Multivariable Nonlinear Pid Controller for Distillation Column
Loading...
Date
item.page.authors
Journal Title
Journal ISSN
Volume Title
Publisher
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