Strategy towards model-based design and testing of the ITER Plasma Control System

Abstract

The plasma control system (PCS) forms the backbone of safe and high performance plasma operation in all tokamaks. The ITER tokamak will be significantly larger than its predecessors and will become a nuclear environment at a later stage of its operation. Therefore, new control algorithms for the ITER PCS have to follow a systematic, rigorous and formal design and testing procedure. Then, controllers have to be formally commissioned preceding their ultimate deployment on the PCS for real-time operation. Considering the significant cost of operating the ITER machine and its tight operational schedule, limited time is available for controller commissioning. Therefore, the design team for the ITER PCS is implementing a model-based design and testing approach to exploit models not only for the design and tuning of controllers, but also the verification and validation (V&V) tests. To this end, controllers are first designed in the Plasma Control System Simulation Platform (PCSSP) in Simulink. The PCSSP has been deployed on a continuous integration (CI) server on which V&V and functional performance tests are automatically run and analysed to scrutinize the design. Once controllers have passed the design phase in PCSSP, automatic code generation or manual coding allows implementation on the real-time framework (RTF) forming the real PCS. Components on the RTF are continuously validated against their PCSSP counterparts to make sure the implementation is bug-free. The increasing complexity of the interplay between system requirements, modules, controllers, and tests results is managed in a structured Plasma Control System Database (PCSDB). This PCSDB ultimately allows the monitoring of the PCS design and implementation progress. In this work, we present the stages ITER controllers have to pass and the new tools that are being implemented to facilitate and structure this model-based design and testing workflow.