Extended database of 2D UEDGE simulations for KSTAR detachment control with variations of plasma currents*

In this work, we extend the database of 50,000 2D UEDGE simulations [1] for Machine Learning surrogate models [2] designed for KSTAR detachment control. These simulations explore five control parameters: core boundary (at PsiN = 0.85) plasma density, SOL input power, carbon impurity fraction, perpendicular anomalous transport coefficients and the plasma current. Specifically, core boundary density ranges from to , and input power varies from 1 to 8 MW, covering the typical KSTAR operation regime. The simulations assume a spatially constant concentration of carbon impurities using a fixed-fraction model and include the plasma potential equation and cross-field drifts with the Braginskii plasma and neutral fluid model. The anomalous transport coefficient profiles are varied based on fits to SOLPS-ITER simulated upstream and from KSTAR discharge #22849, providing flexibility for the surrogate model considering real-time plasma transport uncertainty. The database now includes plasma currents of 300 kA, 500 kA, 600 kA, 700 kA, and 800 kA. The data reveal that for specific power input, impurity fraction and plasma current, the ion saturation current at the target plates exhibits a roll-over phenomenon as upstream collisionality increases, consistent with the experimental trend for divertor-plasma detachment onset. Additionally, the heat flux width on the outer target plate at a low collisionality (attached plasma) aligns with the Eich's scaling [3]. The surrogate model based on the database is being validated against SOLPS-ITER simulations and KSTAR experimental data and will be presented in [2].