Project Methodology For Multivariable Optimal Controllers And State Observers Applied To Electrical Drive Systems.

Abstract

This works presents, the problem of estimate of states of multivariable systems,
using complete order state observer optimal discretes. For study, it is considered the twophase
mathematical equivalent model of the three-phase asynchronous machine, their due to
your characteristics (sixth-order model, non linear and hardly coupled), making possible to
generalize for application in other systems. Due to the computing time necessary for
discretization of the model, the procedure is realized off-line, resulting in a discrete model
containing algebraic relations. The feedback matrix (gains) it is calculated off-line, using
technique of optimal control. The gains are calculated for various frequencies of machine
operation, resulting at various gain matrixes operating in a gain scheduling.