Sliding-mode MRAS speed estimators for sensorless vector control of induction machine

This paper presents two novel sliding mode (SM) model reference adaptive system (MRAS) observers for speed estimation in a sensorless-vector-controlled induction-machine drive. Both methods use the flux estimated by the voltage model observer as the reference and construct SM flux observers that allow speed estimation. Stability and dynamics of the two proposed SM flux observers are discussed. The observers are compared with the classical MRAS observer. The proposed estimators seem very robust and easy to tune. Unlike the classical MRAS, the speed-estimation process is based on algebraic calculations that do not exhibit underdamped poles or zeros on the right-hand plane. Simulations and experimental results on a 1/4-hp three-phase induction machine confirm the validity of the approaches.