperature profile in plasma termination on JET
The capability to terminate plasma pulses safely is an important goal towards the optimization of operational scenarios in tokamaks, so it is of great importance to study the physical phenomena involved in plasma disruptions and to develop precursors for avoidance and/or mitigation actions. The development of tearing modes (TMs) inside the plasma is a major cause of disruptions. It has been shown that there are two main TM destabilization paths in plasma termination on JET, connected to the problem of impurity control: the core accumulation of high-Z impurities, leading to a temperature hollowing (TH) and to a broadening of the current density profile, and the influx of low-Z impurities, which are mainly radiating at the edge, leading to an edge cooling (EC) and to a shrinking of the current density profile. The formation of an “outboard radiating blob” due to high-Z impurities accumulated in the low-field side can also be responsible for EC.
Following the picture of TMs generated by changes in the current density profile, reflecting the changes in the electron temperature profile, two parameters have been defined from ECE radiometry to highlight the occurrence of TH and EC, and the possibility of identifying locked mode precursors based on the two parameters has been preliminary explored, by evaluating, for a large dataset of pulses, the characteristic time intervals between the increase of such parameters and the mode lock, which is widely adopted as disruption precursor to trigger mitigation actions. The obtained advances with respect to mode lock signal (1 s for TH, 100 ms for EC) are associated to the effective resistive diffusion time linking the changes in the electron temperature profile and the changes in the current density profile leading to large mode amplitudes. This is the reason why the real-time implementation of such parameters (as already done for a temperature hollowness indicator during the current ramp-up phase) would offer the capability of obtaining alerts before the mode onset, when stability analysis indicate a stable MHD scenario. In particular, the TH parameter could provide alerts useful to attempt to correct the termination, avoiding the disruption, whilst the EC parameter could provide alerts useful to anticipate mitigation actions.
Additional information are provided by the dynamics of mode lock signals. Mode saturation is quite general for EC in peaked electron temperature profile and the thermal quench (TQ) is usually induced by disruption mitigation valve (DMV) intervention, so it is not crucial to anticipate DMV. However, an explosive growth of the mode amplitude is sometimes observed for EC in hollow electron temperature profiles, leading in some cases to unmitigated TQ, so even a small advance in DMV intervention would be essential. A detailed analysis of this behaviour is planned in view of ITER, where the unmitigated disruption rate should be reduced as much as possible. The possibility of combining information on electron temperature and radiation profiles is also discussed, and examples from recent DT and isotopic experiments on JET are shown.Speaker: Dr Gianluca Pucella (ENEA, Fusion and Nuclear Safety Department, C.R. Frascati, Italy) -
38
Disruption Event Characterization and Forecasting Results and Initial Real-Time Application
Disruption prediction and avoidance is critical for ITER and reactor-scale tokamaks to maintain steady plasma operation and to avoid damage to device components. Physics-based disruption event characterization and forecasting (DECAF) research determines the relation of events leading to disruption, and forecasts event onset. The analysis has access to data from multiple tokamaks to best understand, validate, and extrapolate models. Recent code improvements allow fully automated analysis spanning an entire device run campaign or even the entire device database. Such analysis is showing very high true positive success rates, in some cases over 99% with early forecasting (on transport timescales) well before the disruption. While this result is very encouraging over broad ranges of shots, analysis continues to ensure causality of the computed DECAF events with the detected disruption, rather than just correlation. This is a critical question to be answered for any disruption prediction analysis. Significant new hardware and software for real-time data acquisition and DECAF analysis are being installed on the KSTAR superconducting tokamak. Real-time magnetics, electron temperature, Te, profiles from electron cyclotron emission (ECE), 2D Te fluctuation data from ECE imaging, and velocity and Ti profiles show excellent agreement with offline data/analysis. An MHD mode locking forecaster has been developed for off-line and real-time use using a torque balance model of the rotating mode. Early warning forecasts on transport timescales potentially allow active profile control to avoid the mode lock. Mode stability alteration by ECCD is examined and recent experiments have shown the ability to avoid mode lock-induced disruption by applying rotating 3D fields. Innovative counterfactual machine learning is used to examine hypothetical RWM stabilization scenarios with rotating MHD. An ELM identification event module includes the ability to distinguish localized and global MHD events. Fully non-inductive current scenarios in KSTAR are examined by “predict-first” analysis of already highly (~75%) non-inductive plasmas. Resistive stability analysis including delta’ computed by DCON is evaluated with comparison to experiment examining sensitivity to localized variations of kinetic equilibrium reconstructions of the q profile using MSE magnetic pitch angle data.
Supported by US DOE Grants DE-SC0016614 and DE-SC0018623.Speaker: Steven Sabbagh (Columbia University) -
39
Transfer Today’s Disruption Prediction Model to Future Tokamaks
Disruption is a major obstacle for tokamaks to be commercially viable reactors. Accurately predicting an incoming disruption and deploying disruption mitigation system is one of the keys to solve this problem. Today’s machine learning based disruption predictors do have great performance if given good enough data to train. But future tokamak will not provide good enough data before damaging itself. The way to exploit the limited data from future machines and the abundant data from existing machines is the key to solve this problem. This work presents multiple attempts to leverage on the transfer learning technique to solve the above problem. Firstly, a transferable deep neural network tokamak diagnostic feature extractor is proposed and demonstrated on 2 tokamaks. Secondly, a domain adaptation method called covariance alignment was applied to manually extracted features to align feature spaces from 2 tokamaks. Thirdly, an adaptive anomaly detection method was introduced to bootstrap a pre-trained disruption prediction model on a new machine. Lastly, we tried to transfer a model trained with low performance discharges to high performance scenarios. Those above methods displayed variant degrees of success. The results and the models was analyzed and revealed some hints about transferring existing disruption prediction models to new tokamaks.
Speaker: Wei Zheng (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, Huazhong University of Science and Technology)
-
37
-
12:45
Lunch
-
Prediction & Avoidance Council Room
Council Room
ITER Headquarters
-
40
Control and machine learning algorithms for disruption avoidance and proximity control
Achieving acceptably low disruptivity on ITER and future reactors will require active monitoring and control of the proximity of operating points to unsafe regions. Although mitigation strategies can protect the infrastructure from disruptions and engineering limits, maximizing scientific or economic output of a device demands avoiding triggering mitigation systems while optimizing performance. Model-based control combined with system identification, AI/ML, and other data science techniques offers a potential solution to this problem. Model Predictive Control (MPC) strategies use dynamic models to repeatedly plan actuator trajectories over a finite horizon in real time, enabling optimization of operating points within actuator and state constraints. Two main challenges to applying these techniques to tokamak control and disruption avoidance are the need for suitably fast and accurate models, and selection of appropriate constraints. Novel algorithms have recently been developed to address these issues, enabling (1) accelerated, real-time forecasting of plasma behavior and (2) identification of the local safe operating region. These new tools, coupled with model predictive control techniques, will enable real-time optimization of tokamak operating points within a set of constraints approximating the safe operating region.
Accelerated modeling: Surrogate modeling approaches are used to identify fast, local (to the distribution of training data) approximations of physics models. This technique has been used to develop accelerated approximations of actuator models (e.g., NUBEAM), free boundary equilibrium solutions, and predictive integrated modeling simulations (e.g., TRANSP). Where physics models are lacking, the same modeling techniques can be applied to empirical data. Combining these approaches enables physics-informed modeling that can learn from empirical data to improve predictions, while maintaining real-time-relevant execution times.
Identification of constraints: The proposed safe operating region identification algorithm enables use of state-of-the-art disruption prediction models as tools for active disruption avoidance. A GPU implementation enables rapid (~10ms) evaluation of the disruption prediction model at many points in a region around the local operating point, followed by identification of a set of convex linear constraints bounding the local safe operating region. The identified constraints are in a form that enables calculation of the proximity to disruption boundaries (in terms of controllable physical quantities) and can be used within fast model predictive control algorithms.
Simulations and experiments: Closed loop TRANSP simulations have been used to test constrained model predictive control strategies. For example, simultaneous control of stored energy, internal inductance, and loop voltage while actively avoiding vertical stability limits has been simulated for KSTAR. Initial progress toward experimental implementation of the proposed methods on DIII-D will be discussed. Demonstration of these approaches is expected to inform development of control strategies for ITER and future burning plasma experiments.Speaker: Mark Boyer (Princeton Plasma Physics Laboratory) -
41
Proximity-to-Instability Control and Cross-Machine VDE Stability Metrics on DIII-D and KSTAR
ITER will require exceptionally low disruptivity while pushing the limits of plasma performance. Ensuring robust stability will require a comprehensive strategy, and must include the continuous regulation of the proximity to stability and controllability limits, also called “Proximity Control.” DIII-D has been developing a Proximity Control architecture [1] which modifies control targets and actuator constraints based on stability metrics in real-time. The system has been applied and tested for robust Vertical Displacement Event (VDE) avoidance and preventing un-intended H-L back-transitions. Despite a pre-defined (intentionally) increasing trajectory for elongation (κ) set to induce a VDE, the Proximity Control algorithm successfully over-rode original control inputs when growth rates exceeded a threshold for intervention and adjusted plasma shaping targets (both κ and plasma distance to the inner wall) to ensure stability. The controller regulated the VDE growth-rate at 800 rad/s for more than 3s [1]. Unintended H-L back-transition protection integrated with β_N feedback control was demonstrated with the real-time modification of the minimum input power based on two different real-time stability monitors. The first monitored the radiated power, and the second used a machine learning (ML) predictor for the likelihood of H-/L-mode. Both tools prevented an H-L back-transition with a user-set, low β_N feedback target trajectory, which would have otherwise led to reduced heating and back-transition. The J|| gradients in the edge of ITER Baseline Scenario plasmas have been identified as indicators of tearing mode stability [2]. Recent experiments are described, which demonstrate first tests and limited controllability of edge gradients in J|| in the ITER Baseline Scenario using real-time adjustments in the plasma triangularity and squareness.
Experiments on KSTAR have diagnosed the limits of its vertical control [3-4] using machine-scalable metrics, including measured VDE growth rates and the maximum Z-excursion that the vertical control system can restore (dZmax). A maximum κ > 2.2 was sustained for more than 3s in experiment. Triggered VDEs were used to directly diagnose the vertical growth rates, measured to reach up to 300 rad/s at these high elongations. dZmax was diagnosed under two different ELMing conditions: one with slower, larger ELMs, and the second with faster, smaller ELMs. dZmax was measured to be 2cm and 2.3cm in each of these conditions, respectively. These correspond to machine-scalable metrics of dZ_max/〈dZ〉_noise~15x-10x and dZ_max/a_minor~4.5-5%, consistent with previous studies on DIII-D and similar to expectations of marginal stability on ITER [5].
In future work, the DIII-D Proximity Control algorithm will be expanded to addition disruption prevention applications as well as ported to KSTAR for direct cross-device studies. On DIII-D, the controller has been integrated with the Disruption Prediction using Random Forests (DPRF) [6] interpretable ML algorithm, to be tested in upcoming experiments.
This work was supported in part by the US Department of Energy under DE-FC02-04ER54698 and DE-SC0010685.
[1] J.L. Barr 2020 NF 61 126019 [3] D. Mueller 2018 FED 141 9-14
[2] F. Turco 2018 NF 58 106043 [4] D.A. Humphreys 2009 NF 49 115003
[6] S.-H. Hahn 2020 FED 156 111622 [5] C. Rea 2019 NF 59 096016Speaker: Jayson Barr (General Atomics) -
42
The Desirability of Locking for RF Stabilization of Islands in Large Tokamaks
Although the stabilization of locked islands using RF-driven currents has been demonstrated experimentally in a pioneering series of experiments [1], experimental and theoretical research on RF island stabilization has continued to focus almost exclusively on stabilization during the rotating phase, before locking occurs. An emphasis on avoiding island locking has emerged from a concern about the damaging impact of locked islands on plasma performance, and in particular from a concern about the impact on disruptivity. However, as we discuss, there are circumstances under which it will be desirable to allow an island to lock before stabilizing it [2]. In large tokamaks such as ITER, the m=2, n=1 islands are expected to rotate slowly and to lock at relatively small width, impacting the ECCD power required to stabilize the island before it locks. Continued emphasis on stabilizing the island before it locks could have a significant negative impact on Q, the fusion gain. The impact on Q may be greatly diminished by adopting a strategy of allowing the island to lock before stabilization. The threat posed to the plasma performance by a locked island is ameliorated for sufficiently small islands. The rate of growth of the island may increase after locking, but that rate will remain on a slow resistive time scale. Experimental studies find that islands do not trigger disruptions until they become quite large [3]. Experimental results also indicate that loss of an H-mode can be avoided if the island is suppressed on a momentum confinement time scale after it locks. It will be necessary to use an externally imposed nonaxisymmetric field to control the phase of the locked island, so that the O-point will lie in front of the EC launcher. Contemporary tokamaks, in any case, have nonaxisymmetric coils for the purpose of controlling error fields. ITER will have a set of nonaxisymmetric coils for the purpose of compensating for n=1 field errors, and it is expected that the coils will be used to decrease these resonant field errors by a factor of about 4. For ECCD stabilization of the locked island it is only required that the currents in these coils be tuned such that the residual error field after imposition of the field error correction has the appropriate phase. No separate applied perturbation is required. In any case, if the phase of the error correction field is not controlled in this way then it will not be possible to employ ECCD stabilization if an off-normal event produces a locked island. These considerations suggest that it will be desirable to do locked island stabilization experiments for small islands in advance of ITER to provide validated modeling of potential stabilization scenarios.
[1] F. Volpe et al,Phys. Rev. Lett. 115, 175002 (2015).
[2] R. Nies et al, arXiv:2106.06581.
[3] P.C. de Vries et al, Nucl. Fusion 56, 026007 (2016).Speaker: Allan Reiman (Princeton Plasma Physics Lab)
-
40
-
Posters Entrance Lobby
Entrance Lobby
ITER Headquarters
All posters will be displayed on both days.
Please choose among the following formats for your poster presentation:
a) In person: please upload your e-poster in Indico and display your printed poster at the ITER site during the two posters sessions (19-20 July).
b) Remote: please upload your e-poster in Indico with audio recording as well as a summary slide (which may be presented during the appropriate discussion session). -
Prediction & Avoidance Council Room
Council Room
ITER Headquarters
-
43
P&A Discussion
-
43
-
Dinner Aix-en-Provence
Aix-en-Provence
-
-
-
ITER Site Tour
-
11:00
Coffee Council Room Lobby
Council Room Lobby
ITER Headquarters
-
Mitigation Council Room
Council Room
ITER Headquarters
-
44
Cryogenic Pellet Ablation Physics and Integrated Modelling of Shattered Pellet Injection
While experimental dependence of the pellet penetration depth can be approximated using simple scaling laws, underlying physics is extremely complex and requires challenges for multiscale and multiphysics modelling. The electron density of an ablation cloud changes over eight orders of magnitude from solid density ($10^{28}$ m$^{-3}$) to background one ($10^{20}$ m$^{-3}$) as it expands along the magnetic field line. Because the ablation cloud is over-pressured by heating from hot ambient electrons, the materials can drift down the magnetic field gradient during assimilation. For Shattered Pellet Injection (SPI), the situation is further complicated by that many fragments simultaneously penetrate into core plasma, and massively injected materials trigger global instabilities and thermal quenches. Although we have not yet acquired the SPI simulations that can consider all the relevant mechanisms self-consistently, our understanding has made a constant progress through the ITER DMS design validation activities.
In ITER, hydrogen and neon pellets will be used to fulfill the requirements for disruption mitigation. For Massive Gas Injection (MGI), it has been widely acknowledged that a radiative cold front is formed in the peripheral region, which destabilizes tearing modes and triggers a thermal quench. While the formation of the radiative cold front is followed by gas penetration during MGI, solid shards can precede it in the case of SPI and fuel particles in the core region. Runaway Electron (RE) avoidance may require rising the electron density by a factor of 20-40 or more. Therefore, a guideline for optimizing the SPI parameters is to raise the core density by the arrival of pellet shards at the plasma center before the trigger of a thermal quench. While larger fragment size, higher injection velocity, and higher velocity dispersion support this trend for one-stage neon mixed pellet injection, the same idea motivates us staggered injection of pure H$_2$ and neon mixed pellets for the avoidance of hot-tail RE mechanism. However, recent observations show that a poor assimilation efficiency of pure H$_2$ SPI poses a trade-off. When comparing the non-shattered pellet injection between different compositions, although the penetration depths are comparable between pure H$_2$ and 5 % neon mixed pellets, the density rise at the ablation position is only observed for the neon mixed pellets; in contrast, the pure H$_2$ injection exhibits a hollow density profile and particle loss to the scrape-off layer. The observation is explained by the dependence of the ExB drift displacement on the ablation cloud pressure, and less radiative pure hydrogen pellets lead to the poor assimilation efficiency, as observed for pure D$_2$ SPI at DIII-D. Taking such dependence of the ExB drift displacement into account, major trends of the ITER DMS functions have been analyzed for variations of thermal energy content, electron temperature, and magnetic field strengths through staged upgrade of the plasma performance of an ITER pulse using the 1D integrated disruption code INDEX.
Speaker: Akinobu Matsuyama (National Institutes for Quantum and Radiological Science and Technology) -
45
Non-axisymmetric Radiation Modeling of JET SPI Discharges Using Emis3D
Precise values for radiated power during tokamak disruptions are required to predict the effectiveness of mitigation techniques at preventing damage in net energy tokamaks like ITER and SPARC. Conventional approaches to calculating $ P_{rad} $ on JET assume a toroidally symmetric radiation source to allow fitting from foil bolometer arrays at two toroidal locations, which is otherwise an underdetermined problem. In mitigated disruptions, 3D MHD modes and localized impurity sources may break toroidal symmetry. To incorporate toroidally asymmetric structures in $ P_{rad} $ calculation, the Emis3D radiation modeling code introduced here adopts a physics motivated guess-and-check approach, comparing experimental bolometry data to synthetic data from absolutely calibrated candidate emissivity structures. Candidates are observed with the Cherab modeling framework [M. Carr et al. EPS 2017] and a best fit chosen using a reduced $ \chi^2 $ statistic. An uncertainty bar is derived by considering all models within two standard deviations of the best fit (the “uncertainty pool”), using bolometer channel errors of 10% of the maximum brightness at each timestep. The resulting uncertainty pool is sensitive to changes in the uncertainty quantification formalism, but the best fit is a robust finding. 2D tomographic inversions are tested, as well as helical structures and 3D MHD simulated distributions from the JOREK code [Huysmans et al. NF 47, 2007]. Two discharges in JET terminated by nominally identical pure neon shattered pellets carrying $2.46\times 10^{22}$ Ne atoms are analyzed using Emis3D (95709 and 95711). 2D tomographic inversions are built in 1ms intervals from the time of peak radiation through the current quench. These inversions are the best fit at these times, although helical distributions are within the uncertainty pool, introducing large lower error bars on $ P_{rad} $. Helical radiation structures are found to fit the pre-thermal quench (pre-TQ) best and exhibit a parallel flow towards the high-field side consistent with the magnetic nozzle effect and with JOREK simulations. Candidates from JOREK simulations are not found within the uncertainty pool, although this result is sensitive to the uncertainty quantification formalism. Improvements in the agreement between JOREK and the pre-TQ bolometry data were found by reducing the toroidal extent of the impurity source used in the simulation. On 95709 and 95711, radiated fractions of $f_{rad}= 0.95 +0.05/-0.31$ and $f_{rad}=0.98 +0.02/-0.26$, of the plasma stored energies 11.5 and 11.7 MJ respectively, are found, suggesting that the thermal and magnetic energy may have been fully mitigated in both discharges, in contrast with the standard weighted sum analysis that finds $f_{rad}=0.77$ and 0.90 respectively. However, both results are within the uncertainty margin, and $f_{rad}$ as low as $\sim65\%$ is possible for 95709. Time-dependent toroidal peaking factors will be presented.
Speaker: Benjamin Stein-Lubrano (Plasma Science and Fusion Center, Massachusetts Institute of Technology) -
46
Magnetohydrodynamic Modeling of Shattered Pellet Injection and Runaway-Electron Dynamics
Future tokamaks will require robust disruption mitigation to prevent machine damage from thermal loads, electromagnetic forces, and runaway-electron (RE) impact. The leading-candidate for this is shattered-pellet injection (SPI), which is being tested experimentally on several tokamaks and will be used on ITER. Verified, predictive models are needed to project the performance of these systems on future devices. We present an overview of disruption-mitigation and disruption-dynamics modeling performed with the M3D-C1, NIMROD, and MARS-F magnetohydrodynamics (MHD) codes. M3D-C1 and NIMROD have been coupled to a coronal non-equilibrium model for impurity ionization, recombination, and radiation along with a state-of-the-art model for pellet ablation. A 3D benchmark between M3D-C1 and NIMROD for an injected pellet in DIII-D has been improved due to a number of code enhancements. The codes agree on the peak radiated power as well as time scales for thermal quench, current quench, and onset of macroscopic MHD instability. Understanding of remaining discrepancies will be considered. The agreement found gives confidence in the ability of both codes to perform high-fidelity, predictive modeling for ITER and other future devices. M3D-C1 modeling has been performed for realistic SPI plumes based on JET experiments. Pure-neon and neon-deuterium pellets are considered, which vary in speed and shatter distribution due to the differing composition. Simulations with the velocities swapped show that at low speeds, the quench dynamics are similar for the two compositions, while at high speeds, the mixed pellet travels further into the plasma before complete thermal quench. These results show a competition of time scales between the traversal of the pellet and outside-in radiative collapse. Similar NIMROD modeling of JET finds that reducing viscosity increases MHD activity and decreases thermal quench time slightly. Initial results from KSTAR modeling with M3D-C1 and NIMROD will also be considered, with particular focus on comparing single versus dual, symmetric injection. In addition, validation of M3D-C1 and NIMROD DIII-D SPI modeling with experimental data and EFIT reconstructions during the early thermal quench will be considered. A coupling between NIMROD DIII-D SPI quench simulations and the Fokker-Planck CQL3D kinetic code has been established to simulate the production of REs and their radial transport. It is shown that without the radial transport, a large RE current is generated, up to 30% of the pre-pellet Ohmic current. However, when the radial transport is included in CQL3D, the RE current is reduced to undetectable levels, consistent with experiment. The MARS-F code has been updated with various modules capable of modeling RE loss due to 3D field perturbations (with REORBIT), runaway avalanche, and interactions between REs and MHD instabilities (MHD-RE hybrid). Important examples will be reported utilizing these new tools.
Work supported by US DOE grants DE-SC0018109, DE-SC0020299, DE-SC0016452, DE-FC02-04ER54698, & DE-FG02-95ER54309, the ITER Organization under Contract # IO/19/CT/4300002130, and is contributing to the ITER-Organization Disruption-Mitigation Task Force.
Speaker: Brendan Lyons (General Atomics) -
47
Non-equilibrium JOREK simulation of ITER L-mode SPI process
Simulation of Ne/H Shattered Pellet Injection (SPI) into an ITER L-mode plasma is carried out by JOREK non-equilibrium impurity model with separate electron and ion temperature. The focus of this report is on the MHD dynamics, the density transport, the temperature profile evolution and relaxation during the loss of the core confinement, as well as the impact from the non-equilibrium impurity, the neon mixture ratio, the nonlocal effect on ablation and the numerical toroidal elongation of the density source.
The non-equilibrium result is first compared against the Coronal Equilibrium (CE) result. The non-equilibrium treatment is shown to capture the early phase cooling due to the lowly charged impurities which is missed by the CE model, although in the late phase radiation is comparable between the models and the characteristic MHD dynamics are similar.
Meanwhile, SPIs with neon mixture ratio ranging from 0% to 20% exhibit only a slight difference in the characteristic MHD dynamics despite the difference in the radiative power. The relationship between this behavior and the lack of radiative collapse is discussed.
Further, ``$T_e$ hole'' is found to develop along the field line with Braginskii thermal conduction due to the competition between local thermal sink and parallel conduction, resulting in a lower ablation rate. Considering the mean-free-path of the electrons, flux-averaged electron temperature is used in ablation rate calculation, results in slightly enhanced ablation, broader MHD spectrum and stronger stochasticity.
Last, the numerical toroidal elongation of the ablation source is found to artificially lower the radiation power, the extent of such effect and its implication to numerical result is discussed.Speaker: Di Hu (Beihang University)
-
44
-
13:05
Lunch
-
Mitigation Council Room
Council Room
ITER Headquarters
-
48
Extrapolation of the Runaway Electron Benign Termination Scenario to ITER
D2 injection into mature runaway electron (RE) beams is found to enable access to a benign termination scenario that can mitigate MA-level RE currents without measurable wall heating. This result is enabled by the excitation of large and sudden MHD events (dB/B ~ 5%) that are found to promptly disperse the entire RE population over a large wetted area, with MHD accelerated by a recombined background plasma [1,2]. Fast RE loss timescales (<< ms) also prevent magnetic to kinetic energy conversion. We review benign termination phenomenology with supporting published data and focus on extrapolation to ITER, specifically: 1) the required D2 or H2 injection to recombine the background plasma, 2) vertical displacement event evolution and MHD instability access; 3) the required wetted area enhancement to disperse the kinetic energy; 4) the impact of the increased avalanche gain. Using the DINA code, we find that high current ITER RE beams should robustly access edge q of 3 & 2, where instability is expected. Using the MARS-F code, we find that the large-scale dispersal of RE kinetic energy is expected if dB/B is large, as was found in existing experiments [3]. The large avalanche gain in ITER poses a severe challenge, likely requiring multiple cycles of the benign loss to fully terminate a high current RE beam. [1] C. Paz-Soldan et al PPCF 2019 & NF 2021 [2] C. Reux et al PRL 2021, [3] Y. Liu et al, NF 2022
Work supported by the US DOE under DE-FC02-04ER54698, DE-SC0020299, DE-SC0022270.
Speaker: Prof. Carlos Paz-Soldan (Columbia University) -
49
Self-Consistent Kinetic Simulations of Runaway Electron Termination Schemes
Benign termination of mega-ampere level runaway current has been
convincingly demonstrated in JET [1] and DIII-D [2],
establishing it as a leading candidate for runaway mitigation on
ITER. This comes in the form of a runaway flush by parallel streaming
loss along stochastic magnetic field lines formed by global
magnetohydrodynamic (MHD) instabilities, which are found to correlate with a
low-Z (deuterium) injection that purges the high-Z impurities from a
post-thermal-quench plasma. After the runaway flush, there are two
scenarios determining whether the runaway current reforms.The scenario of no-runaway-reconstitution is enabled by (1) high
degree of high-Z impurity purge by massive deuterium injection and (2)
limited assimilation of deuterium in the post-purge plasma. Both
conditions are conducive to reduce radiative losses that would
have otherwise prevented electron reheating by Ohmic dissipation of
the plasma current, which is critical to establish and sustain a
parallel electric field below the runaway avalanche threshold.
Although electron reheating is most efficient at vanishing impurity
density and low deuterium density, the current ITER target for current
quench duration demands effective electron cooling to cap the electron
temperature to a few tens of eV. This translates into a limit on how
high the deuterium density must be, the exact magnitude of which is
controlled by the residue impurity content.In the likely scenario of insufficient electron reheating, the parallel
electric field stays above the runaway avalanche threshold and runaway
current can be reconstituted. Here the quantity of interest is the
plasma current drop before the Ohmic-to-runaway conversion is
completed. This critically depends on runaway seeding, which is
dramatically different for a post-flush plasma. Specifically, the
trapped ``runaways'' dominate the seeding for runaway reconstitution,
in sharp contrast to that in a post-thermal-quench plasma where much
lower energy hot tail, Dreicer mechanism, and tritium decay/Compton
scattering are at play. The high-Z impurities previously injected to
mitigate the thermal quench can greatly enhance such a trapped "runaway"
population, which remains well confined in the presence of stochastic
magnetic fields, and thus survives the 3D MHD flush. Furthermore,
their relatively high energy, around 1 MeV, ensures a low collisional
detrapping rate while the magnetic surfaces reheal at low electron
temperature. Most interestingly, the incomplete purge of high-Z
impurities helps drain the seed but produces a more efficient
avalanche, two of which compete to produce a 2-3~MA step in current
drop before runaway reconstitution of the plasma current.The different post-flush scenarios and the corresponding plasma
parameter regimes demarcated here, help place the existing experiments
in perspective in relation to ITER requirements. The present work will
also elucidate strategies through which the phase space distribution
of runaways during the current plateau can be tailored to minimize
this trapped population of electrons, along with determining critical
parameters for expediting the decay of the remnant electron population
before the magnetic flux surfaces are able to reform. This work was
supported by DOE OFES and OASCR.[1] Reux et al., Phys. Rev. Lett. (2021), [2] Paz-Soldan et al. Nucl. Fusion (2021)
Speaker: Christopher McDevitt (University of Florida) -
50
Mitigation of runaway electrons with a passive 3D coil
A 3D coil, which would be passively driven by the current quench (CQ) loop voltage during a disruption, has been incorporated into the design of SPARC, a high-field tokamak under development by Commonwealth Fusion Systems, and a similar coil is planned for installation on DIII-D. The effects of each of these runaway electron mitigation coils (REMC) on magnetic flux surfaces and runaway electron (RE) confinement have been modeled with the 3D MHD code NIMROD. For SPARC, an n=1 coil geometry was found to be much more effective at promoting RE losses than n=2 or n=3 coil designs. Using the ASCOT5 code to calculate RE advection and diffusion coefficients and the DREAM code to calculate runaway current evolution, the n=1 coil, with a maximum current of 590kA (~7 % of the flattop Ip), was found to fully suppress the runaway current with only the coil-induced MHD activity during the CQ accounted for. When the coil current is clamped at only 250kA, a runaway current plateau develops. Inclusion of the thermal quench (TQ) MHD was found to produce significantly faster losses of REs at early times, but also to modify the current profile such that flux surface reformation occurs earlier in time and allow avalanching to begin. The large avalanche multiplication factor in SPARC necessitates extremely good deconfinement to prevent a RE plateau. The effects of the close ideally-conducting wall were also studied in SPARC disruption simulations without the REMC and found to stabilize the MHD modes responsible for RE deconfinement. This suggests that when the REMC simulations are extended to a more realistic wall model, more optimistic results for the coil operation may be obtained.
In DIII-D CQ modeling with both inner-wall-limited (IWL) and lower-single-null (LSN) diverted initial conditions, a majority of RE test-particles were lost with maximum coil currents of 100 or 200kA (~5-10% of the flattop Ip). The IWL case was found to be insensitive to the coil current amplitude, with the evolution of the q-profile playing a dominant role in determining the timing and extent of island overlap. While the retained fractions of test-particles in the DIII-D simulations is larger than in the SPARC modeling, a measurable effect on the final RE current is easily expected because of the much lower avalanche multiplier in DIII-D plasmas. The importance of the q-profile evolution, which determines resonance with the coil, is made clear from simulations of both devices. Predicting this more accurately will also require moving beyond the ideal-wall boundary condition. Ongoing work is focused on NIMROD simulations with a resistive wall.
This work was supported in part by each of the following: Commonwealth Fusion Systems; DOE under Award Number DE-FG02-95ER54309; Swedish Research Council (Dnr. 2018-03911).
Speaker: Valerie Izzo (Fiat Lux)
-
48
-
15:35
Coffee Council Room Lobby
Council Room Lobby
ITER Headquarters
-
Mitigation Council Room
Council Room
ITER Headquarters
-
51
Characterization of the ASDEX Upgrade shattered pellet injector
A highly flexible shattered pellet injection (SPI) system was installed in December 2021 on ASDEX Upgrade. This system provides three independent guide tubes (GTs) and the pellet velocity can be varied between 100-900 m/s depending on pellet size and composition. At the end of each GT, different shatter heads had been installed for characterization of the disruption behaviour for different shatter distributions.
In this talk we will discuss the results of the laboratory characterization, during which over 1300 different pellets were fired and their post-shatter shard distributions analysed. In total, 10 different shatter head geometries were compared and the fragment distributions were analyzed with the help of fragment tracking algorithms. Figure 1 shows a comparison between two pellets fired into a shatter head with a rectangular and circular cross-section, respectively.
Our results also show that there is a significant discrepancy (for the largest and smallest fragments) between measured distributions and those predicted by the currently employed pellet break-up models. We are going to share our experience with pellet creation and launching for in total over 1900 pellets and also first results of the ongoing ASDEX Upgrade SPI campaign.Acknowledgements:
We would like to thank I. Vinyar for fruitful discussions.
Speaker: Mr Paul Heinrich (Max-Planck-Institut für Plasmaphysik) -
52
Dual SPI experiments with Identical pellets on DIII-D
Shattered Pellet Injection (SPI) has been chosen as the baseline disruption mitigation system for ITER. However, many questions remain regarding its operation, particularly under the presently envisaged operating scenario where several simultaneous and staggered SPIs are needed to inject high-Z radiating impurities prior to the thermal quench. Experiments on DIII-D used two SPIs with pellets of equal composition (~200 torr-L of pure Ne each) simultaneously or with a slight time delay between injections. Simultaneous injection exhibits a reduction in the pre-thermal quench time (time from when SPI fragments reach the plasma edge until the start of the thermal quench), relative to similar single SPI mitigated shutdowns. Despite the decreased time to assimilate the injected impurities, radial density measurements have similar increases in electron density while a vertical array shows a much faster electron density increase in the plasma core for simultaneous SPIs. Total radiated energy during the thermal quench, determined through summing radiated energy at three toroidal locations, and the current quench (CQ) duration are approximately the same for single or simultaneous injection. Additionally, fast visible camera images and analysis of impurity radiation from fast bolometer fan arrays show the injected impurities spread primarily in the parallel direction, away from the injection location in two distinct regions, corresponding to each of the SPIs. This separation of radiative zones suggests a lower radiation peaking factor, which is a promising result towards the success of the massively parallel ITER SPI system.
*Supported by the US DOE under DE-AC05-00OR22725, DE-FG02-07ER54917, and DE-FC02-04ER54698.
This manuscript has been authored in part by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).
Speaker: Jeffrey Herfindal (UsOakRidge) -
53
Benign Termination of Runaway Electron Beams on ASDEX Upgrade and TCV
Runaway electrons (RE) pose a significant threat for ITER and a mitigation technique is yet to be validated. Preliminary experiments on DIII-D and JET showed that reduced damage to the vessel can be achieved by accessing MHD instabilities that lead to spreading of the RE beam impact area [Paz-Soldan 2019, Reux 2021]. These instabilities were attained by “flushing” impurities and recombining the companion plasma, thus reducing the resistance and electric field (E), and sustaining beams until a low edge safety factor (qa) was achieved. Regeneration of the REs did not occur during the current quench due to the low E. This approach was successfully applied on ASDEX Upgrade and TCV this year. On-going experiments have been exploring the possible operational space and the underling physics involved and this paper will report the preliminary findings and outlook of this program.
RE beams were generated via massive gas injection (MGI) of argon on ASDEX Upgrade and neon on TCV. Secondary injections of D2 to flush impurities and recombine the companion plasma were executed with combinations of MGI, fueling pellets and fueling valves. Experiments on TCV showed that neutral pressure was more directly linked to plasma recombination than injected quantity. This was due to pumping effects that become non-negligible with beam durations in the order of seconds. It was found that neutral pressures of 0.10-0.15 Pa were required on both machines to recombine companion plasmas with RE currents of 120-600 kA. This study was further extended on TCV through variations in injected impurity gas quantity and the use of H2 instead of D2 as the secondary injection gas. It was found that higher neutral pressures were required to recombine the plasma when higher neon quantities were injected and no significant differences between H2 and D2 secondary injection were observed.
Fueling valves were used to maintain the neutral pressure required to prevent re-ionisation of the companion plasma and access a low E. This led to a record RE beam duration of ~4s on ASDEX Upgrade and the ability to increase the RE current on both machines. Reduction of qa was achieved via plasma current ramps on TCV and compression of the plasma onto the central column on both machines. Beam currents, current ramp rates and compression rates were scanned. Most beams terminated at a qa of ~2, with the exception of a 200kA beam on AUG, compressed over 500 ms, which produced a benign termination at a qa of ~3. Experiments compressing partially recombined plasmas have also begun on both machines and differences in wetted area and RE beam energy for flushed and unflushed plasmas have been measured on both machines. Preliminary results show significantly reduced heat fluxes to the inner wall and a reduction in beam energy during the recombined companion plasma phase.
Further experiments are planned on both machines for later this year. These experiments will focus on benign termination of partially recombined companion plasmas, scans of compression rates and variations in secondary injection gas species and concentration.
ImagesSpeaker: Umar Sheikh (SPC-EPFL) -
54
MIT Discussion
-
51
-
-
-
Mitigation Technology Council Room
Council Room
ITER Headquarters
chair: M. Lehnen
-
55
The ITER Disruption Mitigation System design progress and validation
The Disruption Mitigation System (DMS) is an essential plant system of ITER to reduce the deleterious effects of a disruption on in-vessel components. The design of the ITER DMS is based on shattered pellet injection (SPI) technology. This technique works by forming large cryogenic pellets of up to 28mm in diameter and firing them at high speeds towards a shattering section, where they fracture into showers of small fragments and sprayed into the plasma to provide high mass assimilation in the plasma as needed for the mitigation process.
The ITER DMS consists of 27 toroidally and poloidally distributed barrel type single shot pellet injectors. The majority of these injectors are located in three port cells on the equatorial level, while three injectors are installed in three port cells on the upper level. The integration of the injectors, which are located in the port cells inside the bio-shield where they are exposed to excessive neutron flux and radiation, had to overcome major challenges not only in terms of functionality but also for accessibility and maintainability. The port cells will be equipped with a large number of highly integrated components which provide the conditions and the monitoring needed to form and maintain pellets in the injectors.
In addition to these technical and integration solutions, the remaining technological challenges to make this first of a kind system highly effective and reliable are addressed by the ITER DMS task force through a number of dedicated technology studies. For example, the ideal de-sublimation process for protium which requires careful design and operating conditions, such as the thermal environment, the mass flow and the barrel pressure, is studied to form pellets with high integrity. Pellet launching methods are tested to dislodge and accelerate pellets intact, reliably and with minimum jitter. Novel propellant gas suppressors with internal structures are investigated to prevent the gas entering the torus ahead of the pellet fragments. An optical pellet diagnostic capable of withstanding the harsh tokamak environment is being developed to measure pellet parameters such as velocity and integrity. Flight line components are optimised to control the pellet flight path. In order to determine the ideal shattering geometry to achieve the desired fragment size distribution and velocities, laboratory experiments are carried out and the complicated fragmentation process is simulated using codes based on discrete element methods.
This presentation will describe the ITER DMS design and summarise the DMS task force technology activities to support and validate the design.
Speaker: Stefan Jachmich (ITER Organization) -
56
Modelling and simulation of the pellet shattering process related to the SPI technology for the ITER DMS
The ITER disruption mitigation system (DMS) is based on shattered pellet injection (SPI) technology. The principle of operation is to form and accelerate cylindrical cryogenic pellets to high speeds towards a shattering unit, where the pellets disintegrate into a shower of fragments of different sizes, which enter the plasma for the mitigation process. The effectiveness of this mitigation process is strongly dependent on the optimal fragment size distribution and velocities produced by the shattering process. In order to achieve these, it is important to know how the impact conditions, namely pellet material, velocity, impact angle, orientation etc., influence the fragmentation characteristics. The ITER DMS task force has launched a program to characterize and study the fragmentation experimentally.
As part of this program, Fraunhofer EMI is developing numerical models and computer codes to simulate and analyze the complex fragmentation process within the framework of an ongoing research project. The above-mentioned experiments serve, among other things, to calibrate and validate the developed models and procedures. The validated models will then be used to significantly extend the experimentally determined database. The goal is to optimize the shatter unit design as well as to derive guidelines for optimized impact conditions in order to get the desired fragment characteristics.
Our presentation gives an insight into the modelling approach, which combines elements from the peridynamic theory with the discrete element method into a software to simulate cracks and fracture of brittle materials such as cryogenic pellets. For exemplary cases we show the current status of results from simulated pellet shattering processes, and compare them to experimental data. Furthermore, we present how synthetic diagnostics, fragment detection and optimization methods are used to systematically perform the model calibration and validation process.Speaker: Pascal Matura (Fraunhofer Institute for High-Speed Dynamics, Ernst-Mach-Institut, EMI) -
57
Pellet production and shattering studies in the ITER DMS support laboratory
The ITER disruption mitigation support laboratory is part of the ITER Disruption Mitigation System (DMS) Task Force programme to establish the physics and technology basis for the ITER DMS. The laboratory is located at the Centre for Energy Research, Budapest Hungary.
The aims include production, launching and shattering of 19x39 mm and 28.5x57 mm (d x L) H, D, Ne and mixture pellets, and diagnosing the fragment plumes resulting from shattering.The pellets are desublimated in a stainless steel barrel cooled by liquid Helium evaporation to 5-12 Kelvin. The pellet production process can be monitored in detail by a camera looking along the barrel. The gas for the pellet is supplied through a precooler by a fully computer-controlled gas system capable of controlling the barrel pressure and/or gas feed rate using different algorithms. The finished pellets are launched by a propellant gas pulse injected by a specially built valve operating at up to 150 bar pressure and opening in 1.5 ms. After launch the pellet is diagnosed by two orthogonally viewing fast cameras. The propellant gas expands in a large vacuum volume and the pellet flies along a 60 mm diameter flight tube to about 4 m to the shattering head. This distance is identical to the injector-shattering head distance in the ITER SPI design. The shattering process is observed by a fast camera. The resulting fragments are diagnosed by a double laser curtain diagnostic. Each of them consists of a line laser illuminating a plane roughly perpendicular to the flight direction. To cameras view close to parallel to the flight direction and observe light scattered by the fragments. The two laser curtains operate at different wavelengths, therefore the cameras can operate in parallel and reveal the fragment size, velocity and flight direction information.
Up to now 19 mm diameter H, D, Ne and H-Ne mixture pellets were produced. At 5 K when the barrel pressure at the start of the pellet production process is raised to 50 mbar with a rate of about 10 mbar/min a thin dark layer (presumably snow-like) forms on the pellet-barrel interface. This enables launching of all pellet types, even Neon, with maximum 100 bar valve pressure, thus a punch mechanism is not needed to dislodge the pellet. The desublimation process for H and D pellets takes about 20-25 minutes, for Neon up to 45 minutes. These times agree well with results of a 2D pellet desublimation model.
It was found that all intact pellets reach the shattering head. The fragment clould was found to be highly asymmetric, about 3 times wider along the shattering plate then across it.
The results confirm that 19 mm H, D, and Ne pellets are suitable for the ITER SPI system. Experiments with the final 28.5 mm pellet diameter will start in June 2022.
Speaker: Dr Sandor Zoletnik (Centre for Energy Research) -
58
Inductively-Driven Electromagnetic micro-pellet Injector for injection of solid impurities into a Tokamak: Design, Development and Results from ADITYA-U
Micro-particle injectors used for quenching a burning plasma in case of a disruptive instability have predominantly relied on well-established pneumatic drives. Pneumatic drives are limited by the slow thermal velocity of the propellent gas molecules. Response time is also limited due to mechanical valves present in the gas-feed system. Such methods of acceleration may be therefore unsuitable for shutting down the plasma discharge with large plasma volume requiring a short warning time scale of less than 10 milliseconds. An electromagnetic pellet injector is expected to overcome these limitations easily as higher velocities can be achieved and can meet short warning timescales –both of which are important on large-sized, fusion grade reactors.
This talk will report the design and development of an Inductively Driven Pellet Accelerator and Injector or IPI -- a technological innovation and advancement providing an alternate means of injection of solid particles. The device has been uniquely engineered combining the fundamental principles of electromagnetic acceleration, pulsed-power technology and principles of impact & fracture mechanics to achieve acceleration and separation of granular pellets from a cartridge and their injection into a system. The driving forces (Lorentz Forces) are generated in a contactless way by pulsed magnetic fields of one or more electromagnets and currents that are induced on a cartridge located within the magnet. The device may therefore be considered as a pulsed, synchronous version of a linear induction motor (for those familiar with Electrical Technology) or adaptation of “Coilgun” - a tubular induction Electromagnetic Launcher - with several distinct novelties. At the core of the invention is a unique design of a cartridge that undergoes in-situ fracture to release the pellets on-the-fly. The stop-and-rupture mechanism is optimally designed to achieve a certain degree of control over the mean velocity of pellets after their release from the cartridge. The device is adapted to operate at conditions ranging from atmospheric pressure to high/ultra-high vacuum; can accommodate wide range of quantity, size and material of pellets; has a magazine that can inject upto six cartridges without interrupting the vacuum or opening the system; can vary the pellets’ velocity over a coarse and fine range and has a modular design such that the maximum velocity can be increased by increasing the number of modules. The IPI therefore brings to table a certain kind of versatility not commonly attributed to contemporary injectors. Higher velocities and short warning time scales can be achieved with IPI. It is now possible to inject micro-granular pellets obviating the need for shattering for easy dispersion/ablation. This paper will discuss a set of experiments from ADITYA-U where pellets of Li2TiO3/ Li2CO3 particles have been injected with a mean velocity of 200 m/s and a thermal quench followed by a rapid termination of plasma current has been achieved within 6 ms. A diverse range of experiments can now be performed and controlled with relative ease to explore the mass, material, velocity of pellets to be used for heat and energetic particle mitigation during disruption.Speaker: Dr Sambaran Pahari (Bhabha Atomic Researh Centre)
-
55
-
10:40
Coffee Council Room Lobby
Council Room Lobby
ITER Headquarters
-
Mitigation Technology Council Room
Council Room
ITER Headquarters
chair: M. Lehnen
-
59
Hydrogen Pellet Formation, Release, Survivability, and Shattering Research Supporting the ITER SPI System
Large hydrogen pellets will be the primary injected material for the ITER disruption mitigation (DM) system, based on the shattered pellet injection (SPI) technique. Shattered pellet injection utilizes cryogenic cooling to desublimate gas into the barrel of a pipe gun, forming a solid cylindrical pellet. SPI systems deployed worldwide usually operate with deuterium as the primary pellet material so the issues pertaining to utilizing much larger hydrogen pellets for SPI have not been encountered until recently. The physical properties: material strength, heat of sublimation, and thermal conductivity of hydrogen are less favorable than deuterium for large pellets (>16 mm in diameter) resulting in limitations on the formation duration and release of large intact hydrogen pellets. The current method of releasing pellets from a barrel consists of using a fast-operating valve to deliver a high-pressure pulse of gas to the rear of the pellet. The resulting warm gas pulse delivers a large force, which can have deleterious effects on a large hydrogen pellet.
A series of experiments were conducted to study the formation and release requirements for various sizes of hydrogen pellets. Experiments showed that 20 mm diameter pellets are approaching the largest hydrogen pellets that can be released intact using a pulse of high-pressure propellant gas. Pellets with diameters of 28.5 and 23.5 mm were formed and catastrophically fractured during the release process, independent of propellant pressure and amount of propellant gas delivered (within relevant conditions). Release of intact pellets is essential for the reliability of ITER’s DM system. To better understand the survivability and fragmentation mechanics of hydrogen, low angle impact tests were conducted using 10 and 20 mm H2 pellets impacting a flat angled plate to determine the highest possible impact velocity these pellets can survive, known as the fracture threshold velocity. This measured parameter is utilized in the statistical fragmentation model to determine the ratio of impact energy versus the minimum impact energy survivable [1]. Since the use of hydrogen as the base pellet material for SPI on ITER is a relatively new development, it was not included in the original studies dedicated to measuring and understanding the survivability limits and fragmentation.[1] T. E. Gebhart et al., IEEE Trans. Plas. Sci. (2020)
Speaker: Dr Trey Gebhart (Oak Ridge National Laboratory) -
60
MIT Technology Discussion
-
59
-
General Discussion & Closure Council Room
Council Room
ITER Headquarters
-
61
General Discussion
-
62
Closure
-
61
-
12:45
Lunch (optional)
-
