Design of multivariable controllers Using relay feedback responses

Abstract

Most of the unit operations in process industries have more than newlineone control loop Systems with more than one control loop are known as newlineMulti Input Multi Output MIMO or multivariable systems Interaction newlinebetween input and output causes a manipulated variable that affects more than newlineone controlled variable therefore multivariable processes are difficult to newlinecontrol at independent set points As loop interactions can lead to instability newlinewhich has to be taken into account for proper choice of input output pairing in newlinecontrol system design so as to minimize the interactions newlineIn this research work it is proposed to develop a comprehensive newlinemathematical modeling framework for the design of multivariable control newlinesystem using relay feedback responses Also proposed to design the newlinerespective multi loop controllers and to describe the system behavior during newlineset point and load changes In order to attain better performance input output newlineinteractions are reduced and the accuracy level is maintained as per the design newlinespecifications newlineDue to interactions between input and output closed loop newlinemultivariable systems yield undesired responses To characterize these newlineresponses a novel approach is followed to model off diagonal closed loop newlinetransfer function of low order MIMO systems in time domain using ideal as newlinewell as biased relay feedback tests Analytical expressions are derived using newlineideal and biased relay feedback tests for square transfer function models of 2 newlineby 2 and 3 by 3 MIMO systems having individual transfer function elements newlineof First Order Plus Dead Time FOPDT structures After simulating 2 by 2 newlineand 3 by 3 MIMO systems relay responses are obtained and validated with newlinetheoretical model responses The various MIMO benchmark processes are newlinediscussed in the later section newline newline