High-beta Disruption Research Articles

High-beta disruption prediction study on HL-2A with instance-based transfer learning

Unmitigated disruptions pose a much more serious threat when large-scale tokamaks are operating in the high performance regime. Machine learning based disruption predictors can exhibit impressive performance. However, their effectiveness is based on a substantial amount of training data. In future reactors, obtaining a substantial amount of disruption data in high performance regimes without risking damage to the machine is highly improbable. Using machine learning to develop disruption predictors on data from the low performance regime and transfer them to the high performance regime is an effective solution for a large reactor-sized tokamak like ITER and beyond. In this study, a number of models are trained using different subsets of data from the HL-2A tokamak experiment. A SHapley Additive exPlanations (SHAP) analysis is executed on the models, revealing that there are different, even contradicting, patterns between different performance regimes. Thus, simply mixing data among different performance regimes will not yield optimal results. Based on this analysis, we propose an instance-based transfer learning technique which trains the model using a dataset generated with an optimized strategy. The strategy involves instance and feature selection based on the physics behind differences in high- and low-performance discharges, as revealed by SHAP model analysis. The TrAdaBoost technique significantly improved the model performance from 0.78 BA (balanced accuracy) to 0.86 BA with a few high-performance operation data.

Likelihood Identification of High-Beta Disruption in JT-60U

Likelihood Identification of High-Beta Disruption in JT-60U

Prediction of high-beta disruptions in JT-60U based on sparse modeling using exhaustive search

Disruption is a critical phenomenon in a tokamak reactor. Although disruption causes serious damage to the reactor, its physical mechanism remains unclear. To realize a tokamak reactor, it is necessary to understand and control the disruption phenomenon. The present research constructs a disruption predictor using experimental high-beta plasma data in the JT-60U tokamak. The predictor was constructed using a support vector machine as a linear discriminant, and we focus on a variable selection problem for the binary classification by sparse modeling, specifically, exhaustively searching the best combinations of variables which maximize the predictor performance. By the sparse modeling, we found that the six input parameters as the best combinations. The selected parameters were the n = 1 mode amplitude |Brn=1| and its time derivative d|Brn=1|/dt, the plasma density (relative to the Greenwald density limit) and its time derivative, and the time derivatives of the plasma internal inductance and plasma elongation. In particular, it was identified that the parameter d|Brn=1|/dt, plays a key role on plasma disruption. We should notice that the combination with other plasma parameters is indispensable and remarkably make it possible to improve the performance of disruption prediction.

Multi-Wavelength Imaging of Solar Plasma - High-Beta Disruption Model of Solar Flares -

Solar atmosphere is filled with plasma and magnetic field. Activities in the atmosphere are due to plasma instabilities in the magnetic field. To understand the physical mechanisms of activities / instabilities, it is necessary to know the physical conditions of magnetized plasma, such as temperature, density, magnetic field, and their spatial structures and temporal developments. Multi-wavelength imaging is essential for this purpose. Imaging observations of the Sun at microwave, X-ray, EUV and optical ranges are routinely going on. Due to free exchange of original data among solar physics and related field communities, we can easily combine images covering wide range of spectrum. Even under such circumstances, we still do not understand the cause of activities in the solar atmosphere well. The current standard model of solar activities is based on magnetic reconnection: release of stored magnetic energy by reconnection is the cause of solar activities on the Sun such as solar flares. However, recent X-ray, EUV and microwave observations with high spatial and temporal resolution show that dense plasma is involved in activities from the beginning. Based on these observations, I propose a high-beta model of solar activities, which is very similar to high-beta disruptions in magnetically confined fusion experiments.

Proposal of Steady State Superconducting Spherical Tokamak Experiment

This paper presents a proposal of a spherical tokamak (ST) with the superconducting center solenoid and toroidal coil system having the total magnetic stored energy comparable to that of the Large Helical Device. The parameters are as follows: R=1.8-2.1 m, Bt0=2-2.4 T, Ip=8-12 MA (IOH=4-6 MA), Te0=8-12 keV, ne0-4x1020 m-3 and Q(equivalent)=0.3-infinite depending on the transport scaling. This machine is suitable to study the role of the aspect ratio in the physics (cf. the stability condition of the high beta disruption) of the steady state plasma, which is driven by the bootstrap current.

A new solar flare scenario: - High-beta plasma disruption -

The solar corona is believed to be generally low beta. In a low-beta plasma, energy is mainly stored as magnetic energy or current. To dissipate the stored magnetic energy in the highly conductive corona, anomalous resistivity is needed. Activities of outer layers of X-ray loops in solar flares have been interpreted as the result of reconnection of magnetic field due to anomalous resistivity in a current sheet located above the flaring loop.However, recent high-cadence EUV observations show many flares without opening up of magnetic field. This means that existence of current sheets above flaring loops is not a necessary element for flares. It is also found for many flares that low temperature prominence plasma are heated in the very early phase of lifting. These observational facts require a new solar flare scenario different from reconnection.Recent multi-wavelength high-spatial and -temporal resolution observations show that flares originate in the lower atmosphere. We can expect high beta plasma and small curvature magnetic loops in the lower atmosphere. It is also observed high-velocity flows along magnetic loops associated with flares. Bounded motions of charged particles (both thermal motion and flow) along the magnetic lines of force create the centrifugal force. They exceed gravity force under certain conditions (small curvature and high velocity). They can even exceed magnetic tension force depending on beta values of the plasma in the loop.In the case of the upward acceleration (centrifugal - gravity ) plasma in a loop, whose ends are anchored at the photosphere, is unstable against the localized interchange (BALLOONING) mode. It can develop into non-linear phase and explosive phenomena are expected. This is called HIGH-BETA DISRUPTION. The high-beta disruption has many common characteristics to solar flares. Various phenomena related to solar flares such as, high-energy particle acceleration, plasma turbulence, plasma ejections, plasma heating, are interpred as the result of high-beta disruption. High spatial and temporal resolution observations of flares at various wavelengths are used to see the actual instability process. Shibasaki (2001)To search for other articles by the author(s) go to: http://adsabs.harvard.edu/abstract_service.html

Spectrum of Global Ideal-Magnetohydrodynamic Three-Dimensional Ballooning Modes

The class of pressure-driven plasma instabilities known as ballooning modes may be responsible for such diverse phenomena as high-beta disruptions in tokamaks, solar flares and magnetospheric substorms. In this paper the theory of the spectrum of unstable eigenvalues of the linearized ideal magnetohydrodynamic (MHD) equations of motion in non-axisymmetric toroidal equilibria is sketched, comparing and contrasting systems with open field lines and systems with toroidally confined field lines. The need to regularize ideal MHD to keep the wavenumber finite, and the relevance of quantum chaos theory to understand the structure of the spectrum, is pointed out.

High‐Beta Disruption in the Solar Atmosphere

The outer layers of X-ray loops in solar flares are known to be more active: they are hotter than the lower loops and above the loop-top hard X-ray sources are formed there. These phenomena are interpreted as the result of the reconnection above the loop, which converts magnetic energy into thermal and nonthermal energy of plasma. However, little direct evidence for the reconnection has been presented so far. This paper interprets the activity in the outer layer of flaring loops on a different scenario. Coronal loops filled with hot and dense plasma (high beta) or with fast plasma flow, surrounded by the low-beta corona, are unstable at their outer boundary, where the curvature is convex outward and the density gradient is inward. The centrifugal force acting upward on the plasma in the loop can exceed that of gravity. This condition is favorable for localized interchange instability called ballooning instability, and the plasma in the loop is ejected when the instability has developed into a nonlinear phase (high-beta disruption). This is a natural consequence of the high-beta (and/or the high-velocity) plasma confined in a curved coronal loop. The high-beta disruption has many elements common to solar flares. In this paper, importance of the high-beta plasma is stressed, which is a neglected part of the solar activity so far.

Bursts of electron cyclotron emission during ELMs and high disruptions in TFTR

Large ELM events in TFTR are often accompanied by short, intense bursts of electron cyclotron emission. A unique combination of two fast grating polychromator instruments located at different toroidal positions is used to measure the emission and characterize these bursts, which exhibit strong toroidal asymmetries. Bursts are compared to those which occur early in the thermal quench phase of high beta disruptions, the only other time that bursts occur with comparable robustness. In these disruptions, bursts are localized to the vicinity of the ballooning mode, a medium toroidal mode number (n = 10 - 20) precursor, localized toroidally, poloidally, and radially, which triggers the disruption. Rapid loss of particles and sudden rapid cooling occurs. In both cases, bursting can be explained not in terms of excitation of enhanced emission but rather in the reduction of absorption of thermal emission. Bursting is consistent with a modification of the electron distribution function due to a rapid energy or particle exchange between hot electrons and cold electrons from the edge, momentarily reducing the velocity gradient of in the thermal region. This model is qualitatively and quantitatively consistent with experiment, predicting for example emission enhancement factors of .

Characterization of alpha particle loss during disruptions in TFTR during deuterium-tritium operation

The timing of alpha losses, with respect to the various phases of a disruption, and the impact location of the losses are characterized during high fusion power operation of TFTR with deuterium and tritium fuels. Characterization of alpha losses is important for the design of future fusion devices such as ITER. In addition, characterization of the alpha losses with respect to the disruption evolution helps in the understanding of the dynamics of the disruption process and related MHD events such as sawtooth crashes. Disruptions are characterized as having several standard phases, applicable to most disruptions in all tokamaks: precursor, thermal quench(es) and current quench. Most of the losses are observed to occur during the thermal quench phase(s). In high beta disruptions, alpha losses start abruptly during the growth of MHD precursors, just before the onset of the thermal quench. The initial burst of losses, lasting as little as 100 mu s, can release a significant fraction (one third) of the total disruption induced losses during the thermal quench. An inventory of alpha particles suggests that the alpha loss distribution during disruptions might be quite different from that expected during non-disruptive discharges. There are no obvious differences between fast fusion product losses during deuterium-deuterium (DD) and deuterium-tritium (DT) disruptions, aside from the large alpha loss component in DT discharges

Finite pressure effects on sawtooth oscillations

Nonlinear 3-D simulation of sawtooth oscillations shows that the helically twisted hot spot has a prominent toroidal bulge at the large major radius side, characteristic of a high-εβp instability. This bulge drives other m/n islands just outside the mixing radius resulting in a stochastic annular region. A similar effect is also found in shaped plasmas, even at low beta. These effects agree with experimental data, giving detailed support to the Kadomtsev reconnection model. This effect can also be used to explain the experimental sawtooth heat pulse ‘‘anomaly.’’ At high εβp, the stochastic region can essentially fill the whole plasma, and can result in a ‘‘high-beta disruption.’’