An adaptive neuro-fuzzy based speed sensorless induction motor drives

A new algorithm for speed observer based on Model Reference Adaptive System (MRAS) is proposed for high performance induction motor drive. It uses stator current error based MRAS speed observer. The reference model of the stator current error based MRAS is the measured stator current components and the adaptive model is neuro-fuzzy based stator current observer. The adaptive model also needs the use of rotor flux. This rotor flux is estimated by neural network based rotor flux observer. Since both the reference model and the adaptive model are free from the use of pure integrator, there will not be any problem of saturation and d.c. drift in the observed state variables. Also the neuro-fuzzy based stator current MRAS speed observer is insensitive to parameter variations. Five tests are performed which shows effectiveness of the proposed scheme. The results shows faster and better response of indirect vector controlled induction motor drive system with neuro-fuzzy based stator current MRAS speed observer.