Process design and control studies of an elevated-pressure air separations unit for IGCC power plants

In this study, a supervisory controller for an ASU (Air Separations Unit) has been designed based on a MPC (Model Predictive Control) approach. This has been applied to a detailed pressure-driven Aspen Plus/Aspen Dynamics model to understand the performance of a multivariable MPC when adopted for such a complex process with critical targets. In addition, an effective PI-based secondary-layered controller design has been proposed for open-loop stability and maximizing oxygen yield. The rigorous Aspen model is interfaced with Matlab/Simulink and has been used as a “real plant” when performing the step tests for the identification of the simplified linear model and application of supervisory control. The effectiveness of the designed controller has been proved through the comparison between the linear MPC approach, that handles absolute and rate-of-change constraints, and a more conventional control configuration. The results show a significant reduction in overshoots and settling time during load variations. The paper clearly shows how the MPC approach for a supervisory control layer is reliable, easy to design and of real value for practical purposes.