A predictive model for the techno-economic assessment of CO[formula omitted] chemisorption processes applicable to a large number of amine solvents

Abstract

With the growing need to reduce carbon dioxide (CO2) emissions, there have been substantial efforts to identify new solvents that can improve the overall performance of chemical-absorption CO2 capture processes. Given the large number of potential solvents, computer-aided molecular and process design (CAMPD) approaches can play a critical role in accelerating the search for optimal solvents by enabling the systematic exploration of solvent candidates and process conditions. One of the challenges in developing such a framework is the requirement for a process model that can be used to capture the interactions between process performance and solvent structure without significantly increasing its numerical complexity. In the current work, a model for the absorption–desorption of CO2 is developed using the predictive SAFT-γ-Mie group-contribution approach, allowing the performance of numerous solvents to be assessed without the need for extensive experimental data. In order to avoid convergence difficulties when solving this highly nonlinear model, a tailored initialization strategy is established, using an adapted inside-out algorithm to prime a nonlinear equation solver for each column, providing a good initial guess for the whole flowsheet. Tests on three solvents confirm the robustness of the approach. Building on this enhanced numerical stability, the model is validated by comparison against pilot-plant data, showing good accuracy. A detailed parametric study of the effect of the key process variables is undertaken; the important role of the CO2 capture rate, of the lean solvent temperature and loading, and of the desorber pressure is highlighted. The results of the parametric study are used to formulate an optimization problem which is successfully solved for four solvents. A large reduction in total annualized cost and energy requirements is achieved by tuning the operating conditions to each solvent considered. The predictive capability, robustness, and reliability of the proposed model and associated initialization strategy open the way for the evaluation of a large number of novel solvents.