Abstract:Since noise degrades the control precision of International Thermonuclear Experimental Reactor (ITER) poloidal field (PF) ac/dc converter control system, a real-time wavelet filter algorithm for feedback signal denoising is studied. According to the requirements of ITER PF ac/dc converter controller, a real-time digital filter based on a wavelet transform is designed in this paper. The Daubechies wavelet is used as the wavelet basis and the mandatory denoising method is taken to remove the noise. In order to improve the boundary effects of wavelet transform when the width of sliding window and decomposition scale is quite small, the left and right symmetric boundary extension method is proposed. Three different boundary extension methods are compared, and the effectiveness of the proposed boundary extension method for denoising are described. The experiments are carried out on ITER PF ac/dc converter sequential control test platform to prove that the designed real-time wavelet filter eliminates the noise effectively and guarantees the real-time processing of feedback signal.

The Design on the Real-Time Wavelet Filter for Iter Pf Ac/Dc Converter Control System