(482f) Design of Advanced Controllers for Post-Combustion CO2 Capture Process Integrated with a Supercritical Pulverized Coal Plant
Design of Advanced Controllers for Post-Combustion CO2 Capture Process Integrated with a Supercritical Pulverized Coal Plant
Qiang Zhang, Debangsu Bhattacharyya, and Richard Turton
Department of Chemical Engineering, West Virginia University,
Morgantown WV, 26506, USA
Focus of this presentation will be on development of advanced control strategies for an MEA-based, post-combustion CO2 capture and compression unit integrated with a 550MWe commercial supercritical pulverized coal power plant. The steady-state process design indicated that multiple parallel trains of CO2 capture processes are required for treating the flue gas from a commercial sized power plant. Thus the dynamic model used in this work comprises of two parallel trains of CO2 capture processes integrated with the steam-side of a supercritical coal fed power plant. The dynamic model of the CO2 capture unit is rigorous with user models incorporated into the Aspen Plus Dynamics® environment for improving the model accuracy under design and off-design operations.
Linear model predictive control (LMPC) is found to provide good control performance of the CO2 capture rate at a set-point of 90% in the face of typical disturbances such as change in the flue gas flow rate and composition. However, when the integration of the power plant and the CO2 capture processes is considered then the situation becomes more complicated. For example, for a load-following power plant with sliding pressure operation, as the flue gas flowrate changes in response to the load changes, there is a corresponding change in the pressure and temperature of the low pressure steam available for the CO2 stripper reboiler. Since the time constants for the CO2 capture process are much longer than for the power plant and the loading of the MEA is strongly influenced by the stripper reboiler temperature the response of the overall system becomes highly non-linear. In addition, considerable time delay in the process and strong interaction between multiple variables result in a challenging problem. The control problem becomes far more difficult when uncertainties in the input and outputs as well as gradual loss in the absorber and stripper performances are considered. Performance of the LMPC is found to be poor under these scenarios. A number of advanced controllers such as nonlinear MPC (NMPC), H∞ robust controller, and robust MPC are designed with novel strategies for uncertainty quantifications and rigorous analysis of tradeoff between performance and stability. In addition, a number of strategies has been developed to reduce the computational expense of the control system.