Session VO08: Magnetic Confinement: KSTAR Tokamak

Live

Recently, stationary ion internal transport barrier (ITB) discharges have been established in a diverted configuration with two or three neutral beam injection sources at q95\textasciitilde 4. To obtain the stable ion ITB with the diverted configuration, the H-mode transition was avoided by keeping a low density (\textasciitilde 1x10$^{\mathrm{19}})$ and upper single null configuration which are unfavorable to H-mode in KSTAR [1]. The performance and neutron rate of the discharge is comparable to hybrid discharges with $\beta_{\mathrm{N}}$ \textless 2.6, W$_{\mathrm{MHD\thinspace }}$\textless 340 kJ, H$_{\mathrm{89L\thinspace }}$\textless 2.3. Fully non-inductive current drive is obtained in addition to this high performance. This ITB discharges show high stability against the locked mode so to survive without disruption while with the locked mode. Moreover, the ion ITB could be re-formed following the natural mode unlocking. The locking threshold is investigated with the density scan. The ASTRA simulation coupled with the NUBEAM code was conducted for the transport analysis and the linear gyrokinetic analysis is performed and compared with fluctuation measurements to study the ITB dynamics. This scenario is suggested as a new advanced scenario for steady-state high performance with low divertor heat load at ITER-relevant collisionality which doesn't require a delicate profile feedback control. Keywords : Ion Internal Transport Barrier, Diverted Configuration Reference [1] S.W. Yoon et al. 2011 Nucl. Fusion 51 113009   

Live

Over the past several years, KSTAR has successfully performed the high beta steady-state experiments with beam heating at relatively low Greenwald fraction regime. This study reports the recent results on KSTAR high $\beta $N experiment results aiming long-pulse $\beta $N \textasciitilde 3. Owing to relatively long fast-ion slowing time characteristics, optimizing fast-ion confinement was the key reaching high beta operation at KSTAR. Based on database on fast-ion related MHD mode control experiments, it is found that the plasma performance degradation is often accompanied by Alfv\'{e}nic activities and tearing modes. Numerical investigation with TRANSP / NOVA / Kick-model performed, so that the impact of the mode on plasma performance is analyzed. The q profile of the highest $\beta $N is evaluated, and the future prospects are shown with near-term heating upgrade.   

Live

In addition to ELM mitigation, resonant magnetic perturbations (RMPs) may be used to control the fast particle population and possibly as a means of reducing the drive for other deleterious instabilities such as TAEs [1], through degraded orbit dynamics (e.g. bounce-tip stochastization). Earlier experiments were carried out on KSTAR with the application RMP fields on bursting TAE modes during early single-beam heated discharges [2]. New experiments have focused on examining the influence of the amplitude and phasing of n$=$1 fields in similar discharges with high frequency modes in the Alfven frequency range. Such fields are observed to cause a pronounced reduction in amplitude of these modes. However, early application of RMP fields can lead to substantial density pump out and rotation braking, both of which can modify the continuum structure, as well as the ionization rate of beam neutrals. Both indirect and direct effects will be assessed, with a range of fast ion diagnostics including neutron flux monitors. Initial simulations will focus on the continuum structure, 3D equilibrium and particle orbits. Results will help plan experiments on MAST-U. [1] M Garcia-Munoz et al 2019 PPCF 61 054007 [2] M.J. Hole et al PPCF 61 (2019) 025016   

Live

Characteristics of low-k (k$_{\mathrm{\theta }}$r$_{\mathrm{i}}$ \textless 0.5) electron temperature fluctuations has been studied in the edge region of ECH L-mode plasmas with different line averaged densities (n$_{\mathrm{e}})$. The spectral analysis of electron temperature fluctuations revealed the existence of dual modes having opposite phase velocities in the lab frame (v$_{\mathrm{L}})$. Modes with v$_{\mathrm{L}}$ \textless 0 (in the ion diamagnetic direction) are observed in the broad radial region (r/a $=$ 0.6---0.9), while modes with v$_{\mathrm{L}}$ \textgreater 0 (in the electron diamagnetic direction) are observed locally in the edge region (r/a \textasciitilde 0.9). The amplitude of the ion modes depends strongly on n$_{\mathrm{e}}$, i.e. it decreases as n$_{\mathrm{e}}$ increases with the additional gas puffing and becomes undetectably small when n$_{\mathrm{e}}$ is high (3.0e19 m$^{\mathrm{-3}})$, while the amplitude of the electron modes is affected little by n$_{\mathrm{e}}$. Therefore, in the edge region the ion modes are dominant when n$_{\mathrm{e}}$ is low (1.5e19 m$^{\mathrm{-3}})$ and vice versa. The co-existence of dual modes is observed in the L-mode edge with the intermediate n$_{\mathrm{e}}$ (1.5 \textless n$_{\mathrm{e}}$ \textless 2.0e19 m$^{\mathrm{-3}})$. This observation shows that the turbulence distribution in the ECH L-mode edge can change with n$_{\mathrm{e}}$ (or T$_{\mathrm{e}}$ since the ECH power was kept constant during the gas puffing).   

Live

Significant progress has been made in controlling edge-localized-modes (ELM)-crashes using resonant magnetic perturbation (RMP), preventing transiently excessive particle and heat fluxes from damaging plasma facing components and divertor. In that regard, an optimal RMP-driven, ELM control may be defined based on minimal thermal loading, given otherwise similar plasma conditions. Taking advantage of the ITER-like 3-row in-vessel control coils in KSTAR, a series of intentionally misaligned configurations (IMC) driven, ELM-crash-suppression have been explored in terms of divertor heat flux footprints [1]. So far, 3-row IMC-driven divertor heat flux broadening has not been observed on 2-row IMC attempts, even when the mid-row (that has stronger coupling than off-midplane rows) has been energized. Also, when the plasma response gets stronger during RMP-driven, ELM-crash-suppression, a much more peaked heat flux poses a higher risk on divertor. In a sense, RMP-driven, ELM-crash-mitigation might be more effective in diffusing the thermal loads rather than RMP-driven, ELM-crash-suppression. Overall, an optimal RMP-driven, ELM control may require us to consider multi-faceted aspects beyond RMP-driven, edge pedestal changes. [1] Y. In et al, Nucl. Fusion \textbf{59}, 126045 (2019)   

Live

KSTAR experiments showed the path dependence and hysteresis of plasma confinement and performance recovery: even for the same final perturbing 3D currents, starting with higher initial 3D currents leads to lower recovery down the path. This demonstrates the need for a control system to keep the ITER RMP perturbations close to the ELM suppression threshold at all times. To this aim, a comprehensive adaptive real-time (rt) ELM control system that exploits key properties of ELM physics, RMP ELM suppression physics, and set of diagnostic inputs to make real-time decisions about the control of multiple actuators to sustain ELM suppression / mitigation is implemented at KSTAR. MHD spectroscopy allows finding the plasma response to 3D coils which can then be used to calculate plasma stability. Historically, multiple sine waves with different frequencies in a feedforward fashion are applied which is time-consuming (\textasciitilde seconds) making this method unsuitable for rt-control. We implemented a novel feedback based MHD spectroscopy algorithm at KSTAR, the system uses the rt-measurements and changes the applied 3D perturbations as the plasma responds to these perturbations. This leads to fast and stable plasma response identification. We show the results of the initial tests of these controls in KSTAR experiments and discuss the applicability to ITER.   

Live

To fully suppress edge-localized-modes (ELM) via resonant magnetic perturbation (RMP) is essential to reach and sustain high-performance steady-state H-mode plasmas. Using the nonlinear 3D MHD code JOREK [1], we have simulated a recent RMP-driven ELM-crash-suppression in KSTAR. In this study, we have found that the pedestal degradation by RMP can be explained to some extent by the radial transport from the combined effects of the kink-peeling response, tearing response [2], and neoclassical toroidal viscosity (NTV) [3]. Interestingly, ELM-crash-suppression was not only attributable to the degraded pedestal but also to direct coupling between peeling-ballooning mode (PBM) and RMP-driven plasma response [4]. While the linear stability of PBM improved owing to the degraded pedestal, it was not a sole contributor to ELM-crash-suppression, in that the coupling between PBM and RMP increased the spectral transfer between edge harmonics preventing catastrophic growth and crash of unstable modes. In addition, the locking of PBMs has been numerically reproduced during the ELM suppression phase, which supports the experimentally observed importance of V$_{\mathrm{ExB}}\approx $0 on the onset of ELM-crash-suppression. [1] G. Huysmans et al., PPCF (2009) [2] F. Orain et al., Phys. Plasma (2019) [3] N. Logan et al., Phys. Plasma (2016) [4] M. Becoulet et al., PRL (2014)   

Live

Boron nitride (BN) injection into ELMy H-Mode KSTAR discharges with 500 kA \textless I$_{\mathrm{p}}$ \textless 700 kA and with up to 3 MW of neutral beam heating power demonstrated the ability to improve wall conditions, reducing recycling, over the course of several discharges in some regimes. These results suggest that powder injection may be a useful technique for intra-shot wall conditioning. The Impurity Powder Dropper (IPD) was used to deliver 60 $\mu $m BN powder at dose rates between 1 mg/s and 20 mg/s, and for various total durations, to quantify the effect of the powder on the plasma. In the KSTAR experiments, BN was dropped into 10 s duration discharges and was observed to gradually reduce the baseline D$\alpha $ level over six shots, indicating improved wall conditioning. Moreover, while the initial D$\alpha $ signals included mixed ELMs, the smaller amplitude ELMs disappeared after the powder injections. Previous results showed improved wall conditioning with pure boron powder, indicating that the wall conditioning effect of BN is due to its boron content. As the previous results also demonstrated ELM mitigation with BN, BN powder injection appears to offer the simultaneous benefits of ELM mitigation and recycling reduction.   

Live

The application of RMP is one promising way to control ELM, which is the potential challenge of tokamak reactors. On the other hand, the RMP can also drive the core response that could lead to devastating instabilities. A systematic approach can isolate the edge from core resonant fields for safe ELM suppression, and give edge-localized RMP. A robust feature of the edge-localized RMP is the curtailed response to the field at the LFS midplane, as opposed to typical RMPs which strongly resonate with the LFS fields. Sensitivity studies in various target plasmas justify this robust feature, for both n$=$1 and n$=$2 toroidal modes. This improved understanding of edge localized resonant RMPs has been then utilized to explore optimizations of the 3D coils. We propose an efficient way to probe the accessibility of RMP ELM suppression with given target plasmas and engineering constraints. The optimization proposes the unconventional 3D coil parameters in KSTAR implying that more intelligent RMP is possible. This new insight is also utilized for the design of ELM control coils to enhance the efficiency of our ELM suppression capabilities. Simple window-pane coils matching the edge-localized RMP structure can substantially expand in the ELM suppression window beyond the existing coil. Further optimization using the FOCUS code leads to additional enhancement in the edge-localized control \newpage   

Live

We report an investigation of the confinement of magnetic braking experiments in KSTAR. A set of discharges are developed injecting neutral beams (NBs) that supply strong toroidal torques to produce fast rotating H-mode plasmas. We utilize non-axisymmetric (3D) magnetic field to drive toroidal rotation braking and electron cyclotron heating (ECH) to explore the lowest rotation level. In those discharges, toroidal rotation over a range of 120-300km/s is achieved at the core depending on the combination of NB, 3D field, and ECH. Improved energy confinement triggered by 3D field driven toroidal rotation braking is observed, where increase of stored energy by up to 15% is achieved in spite of increased particle transport by 3D magnetic field. ECEI measurement shows that broadband turbulent fluctuations of ~200kHz near the pedestal are largely mitigated and suppressed in the improved confinement phase. FIDA measurement indicates improved fast ion confinement during the same period. Confinement data are collected for magnetic braking H-mode plasmas in the range of plasma parameters of B$_{T}$=1.6-1.8T, I$_{P}$=500-700kA, and q$_{95}$=3.7-5.4. The analysis result will be presented with a focus on the correlation between the confinement and the toroidal rotation.   

Live

Compatibility of divertor plasma detachment with application of resonant magnetic perturbations (RMPs) for control of edge localized modes (ELMs) is a key challenge for magnetic confinement fusion. The Korean Superconducting Tokamak Advanced Research (KSTAR) facility is equipped with a flexible set of perturbation coils that allow scanning of the operation space (phasing and amplitude of individual coil rows) and fine-tuning of the ELM control window for optimal heat load spreading. We discuss the modeling framework required for this endeavor: a 3D boundary plasma model (EMC3-EIRENE) linked with a magnetohydrodynamic plasma response model (GPEC, MARS-F, ...). We show that uncertainties in plasma response model parameters propagate to the boundary model resulting in significant changes to the magnetic footprint which determines heat loads. The boundary plasma model is extended to include low collisionality corrections to the classical parallel heat conduction (heat flux limit).   

On Demand

ITER adopts a strategy for evenly distributing radiated power during disruption mitigation and reduces the time to prepare pellets using multiple shattered pellet injections (SPI). However, as there was no device with a completely symmetric SPIs, sufficient studies have not been conducted on the effect of multiple injections. To confirm the feasibility of mitigation strategy in ITER, KSTAR installed two SPIs in toroidally opposite positions. We investigated the difference in the disruption mitigation by changing the arrival times of the two SPIs to assess the jitter effect among SPIs. The current quenching rate changes as the time difference changes from a few percent to tens of percent of the thermal quenching (TQ) duration. The results show that more energy can be released when multiple SPIs are injected simultaneously. For dual SPIs, the measured peak density is 1.2x10$^{\mathrm{21}}$ m$^{\mathrm{-3}}$ near the end of the TQ, which is almost twice the value of a single SPI. We will first focus on multiple injections of different toroidal locations as well as multiple barrel injections of the same poloidal and toroidal location. Through this, we plan to provide data that is the basis of ITER DMS design.   

On Demand

Toroidal rotation from pure ohmic discharges without any external momentum sources is one of the most fundamental types of self-generated intrinsic rotation for magnetic fusion researches. There have been reported wide ranges of magnitude, direction and abrupt rotation reversal for ohmic toroidal rotation studies, no clear physical mechanisms are concluded to explain these intrinsic ohmic rotation behaviors. The core ohmic toroidal rotation at the early plasma current ramp-up phase has been measured mostly in the counter-current direction from KSTAR. The possible explanation for the counter-current rotation is speculated from the momentum transfer between neutrals and plasma particles. The calculated toroidal rotation based on the momentum transfer agrees well with the experimental measurements. The core ohmic toroidal rotation at the plasma current flat-top phase is well fitted with the ion temperature divided by plasma current from KSTAR. Recently, we expanded the ohmic rotation scaling to the co-current direction utilizing lower electron density regimes. In this presentation, we will report the extended scaling and analysis results for ohmic L-mode plasmas in KSTAR This work was supported by the Korea Ministry of Science and ICT under the KSTAR project contracts.   

On Demand

In the frame of the ITER Disruption Mitigation Task Force program, a new Dispersion Interferometer (DI) system on KSTAR is developed and installed for the Shattered Pellet Injection (SPI) experiments. The DI is essential for assessing the efficiency of multiple SPI to raise the density to values sufficient for avoiding runaway electron formation in ITER. Since the SPI induced density rise is 20 times higher than the routine electron density in KSTAR, an interferometer with shorter wavelength is required. The dispersion interferometer can avoid the vibrational noise at short wavelength. A vertical 3-channel DI system using 1064 nm base laser is designed and installed. The detailed description of the system including the second harmonic generation, the modulation, the detection, and the vibration isolation will be presented along with the first result from the test channel, which shows a fast rise of the electron density up to 7x10$^{\mathrm{21}}$/ m$^{\mathrm{3}}$ in 1 msec during SPI.   

KSTAR Overview

The recent progress of the KSTAR tokamak is summarized. First of all, new advanced scenarios were developed targeting steady-state operation with the fine tuning of diverting and heating and significant progress in shape control capability. The stationary internal transport barrier is successfully reproduced with comparable confinement as H-mode level both in limited and USN configuration and a low q$_{\mathrm{min}}$ scenario is developed based on early diverting and delayed core heating scheme. Recent 3D experiments have focused on several ITER-relevant issues, such as divertor heat flux broadening in resonant magnetic perturbations on ELM-crash suppression, RMP-driven ELM-crash-suppression on ITER-like low q$_{\mathrm{95\thinspace }}$(\textasciitilde 3.2-3.4) and the characterization of ELM-crash suppression window in terms of normalized electron collisionality and plasma toroidal rotation at pedestal top. Cross-validation between the advanced diagnostics and the modeling provides new insight on the basic transport process, i.e., non-diffusive avalanche-like electron heat transport events are observed by the ECEI and these observations have been successfully reproduced by gyrokinetic simulations indicating the broad range of spatial scales up to the minor radius. Finally, performance of symmetric multiple Shattered Pellet Injections (SPIs) is demonstrated that peak density was increased twice with dual SPIs compared with a single SPI and energy can be radiated effectively when multiple SPIs are injected simultaneously, as planned in ITER.