An improved off-line parameter identification algorithm for induction motors

In order to implement self-commissioning of the frequency-variable speed-adjustable system, an improved off-line parameter identification algorithm for induction motors (IMs) is proposed in this paper. Most of motor parameters which directly related to speed sensorless stator-flux-oriented vector control are estimated using the improved algorithm. Stator resistance is estimated by modified DC test. The equivalent control circuit of the DC test is obtained by analyzing switch states of the inverter. Rotor resistance and total leakage inductance are estimated by single-phase locked-rotor test. The stable operating condition of the single-phase test is also deduced by the phasor graph of the induction motor. Stator self-inductance is estimated by constant V/f no-load test. Stator frequency is compensated to eliminate possible oscillation of stator currents in V/f no load test. The stator voltages are constructed by DC voltage and PWM signal’s duty ratios. The voltage errors due to inverter’s dead time are also compensated according to motor current polarity. The experiments of parameter identification and speed sensorless vector control are performed on two induction motors with different rated power. All results indicate that the identification results are insensitive to DC voltage and furthermore the exactness of the estimated parameters can meet the requirements for the speed sensorless vector control.