JKJae-Min Kwon

Jae-Min Kwon

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Jae-Min Kwon has a total of 106 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Development of a neural network-based surrogate model for neutral beam injection in KSTAR

    by Young-Hoon Lee, Chanyoung Lee, Hyun-Seok Kim, Minho Woo, Sang-Hee Hahn, S.S. Kim, Jae-Min Kwon

  • Open Access

    Linear stability analysis of non-resonant kinetic ballooning mode in comparison with the ion temperature gradient/trapped electron mode case

    by J.Y. Kim, J.H. Seo, J.M. Kwon, Y. Kishimoto, H.S. Han

    Unlike in the finite magnetic shear region where turbulent transport is typically dominated by the electrostatic ion temperature gradient and/or trapped electron mode (ITG/TEM), in the weak shear region it is possible for the electromagnetic non-resonant kinetic ballooning mode (KBM), also known as the kinetic infernal mode (KIM) in the low- n limit (here, n is the toroidal mode number), to play a significant role with its very small excitation threshold. Considering the potential relevance to turbulent transport or internal transport barrier (ITB) formation in reverse-shear plasmas, a more systematic study using the gKPSP gyrokinetic code is performed on its linear mode characteristics and excitation threshold, in comparison with the ITG/TEM case. It is shown that, when the magnetic shear decreases, the transition from the resonant to the non-resonant type occurs relatively smoothly in the gradual reduction of global width and coupled slab harmonics number. While this transition feature is similar for ITG/TEM and KBM, the non-resonant mode structure has a notable difference, with the KBM more dominated by a single slab harmonic which then makes it sensitive to the value of nq , allowing clear oscillatory behaviors of its mode center and eigenvalue when n or q is varied (here, q is the safety-factor). Also, while the global radial width typically has a large reduction with the transition, it still appears to have a significant value in the low- n KIM limit. Meanwhile, in accordance to the destabilization of the ideal ballooning mode with decreasing magnetic shear, the non-resonant KBM is found to be more unstable than the resonant KBM, particularly with its threshold temperature gradient (in the broad density profile condition) being smaller than the ITG/TEM one if q is relatively high (>1.5–2.0). These results thus demonstrate that the non-resonant KBM can play a significant role in the weak shear region, emphasizing the necessity of its stabilization for triggering the ITB in reverse-shear plasmas.

  • Impurity modelling and transport coefficient reconstruction in tokamak core plasmas

    by Yoon Seong Han, Junhyeok Yoon, Inwoo Song, Junghoo Hwang, Hyungho Lee, Siwoo Yoon, Jae-Min Kwon, Wonho Choe

    The developed KAIST impurity modelling (KIM) code analyses impurity transport in the poloidal cross-section of tokamak plasma in both prediction and reconstruction modes. The prediction mode utilizes a numerical solution of the one-dimensional radial continuity equation to simulate impurity transport based on the given plasma parameters and extends the resulting radial impurity density distribution to a two-dimensional poloidal distribution by incorporating centrifugal-force-driven asymmetry. Synthetic diagnostic data, such as radiation power distribution and spectral emissions, are generated using atomic databases for direct comparison with experimental measurements. In reconstruction mode, the code reconstructs the impurity transport coefficient profiles from the experimentally obtained impurity density evolution over time, to overcome the limitations of forward-modelling-based transport coefficient inference. The developed reconstruction algorithm is based on numerical impurity flux calculations and Bayesian techniques and was validated using six reference impurity transport coefficient profiles from the KSTAR, JET, and ASDEX-U tokamaks. The reconstruction accuracy evaluated using the R 2 value, demonstrated good agreement ( R 2 > 0.87) between the reconstructed and reference profiles in the core regions ( r / a < 0.8), even in the presence of 5% random noise. The code was applied to a stationary KSTAR H-mode discharge to demonstrate its capability in analysing experimental W transport. The reconstructed transport coefficients were compared with neoclassical predictions to assess their physical validity. The results show that the total diffusion coefficient follows a radial trend similar to the neoclassical prediction but at a higher level consistent with the expectation of turbulent transport enhancement. This trend confirms that the KIM code yields physically reasonable transport coefficients from experimental data. This study details the theoretical framework, numerical validation, and experimental efficacy of the KIM code in advancing impurity transport research.

  • Open Access

    Bayesian optimization for design space exploration of tokamak fusion reactors

    by Boseong ­kIm, Jae-Min Kwon, GaHyung Jo, Bong Guen Hong, Sang Hee Hahn, Darian Figuera-Michal, Eisung Yoon

    Designing a tokamak fusion reactor requires exploring a high-dimensional parameter space where multiple coupled variables must simultaneously satisfy physics constraints and engineering limits. Conventional exhaustive search approaches become computationally prohibitive as dimensionality increases. This paper demonstrates the application of multi-objective Bayesian optimization (MOBO) to tokamak design space exploration using a low-fidelity system analysis code.We compare state-of-the-art acquisition functions against an exhaustive search baseline for bi-objective optimization minimizing major radius (a capital cost proxy) and maximizing net electric power. Results show that all MOBO methods recover Pareto fronts comparable to exhaustive search while requiring orders of magnitude fewer function evaluations. We investigate the effects of acquisition function choice, batch size for parallel evaluation, and constraint handling strategies on convergence behavior. The methodology provides practical guidelines for efficient design optimization and establishes a framework applicable to higher-fidelity,computationally expensive plasma simulations.

  • Top Cited Papers

    State of the art, gaps, and prospects in fusion materials theory and modelling

    by Luca Reali, Mark R. Gilbert, David Cereceda, Krishna Chaitanya Pitike, Sergei L. Dudarev, Philip Edmondson, Shin Kajita, Jae-Min Kwon, Byeongchan Lee, Jaime Marian, et al.


  • Open Access

    Limit-cycle oscillations in turbulence–zonal flow dynamics: Observations in toroidal gyrokinetic simulations and sparse regression modeling

    by Jung-Min Suh, Sumin Yi, Jae-Min Kwon, Gahyung Jo, Chanyoung Lee, Sang-Hee Hahn

    We investigate the dynamics of turbulence–zonal-flow interaction in toroidal ion-temperature-gradient (ITG) turbulence using global particle-in-cell gyrokinetic (GK) simulations. Under the conditions considered, we observe long-lasting, regular limit-cycle oscillations (LCOs) for different temperature gradients. Two factors are crucial for obtaining long-lasting and apparent oscillations: the narrow radial width of the turbulent region and the application of aggressive noise control. The former enables zonal flows to grow via modulational instability and interact locally with ITG modes, while the latter suppresses irregularity in the LCOs by reducing numerical particle noise. To assess whether these dynamics conform to the self-regulation framework, we adopt predator–prey-type equations for turbulence, zonal flow, and temperature gradient. The coefficients are identified from GK data using a regression method. The reduced model successfully reproduces the LCOs near the nonlinear threshold, thereby verifying the paradigm of zonal-flow regulation of turbulence in first-principles simulations. As the temperature gradient increases, however, we find the discrepancies between the model and GK simulations, implying modifications (or limitations) of the model equations in the strong-turbulence regime.

  • Open Access

    Predictive study of non-axisymmetric neutral beam ion loss on the upgraded KSTAR plasma-facing components

    by Taeuk Moon, Tongnyeol Rhee, Jae-Min Kwon, Young-Mu Jeon, Eisung Yoon

    We simulate ion loss induced by neutral beam injection (NBI) in three-dimensional (3D) space with high fidelity on the plasma-facing components (PFCs) of the Korea Superconducting Tokamak Advanced Research (KSTAR) device. Utilizing a 3D collision detection routine added to the NuBDeC code and computer-aided design data reflecting the recent upgrade to a tungsten divertor, we have characterized 3D heat flux distribution and patterns over PFCs due to the NB ion loss throughout parameter scans. First, we identify axial asymmetry in the heat flux distribution. The plasma-wetted areas extend along the direction of the plasma current and deviate diagonally in the direction of the NB ion's poloidal turn. Additionally, on the PFC surfaces, local heat flux peaks are observed along both poloidal and toroidal directions. These local peaks emerge on the surfaces of the PFCs that protrude toward the regions swept by the NB ions. Second, through a case study of NB1-C (beam source C of neutral beam 1) that results in the most loss, we analyze several changes in the heat flux patterns observed on the divertor and the poloidal limiter. Such analysis allows us to examine how variations in the parameters affect the movement of the peak heat flux position and the extent of the plasma-wetted area, if formed. We observe that the ion loss increases under several conditions: shallower beam deposition, higher beam energy, larger poloidal beta, and lower plasma current. Ions born near the plasma edge due to shallow beam injection follow orbits with a larger minor radius. Flux surface shifts and ion drifts from changes in beam energy, poloidal beta, and plasma current move these orbits closer to the wall. These factors increase the chance of ion loss through wall collisions. This study is believed to help optimize design and operation of NBI systems.