Design of Decentralized Fractional Order PIʎ Controller for Pilot Plant Binary Didtillation Column

Abstract

This paper presents the application of fractional order operators to improve the control quality of multivariable systems. The basic ideas of this tuning method are based, in the first place, on the existed tuning methods for setting the parameters of the decentralized fractional order PIʎ controller for ʎ=1, which means setting the parameters of the classical decentralized PI controller, and the minimum integral criterion by using Particle Swarm Optimization (PSO) algorithm for setting the fractional integration action order ʎ. The integral criterion is formulated to improve the dynamic response of the system, while causing a good decoupling between control loops. The Distillation Column, which is a multivariable system with two inputs and two outputs (TITO), in a decentralized control structure, is analyzed. Simulation results are presented to show the control quality improvement of this proposed decentralized fractional order PIʎ controller tuning method compared to the decentralized PI controller tuned using any existed tuning method.