engleski

Estimation of the induction motor stator current frequency

This paper provides a study of the estimation of the induction motor stator current frequency, based on the time-frequency representations of the nonstationary stator current signal. The study includes following time-frequency distributions: the spectrogram, the Wigner-Ville distribution and the Zhao- Atlas-Marks distribution. Experimental current measurements were conducted during the motor acceleration and the motor operation in the steady state. Measured signal was also corrupted by adding the white Gaussian noise, thus obtaining 4 dB signal-to-noise ratio. After the signal preprocessing, three different time-frequency distributions were applied to the obtained signals. The maximum values of the computed distributions were then used to estimate the instantaneous frequency of the current signal. Calculated estimations were evaluated with respect to the reference instantaneous frequency obtained from the original signal by a zero-crossing algorithm and a cubic interpolation. Results suggested that the Zhao-Atlas-Marks distribution provided the lowest root mean squared estimation error, while the performance of the estimation based on the Wigner-Ville distribution was not acceptable. Moreover, the estimations computed from the original signal and the estimations computed from the noisy signal offered similar performance, with the errors of estimations from the original signal being just slightly lower than those of estimations from the noisy signal.

frequency estimation ; induction motor ; time-frequency distributions ; spectrogram ; Wigner-Ville distribution ; Zhao-Atlas-Marks distribution

nije evidentirano