Neural networks based speed estimation of vector controlled asynchronous motors

High dynamic performance that can be obtained from direct current (DC) motors has also proved obtainable from asynchronous motors together with the advances in power semi-conductors, digital signal processors (DSP) and control techniques. With the use of field-oriented control in asynchronous motors, flux and torque can be controlled separately like in DC motors.Because systems like asynchronous motors are generally non-linear, such systems are obstacles to the implementation of successful applications of linear control theory. Naturally, Artificial Neural Networks (ANN) that have non-linear transformation capability have offered a more preferable and alternative solution to non-linear identification and control problems.In this study, ANN was used as sensorless speed observer in controlling indirect vector-controlled asynchronous motors. Data obtained from experiments on DSP application board were analyzed and in the end the success of ANN speed observer was evident.