Design of Robust Decentralized Control Systems for Distillation Columns

This paper is devoted to the development of a decentralized control system design method with improved robustness, applicable to distillation columns control problem. The aim is attained by solving the following tasks: selection of representative models of columns; development of an effective transfer function identification algorithm; development of a decentralized controller design algorithm, which takes into account the cross-links impact; verification of the developed method for the selected models of columns. The most important result is the developed design method, which includes two algorithms. The first algorithm can be used to identify simplified models of the effective transfer functions. The second algorithm makes it possible to design a decentralized control system based on the ratio between the diagonals of the original model and effective transfer functions. Modified PI controller is used. The achievement of the distillation column control system robustness to the synchronous technologically conditioned deviation of the model parameters by up to three hundred percent of the nominal value is the paper significant result. The digital implementation of the designed controllers is simple. The analysis of the literature shows that none of the previously proposed methods for designing decentralized automatic control systems solved such a problem. The efficiency of the developed algorithms is illustrated using the known models of dynamics of distillation columns. The analysis of limitations of the method as well as its significant differences to the Biggest Log Tuning and H∞ loop shaping methods is presented.