Model-Based Control of the Dimensionless Gain in KSTAR by Leveraging Real-time Estimation of the Confinement Time*

Controlling Q, which is defined as the ratio of the fusion power to the auxiliary power, will be critical in ITER and future nuclear-fusion power reactors. By writing the global energy confinement time as τ = H_sτ_s, where τ_s is a particular confinement scaling (e.g. ITER89P) and H_s is a ratio defining the confinement quality, it is indeed possible to show [1] that Q ∝ G (= β_NH_s/q₉₅) where β_N is the normalized beta and q₉₅ is the safety factor at ρ = 0.95. Thus, Q is maximized by requiring high pressure (β_N) , high confinement quality (H_s), and high current-carrying capacity (inverse of q₉₅). Therefore, the dimensionless gain G plays a key role in defining a stable plasma scenario with a specific target fusion gain. As part of the effort by the KSTAR program on developing Q = 10 scenarios for ITER, a model-based control algorithm is proposed for the regulation of G by leveraging real-time estimation of the global energy confinement time and the associated confinement-quality factor [2]. The controller is designed by embedding the 1D magnetic diffusion equation and a 0D energy balance equation in the synthesis process. The performance and robustness of the proposed controller are tested in higher-fidelity nonlinear simulations for KSTAR based on COTSIM.

[1] T. Luce, Fusion Science and Technology 48:2 (2005) 1212-1225.

[2] K. E. J. Olofsson et al., IEEE CCTA, Montreal, Canada, 2020.