Optimal Operation of Kaibel Columns

Abstract

Kaibel distillation column performs the separation of 4 products in a dividing-wall arrangement. This column provides significant energy and capital savings compared to conventional arrangements. The focus in this thesis is on optimal operation and control of these columns with two operational objectives: maximizing the product purities with fixed boilup, and minimizing energy with specified product purities.

The first part deals with optimal design and operation of Kaibel column both by using Vmin diagram as a shortcut tool and rigorous simulations. Throughout the thesis, it’s assumed that vapour split is a degree of freedom. It is shown that the energy saving potential may be lost if the column is operated away from its optimal point, due to disturbances.

A control structure is designed based on the plant-wide control procedure by Skogestad. This is done for the case of maximizing product purities. For the second operational objective, a soft sensor is designed by reformulating the self-optimizing control method. The performance of the estimators depend on whether it is used for monitoring (open-loop) or for closed-loop control applications. In this work, estimators are derived for each case. The approach is to minimize the estimation error for expected disturbances and measurement noise. The main extension compared to previous work is to include measurement noise and to provide explicit formulae for computing the optimal static estimator. The results are compared with standard existing estimators, e.g. Partial Least Squares (PLS).

One of the drawbacks of combining measurements is that measurements with slower and faster dynamics has weights with negative and positive signs and this may lead to an RHP zero. Different possibilities to overcome the inverse response problem are studied. In the end, control of Kaibel column by estimating the compositions of the light and heavy keys at the ends of the prefractionator is studied.