Ensuring consistency in VLE and density modelling of (CO2 + H2) mixtures that are relevant in CCS pipeline transport: fitting BIPs for various EoS implemented in ‘off-the-shelf’ property packages

Abstract

The impact of impurities in the physical properties of CO2 is now widely acknowledged by industry. In particular, the presence of H2 in CO2 mixtures has a significant effect on the saturation pressure (bubble point) and bulk (liquid) density. Both physical properties are key in the many engineering tasks, for example, in calculating the closed-in tubing head pressure of an injection well or in defining operating pressures to avoid two-phase region in transport systems. If there are large uncertainties in the estimated properties, then, significant design margins will be required making the system less cost-effective.The reliability of Equations of State (EoS) implemented in property packages used in industry is limited to the availability and quality of the experimental data that was used to calibrate the model. For the pair CO2H2, there is lack of validation at conditions of CCS systems. Thus, the performance of EoS is unsatisfactory for many CO2H2 mixtures. An additional complication is that there is large discrepancy not only between models but also between the ‘same’ model implemented across different platforms. For the practitioner, this poses quite a conundrum: ‘What EoS and package should be used?’.In this work, an attempt to homogenize various EoS implemented in commercial software is presented. To calibrate the models, a subset of available of experimental data with content of up to 5%mol of H2 is used. A different subset of data is used to assess the performance of the fitted models. In the parameter estimation problem, the EoS are tuned via Binary Interaction Parameters (BIPs) or departure parameters. To improve liquid density calculations, an additional step is used to regress cubic EoS via volume shift parameters.The resulting optimized EoS can be readily available in commercial simulators, and it is hope that they will be part of a growing databank/toolbox for engineering design activities of CCS systems.