Active magnetic bearings (AMBs) have the unique capability to act concurrently as support bearings and load cells for measuring shaft forces. Current state-of-the-art methods for force measurement rely on models with limited accuracy due to effects which are difficult to characterize such as fringing, leakage, and variations in material properties. In addition, these effects may be a function of actual air gaps which are difficult to determine in a dynamic operating environment. This paper discusses a new force measurement methodology that inherently accounts for these types of effects and other system uncertainties by utilizing multiple sets of current pairs in opposing actuators, in conjunction with a calculation algorithm, to accurately determine the force applied by the AMB. This new multi-point methodology allows for the determination of bearing forces from information on basic actuator geometry and control currents only, with no knowledge of actual operating air gaps required. The inherent nature of the methodology accounts for model uncertainties such as fringing, leakage, and other system unknowns. Initial static experimental test results are presented demonstrating 3% error in measuring the nominal determined bearing load, and a variation in calculated forces of less than 5% in most cases (8% in one case) when the location of the rotor within the bearing stator is modified. For the analogous conventional single-point measurements, the results show 15% error and 23% variation.

This content is only available via PDF.
You do not currently have access to this content.