(663b) Rigorous Modeling of a MEA-Based CO2 Capture Process with Uncertainty Quantification and Validation Using Large Scale Pilot Plant Data
The MEA model in this work is developed through a compilation of submodels, including physical property, reaction kinetics, mass transfer, and hydraulic models. The model parameters are regressed with experimental data that span the temperature, pressure, and composition ranges of interest. For each submodel, a Bayesian inference methodology is used to quantify the parametric uncertainty.
This model is validated with high quality steady-state data (23 total trials) collected from the National Carbon Capture Center (NCCC) in Wilsonville, Alabama. While most of the steady-state test runs reported in the literature are focused on narrow operating range, the key manipulated variables such as solvent flowrate and reboiler steam flowrate and disturbance variables such as the flowrate and CO2 concentration of the inlet flue gas were varied widely in these test runs. Moreover, the number of beds, and thus packing height, and the presence of intercooling in the absorber are variable in the test runs. For each of the 23 cases, uncertainty in the key variables of interest, including percentage of CO2 capture in the absorber and required reboiler duty in the regenerator, are determined by propagating parametric uncertainty through the process model. This rigorous methodology for obtaining uncertainty ranges on key operating variables is not only novel, but also provides insight into the effects of the operating variables on output uncertainty and helps to identify process submodels that have significant impact on output uncertainty. The insights gained through this developed methodology can play a key role in reducing scale up uncertainty of CO2 capture processes.