Model based estimation of inertial parameters of a rigid rotor having dynamic unbalance on Active Magnetic Bearings in presence of noise

摘要

Unbalance in rotors is known as the primary source of excitation, the state of unbalance changes over time and, therefore, industrial rotors require online or smart identification and attenuation of unbalance for smooth operation. To begin with, this paper proposes the use of Active Magnetic Bearings (AMBs) to identify the unbalance. Unbalance is defined as the location of centre of mass of the rotor and orientation of the axes of principal mass moments of inertia, both with respect to the spin-axis, assumed as the geometric axis of symmetry of rotor. Non-uniform distribution of mass makes a rotor dynamically unbalanced. The centre of mass shifts from the spin-axis and none of the principal axes of inertia is aligned with it. Dynamic behaviour of such a rotor is modelled by a set of differential equations with time varying coefficients. A rigid rotor is assumed to be supported on two AMBs with all known characteristics. A methodology based on modelling and simulation is proposed to determine the eccentricity and the orientation of the principal axes of inertia (angles of its skewness) by utilizing the force and response of the rotor recorded by each AMB. The median of simulated inertial properties are shown to be in close agreement with their actual values. The proposed methodology is also found robust enough in the presence of different types of simulated noise used to pollute signals from the bearings. Therefore, the methodology is suitable for online identification of inertial properties of a rotor.