Estimation of magnetic bearing constant in active magnetic bearings

Abstract

Active magnetic bearings (AMB) provide non-contact support to rotors and require closed-loop control systems as they are inherently unstable. The methodology to obtain parameters of the force model of an AMB from closed-loop control signals and air gap signals using a successive integration-based algebraic technique is presented in this work. This transient data-based approach does not require additional load cells or differentiation of control and gap sensor signals which lead to noise amplification. This estimation methodology is shown to apply to AMB with unequal bias currents in both current control and voltage control modes. This methodology is also shown to be robust in the presence of noise in sensor signals. Also, it can estimate the force model parameters in the presence of harmonic force disturbances on the AMB. This estimation approach is developed using linear and nonlinear AMB simulation models and then applied to initial transient data from the AMB test rig. The estimated magnetic bearing constant (Km) is compared to Km values obtained from the experimental measurements on the AMB actuators. This direct force measurement is further supported with steady-state closed-loop control signal data from the AMB test rig for a range of levitated positions within the air gap.