Full text loading...
The arrangement of the conducting wall sections, radial control coils, and poloidal sensor coils on HBT-EP.
Evolutions of the plasma current and edge safety factor for a typical discharge in the feedback experiments. The external kink mode appears when surface is near the edge of the plasma, in the range .
Poloidal field fluctuations measured from the plasma surface for (a) no feedback, (b) positive feedback using the Kalman filter, and (c) negative feedback with the Kalman filter.
Frequency spectrum of poloidal field fluctuations in arbitrary units as a function of the phase angle between feedback and the measured mode. Feedback may be phased to either suppress or excite the mode near , but the Kalman filter prevents excitation at higher frequencies.
Adding flat-spectrum, Gaussian noise to the mode calculated by the FPGA algorithm showcases the Kalman filter’s noise removal capability. Plot (a) shows a simulation of the cosine mode as calculated in the DFT step in the algorithm (black) and the sum of this signal with the cosine component of the added noise (gray). Plots (b) and (c) show one of five FPGA outputs to the control coils in the added noise experiments for algorithms with and without the Kalman filter, phased to excite the mode.
A comparison of the frequency spectrum of the poloidal field for added noise experiments, for feedback off (solid), Kalman filter mode ID feedback (dotted), and mode ID feedback (dashed) discharges. Plots (a) and (c) show suppression and excitation experiments with no added noise. Plots (b) and (d) show the suppression and excitation cases with extra noise added to the feedback algorithms. With the added noise, the algorithm without the Kalman filter loses its ability to suppress and excite the mode, while the Kalman filter algorithm does not.
Article metrics loading...