Parameter identification procedures and model validation are major steps toward intelligent machines supported by active magnetic bearings (AMB). The ability of measuring the electromagnetic bearing forces, or deriving them from measuring the magnetic flux, strongly contributes to the model validation and leads to novel approaches in identifying crucial rotor parameters. This is the main focus of this paper, where an intelligent AMB is being developed with the aim of aiding the accurate identification of damping and stiffness coefficients of active lubricated journal bearings. The main contribution of the work is the characterization of magnetic forces by using two different experimental approaches. Such approaches are investigated and described in detail. A special test rig is designed where the four pole AMB is able to generate forces up to 1900N. The high-precision characterization of the magnetic forces is conducted using different experimental tests: (i) by using hall sensors mounted directly on the poles (precise measurements of the magnetic flux) and by an auxiliary system, composed of strain gages and flexible beams attached to the rotor, (ii) by measuring the input current and bearing gap variations, monitoring the bearing input signals. Advantages and drawbacks of the different methodologies are critically discussed. The linearity ranges are experimentally determined and the characterization of magnetic forces with a high accuracy of <1% is achieved (percent error is normalized with respect to the instantaneous measured force obtained from the strain gauges signals).
Primary resonance analysis of a nonlinear flexible shaft supported by active magnetic bearings using analytical method
