Sensorless explicit model predictive control of permanent magnet synchronous motors

Abstract

The present paper deals with sensorless model predictive control of permanent magnet synchronous motors. The proposed explicit controllers consist in precomputed (optimal) state feedbacks that are selected according to the measured state using binary search-trees. This type of controller is well suited to obtain very fast control algorithms that impose high dynamic performance. The control scheme is based on two cascaded explicit model predictive controllers, one for the torque control another for the speed control. The nonlinearities of the motor dynamics that are due to the speed are taken into account into the torque controller derivation. The resulting adaptive closed-loop system presents a high dynamic performance for all operating points. A rotor speed and position observer is used to orient the control variables on the rotor flux without position sensor. The overall control scheme allows optimal use of the inverter and drive dynamic capabilities. The proposed concepts are validated experimentally.