The U.S. Department of Energy’s Carbon Capture Simulation Initiative (CCSI) is focused on developing computational tools and models for accelerating the development and commercialization of post-combustion CO2 capture technologies. The process modeling team of CCSI is developing high-fidelity process models for a monoethanolamine (MEA) solvent system, which has high potential for commercial applications, for rapidly designing and scaling up such processes while also improving their operability. Previous works of the authors have determined key thermodynamic and transport models (density, viscosity, surface tension, heat capacity and VLE models) and are applied in the current investigation. One of the focuses of this work is to develop an accurate mass transfer model. In the existing literature, typically the mass transfer models and their parameters such as the liquid and gas-side mass transfer coefficients, diffusivity, and interfacial area are regressed using the data obtained from different experimental set-ups and scales, often in a sequential and sub-optimal way, leading to increased parametric uncertainties. For minimizing these uncertainties, a simultaneous regression of mass transfer coefficients, diffusivity and interfacial area was performed by using FOQUS, a tool developed in CCSI, leveraging the operating data from the commercial packings and wetted wall columns.
In the open literature, there is a clear lack of in-depth dynamic validation studies. In this work, the rigorous dynamic models developed in both Aspen Plus Dynamics® and gPROMS® platforms are validated using dynamic data collected from the US DOE’s National Carbon Capture Center (NCCC) in Wilsonville, Alabama. The dynamic test runs were conducted by ensuring persistent of excitation to give key insights into model-form and parametric discrepancies of the process. The data include solvent composition and loading data that were collected every 5 minutes and analyzed in the NCCC laboratory. Due to the noisy and erroneous data and missing measurements of a number of key variables, a dynamic data reconciliation problem is solved. The transient behavior of the process and model validation provide strong insights that will be discussed in this presentation.