Sensorless control of Interior Permanent Magnet motor using a Moving Horizon Estimator based on a linearized motor model

This paper presents a sensorless control solution for Interior Permanent Magnet synchronous motor using a Moving Horizon Estimator. The active flux concept is used to describe the motor model. The position estimation problem is non linear for the considered type of synchronous machine, thus a linearization of the model is performed using a Taylor first order approximation. Thanks to the linearization, the problem assumes the form of an equality constrained Quadratic Programming that can be solved directly, without using any iterative method. The estimator is real-time implemented, coupled with a standard PI speed controller and a Model Predictive Control for the current loop, highlighting the feasibility of this innovative control architecture. Several steady state and dynamic tests are provided in this work to show the drive performances. In details, steady state operation, torque step, speed inversion and a speed ramp are included in the study.