Autotuning of decentralized and centralized controllers for square and nonsquare mimo systems using relay feedback

Abstract

Industrial chemical processes are mostly large, challenging and newlinenon-linear in nature. Each manipulated variable affects more than one or all newlinethe controlled variables because of the presence of interaction. As newlinethe intricacy of the process increases, it becomes more complicated to control newlinethe process accurately. In the process industry, MIMO systems having equal newlineand unequal numbers of inputs and outputs are common. A square system has newlinean equal number of input variables and output variables; whereas a newlinenon-square system has an unequal number of input variables and output newlinevariables. Based on the interaction among the input variable and output newlinevariables, the control structure is chosen for MIMO systems. newlineMultivariable control design methods lead to high complexity due newlineto the inherent interaction between input and output behavior, parameter newlineuncertainty, unsteady state process behavior and dynamic nature of the newlineprocess. The implementation of controllers with simpler forms is strongly newlinemotivated, in practice. The conventional methods of relay autotuning do not newlineconsider the presence of higher-order harmonics for the controller design. newlineHence, it affects the performance of the controller. A proper method is needed newlineto compute the improved ultimate gain value. In general, industry frequently newlineprefers simple decentralized PI/PID controller design. newline newline