Reducing the parameter dependency of Encoderless Predictive Torque Control for reluctance machines

Encoderless Predictive Torque Control without Signal Injection utilizes internal information of a predictive torque controller to obtain the angle of the rotor saliency. When applying this approach to the reluctance machine the circulation of the current caused by the rotation of the strong saliency needs to be compensated by a parameter dependent term. This paper presents a new technique for compensating the circulation without knowledge of machine parameters. As a result the encoderless control scheme is free of motor parameters and the Predictive Torque Controller gains robustness. The approach is derived mathematically for a reluctance machine and verified with measurement results.