CONSTRUCTIVE ESTIMATOR DESIGN FOR BINARY DISTILLATION COLUMNS

The problem of jointly designing the estimation structure and algorithm for binary distillation columns is addressed within a constructive framework that combines structural and robustness concepts in the light of system characteristics. The model sensor location and number, the innovated-noninnovated state partition, and the model dimension are regarded as structural design degrees of freedom. The geometric and extended Kalman filter estimators are regards as algorithms to perform the data assimilation task. The high-order Lie derivation applicability obstacle of the geometric estimation technique is removed, and the equivalence between geometric and extended Kalman filter is established. The methodological findings and the estimator functioning results are illustrated and tested with experimental data.