Multivariate adaptive control of a pulsed liquid-liquid extraction column

This engineering research thesis reports the study and development of the automatization of a liquid-liquid extraction column. The author reports the hydrodynamic study, optimisation, and selection of command variables of such a column. Then, he presents the adaptive command with generalised minimum variance, and an algorithm of adaptive command with reference model for the simultaneous performance of regulation and tracking objectives. Results obtained for a single-variable adaptive command of the extraction column are presented, and the implementation of two adaptive command loops is reported (the first loop concerns the column hydrodynamic control, and the second one concerns the matter transfer control). In the last part, the author addresses the implementation of a multivariate adaptive command algorithm with generalised minimum variance. The objective of the command is to maintain the column in its optimal operation range while optimising the quantity of used solvent. Parameters of the model adopted to represent the dynamic operation of the column are assessed in real time.