Real-time estimation of the electric parameters of an induction machine using sinusoidal PWM voltage waveforms

This paper studies the parameter estimation of induction machines. The problems associated with the parameter estimation of induction machines are analyzed through sensitivity functions. Then, models and a procedure to estimate a set of four basic parameters (r_s, σl_s, l_s, and τ_r) are proposed. The discrete-time parameter estimation models are written in the δ operator specifically because it provides good numerical properties at high sampling rates, where the discrete-time model approaches its continuous-time equivalent. This feature permits the direct estimation of the continuous-time parameters. The main feature of the proposed procedure is the possibility to estimate the parameters by using sinusoidal signals, and without the measurement of the machine speed. The experimental results obtained with the proposed estimation procedure are presented and demonstrate that it is possible to map the parameters in terms of the operating conditions of the induction machine