This article provides results of the in situ monitoring of carbon dioxide (CO₂) removal under high pressure. An aqueous solution of methyldiethanolamine (MDEA) promoted by piperazine (PZ) for absorption rate acceleration was used. This system is promising for natural gas purification. A predictive statistical model was built using the chemometrics method and measurements of density, pH, conductivity, sound velocity, refractive index, and Near Infra-Red (NIR) spectroscopy. The CO₂ capture rate ranged from 60% up to 96% at pressures in the absorber column ranging between 15 and 20 bar. The liquid stream composition was monitored at the low pressure part of a pileline at a location before a compressor stage. The concentrations of MDEA, PZ, and CO₂ were predicted in-line using different sets of the measurement devices during 3 days of the measurement campaign. The developed approach allowed for prediction of the concentrations with accuracies of 0.7% for MDEA, 0.4% for PZ, and 2.5% for CO_2.@en

The solvent and the dissolved CO₂ concentrations are two essential properties of CO₂ absorption processes. Currently, they are typically monitored using time-consuming offline analytical techniques. Initial development efforts aiming at a cost-effective and reliable inline monitoring system are described. Using a fractional factorial experimental design covering six solvent properties and five analytical techniques, it is demonstrated that the CO₂ and solvent concentrations correlate well with the density and refractive index. Applying a simplistic multivariate model to solvent samples from an industrial pilot plant indicated good potential for predictive use. Future work should focus on the development of a more accurate multivariate model, on the modeling of more solvents, and on combining the results with available process data.

@en