München 2019 – wissenschaftliches Programm
P 18.86: Poster
Donnerstag, 21. März 2019, 16:30–18:30, Foyer Audimax
Towards Applications of Deep Learning Techniques to Establish Surrogate Models for the Power Exhaust in Tokamaks — •Martin Brenzke1, Sven Wiesen1, Matthias Bernert2, and The ASDEX Upgrade Team2 — 1Forschungszentrum Jülich, Institut für Energie- und Klimaforschung, 52425 Jülich, Germany — 2Max Planck Institute for Plasma Physics, 85748 Garching, Germany
One of the main challenges in the design of an economically viable fusion reactor are the thermal loads experienced by the plasma facing components, especially the targets in a divertor-based design, on which this work and current developments focus. These thermal loads cause degradation of the target material and might severely damage the machine, resulting in longer downtime for maintenance. Modeling these thermal loads is one of the most important points in determining the operating scenarios for future fusion devices. Under attached conditions, simplified analytical models, such as the two-point model, are sufficient to determine the thermal load experienced by the divertor targets for given conditions of the main plasma. However, modeling and predicting thermal loads is a challenging yet crucial task for future devices. In light of current developments and successes in the field of machine learning techniques, data-driven modeling is an interesting option for this problem. We present first steps towards modeling the power exhaust of tokamaks using deep learning methods (neural networks) and experimental data from the ASDEX Upgrade experiment. The work focuses on data selection and initial approaches to the problem of modeling the power exhaust from experimental data.