Active control of radial rotor vibrations. Identification, feedback, feedforward, and repetitive control methods

Active vibration control methods for rotors were studied in order to develop solutions to enhance machines' dynamic behaviour, durability, and operating range. The aim of the thesis was to develop identification and control methods for active vibration control of a rotor. The identification method developed in the thesis improved run-time rotor identification by compensating rotation-related disturbances before the actual identification procedure. The control system design comprised an inner feedback loop and an outer loop for compensation for harmonic excitations due to mass unbalance and other rotation-related excitations. The feedback loop was shown to be essential in terms of providing favourable conditions for the other compensation algorithm in the outer loop. For the outer loop, three algorithms were tested: two feedforward control methods and a repetitive control method. The algorithms were validated and compared using an experimental set-up. Concerning the feedforward methods, the Convergent Control algorithm was found to be a more effective and simpler algorithm for the purpose than the adaptive FIR filter with the LMS algorithm. The adaptive gradient-based repetitive control, developed in this thesis, was found to have a poorer performance than the feedforward control methods, but to provide benefits for applications where excitation frequencies are not as predictable as in the current application.