Process Modeling and Optimization of Post-Combustion Carbon Capture With N2-Selective Membranes

Membrane separation has been proposed as a post-combustion carbon capture technology for its potential advantages such as a small footprint, environmental friendliness (i.e., no solvent involved in the process), and a lower energy requirement. Current modeling and optimization work has focused on post-combustion capture using polymeric, CO₂-selective membranes, where the CO₂ in flue gas is selectively permeated through the membrane and is concentrated in the permeate stream. One of the limitations of CO₂-selective membranes is that the low CO₂concentration in flue gas (12–15 vol% from coal-fired power plants, 5–8 vol% from natural gas–fired power plants) provides a low driving force for separation, which requires that multistage separation and/or stream recycling be employed to achieve DOE’s capture target of 90% capture (removal efficiency) and 95% product purity.

The concept of N₂-selective membranes has been devised as a way to further lower the land and energy requirements of membrane separation as a post-combustion carbon capture technology. Two factors may contribute to the potential energy savings. First, the predominantly higher N₂ concentration in flue gas may provide a higher driving force for membrane separation, eliminating part of the feed compression requirements. Second, the concentrated CO₂ stream exits from the high-pressure retentate stream, which may displace part of the compression energy required for pipeline transport. The concept of N₂-selective membranes has been experimentally proven with pure vanadium (V) and their alloys with ruthenium (Ru) as membrane materials. In our modeling work, flue gas from a coal-fired power plant containing ~15 vol% CO₂ is treated with an ideal N₂-selective membrane (assuming minimal defects present in the material and only N₂ passes across the membrane) in the countercurrent flow scheme. The membrane optimization literature mostly deals with single objective optimization. However, a capture plant should satisfy two objectives of low parasitic energy consumption and a small footprint, which have direct implications on operating and fixed costs, respectively. In the present study, we performed multi-objective optimization using Genetic Algorithm (GA) to minimize energy consumption and membrane area simultaneously. The results of these studies will be presented.