Compositional Model Studies - CO2 Oil-Displacement Mechanisms

Abstract

Abstract This paper presents matches, using a fully compositional model, of the performances of seven laboratory CO2 displacements of a 10-component synthetic oil. The criteria for achieving a match of laboratory performance include (1) comparisons of predicted and experimentally determined oil recovery and (2) effluent compositional profiles for each component as functions of hydrocarbon pore volumes (HCPV) of CO2 injected.An equation of state was tuned to predict single-contact (PVT) phase equilibria for CO2/synthetic-oil mixtures. The model incorporates this equation of state to predict the multiple-contact phase equilibria during a CO2 displacement test. Input to the model were independently determined gas/oil relative permeability characteristics and - for each laboratory displacement - injection rate, effluent pressure, pore volume, and temperature.The experimental displacements were conducted in linear Berea core systems using a synthetic (C1-C14) oil at 120 and 150 degrees F. Three displacements at 120 degrees F have been published by Metcalfe and Yarborough. Previously, it was thought that these displacements were conducted at selected pressures so that oil displacements encompassed immiscible, multiple-contact miscible (MCM), and contact miscible mechanisms. However, the model results show that only contact miscible and MCM displacement mechanisms were involved. To confirm the mechanistic understanding at 120 degrees F, three additional laboratory displacements were conducted at 150 degrees F. These encompassed pressures such that the displacement was controlled by an immiscible, an MCM, and a contact miscible mechanism, respectively. The model results at 150 degrees F match the experimental data and confirm the mechanistic understanding. The experimental and numerical results are in agreement with the minimum miscibility pressure theory of Yellig and Metcalfe.The results of this study confirm the importance of experimentally determined effluent compositional profiles and fully compositional models for CO2 mechanism studies. Introduction A mechanistic understanding of oil displacement by CO2 is basic to establishing the CO2 requirements and predicting performance for field projects. The petroleum engineering community relies heavily on fully compositional models and sophisticated laboratory experiments to acquire this mechanistic understanding.Currently, it is not deemed feasible to use fully compositional models to simulate performance of field-wide miscible floods. However, these models should be capable of predicting performances of laboratory-scale displacements. The results of such predictions will identify important variables that control oil recovery and that must be incorporated in mechanistically simpler field performance simulators. Further, confidence in field models will be enhanced greatly by the demonstrated ability to predict laboratory floods.Studies to improve the effectiveness and efficiency of multicomponent compositional simulators have been reported. A cell-to-cell flash model has been used to study mechanisms in rich-gas drives. However, no investigations have been reported previously that demonstrate that a fully compositional model can predict results of rich gas or CO2 laboratory displacements, including prediction of the phases and compositions developed in situ. SPEJ P. 89^