SMuK 2021 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

P: Fachverband Plasmaphysik

P 5: Poster I

P 5.1: Poster

Tuesday, August 31, 2021, 14:00–16:00, P

Introduction of quasilinear transport models to the Integrated Data Analysis framework — •Michael Bergmann, Rainer Fischer, and Frank Jenko — Max-Planck-Institute für Plasmaphysik, Boltzmannstr. 2, 85748 Garching

Combining the analysis of multiple diagnostics, well-chosen priors and Bayesian probability theory the Integrated Data Analysis (IDA, see [Fischer 2010]) is capable of providing density and temperature radial profiles of fusion plasmas. These profiles are then used for further analysis such as the simulation of discharges. Since IDA considers uncertain measurement data from a heterogeneous set of diagnostics but no transport physics, the estimated profiles and their gradients can be in contradiction to the profiles from transport solvers. Using transport solvers such as QLK and their much faster neural network models e.g. QLKNN we have created a loop in which simulated profiles are fed back into IDA as another prior thus providing constraints about the physically reasonable parameter space. This work feeds into a broader effort to make IDA more robust against measurement uncertainties by combining multiple transport solvers with different accuracies and computing costs in a multi-fidelity approach.