Investigations on model order reduction and design of speed controller for electric drive systems

Abstract

Speed controller design of electric drive systems gives a challenge to newlinemany researchers due to complexity in drive system. The speed of the electric newlinedrives can be controlled from below and up to rated speed using a converter. newlineThe converter firing circuit receives signal from controller and then the newlineconverter gives variable voltage to the armature (or rotor) of the motor for newlineachieving desired speed. There are two control loops, one for controlling newlinecurrent and another for speed. Conventionally the Proportional Integral type newlinecontroller was used by Symmetric Optimum approximation method. This newlinemethod eliminates the effects due to disturbance very rapidly compared to newlineother optimum techniques such as linear or modulus optimum techniques. newlineApproximation method was used based on the assumption that the time newlineconstants of the systems are equal. The controller used here is Model Order Reduction (MOR) technique newlinewith Genetic Algorithm (GA) based Proportional Integral (PI) type which newlineremoves the delay time and provides fast response. In modeling electric drive newlinesystems, the order of the drive system gives an idea of the measure of newlineaccuracy of the modeling of the system. The higher order system model will newlinebe more accurate in describing the drive system. But in several cases the newlineamount of information contained in a complex model may obfuscate simple, insightful behaviours, which can be better captured and explored by a model newlinewith a much lower order. newlineThus, by approximating a higher order system to a suitable lower order newlinesystem, a much better understanding of the system is reached. Hence, the newlineprocess of model order reduction involves studying the properties of a newlinecomplex dynamic system in application for reducing its complexity, while newlinepreserving (to the maximum possible extent) its input-output behaviour. newlineDepending upon the practical application, certain specific properties of the newlinehigher order complex system have to be preserved in the reduced order newlinemodel. newline