Fig. 5: System identification experiments.
From: Real-time feedback control of the impurity emission front in tokamak divertor plasmas
Overview of system identification results as obtained on the TCV tokamak. Each separate measurement plot contains as a top trace the time evolution of the multi-sine perturbation (black) and response in the emission front Lpol (red). In the experiment, the perturbation is summed with a feedfoward waveform to yield the total (always positive) gas valve command. Each bottom plot shows the Discrete-Fourier transform (DFT) of the perturbation (black) and the plasma response (red), a dashed black line shows an estimate of the noise floor. This noise floor in the unbaffled L-mode experiment (a) is high due to the lower (200 Hz) MANTIS frame rate. The baffled L-mode experiment (b) has the highest SNR with MANTIS running at 800 Hz. In contrast, the H-mode experiment (in c) shows a high measurement noise floor due to the additional disturbance in the form of ELMs, which also cause significant contributions at non-excited frequencies. Furthermore, the plasma response is overall evidently smaller compared to the L-mode cases in (a and b). Finally, the less favorable experimental conditions in H-mode lead to a more pronounced drift of Lpol over time in (c), which is compensated in further analysis.
