나용수 교수

In this work, we address a new feedforward control scheme for the normalized beta ( β N ) in tokamak plasmas, using the deep reinforcement learning (RL) technique. The deep RL algorithm optimizes an artificial decision-making agent that adjusts the discharge scenario to obtain a given target β N from the state–action–reward sets explored by its own trial and error in a virtual tokamak environment. The virtual environment for the RL training is constructed using a long short-term memory (LSTM) ne

On hybrid scenarios in KSTAR, Na, Yong-Su, Lee, Y.H., Byun, C.S., Kim, S.K., Lee, C.Y., Park, M.S., Yang, S.M., Kim, B., Jeon, Y.-M., Choi, G.J., Citrin, J., Juhn, J.W., Kang, J.S., Kim, H.-S., Kim, J.H., Ko, W.H., Kwon, J.-M., Lee, W.C., Woo, M.H., Yi, S., Yoon, S.W., Yun, G.S., KSTAR team

Abstract This work develops an artificially intelligent (AI) tokamak operation design algorithm that provides an adequate operation trajectory to control multiple plasma parameters simultaneously into different targets. An AI is trained with the reinforcement learning technique in the data-driven tokamak simulator, searching for the best action policy to get a higher reward. By setting the reward function to increase as the achieved β p , q 95 , and l i are close to the given target values, the

Although gas breakdown phenomena have been intensively studied over 100 years, the breakdown mechanism in a strongly magnetized system, such as tokamak, has been still obscured due to complex electromagnetic topologies. There has been a widespread misconception that the conventional breakdown model of the unmagnetized system can be directly applied to the strongly magnetized system. However, we found clear evidence that existing theories cannot explain the experimental results. Here, we demonstr

High confinement and stability are obtained simultaneously in stationary conditions in improved H-mode discharges at ASDEX Upgrade. The improved H-mode discharges are typically composed of two different phases: 'lower heating phase', where H98(y, 2) is similar to standard H-modes (H98(y, 2) ∼ 1), and 'fully developed improved H-mode phase', where H98(y, 2) is higher than standard H-modes (H98(y, 2) up to 1.4). In this paper, the confinement physics is studied by comparing these two different pha

A tokamak, a torus-shaped nuclear fusion device, needs an electric current in the plasma to produce magnetic field in the poloidal direction for confining fusion plasmas. Plasma current is conventionally generated by electromagnetic induction. However, for a steady-state fusion reactor, minimizing the inductive current is essential to extend the tokamak operating duration. Several non-inductive current drive schemes have been developed for steady-state operations such as radio-frequency waves an

ECH-assisted start-up using trapped particle configuration (TPC) is firstly studied in a superconducting, conventional tokamak, KSTAR. First, improved and efficient start-up using TPC than conventional field null configuration (FNC) is achieved by enhanced pre-ionization plasma quality. TPC shows the broader operation window in terms of the poloidal field quality and the deuterium prefill pressure than that of FNC. Surprisingly the particle trapping enhances the plasma start-up performance even

Abstract Two types of experiments were carried out to conduct an intrinsic rotation study in KSTAR. The first was a density ramp-up experiment without neutral beam injection, and the second was an experiment with beam blip technique. In these experiments, some characteristics of the intrinsic rotation were observed in the KSTAR Ohmic L-mode plasmas including: (i) a non-monotonic dependence of the core intrinsic rotation, called U-curve behaviour, with respect to the electron density and the coll

We report the results of predictive modelling of high performance steady state operation scenarios in KSTAR. Firstly, the capabilities of steady state operation are investigated with time-dependent simulations using a free-boundary plasma equilibrium evolution code coupled with transport calculations. Secondly, the reproducibility of high performance steady state operation scenarios developed in the DIII-D tokamak, of similar size to that of KSTAR, is investigated using the experimental data tak

We report results of benchmarking of core particle transport simulations by a collection of codes widely used in transport modelling of tokamak plasmas. Our analysis includes formulation of transport equations, difference between electron and ion solvers, comparison of modules of the pellet and edge gas fuelling on the ITER baseline scenario. During the first phase of benchmarking we address the particle transport effects in the stationary phase. Firstly, simulations are performed with identical

International audience

Abstract This paper deals with one of the origins and trigger mechanisms responsible for the observed performance enhancements in the hybrid scenario experiments conducted in Korea Superconducting Tokamak Advanced Research (KSTAR). The major contribution to the performance improvement comes from a broader and higher pedestal formation. The increase of fast ion pressure due to a plasma density decrease also contributes substantially to the global beta. Although the reduced core plasma volume resu