Parametric Numerical Analysis of Restoring Magnetic Forces Dependences in Radial Active Magnetic Bearings with a Given Control Law

The paper proposes a technique for calculating the force characteristics of a radial active magnetic bearing with an eight-pole stator with a given control law. This approach is based on the application of the finite element method and a software package for a general technical analysis of physical processes. Variable data of the technique are: the bearing design parameters, the control law parameters and the initial data of a variant parametric analysis of magnetic forces depending on the position of the rotor in active magnetic bearings and control signals. Geometric and finite element calculation models are three-dimensional. They accurately reconstitute the geometry of the stator and rotor parts of magnetic circuits, pole coils, internal and external air spaces and air gaps between the rotor and the pole. The active magnetic bearing control law is control voltages on electromagnets windings. Their values depend on the instantaneous position of the bearing journal and the instantaneous values of currents in pole coils circuits. In finite element calculation models, the load is specified in the form of current density in the volumes of the coils, taking into account the direction of the turns at each point. Its value is determined in accordance with the control law and the parameters of electrical circuits. The magnetic forces are calculated from the Maxwell tension tensor in a given cloud of points of the rotor position. The force characteristics are dependences of the restoring magnetic forces on the displacement of the bearing journal in two radial directions. The results of computational studies to determine the dependences of magnetic forces on rotor displacements for the design of an active magnetic bearing used in an experimental scale model of a turbocompressor of a gas compressor unit with active magnetic bearings are presented. Validation of the approach and results was performed by comparing the calculated and experimental data.