Practical considerations for sensorless induction motor drive system

Speed information is necessary for high performance vector controlled induction motor drives. This information is generally provided by a sensor which spoils the ruggedness and the simplicity of the induction motor. This paper describes the practical considerations for an implementation of the newly developed speed sensorless drive based on neural networks technique. The back-propagation neural networks technique is used to provide a real-time adaptive identification of the motor speed. The estimation objective is defined in terms of a reference or desired trajectory that the neural networks model output should match or track as closely as possible. The backpropagation algorithm is used to adjust the motor speed so that the neural model output follows the desired trajectory. This backpropagation mechanism forces the estimated speed to follow precisely the actual motor speed. The zero speed crossing phenomena is also described and experimental results are presented and analysed