In-Vehicle System Identification of an Induction Motor Loss Model

Abstract

The influence of induction motor model parameter deviations on field-oriented control performance has been widely investigated. Various methods have been introduced to track variations of magnetic and resistive parameters of the so called T-equivalent circuit model. Online methods on embedded systems exist and are successfully used in modern motor controls. For the use in vehicle motion control and particularly supervisory controls, however, state-of-the-art identification methods may not be applied directly, due to restricted communication interfaces or a limited amount of available measurements at sampling frequencies above the required rates. To cope with these limitations, a moddeling approach is introduced which is based on the equivalent flat system representation of the induction motor, stationary operation conditions, and the incorporation of the vehicle specific field-oriented control strategy. The inclusion of the control strategy allows for derivation of least-square error formulations which are used to identify a selection of induction motor model parameters from low frequency vehicle measurements. An experimental study demonstrates the accuracy of the proposed method and shows how effectively the introduced model can reproduce the measurment of the rms phase current and electric power.