Model-predictive kinetic control with data-driven models on EAST
Résumé
In this work, model-predictive control (MPC) was combined for the first time with singularperturbation theory, and an original plasma kinetic control method based on extremely simpledata-driven models and a two-time-scale MPCalgorithm has been developed. A comprehensivereview is presented in this paper. Slow and fast semi-empirical models are identified from data,by considering the fast kinetic plasma dynamics as a singular perturbation of a quasi-staticequilibrium, which itself is governed, on the slow time scale, by the flux diffusion equation.This control technique takes advantage of the large ratio between the time scales involved inmagnetic and kinetic plasma transport. It is applied here to the simultaneous control of thesafety factor profile, q(x), and of several kinetic variables, such as the poloidal beta parameter,βp, and the internal inductance parameter, li, on the EAST tokamak. In the experiments, theavailable control actuators were lower hybrid current drive (LHCD) and co-current neutral beaminjection (NBI) from different sources. Ion cyclotron resonant heating (ICRH) and electroncyclotron resonant heating (ECRH) are used as additional actuators in control simulations. Inthe controller design, an observer provides, in real time, an estimate of the system states and ofthe mismatch between measured and predicted outputs, which ensures robustness to modelerrors and offset-free control. Based on the observer information, the controller predicts thebehavior of the system over a given time horizon and computes the optimal actuation by solvinga quadratic programming optimization problem that takes the actuator constraints into account.A number of control applications are described in the paper, either in nonlinear simulations withEAST-like parameters or in real experiments on EAST. The simulations were performed with afast plasma simulator (METIS) using either two control actuators (LHCD and ICRH) in a lowdensity scenario, or up to four actuators at higher density: LHCD, ECRH, and two NBI systemsdriven in a on/off pulse-width-modulation (PWM) mode, with different injection angles. Thecontrol models are identified with the prediction-error method, using datasets obtained fromopen loop simulations in which the actuators are modulated with pseudo-random
Origine | Publication financée par une institution |
---|---|
licence |