Model reduction techniques in tokamak modelling

In present tokamak experiments, there is the need of sufficiently detailed models describing the plasma behaviour and its electromagnetic coupling with the active and metallic structures for simulation and control design. The required level of detail yields linearized model of the system at particular working points of very high order, thus complicating the design of controllers of the position and shape of the plasma. In this paper, several model reduction techniques which are available from linear state-space control theory are considered with application to tokamak modelling. In particular, a technique based on selective modal analysis is proposed, and its effectiveness in approximating the overall system behaviour while retaining the physical meaning of the state variables is shown.