Variable Bias Current Control and Adaptive Parameter Estimation in Active Magnetic Bearings

Reducing the energy consumption of active magnetic bearings is important for some appli-cations to reduce operating cost, expand the operating period, and avoid heat generation. The power loss of magnetic bearings consists of ohmic, electronic circuit, windage, eddy current, and hysteresis losses. With the exception of windage, these losses are functions of the coil currents, therefore, minimization of these has a significant effect on the overall energy efficiency of a mag-netic bearing. In this paper, a variable bias current scheme is introduced to minimize the energy consumption without altering the dynamic performance. Since the current stiffness and negative stiffness of the magnetic bearing are functions of the bias current, it is possible to achieve invari-ant dynamics for various bias current settings. The optimum bias current is calculated from the demand bearing force and instantaneous rotor displacement, but errors in model parameters may result in changes in system dynamics. Therefore, correct parameter values are required for invari-ant dynamics. In this paper, estimation methods for various parameter errors are introduced and verified experimentally. The proposed variable bias current controller and parameter estimation methods yield significant energy savings, are simple to implement and applicable to a wide range of magnetic bearing systems.