Sensorless active magnetic dampers for the control of rotors

The aim of this paper is to present self-sensing Active Magnetic Dampers (AMDs) for the vibration control of rotating machines and evaluate their performance and advantages with respect to standard sensed solutions. The technique is implemented on a rotor reproducing the typical dynamic behaviour of an aero-engine gas turbine shaft. The proposed technique is based on a Luenberger observer that estimates the mechanical states of the system. The observed states are fed back in closed-loop to introduce damping into the system and to reduce the vibrations during critical speed crossing. The rotordynamic and electromechanical modeling is illustrated taking into account the anisotropy of rotor elastic supports. The control design is described along with a sensitivity analysis on the most critical model parameters and a study of electromagnet nonlinear effects on the closed loop behaviour. The importance of the inherent collocation in the self-sensing configuration during control design is discussed analysing modal shapes and sensor/actuator transfer functions. A phase of experimental identification of actuator parameters is performed on the open-loop system response to improve the reliability of the model. The effectiveness of the proposed method is evaluated experimentally by measuring unbalance response in open and closed-loop configuration showing a reduction of displacement during critical speed crossing from 0.35 mm to 0.04 mm. Furthermore, a classical AMD realized with the use of position sensors is implemented on the same rotor. The results obtained with sensed and self-sensing controls are compared to show the good quality of the damping performance reached with the proposed self-sensing technique.