Virtual off-line auto calibrating position and current sensors for switched-reluctance-motor drives

To obtain high-performance motor drives, modern control strategies should be employed in the drive system. These techniques are inherently dependent to the measurement devices, or sensors, operating properly. In particular, switched-reluctance-motor drives need position and current sensors to demonstrate full torque-control capability. This requirement yields more expensive and complex drive systems. On the other hand, when these sensors fail, the controller should be able to recover and handle the proper operation of the drive system. This fact necessitates back-up systems to support the operational drive in case of sensor failure, especially for life-dependent applications such as automotive systems. In the paper, novel virtual position and current sensors are designed based on modelling of magnetic double-saliency characteristics in switched-reluctance-motor drives. This approach does not suffer from unpredicted variations of spatial distribution of magnetic field since a practical autocalibration method is designed to tune the virtual sensors. The proposed virtual autocalibrating sensors are implemented fully on an experimental SRM drive based on a fixed-point digital signal processor. The experimental results verify the capability and high performance of the proposed approach