NEURO-GENETIC OBSERVER SPEED FOR DIRECT TORQUE NEURO-FUZZY CONTROL OF INDUCTION MOTOR DRIVE

In this work, we applied artificial intelligence techniques to improve the dynamic performance of sensorless direct torque control (DTC). First, we replace the conventional comparator and selection table applied to the DTC induction motor with a neural-fuzzy controller to learn the advantages of both methods; learning capacity of the first and readability and exibility of the second. This evaluation is obtained using the electromagnetic torque error, the error module and angle of the stator flux vector. Then we propose a new estimation method rotor speed using an adaptive neural observer. The error between the desired state variable and the actual state variable of a neural model is back propagated to adjust the weights of neural model, so that the actual state variable tracks the desired value. Finally, the proportional and integral controller gains are optimized by a genetic algorithm to ensure the optimum performance of rotor speed. The simulation results show the effectiveness of control methods available.