Applicability and Accuracy of Two Component Characterizations of Reservoir Oils for Numerical Modeling Purposes

Abstract

The applicability and accuracy of a new binary pseudo-component PVT representation relative to conventional black-oil and equation-of-state representations is determined. This PVT representation extends the conventional black-oil treatment by decoupling the reservoir and surface PVT properties and by modeling the properties of oil and gas through the use of K-values and Z-factors for a two component system. Processes considered include depletion, water injection, and gas injection in oil reservoirs. In depletion simulations the ability of the two-component K-value (2CKV) representation to model gas phase condensation in surface separators results in higher and more accurate (relative to equation-of-state predictions) oil recovery predictions than those obtained using a conventional black-oil model. The capability of the 2CKV formulation to model time-dependent separator conditions is shown for a waterflood in a vertical cross section. Neither the 2CKV nor the black-oil formulation can accurately simulated a gas injection process when the composition of the injection gas is significantly different from the composition of equilibrium reservoir gas in the injection cell(s). Introduction Several authors have proposed modifications of black-oil models to approximate compositional effects in gas condensate and volatile oil reservoirs. These models represent various degrees of implementation of a binary pseudo-component PVT property representation. Rovere et al. proposed an "alpha-beta" model which differs from conventional black-oil models in the treatment of surface separation. Differential liberation data without adjustment are assumed to represent PVT behavior in the reservoir. Gas phase condensation during separation is modeled through the use of two pressure-dependent parameters, alpha and beta. Conventional black-oil surface volume rates are given by: (1) (2) Rovere et al. modified these equations as: (3) (4) Alpha represents the fraction of reservoir free gas produced as gas in the separator, and beta represents the oil produced per unit of free gas due to free gas condensation. The parameters alpha and beta are determined by equation-of-state prediction. The two component K-value model presented here differs from previous work in the representation of reservoir and separator PVT properties. The compositional treatment of the gas phase allows modeling of gas phase composition variations in the reservoir and provides a means for modeling gas phase condensation in the separator(s) without requiring any data in addition to that for conventional black-oil simulation. DESCRIPTION OF MODELS In order to guarantee consistency in comparisons of simulation results, the three models used in this study are identical except for the hydrocarbon phase PVT property representations. PVT properties in the compositional model are calculated using the Peng-Robinson Equation-of-State. P. 665^

From BO (Binary) to Ternary Fluid Representation for Simulation of Condensate Recovery by Lean Gas Injection

“…The two last items are used when the reservoir and surface PVT properties are dissociated as suggested, for example, in Ref. 8.…”

“…The various BO formulations 4,5,6,7,8 are basically two component representations : an heavy component 'H' and a and a volatile one 'V', originally presented as corresponding to the 'stock tank oil' and to the 'surface gas', main enhancements from the original BO formulation consisting in:…”

Simulation of Gas Injection Processes in Gas-Condensate Reservoirs Using a Binary Pseudo-Component Representation

“…To try and minimise the material balance error authors have used partial densities 8 , inversion of the 'flash' equations 9 , and additional parameters to account for mass transfer 'between separator products 10 -11 • The first physically consistent binary pseudo-component PVT representation for use in (extended) conventional black oil simula.tors without code modifications was presented by Drohm and Goldthorpe 7 in 1987. Since then Coats et al 11 have used a two component K-value approach for decoupling reservoir and surface fluid properties, but this scheme still uses oil and gas as pseudo~components. Contact or mixing parameters allow correlation of fluid properties with the cumulative pore volume of injection gas in simulator grid blocks.…”