Skip to content

Active magnetic bearings (AMBs) are currently in operation in many commercial applications including industrial turbo machines [1] as well as turbo molecular pumps. Since AMBs have important advantages over conventional bearings, new applications of this technology are developed yearly in fields such as commercial heating and air conditioning and space based rotary equipment. A notable advantage of AMBs is the ability to act concurrently as a means of bearing support and as a shaft force sensor. Sensing the force (or reaction) that is applied to the AMB stator through the shaft rotor is achieved by measuring information about the AMB’s magnetic field. This capability has the potential to facilitate advanced health monitoring techniques as well as facilitate improvements in the continuous control of processes used in the production of microfibers and in detailed metal processing [1]. Quality control for such applications can be directly linked to real-time rotor force evaluation. Some researchers have examined the force sensing capability of magnetic bearings by using Hall effect sensors in air gaps between bearing stator and rotor components. In this approach, measured magnetic flux densities are used in force equations to determine applied bearing reactions. Other researchers have employed methods based on measuring magnetic flux density in the bearing electromagnets from electric current data to determine applied bearing reactions, thus eliminating the need for additional sensor hardware. Limitations with both methods include assumptions in the force model associated with leakage, fringing, heterogeneity of magnetic material, and non-uniformity of axial and radial air gaps due to manufacturing tolerances. Misalignment and thermal issues encountered in the field also affect model accuracies [2]. Most magnetic force models assume that these effects are negligible and/or non-measureable. The research presented in this paper introduces a new in situ approach for bearing system identification in an effort to improve magnetic force models and is based on measurement of current and position perturbations. The paper also includes initial proof-of-concept static verification of the proposed method.

Author: | Published:
Booktitle: Proceedings of ISMB12