Modelling and Optimization of Vacuum Pressure Swing Adsorption Co2 Capture Pilot Using Mil-160(Al)

Global warming, driven by increasing CO2 emissions from fossil fuel combustion, necessitates the
development of effective carbon capture technologies. Among various approaches, Vacuum Pressure
Swing Adsorption (VPSA) offers an energy-efficient solution for post-combustion CO2 capture,
especially in power plants and energy-intensive industries. This work focuses on validating a simulation
model using a laboratory VPSA pilot with the Aluminum Metal-Organic Framework (Al-MOF)
MIL-160(Al) and optimizing both lab-scale and industrial-scale VPSA pilots through simulation. Process
modeling in Aspen Adsorption software simulated a 3-bed 6-step cycle using parameters from
experimental adsorption isotherms and breakthrough curves. The simulation model was compared to
previous lab-scale VPSA pilot experiments treating a synthetic 15/85 CO2/N2 mixture at 1 Nm³/h,
showing mean absolute errors of 1.47% for purity and 3.19% for recovery.
Surrogate models, including kriging and artificial neural networks (ANN), were used to optimize
recovery and purity of the lab-scale pilot using a genetic algorithm (NSGA-II). The ANN model proved
more accurate, especially in determining pareto fronts. The model was extended to an industrial VPSA
pilot as part of the MOF4AIR project, designed to treat flue gas of 50 to 100 Nm³/h with three 41 L
adsorption columns. Simulations showed that the pilot could achieve 95% purity and recovery for CO2
concentrations ranging from 5 to 15%. The estimated energy consumption and productivity for 15%
CO2 gas were 413.19 kWh/tCO2 and 3.03 tCO2/(m³ads.day) at a gas flow rate of 55.62 Nm³/h,
demonstrating the technology's potential and competitiveness on a larger scale.