Soft Starting of an Induction Motor using Adaptive Neuro Fuzzy Inference System

Abstract

Induction motors are used in industry in variety of applications such as blowers, fans, compressors, pumps, mixers, crushers and grinders etc. Soft starters, involving voltage controllers, are used for starting and speed adjustment of induction motors. This paper presents a neural network based soft starter which controls the speed by the adjustment of firing angles of thyristors. The suggested technique eliminates the starting torque pulsations by triggering the back to back connected thyristors. The control strategy is implemented using microcontrollers and neural networks and performance analysis of the system is carried out with the help of hybrid model of induction motor. The results obtained are satisfactory and promising. The advantage of such a soft starter is its simplicity, stability and high accuracy as compared to the conventional soft starter which uses mathematical calculations of firing angle which is a complex and time consuming task especially in on-line control applications.