Evaluation of SPUNG* and Other Equations of State for Use in Carbon Capture and Storage Modelling

Abstract

In this work, Equations of State (EoS) relevant for carbon capture and storage modelling have been evaluated for pure CO2 and CO2-mixtures with particular focus on the extended corresponding state approach, SPUNG/SRK. Our work continues the search of an EoS which is accurate, consistent and computationally fast for CO2-mixtures. These EoS have been evaluated: Soave-Redlich-Kwong (SRK), SRK with Peneloux shift, Peng-Robinson, Lee-Kesler, SPUNG/SRK and the multi-parameter approach GERG-2004. The EoS were compared to the accurate reference EoS by Span and Wagner for pure CO2. Only SPUNG/SRK and GERG-2004 predicted the density accurately near the critical point (< 1.5% Absolute Average Deviation (AAD)). For binary mixtures, Lee-Kesler and SPUNG/SRK had similar accuracy in density predictions. SRK had a sufficient accuracy for the gas phase below the critical point (<2.5%), and Peng Robinson had a decent accuracy for liquid mixtures (<3%). GERG-2004 was the most accurate EoS for all the single phase density predictions. It was also the best EoS for all the VLE predictions except for mixtures containing CO2 and O2, where it had deviations in the bubble point predictions (∼20% AAD). Even though multi-parameter EoS such as GERG-2004 are state-of-the-art for high accuracy predictions, this work shows that extended corresponding state EoS may be an excellent compromise between computational speed and accuracy. The SPUNG approach combines high accuracy with a versatile and transparent methodology. New experimental data may easily be taken into account to improve the predictive abilities in the two phase region. The approach may be improved and extended to enable applications for more difficult systems, such as polar mixtures with CO2 and H2O.