Health monitoring of a brushless permanent magnet machine using an adaptive Kalman filtering approach

A feasibility study has been conducted to demonstrate the applicability of utilizing an adaptive Kalman filter as a health monitor for the brushless permanent magnet machine. After developing a detailed state variable model of the machine, the discrete Kalman filter is formulated for the motor system. Results from simulation and laboratory experimentation demonstrate the efficacy of the model and filter. Next, the adaptive filter is developed using a parallel implementation of discrete Kalman filters. A strategy for applying this filter technique to detect and predict failure modes for the machine is presented. Results from simulation and laboratory experimentation are presented which support the concept of using this approach as a health monitor for the machine.