Induction machine parameter tracking from test data via PWM inverters

This paper presents a new method that can track induction machine parameters for different operating conditions using the machine power converter. The maximum likelihood estimation method is used for identification. The machine is tested using the power converter in three different modes. In the first test, while the machine is at standstill, a voltage pulse is applied to the stator terminal and then the leakage inductance and the combined stator and rotor DC resistance are estimated. In the second test, the power converter is used in current control mode and the stator voltage is measured and then the stator resistance and the mutual inductance are estimated. In the third test, the stator voltages and currents are measured while the machine is running at different speeds. These signals are used to simultaneously track the rotor resistance as affected by frequency and the mutual inductance based on the parameters obtained in the first and second tests. In spite of noisy operating conditions caused by the inverter, the simulated model response matches the measured response well.