Speed Control of Induction Motor through PI-ANN Controller

Abstract

An induction motor is also called an asynchronous motor – it is an AC electric motor used extensively in applications. This particular type of motor is utilized in the majority of industrial and commercial electrical applications because simple, durable and affordable. Induction motors find application in domestic equipment, automobiles, IT equipment’s, industries, public life equipment, transports, aerospace, defense equipment, power implements, toys, vision & sound equipment & health & medical devices. This has become possible by the following reasons; high efficiency, fast response, light weight, precise and accurate control, high reliability and maintenance free operation, construction with no brushes, high power density and small size. Induction motors found its way easily into industrial systems because of its high speed. And so, when working to achieve stability and productivity of a system it is necessary to stabilize the actual speed of the AC motor existing in the automation system with reference to the set speed and maintain a speed which is higher than the load speed. AC motor is used in a number of industrial applications where large variability in speed and torque is demand. In the past, traditional feedback controllers like PI controller have been applied extensively in industrial processes but tends to exhibit certain shortcomings in handling nonlinearity and parameter fluctuations as well as load disturbances. In response to these challenges, this thesis proposes a new solution where Proportional Integral control is combined with Artificial Neural Networks to provide accurate control of the speed of induction motors. The proposed PI-ANN controller tries to integrate the concept of PI control and ANN to adjust the control parameter when changing the parameter of motor for optimum performance is needed. The system is simulated and validated on a squirrel cage induction motor using a voltage source inverter for voltage regulation. By training the ANN to change its behavioral pattern in terms of motor speed, load condition and system dynamics the proposed method far out performs the traditionally used PI controller. Records from the simulation studies, as well as experimental findings, confirm that the proposed speed control algorithm based on the PI-ANN excels the basic PI controllers in various aspects such as lesser overshoot, quicker time response, and better stability when the electrical load is variable. This class of hybrid control strategy provides a viable solution in high-performance motor control applications and hence forms a platform for enhancing more control strategies for the induction motors in the area of intelligent control systems.