API MPMS 12.3:1996 pdf download
API MPMS 12.3:1996 pdf download.Manual of Petroleum Measurement Standards Chapter 12-Calculation of Petroleum Quantities Section 3-Volumetric Shrinkage Resulting From Blending Light HydrocarbonsWith Crude Oils.
4 Data and Mathematical Model Selection
The data used to derive the correlation presented in this document originated from four separate experimental studies, authored by Childress, Scott, Schuchardt, and Booker. The ranges of density of oil and light hydrocarbon are given in Table 1. The data are described in more detail in Tables B-i to B-4. Available information on the sources and densities of crudes and light components is included, together with the number of data sets and summarized statistical data. Sources of the data are also shown graphically in Figure B- 1.
The selection of an appropriate mathematical model was based on the ability of different correlations to predict the experimental results. As expected from the form of the equation, API 2509C performs very poorly above a 25 percent spike. This is apparent from the large deviations between predicted and experimental values (i.e. the absolute residuals) plotted in Figure B-6.
Improvement was obtained by selecting the form of correlation equation proposed by Ashcroft et al.8 Figure B-6 shows the average absolute residuals determined using the model of Ashcroft et a]., and Figure B-7 plots those obtained by fitting all of the data to an equation of the same form. Statistical differentiation between these latter two equations was achieved using an F-test as suggested by Beck and Arnold.’0 This test confirmed that the new equation presented in this standard gives the best prediction of the available experimental data over the entire range of compositional variation.
Data sets that exhibited the anomalous behavior of expansion rather than shrinkage were not used in deriving the new correlation. This included Booker’s toluene and cyclohexane spiked crude data sets and Schuchardt’s results from crude blends with gasoil.
The available experimental data were obtained at somewhat different temperatures and pressures. The older Childress and Scott data were obtained at 60°F and 100 psig, whereas the Booker and Schuchardt data apply to 15°C and ambient atmospheric pressure. Booker also reported results obtained at 25°C, but corrected these to 15°C using API correlations. The new correlation presented in this standard was derived from the data as presented in the referenced sources, with no attempt to normalize to a standard pressure and temperature. For the Booker data sets obtained at 25°C, the values corrected to 15°C, as given by Booker, were used.
5 Recommended Standard
5.1 STANDARD FOR CALCULATING SHRiNKAGE
The standard is the printed tables in either SI or customary units found in Section 8 below. The equations used to generate the standard are also provided (see Section 5.3) and can be used to develop computer subroutines to replicate the results in the printed table.
6.2 DIFFERENCE BETWEEN CUSTOMARY AND METRIC TABLES
Shrinkage factors obtained from Table 3 in customary units will not necessarily be numerically equivalent to factors determined from Tables 4 and 5 in SI units. This difference is partially the result of the difference in standard temperatures (60°F versus 15°C), but even after volume correction using the API correlation, shrinkage factors may show a relative difference of as much as 0.4 percent. For custody transfer purposes, it is therefore recommended that buyers and sellers agree on which set of tables they will use.
6.3 LIMITATIONS AND ALTERNATIVES
The correlation equation or tables given in this text can predict volumetric shrinkage within the stated precision for mixtures of hydrocarbons over a relatively wide range in density difference and mixture ratios. However, the available experimental data suggest that factors other than density difference affect the volumetric behavior. Current understanding is insufficient to allow accurate prediction of the effect of composition, and thus experimental confirmation is recommended for hydrocarbon types not represented in the existing data sets. It is recommended that when an accurate shrinkage is required for specific mixtures or when there are extenuating circumstances that may make use of the shrinkage tables questionable, the shrinkage may be determined by testing the individual components in a qualified lab using one of the test methods included in Appendix C.
