Nonlinear state estimation and control of a refinery debutanizer column

This article concerns the issue of designing a nonlinear model-based controller for a refinery debutanizer column. First, a nonlinear adaptive state estimator/observer (ASE/ASO) was developed based on a simple observer model structure that mainly consists of only two component balance equations around the condenser-cum-reflux drum and the reboiler-cum-column base. In this observation approach, vapor flow rate of component nC5 (heavy key) leaving top tray, liquid flow rate of component nC4 (light key) leaving bottom tray and distribution coefficient of component nC4 in the reboiler were considered as extra states with no dynamics. Despite process/model mismatch, the proposed state observer estimated the required states of a simulated debutanizer column precisely. Mainly due to the design simplicity, negligible computational effort and fast convergence, the observer is recommended for online implementation. In the subsequent part, the globally linearizing control (GLC) structure, which was consisted of a nonlinear transformer (input–output linearizing state feedback), a linear external controller and an adaptive state observer, has been synthesized. The hybrid GLC–ASO control algorithm provided promising performance compared to the proportional integral (PI) controller that has been thoroughly investigated on the complex refinery debutanizer column.