(734b) Dynamic Modeling and Control of a Post-Combustion Solid-Sorbent Capture System with the Ccsi Models and Tools

Solid sorbent-based post-combustion CO₂ capture processes are a viable alternative to solvent-based capture that may provide better energy efficiency. Integrated dynamic models of these types of processes are essential for understanding transient system behavior and for the development of control strategies to efficiently reject process disturbances and optimize tracking of set points. In this work, a suite of tools and models from the Carbon Capture Simulation Initiative (CCSI) was used to develop a dynamic integrated solid sorbent, post-combustion capture model and implement Nonlinear Model Predictive Control (NMPC). The CCSI Toolset has been released under an open-source license with the goal of accelerating and reducing scale-up risk of carbon capture technologies.

The post-combustion carbon capture flowsheet was developed in Aspen Custom Molder (ACM) and was constructed using the CCSI bubbling fluidized bed (BFB) model¹. The model is one-dimensional with two-phases and an integrated heat-exchanger that is flexible over a wide-range of operating conditions so that it can represent both adsorption and regeneration in either a steady-state or dynamic framework. A dynamic flowsheet was developed that is based on an optimal steady-state design that includes a three stage BFB for CO₂ adsorption, a two stage BFB regenerator, and includes a crossflow heat exchanger between the rich and lean solid streams. The flowsheet also accounts for the pressure flow dynamics and downcomers in the BFB.

The integrated model is computationally expensive and not suitable for direct use in an NMPC framework, hence a reduced order model is required for model-based control. The CCSI data-driven system identification tool, D-RM Builder,² is used to generate computationally efficient dynamic reduced-order models from a set of transient data. D-RM Builder allows for the selection of input/output variables, the generation of a sequence of step-changes, and interfaces with rigorous modeling platforms, such as ACM, to automatically run and collect step-test data. This tool was used to generate a computationally efficient nonlinear Decoupled A-B Net (DABNet) model³ involving multiple time-scales corresponding to “fast” pressure-flow dynamics and “slower” temperature-composition transients. The DABNet model was used as the control model in CCSI’s Advanced Process Control (APC) Framework^4,5. The APC Framework solves the constrained NMPC problem in a computationally efficient manner utilizing “fast” analytical derivative evaluations, implementing the control moves on the rigorous ACM flowsheet model. The control strategy is tested and compared to traditional control for its servo control and disturbance rejection characteristics.

References

1. Lee, A., & Miller, D. C. (2012). A one-dimensional (1-d) three-region model for a bubbling fluidized-bed adsorber. Industrial & Engineering Chemistry Research, 52(1), 469-484.
2. Ma, J., Mahapatra, P., Zitney, S. E., Biegler, L. T., & Miller, D. C. (2016). D-RM Builder: A software tool for generating fast and accurate nonlinear dynamic reduced models from high-fidelity models. Computers & Chemical Engineering, 94, 60-74.
3. Sentoni, G. B., Biegler, L. T., Guiver, J. B., & Zhao, H. (1998). State‐space nonlinear process modeling: Identification and universality. AIChE Journal, 44(10), 2229-2239.
4. Omell, B. P., Ma, J., Mahapatra, P., Yu, M., Lee, A., Bhattacharyya, D., Zitney, S., Biegler, L., & Miller, D. C. (2016). Advanced Modeling and Control of a Solid Sorbent-Based CO2 Capture Process. IFAC-PapersOnLine, 49(7), 633-638.
5. Mahapatra, P., Ma, J. & Zitney, S.E. (2018). Nonlinear Model Predictive Control using Decoupled A-B Net Formulation for Carbon Capture Systems – Comparisons with Algorithmic Differentiation Approach. American Control Conference, Jun 27-29, Milwaukee, WI.