SELECTION OF PVT PROPERTIES FOR BLACK OIL SIMULATION

It is generally accepted that the reservoir depletion process is a dif ferential liberation process where the eas is removed from contact wit h the oil as soon as it is formed, and the production of I fluids thro ugh tubing and separators is a flash liberation process where the gas liberated from solution remains in contact with the oil. Laboratories commonly perform differential liberation and separator tests on the re servoir fluid and then use a correlation described by Amyx et al.((1)) to arrive at a composite liberation curve for an optimum separator pr essure. Engineers are often confused as to which of the two sets of PV T data, i.e., differential or composite, should be used in a black oil simulation of the reservoir. The differences between the two sets of data become more substantial as the API gravity of the oil increases. The black oil simulator used in this report has a compositional formul ation and has the ability to rigorously model the PVT behaviour by ent ering differential liberation data for reservoir conditions and using a surface separator option for the surface Bash process. This paper de scribes a critical comparison of the three methods to enter PVT proper ties in a black oil simulator, The results of using the simulator in t hree ways: a) differential liberation at reservoir, surface separator for flash, b) composite liberation data, c) differential liberation da ta, are described in this paper for a medium gravity crude. It is show n that the composite liberation data most closely represents the rigor ous representation of phase behaviour as in (a). Use of inappropriate liberation data could results in significant error in estimates of oil -in-place as in (c). The paper compares the differences between compos ite liberation (b) and the rigorous representation (a), such as differ ent initial pressure and saturation gradients due to initial density d ifferences.