Page 234 - PVT Property Correlations
P. 234

204  PVT Property Correlations


            problem of estimating the best correlation(s) to use. Many practicing engi-
            neers follow a completely arbitrary process to select PVT correlations.
            Others may follow more structured approaches such as those discussed in the
            following section.



            SELECTION OF PVT OIL PROPERTIES CORRELATIONS

            Numerous PVT correlations exist for each oil (and gas) property. The number
            of oil correlations available exceeds that of gas correlations (Appendix A:
            Oil Correlations Formulae and Appendix B: Gas Correlations Formulae).
               Oil and gas PVT correlations can usually be grouped into correlation
            families. Researchers of PVT correlations usually develop a separate correla-
            tion for each PVT property from their PVT laboratory-measured data.
            For selection of correlation(s) to use for a specific fluid, the following
            approaches are available: (1) use of geographical location, (2) use of the
            type or obvious characteristics of the fluid, (3) use of correlations based on
            application, (4) use of data ranges and ranking, and (5) use of an expert
            system or set of rules.


            Correlation Selection Based on Geographical Location

            Many oil PVT correlations were developed with data sets that belong to a
            specific region. Among the commonly known geographical correlations are
            Al-Marhoun (1988) correlations for Saudi Arabian crudes; Standing (1947)
            for California crudes; Petrosky (1990) for Gulf of Mexico crudes; Hanafy
            et al. (1997) for Egyptian crudes; and Omar and Todd (1993) for Malaysian
            crudes. Appendix C, Oil Correlations Range of Applicability, gives the appli-
            cation ranges of each correlation and identifies the geographical origin of
            each correlation, where applicable.
               In absence of laboratory-measured PVT data, many engineers use specific
            correlation families for crude oils based on geographical origin. Although it
            is logical to assume that oil in different reservoirs in a particular geographi-
            cal location is of the same origin, it is erroneous to assume that this approach
            is consistently reliable. The geographical PVT correlation families may have
            been developed from PVT data collected from relatively shallow reservoirs.
            Deeper reservoirs developed at a later stage usually contain more volatile oil
            and richer gas condensate fluids. New correlations are required for the fluids
            more recently discovered in these traditional regions when sufficient data
            becomes available. Therefore, caution is advisable in use of the geographical
            correlation approach.
               In addition, Al-Shammasi (2001), in a large comparison study, showed
            that geographical location-based correlations are not guaranteed to work.
   229   230   231   232   233   234   235   236   237   238   239