Towards more predictive gyrokinetic turbulence simulations of the tokamak boundary region

Future fusion power plants (FPPs) require computational tools that predict plasma behavior from controllable engineering parameters alone, without relying on input parameters that are adjusted to fit an experiment. We present full-f gyrokinetic simulations that evolve edge and scrape-off layer (SOL) turbulence self consistently (simulating closed and open field line regions together) using only three inputs: magnetic geometry, heating power, and particle inventory. Our novel adaptive sourcing algorithm in the Gkeyll code dynamically balances energy injection with particle losses while emulating neutral recycling, avoiding the use of experimentally informed parameters, though there are of course various approximations made in the simulation. Comparisons with TCV discharge #65125 demonstrate that Gkeyll's simulated density and temperature profiles compare qualitatively well with the experimental measurements. Additionally, simulations accurately reproduce interesting aspects of experimental observations of enhanced confinement in negative triangularity configuration, revealing impacts on E×B shearing rate profile, gradient-B drift transport, and power distribution between wall and limiter. Though more accurate physics is being added to our code (such as more accurate neutrals), these results demonstrate a method of doing direct simulations without adjustable fitting parameters, for more confident predictions of the performance of FPPs.