Accurate Torque Control for Induction Motors by Utilizing a Globally Optimized Flux Observer

Abstract

High-precision torque estimation and control of induction motor drives is an important research field due to the extensive use of these motors in torque controlled applications, for example, in electric vehicles. The open-loop torque control performance depends strongly on the accuracy of both the motor model and the flux estimation. To address this appropriately, an adaptive Kalman filter with offline parameter and observer design optimization is proposed in this article. Thereby, the basic induction motor model is extended by magnetic saturation, iron losses, and skin effect influences. All uncertain motor model and observer configuration parameters are offline identified by a global optimization technique, namely particle swarm optimization. The identification utilizes a comprehensive data set consisting of test bench measurements and leads to an optimized observer enabling precise torque estimation and control. In experimental validation, for both torque estimation and control, the root-mean-square error is below 1 % of the nominal torque over the entire operating range. With the help of an accurate gray-box inverter model for phase voltage estimation and a speed-adaptive Kalman filter tuning scheme, the proposed observer is able to operate also at slow speeds including standstill.