Sensor less low speed control with ANN MRAS for direct torque controlled induction motor drive

This paper presents speed sensorless direct torque control (DTC) of induction motor using Artificial intelligence (AI). The artificial neural network (ANN) MRAS-based speed estimation is used. The error between the reference model and the neural network based adaptive model is used to adjust the weights by on-line Back propagation (BP) training algorithm. To improve low speed operation, the drive uses a full-order sliding-mode without rotor speed adaptation i.e., Due to dual implementation of stationary and rotating reference frames, the proposed observer does not require any speed adaptation and is inherently sensorless. Unlike speed adaptive observers, the proposed observer is immune to speed estimation errors; thus, its performance at very low speed is improved significantly. Experimental results for the sensorless drive are presented. Very low-speed-operation (5 rad/s) capability of the drive with the sensorless observer is demonstrated.