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

P: Fachverband Plasmaphysik

P 9: Poster I

P 9.20: Poster

Tuesday, March 29, 2022, 16:00–17:30, P

Uncertainty Quantification for Multiscale Turbulent Transport Simulations — •Yehor Yudin, Udo von Toussaint, and David Coster — Max Planck Institute for Plasma Physics, Boltzmannstrasse2, 85748 Garching, Germany

One of the challenges in understanding the energy and particle transport processes in the core plasma of a magnetic confinement fusion device is to quantify how it is effected by turbulent dynamics. This work considers a multi-scale approach of modeling this problem, where the numerical solution is obtained for coupled models describing processes on different spatial and temporal scales. Furthermore, the resulting model is used to investigate both epistemic and aleatoric uncertainties in the profiles of the transported quantities. This work proposes application of a surrogate modelling technique to reduce the computational cost of resolving a quasi-steady state solution on the miscro-scale when it is sufficient to capture only statistics of turbulent dynamics. We studied a Multiscale Fusion Workflow that utilizes turbulent energy and particle fluxes computed with a gyrofluid turbulence code GEM in flux tube approximation to calculate the transport coefficients for core transport code ETS. In this work, a data-driven probabilistic surrogate model based on Gaussian Process Regression is used to infer flux values computed by a turbulence code for given core profiles, and to calculate related uncertainties. For that, we use VECMA toolkit to perform uncertainty quantification, as well as to train, test and utilize surrogate models.