Application of Modern Control Methods to a Tokamak Plasma Reactor

FAVEZ Jean-Yves, Joe Lister (CRPP)  BONVIN Dominique  MUELLHAUPT Philippe  SRINIVASAN Bala


All present-day tokamaks routinely use PID controllers that are designed on crude assumptions and tuned by experience. As a result, the control performance is rather limited. Recently, considerable attention has focused on optimal controllers with enhanced robustness like LQG and H (infinity). Although these linear controllers have shown valid results, they usually fail to maintain control in the presence of large disturbances (ELMs). This weakness is due to the saturation of the actuators and can cause considerable damages to the plant. This research work aims at the investigation of control methods that could avoid this problem by maintaining the tokamak into a stable and safe operational domain even when large disturbances occur. A further goal in tokamak control research is to find a way to reduce the power consumption and the AC-losses while maintaining a fast response. This work is targeted at the TCV, JET and ITER tokamaks. Work has so far concentrated on understanding the modelling of the tokamak and on studying the controllability regions of second-order unstable SISO (single input, single output) systems with a saturated input.

This project is carried out in collaboration with de Centre de Recherches en Physique des Plasmas (CRPP) at the EPFL.


Favez J-Y., B. Srinivasan , P. Mullhaupt and D. Bonvin. Condition for Bifurcation of the Region of Attraction in Linear Planar Systems with Saturated Linear Feedback. IEEE CDC02, Las Vegas, USA / December 2002, 3918-3923