| Summary: | Repetitive operation is extensively present in the inverter compressor refrigeration field. Research illustrates that an approximate periodic load disturbance deteriorates its self-sensing and efficiency in a low frequency range. According to an internal model theorem, a repetitive controller can realize periodic signal tracking or suppression; however, ideal repetitive control (IRC) can not be directly employed in practical application owing to its excessive sensitivity to period-time uncertainties or aperiodic component, etc. In addition, the ratio between period delay to sampling time can not always maintain an integer which causes the resonant frequencies to deviate from the interested. Even if the ratio is a fixed integer, the auxiliary function for stabilization in a conventional repetitive controller can result in the steady state tracking error. In this paper, a revised repetitive controller with a novel Savitzky–Golay (S–G) filter is proposed, the S–G filter cascaded with a traditional delay element can approximate the delay element with any ratio. Aiming at a different operating condition, a second order generalized integrator (SOGI) based observer is employed to extract disturbance fundamental component, and Steiglitz–McBride based adaptive notch filter (SM-ANF) is adopted for frequency estimation. The proposed scheme varies the S–G filter according to variable fundamental frequency and maintains its resonant frequencies corresponding with the fundamental and harmonic disturbance. In the end, simulations and experiments are implemented and results illustrate that speed fluctuation can be effectively suppressed by the proposed method.
|
|---|
