Nonlinear Proportional-Integral-Derivative Control of a Multi-input Multi-output Distillation Column Process Incorporating Genetic Algorithms

Abstract

Distillation columns are integral components of chemical processing plants. Their inherent multi-input multioutput (MIMO) configuration, characterized by interdependent inputs and outputs, poses significant challenges for controller design. This paper presents a comprehensive comparative analysis of various control strategies for managing the top and bottom composition for a distillation column, with a focus on the impact of decoupling. Employing the Wood-Berry distillation model in order to facilitate the separation of methanol from water, this study examines the effects of feed flow rate disturbances. With the purpose of mitigating or eliminating the cross-interactions between inputs and outputs, a decoupler was designed in order to transform the analysed system into two independent single-input single-output (SISO) systems. Subsequently, nonlinear proportional-integral-derivative (PID) controllers were developed for controlling the distillation column, their parameters being optimized through a genetic algorithm. The obtained simulation results demonstrate that the proposed control system significantly reduces the total error in comparison with the employed conventional PID control system, particularly in terms of setpoint tracking and disturbance rejection.