Commissioning of a magnetic suspension densitometer for high-accuracy density measurements of natural gas mixtures

摘要

High-accuracy density measurement data are required to validateequations of state (EOS) for use in custody transfer of natural gas through pipelines. TheAGA8-DC92 EOS, which is the current industry standard has already been validatedagainst a databank of natural gas mixtures with compositions containing up to 0.2 molepercent of the heavier C6+ fraction and is expected to predict densities of natural gasmixtures containing higher mole percentages of the C6+ fraction with the same accuracy.With the advances in exploration, drilling and production, natural gas streams containinghigher percentages of the C6+ fraction have become available from the deepwater andultra-deepwater Gulf of Mexico in recent years. High-accuracy, density data for suchnatural gas mixtures are required to check if the AGA8-DC92 EOS covers the entirerange of pressure, temperature and compositions encountered in custody transfer.A state-of-the-art, high pressure, high temperature, compact single-sinkermagnetic suspension densitometer has been used to measure densities of two simulatednatural gas mixtures named M91C1 and M94C1 after validating its operation bymeasuring densities of pure argon, nitrogen and methane in the range (270 to 340) K[(26.33 to152.33) oF, (-3.15 to 66.85) oC] and (3.447 to 34.474) MPa [(500 to 5,000)psia]. Measured densities of M91C1, not containing the C6+ fraction show larger thanexpected relative deviations from the AGA8-DC92 EOS predictions in regions 1 and 2but agree well with predictions from the recently developed REFPROP EOS, implyingthat the AGA8-DC92 EOS may be unreliable in its present state even for natural gasmixtures not containing the C6+ fraction. Measured densities of M94C1 containing morethan 0.2 mole percent of the C6+ fraction deviate from the AGA8-DC92 EOS predictionsby more than the expected values in region 1 which is not surprising but the agreementwith AGA8-DC92 EOS predictions in region 2 is misleading which becomes evidentwhen the measured densities are compared to the REFPROP EOS predictions. Themeasured data can be used to recalibrate the parameters of the AGA8-DC92 EOS or tovalidate an entirely new EOS.