This paper describes methods for calculating the one-dimensional, vertical variation in composition with depth caused by gravity and thermal gradients. The Peng-Robinson (PR) and Soave-Redlich-Kwong (SRK) cubic equations of state (EOS) are used as thermodynamic models. Examples of calculated compositional gradients are given for reservoir fluid systems ranging from black oils to near-critical oils.
A solution algorithm is suggested for solving the isothermal gravity/chemical equilibrium (GCE) problem. The algorithm is simply an adaptation of a method proposed by Michelsen1 for calculating saturation pressure. The problem of false (unstable) solutions is discussed, and the subsequent need for applying phase stability analysis to identify such false solutions. Finally, an algorithm is given for determining the location of a gas-oil contact (GOC).
A model for treating both gravity and thermal gradient has been used to quantify the potential effect of thermal diffusion on compositional grading. The model used was proposed by Belery and da Silva2 . Unfortunately, the physics and thermodynamics of thermal diffusion are not well understood. This model is only one of several approaches which have been suggested for treating thermal diffusion. Examples given in our paper show that thermal diffusion can have a marked effect on compositional grading, with the possibility of enhancing, reducing, or completely eliminating gradients caused by gravity alone.
We illustrate the potential danger of using gradient calculations for defining original hydrocarbon distributions (oil and gas in place) when limited fluid samples and PVT data are available. Furthermore, guidelines are given for when to use gradient calculations, and how to develop an EOS fluid characterization for reservoirs exhibiting compositional variation.