On-line suboptimal feedback control of a trickle-bed absorption column

Abstract

Suboptimal feedback controllers often yield a substantial improvement in control loop performance compared to that achieved with conventional feedback controllers. In this study, the design scheme proposed by Crosby [1] for selecting the best process variables to be used in the feedback control action is further investigated. In this work, the backflow cell model developed by Crosby is used as the model of the trickle-bed absorption column. In the column, an air and C0₂ mixture is contacted countercurrently with tap water. The control problem is to manipulate the inlet water flow rate so as to minimize the changes in the CO₂ gas concentration at the top of the column for load changes. Both in this study and in Crosby's, the load change is a step increase in CO₂ concentration in the inlet gas stream. An integral-time-square-error performance index is used to evaluate the control action. The open-loop optimal state variable and control trajectories to minimize this performance index were previously generated by Crosby and used in the statistical Step-Up procedure to determine the order of importance of state variables. In this work, the open-loop optimal trajectories generated by Crosby are used as inputs to the statistical selection schemes. Both the statistical Step-Up procedure and the Hocking-LaMotte-Leslie procedure are used in evaluating the importance of the state variables in minimizing the performance index. Some of the assumptions and restrictions made by Crosby in the variable selection procedure are modified. Computer simulations are made to tune the gains of the control structures selected. The suboptimal feedback controllers used include the integral mode for the controlled variable plus the proportional mode..