Multivariable Nonlinear Pid Controller for Distillation Column

Abstract

ABSTRACT newlineProcesses which are being controlled by only one manipulated variable are called Single newlineInput Single Output (SISO) systems. Processes with more than one input and output are newlineknown as Multi Input and Multi Output (MIMO) systems. It is also called as Multivariable newlinesystems. Many industrial processes are multivariable systems. The control of MIMO system newlineis difficult since there is strong interaction between the inputs and outputs. However, control newlineof nonlinear industrial process is a risky and challenging task. Then the efficient tuning of newlinecontroller parameter is essential. newlineIn the process industry, PID controllers are commonly used for control applications. newlinePID controllers are simple and easy to construct. It provides more flexibility and stability newlinewhile controlling the processes. The determination of proportional, integral and derivative newlineconstants KP, Ki and Kd are known as tuning of PID controller. Tuning of PID controller newlinegains are easy if the system is a linear system. But many industrial plants are nonlinear newlinesystems. They have the problems such as higher order, instability, time delays, harmonics, newlinepoor damping and time-varying dynamics etc. Hence the perfect tuning of multi-loop newlinemultivariable PID controller for nonlinear process is a challenging work. newlineAn interesting MIMO control problem is given by distillation column process. It s a newlinehighly nonlinear process and serves as an example for ill conditioned plant. At present lot of newlinetuning techniques such as differential evolution, differential evolution combined with chaotic newlinezaslavskii map, μ- synthesis method, H loop-shaping method are used in the linear PI and newlinePID controllers to solve the continuous non-linear optimization problem. But still the linear newlinecontrollers are not fit for the nonlinear control problems since they will take much time to newlinetune in practical cases. Control action must be taken immediately after the occurrence of newlineuncertainty. So nonlinear controllers like self tuning controllers, fuzzy controllers or adaptive newlinecontrollers are added with