Multivariable and multirate self-tuning control: a distillation column case study

This paper is concerned with the evaluation of self-tuning control algorithms for the multivariable control of distillation columns. Effective distillation control remains a key concern of industry because of the high energy consumption associated with distillation processes. The results presented here are taken from a continuing programme of investigations aimed at improving distillation column control performance. The performance of several multi-variable self-tuning strategies is evaluated by online application to a pilot-scale binary distillation column which is subjected to disturbances of ±25% step changes in feed flow rate. The performance of the multivariable algorithms is compared with that achieved using conventional digital multiloop PI/PID control. The results have demonstrated that use of multivariable self-tuning control, in particular multirate-multivariable self-tuning control, can provide better regulatory performance for feed flow rate disturbances than can be achieved using a multiloop PI/PID strategy.