skip to main content

Title: Prediction of plasma simulation data with the Gaussian process method

The simulation of plasma-wall interactions of fusion plasmas is extremely costly in computer power and time - the running time for a single parameter setting is easily in the order of weeks or months. We propose to exploit the already gathered results in order to predict the outcome for parametric studies within the high dimensional parameter space. For this we utilize Gaussian processes within the Bayesian framework and perform validation with one and two dimensional test cases from which we learn how to assess the outcome. Finally, the newly implemented method is applied to simulated data from the scrape-off layer of a fusion plasma. Uncertainties of the predictions are provided which point the way to parameter settings of further (expensive) simulations.
Authors:
;  [1]
  1. Max-Planck-Institute for Plasma Physics, EURATOM Association, 85748 Garching (Germany)
Publication Date:
OSTI Identifier:
22390754
Resource Type:
Journal Article
Resource Relation:
Journal Name: AIP Conference Proceedings; Journal Volume: 1636; Journal Issue: 1; Conference: MaxEnt 2013: 33. International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, Canberra, ACT (Australia), 15-20 Dec 2013; Other Information: (c) 2014 AIP Publishing LLC; Country of input: International Atomic Energy Agency (IAEA)
Country of Publication:
United States
Language:
English
Subject:
70 PLASMA PHYSICS AND FUSION TECHNOLOGY; COMPUTERIZED SIMULATION; GAUSSIAN PROCESSES; PARAMETRIC ANALYSIS; PLASMA; PLASMA SCRAPE-OFF LAYER; PLASMA SIMULATION; TWO-DIMENSIONAL CALCULATIONS; VALIDATION; WALL EFFECTS