Sensorless speed control of induction motors using Adaptive Neural-Fuzzy inference system

This paper presents a new approach to the sensorless speed control of induction motors (IM). The new technique uses an Adaptive Neural Fuzzy Inference Systems (ANFIS) as a rotor speed estimator to avoid using mechanical sensor. This makes the reference model free of pure integration and less sensitive to stator resistance variations. The ANFIS has been trained offline to estimate the rotor speed in wide range of operation and has been implemented online to perform filed-oriented control of IM. The data for training the ANFIS are obtained from experimental measurements based on the current model, avoiding voltage and flux sensors. This has the advantage of considering all drive nonlinearities. Control loops of rotor speed and stator currents employ Fuzzy-PI (FPI) controller. The input-output scale factors of all FPI systems are tuned using Genetic Algorithms (GA) to achieve better results. Experimental results prove that the proposed method has high precision and good dynamic quality in speed estimation and control.