The 4 th KSTAR campaign in 2011 concentrated on active ELM control by various methods such as non-axisymmetric magnetic perturbations, supersonic molecular beam injection (SMBI), vertical jogs of the plasma column, and edge electron heating. The segmented in-vessel control coil (IVCC) system is capable of applying n≤2 perturbed field with different phasing among top, middle, and bottom coils. Application of an n=1 perturbed field showed desirable ELM suppression result. Fast vertical jogs of the plasma column achieved ELM pace making and ELMs locked to 50 Hz vertical jogs were observed with a high probability of phase locking. A newly installed SMBI system was utilized for ELM control and a state of mitigated ELMs was sustained by the optimized repetitive SMBI pulse for a few tens of ELM periods. A change of ELM behavior was seen due to edge electron heating although the effect of ECH launch needs supplementary analyses. The ECEI images of suppressed/mitigated ELM states showed apparent differences when compared to natural ELMy states. Further analyses are ongoing to explain the observed ELM control results.

Along with an expanded evaluation of the equilibrium operating space of the Korea Superconducting Tokamak Advanced Research, KSTAR, experimental equilibria of the most recent plasma discharges were reconstructed using the EFIT code. In near-circular plasmas created in 2009, equilibria reached a stored energy of 54 kJ with a maximum plasma current of 0.34 MA. Highly shaped plasmas with near double-null configuration in 2010 achieved H-mode with clear edge localized mode (ELM) activity, and transiently reached a stored energy of up to 257 kJ, elongation of 1.96 and normalized beta of 1.3. The plasma current reached 0.7 MA. Projecting active and passive stabilization of global MHD instabilities for operation above the ideal no-wall beta limit using the designed control hardware was also considered. Kinetic modification of the ideal MHD n = 1 stability criterion was computed by the MISK code on KSTAR theoretical equilibria with a plasma current of 2 MA, internal inductance of 0.7 and normalized beta of 4.0 with simple density, temperature and rotation profiles. The steep edge pressure gradient of this equilibrium resulted in the need for significant plasma toroidal rotation to allow thermal particle kinetic resonances to stabilize the resistive wall mode (RWM). The impact of various materials and electrical connections of the passive stabilizing plates on RWM growth rates was analysed, and copper plates reduced the RWM passive growth rate by a factor of 15 compared with stainless steel plates at a normalized beta of 4.4. Computations of active RWM control using the VALEN code showed that the n = 1 mode can be stabilized at normalized beta near the ideal wall limit via control fields produced by the midplane in-vessel control coils (IVCCs) with as low as 0.83 kW control power using ideal control system assumptions. The ELM mitigation potential of the IVCC, examined by evaluating the vacuum island overlap created by resonant magnetic perturbations, was analysed using the TRIP3D code. Using a combination of all IVCCs with dominant n = 2 field and upper/lower coils in an even parity configuration, a Chirikov parameter near unity at normalized poloidal flux 0.83, an empirically determined condition for ELM mitigation in DIII-D, was generated in theoretical high-beta equilibria. Chirikov profile optimization was addressed in terms of coil parity and safety factor profile.

A decade-long operation of the Korean Superconducting Tokamak Advanced Research (KSTAR) has contributed significantly to the operation of superconducting tokamak devices and the 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 have achieved improved performance. As examples, a long pulse high-confinement mode operation with and without an edge-localized mode (ELM) crash was well over 70 and 30 s, respectively. The unique capabilities of KSTAR allowed it to improve the capability of controlling harmful instabilities, and they have been instrumental in uncovering much new physics. The highlights are that the L/H transition threshold power is sensitive to the resonant magnetic perturbation (RMP) and insensitive to non-resonant magnetic perturbation. Co-Ip offset rotation dominated by an electron channel predicted by general neoclassical toroidal viscosity theory was confirmed. Improved heat dispersal in a divertor system using three rows of rotating RMP was demonstrated and predictive control of the ELM-crash with a priori modeling was successfully tested. In magnetohydrodynamic physics, validation of the full reconnection model (i.e. q0  >  1 right after the sawtooth crash) and self-consistent validation of the anisotropic distribution of turbulence amplitude and flow in the presence of the 2/1 island with theoretical models were achieved. The turbulence amplitude induced by RMP was linearly increased with the slow RMP coil current ramp-up time (i.e. the magnetic diffusion time scale). The Dα spikes (i.e. ELM-crash amplitude) was linearly decreased with the turbulence amplitude and not correlated with the perpendicular electron flow. In the turbulence area, a non-diffusive ‘avalanche’ transport event and the role of a quiescent coherent mode in confinement were studied. To accommodate the anticipation of a higher performance of the KSTAR plasmas with the increased heating powers, a new divertor/internal interface with a full active cooling system will be implemented after a full test of the new heating (neutral beam injection II and electron cyclotron heating) and current drive (CD) (Helicon and lower hybrid CD) systems. An upgrade plan for the internal hardware, heating systems and efficient CD system may allow for a long pulse operation of higher performance plasmas at βN  >  3.0 with f bs ~ 0.5 and Ti  >  10 keV.

Typical ELMy H-mode discharges have been obtained in the KSTAR tokamak with the combined auxiliary heating of neutral beam injection (NBI) and electron cyclotron resonant heating (ECRH). The minimum external heating power required for the L–H transition is about 0.9 MW for a line-averaged density of ∼2.0 × 1019 m−3. There is a clear indication of the increase in the L–H threshold power with decreasing density for densities lower than ∼2 × 1019 m−3. The L–H transitions typically occurred shortly after the beginning of plasma current flattop (Ip = 0.6 MA) period and after the fast shaping to a highly elongated double-null divertor configuration. The maximum heating power available was marginal for the L–H transition, which is also implied by the relatively slow transition time (>10 ms) and the synchronization of the transition with large sawtooth crashes. The initial analysis of thermal energy confinement time (τE) indicates that τE is higher than the prediction of multi-machine scaling laws by 10–20%. A clear increase in electron and ion temperature in the pedestal is observed in the H-mode phase but the core temperature does not change significantly. On the other hand, the toroidal rotation velocity increased over the whole radial range in the H-mode phase. The measured ELM frequency was around 10–30 Hz for the large ELM bursts and 50–100 Hz for the smaller ones. In addition, very small and high frequency (200–300 Hz) ELMs appeared between large ELM spikes when the ECRH is added to the NBI-heated H-mode plasmas. The drop of total stored energy during a large ELM is up to 5% in most cases.