Active vibration control of rotor in desktop test environment: Licentiate thesis

The goal of this work was to set up a test environment for active vibration control of rotors, to study the dynamics of the system, and to design a control system for controlling rotor vibrations. The principal idea was to use a non-contacting magnetic actuator without a load-carrying function. The test environment consisted of a desktop rotor test rig, a magnetic actuator, and a programmable control unit. The rotor was supported by conventional bearings at the ends, and the control forces were applied at the midpoint of the bearing span. The work reports modal analysis and open-loop measurement results on the test environment. The system was found to vibrate excessively when the rotor was run close to its bending critical speed at 40 Hz. Damping the system, considered as the Jeffcott rotor, by means of a velocity feedback controller was studied. The controller parameters were selected on the basis of estimates derived experimentally from the measured data. According to the simulations and experiments performed, the velocity feedback control system reduced the vibration response significantly. The controller made it possible to run the rotor across the critical speed. Another controller, a feedforward system based on an adaptive finite-impulse-response filter and the use of a reference signal, was designed to compensate disturbances caused by the mass imbalance in the rotor. The filter was adapted by the least-mean-squares algorithm. The simulations and experiments showed that the harmonic response due to the mass imbalance was successfully compensated. The use of the feedback system model improved the performance and extended the operating range of the feedforward system to super-critical conditions. The different roles of the algorithms are pointed out: feedback control increased the damping of the system, while feedforward control compensated the disturbance at the frequency of rotation. The forces required for damping the vibrations were low compared with the mass of the rotor. The use of an adaptive filter led to a considerable reduction in the response, with a minor increase in the active forces used. Results obtained elsewhere in similar test environments are also reported in this work. The data and experience acquired during the construction of the test environment are discussed.