Integração de controladores preditivos de modelos econômicos e otimização em tempo real aplicados uma planta de captura de carbono

Abstract

The mitigation of carbon dioxide (CO₂) emissions resulting from the combustion of fossil fuels represents one of the major technological and environmental challenges of modern society. In this context, Carbon Capture and Storage (CCS) systems have been extensively investigated, particularly post-combustion processes based on chemical absorption with amines, due to their technological maturity and potential for integration into existing industrial plants. However, the high energy consumption and operational costs associated with these systems still limit their large-scale deployment, highlighting the need for advanced modeling, control, and optimization strategies. In this study, an integrated architecture combining Model Predictive Control (MPC) with economic objectives and Real-Time Optimization (RTO) is developed, implemented, and evaluated for a post-combustion CO₂ capture plant based on chemical absorption with monoethanolamine (MEA), rigorously simulated in Aspen Plus Dynamics®. Initially, dynamic simulation of the plant was performed, followed by the application of pseudo-random binary sequence (PRBS) excitation signals to generate data for the identification of linear multivariable process models. Based on these data, output-prediction oriented models (OPOM) were developed, and CLMPC and IHMPCFF controllers were implemented to dynamically track the economic targets defined by the RTO layer. In addition, a Kalman filter was incorporated for the estimation of non-measurable states, providing improved predictive accuracy and eliminating steady-state offset. The MPC controllers were subsequently integrated with a higher-level steady-state model-based economic optimization layer (RTO), responsible for periodically recalculating the optimal operating points while accounting for energy costs, CO₂ pricing, solvent consumption, and environmental penalties. The results demonstrate that the integrated RTO–MPC framework leads to significant improvements in energy efficiency, enhanced dynamic stability, and reduced operational costs under variations in load and flue gas composition. These findings confirm that the integration of predictive control and economic optimization constitutes a promising approach for the sustainable and economically viable operation of industrial CO₂ capture plants.