Search

Mitigation of deleterious heat flux from edge-localized modes (ELMs) on fusion reactors is often attempted with 3D perturbations of the confining magnetic fields. However, the established technique of resonant magnetic perturbations (RMPs) also degrades plasma performance, complicating implementation on future fusion reactors. In this paper, we introduce an adaptive real-time control scheme as a viable approach to simultaneously achieve both ELM-free states and recovered high-confinement (βN~1.91$ and HN~0.9), demonstrating successful handling of a volatile complex system through adaptive measures. We show that, by exploiting a salient hysteresis process to adaptively minimize the RMP strength, stable ELM suppression can be achieved while actively encouraging confinement recovery. This is made possible by a self-organized transport response in the plasma edge which reinforces the confinement improvement through a widening of the ion pedestal and promotes control stability, in contrast to the deteriorating effect on performance observed in standard RMP experiments. These results establish the real-time approach as an up-and-coming solution towards an optimized ELM-free state, which is an important step for the operation of ITER and reactor-grade tokamak plasmas. Notably, the real-time adaptive control scheme introduced here provides a path towards economic fusion reactors by maximizing the fusion gain while minimizing damage to machine components.

A mode identification method for slowly rotating (or non-rotating) n (toroidal mode number) = 1 plasma instabilities has been newly established with the signal compensation method used for an external time-varying magnetic field. The mode identification method is based on the Fourier decomposition scheme and processes the signal of the magnetic probes (MPs) on the passive stabilizer in the Korea Superconducting Tokamak Advanced Research (KSTAR) tokamak. To exclude the non-plasma magnetic perturbation by the external current coils and their secondary eddy currents on the passive stabilizer that may not be easily characterized, the ARX-SISO (autoregressive with exogenous terms-single input single output) method has been introduced in the signal compensation process. Preliminary off-line analysis presented herein confirms that this method can adequately detect the time evolution of the mode-locking, along with a slowly rotating state, even in the presence of external magnetic perturbations.

The plasma-driven inductive voltages were not negligible and should be taken into account for the reliable operation of the voltage-based quench detection of the KSTAR CS coil. For the active compensation of such plasma-induced voltage noise, plasma parameters should be measured in real-time and applied to the quench detection system. The selection of plasma parameters by considering their contribution to the noise is of primary importance. A code, which can estimate the induced coil voltage driven by plasma, is developed by using a simplified plasma response modeling. The numerical analysis has showed that there is an apparent influence from the plasma, while the influence from eddy currents induced in passive structures is negligible. The change of plasma current is the primary noise source and the radial and vertical motions of plasma have an effect, although they are small. The performance of the quench detector is developed and tested during the 2017 KSTAR campaign. The noise rejection performance is discussed with various experimental results and numerical estimations.

Fast vertical control of shaped plasmas is essential for the successful realization of plasma operation near the limit of vertical control to achieve maximum confinement and stable, disruption free operation. In KSTAR, the normal conducting in-vessel vertical control coil (IVC coil) is employed to respond to vertical transients much faster than the superconducting coils are capable. The power supply for the IVC coil is capable of responding in a time commensurate with the expected vertical growth rates on KSTAR. The diagnostics used for the fast vertical control since the first operation with shaped plasmas in KSTAR, however, have a low signal-to-noise ratio which limits the gains that can be used successfully in the proportional and derivative control loops and thus the speed of the control loop. The successful use of relative flux for the Z-position estimate and of the loop voltage difference from a pair of up-down symmetric loops to provide sensitive and less noisy vertical estimate will be discussed. In addition, the control loop for the IVC coil is decoupled from the slow motion controlled by the superconducting coils using a high-pass filter in the control software. Finally the noticeable improvement in plasma control for plasmas near the vertical stability limit will be discussed.

Predictive 3D optimization reveals a novel approach to modify a nonaxisymmetric magnetic perturbation to be entirely harmless for tokamaks, by essentially restoring quasisymmetry in perturbed particle orbits as much as possible. Such a quasisymmetric magnetic perturbation (QSMP) has been designed and successfully tested in the KSTAR and DIII-D tokamaks, demonstrating no performance degradation despite the large overall amplitudes of nonaxisymmetric fields and strong response otherwise expected in the tested plasmas. The results indicate that a quasisymmetric optimization is a robust path of error field correction across the resonant and nonresonant field spectrum in a tokamak, leveraging the prevailing concept of quasisymmetry for general 3D plasma confinement systems such as stellarators. The optimization becomes, in fact, a simple eigenvalue problem to the so-called torque response matrices if a perturbed equilibrium is calculated consistent with nonaxisymmetric neoclassical transport.

Advanced operation scenarios such as high poloidal beta (β P) or high q min are promising concepts to achieve the steady-state high-performance fusion plasmas. However, those scenarios are prone to substantial Alfvénic activity, causing fast-ion transport and losses. Recent experiments with the advanced operation scenario on KSTAR tokamak have shown that the electron cyclotron current drive (ECCD) is able to mitigate and suppress the beam-ion driven toroidal Alfvén eigenmodes (TAEs) for over several tens of global energy confinement time. Co-current directional intermediate off-axis ECCD lowers the central safety factor slightly and tilts the central q-profile shape so that the continuum damping in the core region increases. Besides, the rise of central plasma pressure and increased thermal-ion Landau damping contribute to TAE stabilization. While the TAEs are suppressed, neutron emission rate and total stored energy increase by approximately 45% and 25%, respectively. Fast-ion transport estimated by TRANSP calculations approaches the classical level during the TAE suppression period. Substantial reduction in fast-ion loss and neutron deficit is also observed. Enhancement of fast-ion confinement by suppressing the TAEs leads to an increase of non-inductive current fraction and will benefit the sustainment of the long-pulse high-performance discharges.

A small relaxation of the axisymmetric magnetic field of a tokamak into a non-axisymmetric three-dimensional (3D) configuration can be effective to control magnetohydrodynamic instabilities, such as edge-localized modes. However, a major challenge to the concept of 3D tokamaks is that there are virtually unlimited possible choices for a 3D magnetic field, and most of them will only destabilize or degrade plasmas by symmetry breaking. Here, we demonstrate the phase-space visualization of the full 3D field-operating windows of a tokamak, which allows us to predict which configurations will maintain high confinement without magnetohydrodynamic instabilities in an entire region of plasmas. We test our approach at the Korean Superconducting Tokamak Advanced Research (KSTAR) facility, whose 3D coils with many degrees of freedom in the coil space make it unique for this purpose. Our experiments show that only a small subset of coil configurations can accomplish edge-localized mode suppression without terminating the discharge with core magnetohydrodynamic instabilities, as predicted by the perturbative 3D expansion of plasma equilibrium and the optimizing principle of local resonance. The prediction provided excellent guidance, implying that our method can substantially improve the efficiency and fidelity of the 3D optimization process in tokamaks. A theoretical and numerical approach, validated by experiments at the KSTAR facility, shows how magnetohydrodynamic instabilities in tokamak plasmas can be efficiently controlled by a small relaxation of the confining field into a 3D configuration.

Korea Superconducting Tokamak Advanced Research (KSTAR) program is strongly focused on solving the scientific and technological issues in steady-state high performance plasma operation in preparation for ITER operation as well as the design basis for DEMO. In this regards, KSTAR has made significant advances in developing long pulse and high performance plasma scenarios utilizing the advantage of the fully superconducting tokamak. Ten-year of KSTAR operation showed the outstanding progress in the plasma control extending the operation window of the plasma discharges achieving the H-mode up to 1 MA in plasma current, up to 72 s in flat top duration, and up to 2.16 in elongation. In addition to the long pulse discharge, high performance discharges with high betas (βN ~ 3) could be achieved in the broad range of edge safety factor (q95) without external error field correction. The unique features of the KSTAR device (magnetic accuracy with extremely low error fields, steady-state capable heating systems, in-vessel control coils, and advanced imaging and profile diagnostics) has been fully exploited to explore the unveiled physics as well as to exploring the systematic solution for suppression of edge localized mode (ELM) crash. Achieved examples are the record long pulse of H-mode operation without an ELM crash (~ 30 s up to date), and progress in the fundamental transport physics through systematic study using these unique capabilities. Based on the previous research results, intensive research will be followed to explore the advanced high beta operation (βN ~ 4) with fully suppressed harmful MHD instabilities aiming the integrated solution for DEMO. In this regards, an additional current drive systems and in-vessel structures will be upgraded.

A newly developed plasma response model, combining the nonlinear two-fluid MHD code TM1 and toroidal MHD code GPEC run in ideal mode, quantitatively predicts the narrow isolated q95 windows (Δq95 ∼ 0.1) of edge-localized mode (ELM) suppression by n = 1, 2, and 3 resonant magnetic perturbations (RMPs) in both DIII-D and KSTAR tokamaks across a wide range of plasma parameters. The key physics that unites both experimental observations and our simulations is the close alignment of essential resonant q-surfaces and the location of the top of the pedestal prior to an ELM. This alignment permits an applied RMP to produce field penetration due to the lower E × B rotation at the pedestal top rather than being screened. The model successfully predicts that narrow magnetic islands form when resonant field penetration occurs at the top of pedestal, and these islands are easily screened when q95 moves off resonance, leading to very narrow windows of ELM suppression (typically Δq95 ∼ 0.1). Furthermore, the observed reduction in the pedestal height is also well captured by the calculated classical collisional transport across the island. We recover observed q95, βN and plasma shape dependence of ELM suppression due to the effect of magnetic islands on pedestal transport and peeling-ballooning-mode stability. Importantly, experiments do occasionally observe wide windows of ELM suppression (Δq95 > 0.5). Our model reveals that at low pedestal-top density multiple islands open, leading to wide operational windows of ELM suppression consistent with experiment. The model indicates that wide q95 windows of ELM suppression can be achieved at substantially higher pedestal pressure with less confinement degradation in DIII-D by operating at higher toroidal mode number (n = 4) RMPs. This can have significant implications for the operation of the ITER ELM control coils for maintaining high confinement together with ELM suppression.

A polychrometer-type motional Stark effect (MSE) diagnostic technique, originally developed for the Alcator C-Mod tokamak, has been extended and applied to the Korea Superconducting Advanced Tokamak Research (KSTAR) device, the long-pulse superconducting tokamak, for the first time. It demonstrates a successful in situ subtraction of the polarized reflections off the vacuum vessel wall, sometimes up to half the total signal in some sightlines. To avoid the secondary neutral beam emission that may contaminate conventional beam-into-gas calibrations, a new approach, where the beam-into-gas measurements are made at various torus pressures with fixed vacuum fields, has been devised, which is possible with the stable superconducting coil systems of KSTAR. The validity of this new calibration scheme has been checked via plasma jog experiments. The experimental evidence of the polarized background light and the necessity of its correction in the MSE measurements made in KSTAR are presented as well.

A newly upgraded IVCC (In-Vessel Control Coil) system equipped with four broadband power supplies, along with a current connection patch panel, are introduced with a discussion of their capabilities on various KSTAR experiments. Until the 2014 KSTAR experimental campaign, the non-axisymmetric field configuration could not be changed in a shot, let alone the limited number of accessible configurations. With the installation of the new power supplies, such restrictions have been greatly reduced. Based on the 2015 KSTAR run-campaign, this new system was confirmed to easily cope with various dynamic demands for toroidal and poloidal phases of the 3D magnetic field in a shot. With newly equipped magnetic sensors, this enables us to extend the operational options and further explore the 3D physics for tokamak plasmas.

Excellent reproducibility of KSTAR resonant magnetic perturbation (RMP)-driven, edge-localized mode (ELM)-crash suppression enables us to construct a database reliably for the study of ELM-crash suppression conditions. To establish a high-fidelity database, we have selected one of the frequently used RMP configurations in KSTAR, n = 1, + 90 ° phasing RMP-coil configuration. A series of fitting curves based on edge profile diagnostics data is used for normalized electron collisionality ( ν e *) and plasma toroidal rotation ( V tor) at pedestal top. Since ITER is expected to employ slowly rotating, low-collisionality, high-density plasmas whose conditions are not readily accessible in the existing devices, the exploration and understanding of these two parameters would be important not only for RMP-driven, ELM-crash-suppression physics but also for the success of ITER. The data points for the ELM-crash suppression phase are in the range of 0.2 40 km/s. Suppression thresholds or boundaries in ν e , ped * or V tor , ped are not confirmed in the investigated parameter ranges so far. The KSTAR database still needs additional experimental datasets in ITER-relevant conditions ( ν e , ped * ∼ 0.1 and low-torque low-rotation) to confirm the boundary of the ELM-crash suppression window in KSTAR and reduce the uncertainties of the RMP ELM-crash control technique in ITER. In both ν e , ped * and V tor , ped parameters, the phase-space distribution of the ELM-crash suppression has no clear distinction from that of the ELM-crash mitigation. Linear discriminant analysis provides a linear combination of parameters relevant to ELM-crash suppression, best separating two data categories. Recursive feature elimination indicates that n e , ped and I RMP, as well as ν e , ped * and V tor , ped, are critical variables in the separation of the data groups.

We report the status of hybrid scenario experiments in Korea Superconducting Tokamak Advanced Research (KSTAR). The hybrid scenario is defined as stationary discharges with B N >= 2.4 and H_89 >= 2.0 at q_95 < 6.5 without or with very mild sawtooth activities in KSTAR. It is being developed towards reactor-relevant conditions. High performance of B N

An overview of advances and progress on the KSTAR plasma control improvement is given. The enhancement of the KSTAR plasma control system (PCS) continues in order to enable implementations of more sophisticated control algorithms and capabilities of integrated controls on magnets, gas, heating and off-normal event handling. Features and functionalities on the generic magnetic controls, kinetic controls and new fault scheme design are added for the operation space expansion, mainly regarding recent achievements of high plasma current up to 1.1 MA for 10 s, high-performance long pulse up to 90 s, and discharges relevant to ITER research. Technical challenges and prospective upgrades in the near future are described.

We use a newly developed simulation tool to numerically study fast-ion loading on plasma-facing components (PFCs) at the Korea Superconducting Tokamak Advanced Research facility in the high-poloidal-β plasma operation regime. The new code can calculate neutral beam ionization and follow the guiding center orbit of ionized particles. The results of the simulation indicate that fast ions ionized in the high-field side drift out and strike the PFCs as they rotate poloidally. Momentum projection onto a phase space defined by canonical toroidal angular momentum and magnetic moment leads to a simple criterion to avoid fast-ion loading on poloidal limiters (PLs), which are PFCs that are toroidally localized on the low-field side. Control of fast-ion loss is examined by varying the plasma current and plasma boundary. A larger plasma current and inner shifting of the outer plasma boundary is predicted to substantially reduce fast-ion loading on the PLs. The proposed phase-space criterion is qualitatively consistent with the results of the simulation of fast-ion control.

KSTAR has demonstrated divertor heat flux broadening during edge-localized mode (ELM) crash suppression using ITER-like three-row resonant magnetic perturbation (RMP) for the first time. To address a couple of critical issues in ITER RMP, robust ELM-crash-suppression methodology has been explored at low q 95 and established in KSTAR using low-n RMPs. Taking full advantage of the ITER-like three-row in-vessel control coils (IVCC) in KSTAR, a set of intentionally misaligned RMP configurations (IMC) was tested to investigate whether or not IMC could be compatible with ELM crash suppression, while minimizing electromagnetic loads on RMP coils. As a result, the ITER-like three-row IMCs were found not only to have been compatible with the ELM crash suppression, but also to have broadened the divertor heat fluxes in the vicinity of the outer strike point. In comparison, the two-row RMPs have rarely affected the near scrape-off layer (SOL) heat flux despite slightly broadened profile change in the far-SOL. Since the divertor heat flux broadening reflects the dispersal of the peaked near-SOL heat flux, the experimental outcome is quite favorable to the ITER choice of three rows, instead of two rows. Nonetheless, since the IMC-driven broadening observed in the attached plasmas in KSTAR might appear substantially different in the partially detached plasmas in ITER, additional investigation has been conducted to see if RMP-driven, ELM crash suppression could be compatible with detached plasmas. Although no detached plasmas have been identified with ELM crash suppression yet, significantly reduced divertor heat flux was confirmed in high density, ELM-crash-suppressed plasmas at q 95 = 3.4 using n = 2 RMPs. These new findings elevate the confidence level about the ITER RMP system, while the remaining uncertainties need to be further clarified using the three-row IVCCs in KSTAR.

A decade long operation of KSTAR has contributed significantly to operation of superconducting tokamak device and advancement of tokamak physics which will be beneficial for the ITER and K-DEMO programs. Even with limited heating capability, various conventional as well as new operating regimes have been explored and achieved great performance. As examples, a long pulse H-mode operation with and without the ELM-crash was well over 60 seconds and over 30 seconds, respectively. Unique capabilities of KSTAR allowed to improve control capability of harmful instabilities and have been instrumental for many exciting new physics. The intricate IVCC system has been a great perturbation tool in studying of threshold power of the L/H transition in the presence of non-axisymmetric fields, rotation physics due to NTV effects, heat dispersal in divertor system and predictive control of the ELM-crash with a priori modeling. The state-of-the-art 2D/3D microwave imaging systems uncovered many new exciting new physics in the MHD and turbulence transport physics. They are validation of q0agt;1 right after the sawtooth crash, self-consistent asymmetric distribution of plasma turbulence amplitude and flow in the presence of 2/1 island, and underlying physics of low frequency turbulence induced by 3D resonant fields in suppression of the ELM-crash through non-linear interaction with the ELMs. In the turbulence area, non-diffusive "avalanche" transport event and role of quiescent coherent mode in confinement are studied. To accommodate anticipating higher performance of the KSTAR plasmas with the increased heating powers (NBI and ECH), new divertor/internal interface with full active cooling system will be implemented after the increased heating power systems and new current drive systems are fully tested. Upgrade plan for the internal hardware and efficient current drive system may allow a long pulse operation of higher performance plasmas at βNagt;3.0 with fbs~0.5 and Tiagt;10keV. a#13; a#13

Stainless steel coupons with cavity structure were installed at different poloidal locations in KSTAR during 2009 campaign to study deposition and H/D retention inside the gap of PFCs. The deposition profiles at different poloidal locations indicate that the incoming species have high surface loss probability (b) and they are probably charge exchange neutral particles. The (H + D)/C ratios of layers inside the coupons were in a range from 0.04 to 0.37 at different poloidal locations which might be caused by different contribution of different hydrocarbon species. Campaign integrated, time averaged net C, H/D atom flux density towards outer walls are in the similar range with other carbon dominated machines.

The Korea Superconducting Tokamak Advanced Research (KSTAR) device is an advanced superconducting tokamak to establish scientific and technological bases for an attractive fusion reactor [1]. In-vessel vertical coil (IVC) power supply (PS) is developed for vertical stability and shape control of KSTAR plasma. IVC PS in KSTAR is based on a single-phase full-bridge inverter. The output voltage and current are dc ±1 kV and ±10 kA, respectively. Insulated gate bipolar transistor is used as switching power device in IVC PS and maximal 5 kHz of switching frequency is adopted to meet the requirement of fast control by plasma control system. This paper describes the configuration and engineering of the IVC PS system and analyzes both the IVC coil commissioning and the major performances that D-shape plasma achieved in experiments.

The effect of static $n=1$ resonant magnetic perturbation (RMP) on the spatial structure and temporal dynamics of edge-localized modes (ELMs) and edge turbulence in tokamak plasma has been investigated. Two-dimensional images measured by a millimeter-wave camera on the KSTAR tokamak revealed that the coherent filamentary modes (i.e., ELMs) are still present in the edge region when the usual large scale collapse of the edge confinement, i.e., the ELM crash, is completely suppressed by $n=1$ RMP. Cross-correlation analyses on the 2D images show that (1) the RMP enhances turbulent fluctuations in the edge toward the ELM-crash-suppression phase, (2) the induced turbulence has a clear dispersion relation for wide ranges of wave number and frequency, and (3) the turbulence involves a net radially outward energy transport. Nonlinear interactions of the turbulent eddies with the coexisting ELMs are clearly observed by bispectral analysis, which implies that the exchange of energy between them may be the key to the prevention of large scale crashes.

One of key objectives of tokamak experiments is the exploration of enhanced confinement regimes, and the access of the internal transport barrier (ITB) formation is dealt with an important physics issue in the most of major tokamaks. Also, the advanced tokamak scenario with ITB is expected to lead to a continuous reactor with high fusion power density. From that point of view, the formation of the ITB in KSTAR which is designed for long pulse operation capability is very important although its heating and current drive systems are not fully equipped yet. We have therefore assumed that an early injection of the full NBI power (~5.5 MW) during the current ramp-up would give a chance to form an internal barrier if the plasma could stay in the L-mode. To avoid the H-mode transition, we have produced inboard limited plasmas with detaching from the both upper and lower divertors. Using this approach, an ITB formation during L-mode has been observed which shows improved core confinement. Ion and electron temperature profiles show the barrier clearly in the temperature, and it was sustained for about 7 s in the dedicated experiment. This is the first stationary ITB observed in a full superconducting tokamak. This operation scenario with the ITB could be an alternative way to achieve a high performance regime in KSTAR, and the length of the ITB discharge could be extended even longer. In this paper, we present the formation of the ITB using measured and simulated characteristic profiles.

Toroidal rotation braking by neoclassical toroidal viscosity driven by non-axisymmetric (3D) magnetic fields, called magnetic braking, has great potential to control rotation profile, and thereby modify tokamak stability and performance. In order to characterize magnetic braking in the various 3D field configurations, dedicated experiments have been carried out in KSTAR, applying a variety of static , 3D fields of different phasing of , 0, and . Resonant-type magnetic braking was achieved by phasing fields, accompanied by strong density pump-out and confinement degradation, and explained by excitation of kink response captured by ideal plasma response calculation. Strong resonant plasma response was also observed under phasing at , leading to severe confinement degradation and eventual disruption by locked modes. Such a strong resonant transport was substantially modified to non-resonant-type transport at higher , as the resonant particle transport was significantly reduced and the rotation braking was pushed to plasma edge. This is well explained by ideal perturbed equilibrium calculations indicating the strong kink coupling at lower is reduced at higher discharge. The 0 phasing fields achieved quiescent magnetic braking without density pump-out and confinement degradation, which is consistent with vacuum and ideal plasma response analysis predicting deeply penetrating 3D fields without an excitation of strong kink response.

H-mode plasma operation of the Korea Superconducting Tokamak Advanced Research (KSTAR) device has been expanded to significantly surpass the ideal MHD no-wall beta limit. Plasmas with high normalized beta, βN, up to 4.3 have been achieved with reduced plasma internal inductance, li, to near 0.7, exceeding the computed n = 1 ideal no-wall limit by a factor of 1.6. Pulse lengths at maximum βN were extended to longer pulses by new, more rapid control. The stability of the observed m/n = 2/1 tearing mode that limited the achieved high βN is computed by the M3D-C1 code, and the effect of sheared toroidal rotation to tearing stability is examined. As a method to affect the mode stability in high βN plasmas, the non-resonant alteration of the rotation profile by non-axisymmetric magnetic fields has been used, enabling a study of the underlying neoclassical toroidal viscosity (NTV) physics and stability dependence on rotation. Non-axisymmetric field spectra were applied using in-vessel control coils (IVCCs) with ...

Measuring rotation profiles with a reliable spatial resolution is one of the critical diagnostics in understanding the plasma behavior especially for the edge transport. In the KSTAR experiments, it has been consistently observed from the charge exchange spectroscopy measurements that the magnetic perturbations not only suppresses edge localized modes (ELMs) but also reduces toroidal rotations. In this paper, toroidal velocities of the carbon impurity and their profile evolutions during ELMy and ELM-suppressed phases are presented. The rotation profiles are shown to collapse immediately after an ELM burst and continue to build up until the next burst that accompanies another collapse. Toroidal rotations following the resonant magnetic perturbations applications are observed to be reduced along with the ELMs suppressed.

The variation of a magnetic braking profile by non-axisymmetric magnetic fields has been experimentally demonstrated and numerically validated in the KSTAR tokamak. Two types of n = 2 non-resonant magnetic fields were applied by changing the relative phase of non-axisymmetric field coils. One is even parity, of which non-resonant fields deeply penetrate into the plasma core, and the other is odd parity localized at the plasma edge. The even and odd parity produced significantly different perturbed magnetic field structures, and thereby drove global and edge-dominant toroidal rotation damping, respectively. These distinct braking profiles are consistently reproduced by drift-kinetic particle simulations, indicating the possibility of the predictive utilization of non-resonant magnetic fields for rotation profile control.

Neural network is considered as a parameter estimation tool in plasma controls for next generation tokamak such as ITER. The neural network has been reported to be so accurate and fast for plasma equilibrium identification that it may be applied to the control of complex tokamak plasmas. For this application, the reliability of the conventional neural network needs to be improved. In this study, a new idea of double neural network is developed to achieve this. The new idea has been applied to simple plasma position identification of KSTAR tokamak for feasibility test. Characteristics of the concept show higher reliability and fault tolerance even in severe faulty conditions, which may make neural network applicable to plasma control reliably and widely in future tokamaks.

The change of tokamak plasma behavior by supersonic molecular beam injection (SMBI) was investigated by applying a three-dimensional magnetic perturbation that could suppress edge localized modes (ELMs). From the time trace of decreasing electron temperature and with increasing plasma density keeping the total confined energy constant, the SMBI seems to act as a cold pulse on the plasma. However, the ELM behaviors were changed drastically (i.e., the symptom of ELM suppression has disappeared). The plasma collisionality in the edge-pedestal region could play a role in the change of the ELM behaviors.

The fast ion dynamics and the associated heat load on the plasma facing components in the KSTAR tokamak were investigated with the orbit following Monte-Carlo (OFMC) code in several magnetic field configurations and realistic wall geometry. In particular, attention was paid to the effect of resonant magnetic perturbation (RMP) fields. Both the vacuum field approximation as well as the self-consistent field that includes the response of a stationary plasma were considered. In both cases, the magnetic perturbation (MP) is dominated by the toroidal mode number n = 1, but otherwise its structure is strongly affected by the plasma response. The loss of fast ions increased significantly when the MP field was applied. Most loss particles hit the poloidal limiter structure around the outer mid-plane on the low field side, but the distribution of heat loads across the three limiters varied with the form of the MP. Short-timescale loss of supposedly well-confined co-passing fast ions was also observed. These losses started within a few poloidal transits after the fast ion was born deep inside the plasma on the high-field side of the magnetic axis. In the configuration studied, these losses are facilitated by the combination of two factors: (i) the large magnetic drift of fast ions across a wide range of magnetic surfaces due to a low plasma current, and (ii) resonant interactions between the fast ions and magnetic islands that were induced inside the plasma by the external RMP field. These effects are expected to play an important role in present-day tokamaks.

The Korea Superconducting Tokamak Advanced Research (KSTAR) device is an advanced superconducting tokamak to establish scientific and technological bases for an attractive fusion reactor [1]. In-vessel Vertical Coil (IVC) power supply has been developed for vertical stability and shape control of KSTAR plasma. IVC power supply in KSTAR is based on a single-phase full-bridge inverter. The output voltage is DC ±1kV and output current is ±10kA. IGBT is used as switching power device in IVC power supply and maximal 5kHz of switching frequency is adopted to meet the requirement of fast control by PCS. This paper describes the configuration and engineering of the IVC power supply system and analyzes both the IVC coil commissioning and the D-shape plasma major performances achieved by experiments.

© 2015 IAEA, Vienna Printed in the UK. The structure and evolution of the ion temperature (Ti) and toroidal rotation (V-) profile have been investigated in neutral beam injection (NBI)-heated KSTAR H-mode plasmas, both without and with resonant magnetic pertubations (RMPs). A clear disparity between the width of the ∇Vφpedestal and that of the ∇Ti-pedestal was observed. Also, it was found that there exists a close correlation and weak relative hysteresis between the pedestal ∇Vφand ∇Ti during both L →H and H →L transitions. During the L →H transition, the Vφpedestal is observed to form ahead of the Ti-pedestal, and build inward from the separatrix. Linear gyrokinetic stability analysis of these KSTAR profiles was performed. The results indicate that parallel velocity shear is a relevant drive for pedestal turbulence and transport. This was largely ignored in previous studies of the pedestal micro-stability. Pedestal ion temperature and rotation profiles were also measured during edge localized mode (ELM) suppression experiments on KSTAR using an n = 1 RMPs. It was found that the top values of the ion temperature and toroidal rotation pedestal drop with RMPs when ELMs are suppressed.

Summary form only given, as follows. Summary form only given. In-vessel coils are to be used for the fast plasma position control, field error correction (FEC), and resistive wall mode (RWM) feedback stabilization in the Korea Superconducting Tokamak Advanced Research (KSTAR) device. Recently, a new configuration that incorporates toroidal segmentation concept, has been adopted. The new coil system is found to allow a wider range of plasma control flexibility satisfying the KSTAR advanced physics requirements for the plasma position and FEC/RWM control capability, in addition to engineering advantages. With the modified coil structure, plasma position control has been simulated by using a simple linear /sup 1/RZIP (R, Z, I/sub p/) model. The RZIP model is developed as a simple circuit equation representing a linear plasma response model. This model can be easily constructed without solving plasma, equilibrium and is more explicit in the quantities that define the plasma response. By applying step response simulation and random perturbation simulation for the vertical position control, PID controller gain and power supply limit are compared with those of the previous axisymmetric internal control coils. Results from the simple model are benchmarked with the /sup 2/TSC simulation for axisymmetric internal control coils.

Summary form only given, as follows. Tokamak equilibrium analyses have been routinely carried out by solving the Grad-Shafranov equation with various conventional numerical techniques. As a new approach for the equilibrium analysis a neural network equilibrium solver has been proposed from the functional approximation capabilities of the artificial neural network. The neural network equilibrium solver does not have to perform finite differences and coordinate transformations, so its manipulation becomes easy and simple. It may also save computing time significantly for smoothly evolving equilibria from the equilibrium information of the previous time step. Moreover, the implicitly embedded interpolation function of the neural network solver can make it applicable to real-time equilibrium analysis with the aid of on-line learning algorithms. With comprehensive structural variance and manipulation, characteristics of the neural network solver - the structural dependency on external source profile variations, the number of reasonable iterations for the stable convergence, and computational speed etc. - will be presented and elucidated in view of real-time equilibrium analyses.

We evaluate the characteristics of global energy confinement in KSTAR (?E,KSTAR) quantitatively in three ways; firstly by comparing it with multi-machine scalings, secondly by deriving multiple regression equations for the L- and the H-mode plasmas, respectively, and lastly by comparing confinement enhancement of the H-mode phase with respect to the L-mode phase in each discharge defined as Hexp. The KSTAR database exhibits ?E,KSTAR of ?0.04 to ?0.16?s and of ?0.06 to ?0.19?s in L-mode and in H-mode plasmas, respectively. The multiple regression equations derived by statistical analysis present the similar dependency on PL and higher dependency on Ip compared with the multi-machine scalings, however the dependency on ? in both L- and H-mode plasmas draw the negative power dependency of ??0.68 and ??0.76 for H-mode and for L-mode database, respectively on the contrary to the positive dependency in all multi-machine empirical scalings. It is found that the energy confinement of both L-mode and H-mode of the discharges with Hexp?>?1.5 can be well-predicted by multi-machine scalings, ?E,89L and ?E,92H. Apart from this, the H-mode confinement with 1.5?Hexp?

We show that strong edge magnetic shear, which is generic to divertor plasmas, makes the E × B shearing of turbulence eddys in toroidal geometry more effective. From calculations of the E × B shearing rates for KSTAR edge parameters, we conclude that the enhanced magnetic shear at the diverted KSTAR plasma edge facilitates E × B shear suppression of turbulence and ensuing H-mode transition. (Some figures may appear in colour only in the online journal)

Initial H-mode operation of the Korea Superconducting Tokamak Advanced Research (KSTAR) is expanded to higher normalized beta and lower plasma internal inductance moving towards design target operation. As a key supporting device for ITER, an important goal for KSTAR is to produce physics understanding of MHD instabilities at long pulse with steady-state profiles, at high normalized beta, and over a wide range of plasma rotation profiles. An advance from initial plasma operation is a significant increase in plasma stored energy and normalized beta, with Wtot = 340 kJ, βN = 1.9, which is 75% of the level required to reach the computed ideal n = 1 no-wall stability limit. The internal inductance was lowered to 0.9 at sustained H-mode duration up to 5 s. In ohmically heated plasmas, the plasma current reached 1 MA with prolonged pulse length up to 12 s. Rotating MHD modes are observed in the device with perturbations having tearing rather than ideal parity. Modes with m/n = 3/2 are triggered during the H-mode phase but are relatively weak and do not substantially reduce Wtot. In contrast, 2/1 modes to date only appear when the plasma rotation profiles are lowered after H–L back-transition. Subsequent 2/1 mode locking creates a repetitive collapse of βN by more than 50%. Onset behaviour suggests the 3/2 mode is close to being neoclassically unstable. A correlation between the 2/1 mode amplitude and local rotation shear from an x-ray imaging crystal spectrometer suggests that the rotation shear at the mode rational surface is stabilizing. As a method to access the ITER-relevant low plasma rotation regime, plasma rotation alteration by n = 1, 2 applied fields and associated neoclassical toroidal viscosity (NTV) induced torque is presently investigated. The net rotation profile change measured by a charge exchange recombination diagnostic with proper compensation of plasma boundary movement shows initial evidence of non-resonant rotation damping by the n = 1, 2 applied field configurations. The result addresses perspective on access to low rotation regimes for MHD instability studies applicable to ITER. Computation of active RWM control using the VALEN-3D code examines control performance using midplane locked mode detection sensors. The LM sensors are found to be strongly affected by mode and control coil-induced vessel current, and consequently lead to limited control performance theoretically.

The latest results of confinement and edge-localized mode (ELM) characteristics of Korea Superconducting Tokamak Advanced Research (KSTAR) H-mode plasmas are reported. The estimation of fast ion contribution to the total stored energy, calculated by both the NUBEAM and ASTRA simulations, and of the effective total heating power is used to derive the thermal energy confinement time (τ E,thermal), which is compared with a multi-machine database. The measured power threshold for the L–H transition (P thr) as a function of density shows a roll-over with minimum value at . KSTAR H-mode plasmas exhibit three distinctive types of ELMs: large type-I ELMs, intermediate ELMs and a mixed (type-I and small ELM peaks) ELM regime. Power scans show that the frequency of the large ELMs increases with increasing heating power, a feature of type-I ELMs. The quality of confinement is higher for type-I and mixed ELMy H-mode (H 98(y,2) ∼ 0.9–1) than for the intermediate ELM regime (H 98(y,2) ∼ 0.7). Type-I ELMs have precursor-like signals from the magnetics measurement, while the other two ELM types do not. The low-field side (LFS) profile of electron temperature (T e), from the ECE measurement, and the pedestal profile of the toroidal velocity (V t), from charge-exchange spectroscopy, show a continuous build up on the LFS during the inter-ELM period. However, the pedestal ion temperature (T i) remains unchanged for most of the inter-ELM period until it rapidly rises in the last stage of the ELM cycle (⩾70–80%). The estimated electron pedestal collisionality for a type-I ELMy regime is . The confinement and ELM characteristics for the ELM suppression discharges by the application of an n = 1 magnetic perturbation (MP) have also been investigated for each of the identified stages during the MP application. A second L–H transition during the L-mode phase after the end of first H-mode stage occurs for some discharges when the divertor configuration is restored by the plasma control system. Characteristics of this late H-mode are compared with those for the main H-mode.

A time series of static nonlinear ferromagnetic calculations was performed to mimic the time-dependent modelling of plasma start-up by assessing the effects of the ferromagnetic Incoloy 908 used in the superconducting coil jackets of the Korea Superconducting Tokamak Advanced Research (KSTAR) device. Time-series calculations of a two-dimensional axisymmetric circuit model with nonlinear ferromagnetic effects enabled us to find appropriate waveforms for the KSTAR poloidal field coil currents that satisfied various start-up requirements, such as the formation and sustainment of field nulls, a sufficient amount of magnetic flux for further plasma current ramp-up, sufficiently large Et ?Bt/B? > 1?kV?m?1 contours for successful breakdown, plasma current toroidal equilibria, etc. In addition to the aforementioned requirements, the results introduced in this report also provided the positional stability of the plasma current channel against radial as well as vertical perturbations by compensating the field deformation originating from the ferromagnetic effects. With the improved positional stability, robust plasma start-up was achieved during the 2010 KSTAR campaign under various discharge conditions such as the recovery process from plasma disruptions.

A scheme to obtain relatively scaled profiles of electron cyclotron emission (ECE) temperature directly from uncalibrated raw radiometer data is proposed and has been tested for the 2009 campaign KSTAR plasmas. The proposed method utilizes a position controlled system to move the plasma adiabatically and compares ECE radiometer channels at the same relative radial positions assuming the profile consistency during the adiabatic change. This scaling method is an alternative solution when an absolute calibration is unreliable or too time consuming. The application to the two dimensional ECE imaging data, wherein calibration is extremely difficult, may also prove to be useful.

Numerical simulations are carried out to investigate the applicability of resonant magnetic perturbation (RMP) to KSTAR plasmas for a possible control of edge localized mode (ELM) to suppress or mitigate its damages to divertor materials. For the verification of the feasibility of RMP application, magnetic island configurations, resonant normal fields, magnetic island widths and Chirikov parameters are calculated for two types of KSTAR operation scenarios: steady state and hybrid. Field error correction (FEC) coils in KSTAR are considered to produce externally perturbed magnetic fields for RMP, and the directions of coil currents determine the toroidal mode n and the parity (even or odd). The RMP configurations are described by vacuum superposition of the equilibrium magnetic fields and the perturbed ones induced by FEC coils. The numerical simulations for n = 2 toroidal mode in both operation scenarios show that when the pitches of the equilibrium and perturbed magnetic fields are well aligned, magnetic islands are formed for a series of m poloidal modes and the adjacent islands are overlapped to generate a stochastic layer in the edge region. Even parity turns out to be more effective in making the magnetic islands overlapped to become stochastic field lines in the steady-state operation, while odd parity in the hybrid operation. The formation of the stochastic layer is verified by the calculated Chirikov parameters, which also give basic information on the current requirement of FEC coils. Additionally, lobe structures of stochastic field lines are found in the edge region extended to the divertor plate in the hybrid scenario. Based on the standard vacuum criteria for RMP, the simulation results indicate that the FEC coils will be feasible for control of ELMs and mitigation of divertor heat load by RMP in both steady-state and hybrid operation scenarios.

To improve inherent shortcomings of statistical methods and apply them to the extraction of plasma equilibrium parameters in a fast timescale for real-time plasma control, new concepts of statistical methods such as principal component analysis-basedneural network (NN), functional parametrization (FP)-based NN and double network are introduced by modifying NN and FP. These new methods are benchmarked and compared with the conventional techniques of NN and FP in a simple single-filament system. As a result of their applications to identification of plasma equilibrium parameters in the Korea SuperconductingTokamak Advanced Research tokamak, particularly, the double network concept among them has successfully achieved the improvement of drawbacks in the conventional methods. It is shown that more reliable results from the double network method can be obtained by combining several different statistical treatments as a primary network. Even in the case of nonoptimized methods united as a primary network, quite acceptable results can be achieved in the double network method.