Online-Identification of the Induction Machine Parameters Using the Extended Kalman Filter

The identification of the machine parameters can be executed a priori or during run-time (online) depending on the case of application. If just a controller should be commissioned, simple test bench experiments should be undoubtedly preferred. Nevertheless, online-identification approaches must be prioritized for adapting or monitoring purposes during normal operation. In this paper, the first step towards the monitoring of the induction machine is presented. The performance of the well-known extended Kalman filter is analyzed for different cases of known and unknown machine parameter combinations. It is outlined by simulative and test bench results that the extended Kalman filter is well-suited for the machine parameters identification during run-time. This is the main prerequisite for subsequent analyses of induction machine fault scenarios.