Statistic Errors of Different Inform Evaluation Methods Applied to Magnetic Bearings

Active magnetic bearings offer many advantages compared to conventional bearings. But due to economic reasons their field of application is relatively small. Self-sensing magnetic bearings can contribute to a significant decrease of the system costs, of course at limited accuracy in position detection compared to systems with sensors. The later on described INFORM method is one approach to self-sensing magnetic bearings. The position depending reluctance is measured through high frequent voltage injections and the resulting current slope. This paper discusses how different evaluation methods capable of calculating the current slope from the raw measurement data, differ in sensitivity and noise rejection. Therefore first the theoretical background of the used self-sensing method is explained. Next the error propagation of different evaluation methods is analysed and finally compared to measurements.