An adaptive structure based sensorless position estimator for permanent magnet machines in aerospace applications

A sensorless position estimation scheme based on a modified model reference adaptive structure, and designed specifically for a permanent magnet machine used in an aerospace application, is developed and verified experimentally. The technique, based on two flux models, avoids the need for differentiation and incorporates an adaptation method that compensates integrator wind-up and drift. The method is computationally efficient and is robust to parameter variation. Experimental results show position estimation is accurate to within 3deg_mech enabling the machine to operate with 1% torque ripple.