Sensorless speed estimation for inductions motors using slot harmonics and time-based frequency estimation

This paper proposes a new method for sensorless speed estimation in three phase induction machines based on stator current signal analysis and Sinusoidal Signal Approximation algorithm. Rotor speed is estimated through identification and estimation of slot harmonics frequencies, present in the acquired current signal. In order to estimate the speed with accuracy, precision, and update time similar to a regular digital tachometer (contact or optical), the proposed method applies the Sinusoidal Signal Approximation algorithm in the process of estimating slot frequency, as an alternative to the Fast Fourier Transform (FFT) and spectral estimation techniques. The method was validated by an experimental setup where estimated speed from a 4-pole induction motor, running at different load conditions, is compared to speed measured by a digital optical encoder.