Experiment planning using high-level component models at W7-X

• Published in: Fusion engineering and design 2012-12, Vol.87 (12), p.1949-1953
• Main Authors: Lewerentz, Marc, Spring, Anett, Bluhm, Torsten, Heimann, Peter, Hennig, Christine, Kühner, Georg, Kroiss, Hugo, Krom, Johannes G., Laqua, Heike, Maier, Josef, Riemann, Heike, Schacht, Jörg, Werner, Andreas, Zilker, Manfred
• Format: Article
• Language: English
• Subjects: Component model; Control systems; Design engineering; Encapsulation; Experiment planning; High-level component behavior; Mathematical models; Segment control system; Steady state; Superconductivity; Tasks; Technical and temporal dependencies; Temporal logic
• ISSN: 0920-3796; 1873-7196
• DOI: 10.1016/j.fusengdes.2012.09.002

► Introduction of models for an abstract description of fusion experiments. ► Component models support creating feasible experiment programs at planning time. ► Component models contain knowledge about physical and technical constraints. ► Generated views on models allow to present crucial information. The superconducting stellarator Wendelstein 7-X (W7-X) is a fusion device, which is capable of steady state operation. Furthermore W7-X is a very complex technical system. To cope with these requirements a modular and strongly hierarchical component-based control and data acquisition system has been designed. The behavior of W7-X is characterized by thousands of technical parameters of the participating components. The intended sequential change of those parameters during an experiment is defined in an experiment program. Planning such an experiment program is a crucial and complex task. To reduce the complexity an abstract, more physics-oriented high-level layer has been introduced earlier. The so-called high-level (physics) parameters are used to encapsulate technical details. This contribution will focus on the extension of this layer to a high-level component model. It completely describes the behavior of a component for a certain period of time. It allows not only defining simple value ranges but also complex dependencies between physics parameters. This can be: dependencies within components, dependencies between components or temporal dependencies. Component models can now be analyzed to generate various views of an experiment. A first implementation of such an analyze process is already finished. A graphical preview of a planned discharge can be generated from a chronological sequence of component models. This allows physicists to survey complex planned experiment programs at a glance.