Abstract
This article concerns a classical problem of how to design an inferential control system for enhancing the performance of distillation composition control. The problem is decomposed into two subproblems: 1) the selection of an inferential model and 2) the selection of a control configuration. This article demonstrates that the better control performance cannot always be achieved with dynamic models although dynamic models can outperform static models from the viewpoint of estimation accuracy. The advantage of using static models results from their inherent feed forward effect. In this article, inferential models are proposed for predicting future product compositions, not for estimating the current product compositions. The control scheme, in which the predicted compositions are used as controlled variables, is termed “predictive inferential control.” The detailed dynamic simulation results show that the proposed predictive inferential control scheme integrated with cascade control works considerably better than the other control schemes. Furthermore, the improvement of the control performance through iterative modeling is investigated.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call