Sensorless efficiency optimized induction motor drive with neuro-fuzzy control

Sensorless vector-controlled induction motor drives with neuro-fuzzy based intelligent control show tremendous future promise for high performance applications. This project describes a sensorless induction motor drive with stator flux oriented direct vector control particularly suited for electric vehicle type applications. The drive starts at zero speed and goes to field-weakening mode for all the four quadrants of operation. A start-up method has been developed that gracefully blends into vector control. Speed sensorless vector control at low speed has been possible by a novel method of flux vector estimation. The stator resistance variation effect on the vector control has been compensated by a new on-line fuzzy estimation technique. The efficiency of the drive at light load steady state condition has been enhanced by on-line neuro-fuzzy based search method. The total system has torque and flux control in the outer loops with synchronous current control in the inner loops.

All the proposed control and estimation strategies have been analyzed and then the complete drive system has been simulated by both SIMNON and MATLAB-SIMULINK languages to evaluate the drive performance. The control and estimation strategies have been implemented by state-of-the-art digital signal processors. A laboratory experimental system has been designed for a 5 hp drive and then tested extensively.