Search

This paper is focused on the development and validation of a dynamic model for a monoethanolamine-based CO2capture unit. Starting with a rigorous steady-state process model, the dynamic model is developed in Aspen Plus Dynamics using a modified Murphree efficiency approach and is validated using dynamic data collected from the National Carbon Capture Center in Wilsonville, Alabama. The dynamic test runs were designed with due consideration of process excitation and process nonlinearities constrained by the available time to conduct the test runs and implement the experimental designs in the plant control system. The experimental data include solvent composition and loading. Because the experimental data were found to violate mass and energy balances, filtering algorithms coupled with dynamic data reconciliation techniques were used. Without adjusting any model parameters, the dynamic model satisfactorily predicted the dynamic response of CO2capture due to step changes in the solvent flow rate, flue gas flow rate, and steam flow rates. Transient studies show that the process gain and time constants can considerably change depending on the direction of the disturbance or manipulated variables, indicating process nonlinearities. Because of a storage tank between the absorber and stripper, the time constant of the full plant can be considerably longer than the time constants of the absorber or stripper individually. Furthermore, impact of the control system on the process efficiency while following an optimal CO2capture schedule was also studied.

The U.S. DOE’s Carbon Capture Simulation Initiative (CCSI) has a strong focus on the development of state of the art process models for accelerating the development and commercialization of postcombustion carbon capture system technologies. One of CCSI’s goals is the development of a process model that will serve not only as a definitive reference for benchmarking of the performance of solvent-based CO2capture systems but also as a framework for the development of highly predictive models of advanced solvent systems. In Part 1 of this paper and previous work, submodels for the system were developed, including those for physical properties, kinetics, mass transfer, and column hydraulics, by calibrating model parameters to fit relevant experimental data. For individual submodels, a Bayesian inference methodology was used to refine the estimates of the parameter values and to quantify the parametric uncertainty of the models. This work is focused on incorporating these submodels into a complete process model and validating this model with large scale pilot plant data from the Pilot Solvent Test Unit (PSTU) at the National Carbon Capture Center (NCCC). The model has been validated with data representing a wide range of operating conditions for absorber and stripper columns, including variable packing height and presence of intercooling in the absorber. The uncertainty in the solvent composition is measured by comparing the process measurements at NCCC to standard laboratory techniques of a known uncertainty. Through a sensitivity study, this measurement uncertainty is used to provide insight into some discrepancy between model and data. Parametric uncertainty for various submodel parameters has been propagated through the process model to assess the resulting uncertainty in the key model outputs. Finally, a variance-based sensitivity analysis is used to provide insight into the relative contributions of parameters from various submodels to the overall uncertainty of the process outputs in various operating regimes.

Rigorous process models are critical for reducing the risk and uncertainty of scaling up a new technology. It is essential to quantify uncertainty in key submodels so that uncertainty in the overall model can be appropriately characterized. In solvent-based postcombustion CO2capture technologies, mass transfer and column hydraulics are key factors affecting the performance of the absorber. Developing submodels for mass transfer, column hydraulics, and reactions is a challenging multiscale problem since the phenomena are tightly coupled and it is difficult to design experiments to isolate each properly. In particular, simultaneous mass transfer coupled with fast reaction kinetics makes it difficult to measure the mass transfer rate and reactions rate individually. The typical approach to solving this issue is to use proxy systems to conduct experiments under mass-transfer-limited or reaction-limited conditions. This approach can lead to inaccurate mass transfer submodels. In this paper, a novel simultaneous regression approach is proposed where submodels for mass transfer, diffusivity, interfacial area, and reaction kinetics are optimally identified using experimental data from multiple scales and operating conditions. Since all models have some level of uncertainty, a rigorous uncertainty quantification (UQ) technique is implemented for the hydraulic and mass transfer submodels based on Bayesian inference. Posterior distributions of submodel parameters are propagated through the column model to obtain the uncertainty bounds on critical performance measures.

This work focuses on development of a “gold standard” process model for a MEA-based post-combustion CO2capture process. The steady-state model includes a comprehensive thermodynamic framework in conjunction with the chemistry model. Parameters for all thermodynamic, transport, and physical properties models are regressed using extensive data available in the literature. An integrated mass transfer model is developed and validated using experimental data. The steady-state model is validated using data collected from the U.S. DOE’s National Carbon Capture Center (NCCC) in Wilsonville, Alabama. In addition to the steady-state runs, dynamic test runs were conducted at NCCC by introducing carefully-designed step changes and recording the transients of all key variables. Due to measurement noise and missing measurements for a number of key variables, a dynamic data reconciliation problem was solved to ensure material and energy balance of the collected data. Both the steady-state and dynamic models were validated against plant data for a wide range of operating conditions.