Hydrocarbon Dew Point Measurement and Model Evaluation of Synthetic and Real Natural Gases.

Abstract

Hydrocarbon dew point (HCDP) is one of the most important quality specifications of natural gas. Measuring and predicting the HCDP accurately are essential for the natural gas industry. However, the comprehensive experimental HCDP curve data are still rare, and knowledge about adopting proper prediction models remains unclear. In view of this, HCDP determination work by use of an improved test system and model evaluation based on more than 1000 dew points data have been done to improve the aforementioned dilemma. HCDP curve data of three gravimetrically prepared synthetic natural gases (SNGs) and one real gas (RG) are determined first. Then, one set of data containing 712 dew points from 28 SNGs and 334 dew points from 14 RGs is used to evaluate the performance of eight different HCDP prediction models including Soave–Redlich–Kwong (SRK), SRK–Twu, Peng–Robinson (PR), Twu–Sim–Tassone (TST), predictive SRK (PSRK), GERG-2008, PSRK, and perturbed-chain statistical associating fluid theory (PC-SAFT) models. Considerable prediction deviation of these models in the high-pressure region (pressure above 6.0 MPa) is observed compared to that in the low-pressure region (under 6.0 MPa), and the reasons for that difference are discussed. Evaluation results reveal that among the eight prediction models, GERG-2008 has the best performance (overall average absolute deviation (AAD): 1.44 °C) for SNGs, and PSRK and SRK–Twu fits the experimental data best for RGs (overall AAD: 2.50 °C). Therefore, GERG-2008 is recommended for HCDP prediction of relatively lean gas, whereas PSRK and SRK–Twu are recommended for calculating the HCDP of relatively heavy natural gases in low-pressure and high-pressure regions, respectively.