Real-time Model-based Reconstruction and Control of Tokamak Plasma Profiles

A new paradigm for real-time plasma profile reconstruction is demonstrated in the TCV tokamak. Predictions based on physics models are merged with available real-time diagnostic data to construct a self-consistent profile state estimate compatible with a timedependent model of transport processes in the plasma. This is enabled by a new RApid Plasma Transport simulatOR (RAPTOR), implemented in the new TCV real-time control system. RAPTOR simulates the radial current diffusion including the ohmic coil transformer voltage and non-inductive sources in real-time, while the plasma physically evolves in the tokamak. This makes available an extensive set of quantities which are normally not known in real-time such as the bootstrap current fraction, safety factor, magnetic shear and loop voltage profiles. This approach represents a generalization of existing approaches for real-time equilibrium reconstruction with measurement-constrained current density profile, as transport physics knowledge is now included in the reconstruction. The same rapid transport code is also used in predictive mode, including a model of the electron temperature evolution, for off-line studies of optimal actuator trajectories during plasma ramp-up scenarios. Constraints are included in the optimization to reflect realistic operational limits. These studies show that a plasma current overshoot combined with appropriately timed heating are beneficial for rapidly reaching a stationary q profile with flat central shear. The demonstration of this new paradigm paves the way for further integration of real-time tokamak plasma simulations for prediction, scenario monitoring, disruption avoidance and feedback control.