Design of NDT and PSO based decentralised PID controller for wood-berry distillation column

Abstract

In this paper, a multi variable PID controller design based on Non-dimensional Tuning (NDT) and Particle Swarm Optimisation (PSO) methods are proposed. Multivariable system has to be controlled effectively to achieve high product quality and to minimize energy consumption requirements of the process industries. NDT method for tuning PI controllers for First Order Plus Dead Time (FOPDT) process is used to minimize the integral of absolute error (IAE) for a step change in the set point. PSO method is a robust stochastic optimization technique based on the movement and intelligence of swarms. By considering the Two Input Two Output (TITO) process, first the TITO process is decoupled into two higher order decoupled processes through a decoupler matrix. An FOPDT approximation model is introduced for model reduction. The higher order decoupled processes is determined by using FOPDT model. The FOPDT Model is calculated by using nyquist plots at particular points. Then, a decentralised PID controller is obtained using the NDT and PSO method. Decentralised PID controller is also called as multiloop PID controller. An wood-berry distillation column of TITO system is given to demonstrate the effectiveness and accuracy of the proposed algorithm. It is found very clearly that the PSO based controller reduces the overshoot, settling time, rise time and peak time. Hence PSO outperforms that of the traditionally tuned controller with NDT criterion. These tuning methods has been tested in MATLAB simulink environment.