Multirate self-tuning predictive control with application to binary distillation column

Abstract
An adaptive controller for multi-input multi-output (MIMO) systems is presented, which is able to modify repeatedly the control variables between successive output measurements. This feature makes the proposed approach particularly appealing for the control of composition products of distillation columns when expensive and slow chromatographs are used. The control algorithm is derived by minimizing the predicted error signal, as well as the control increments, over a future control horizon which is suitably selected. Applications to a simulated non-linear model of a binary distillation column demonstrate the effectiveness of the proposed method.