Experimental and modeling studies of density and viscosity behavior of a live fluid due to CO2 injection at reservoir condition

In this study, highly accurate measurements of density and dynamic viscosities of a recombined live oil and its mixture with additional CO2 were performed. The experiments were carried out under pressure and temperature gradients found in Brazilian Pre-salt reservoirs, that is, in the pressure range...

Full description

Bibliographic Details
Main Authors: Santos David C., Lamim Marina N., Costa Daniela S., Mehl Ana, Couto Paulo, Paredes Márcio L. L.
Format: Article
Language:English
Published: EDP Sciences 2021-01-01
Series:Oil & Gas Science and Technology
Online Access:https://ogst.ifpenergiesnouvelles.fr/articles/ogst/full_html/2021/01/ogst210047/ogst210047.html
Description
Summary:In this study, highly accurate measurements of density and dynamic viscosities of a recombined live oil and its mixture with additional CO2 were performed. The experiments were carried out under pressure and temperature gradients found in Brazilian Pre-salt reservoirs, that is, in the pressure range from (27.6 to 68.9) MPa and at (333.15 and 353.15) K. The assumption of volume change on mixing is evaluated from the experimental results, and the influence of pressure and temperature on the volume change upon mixing is assessed. The densities of mixtures are calculated considering (i) the excess volume approach, and (ii) no volume change. The densities are better correlated using the excess volume approach with Average Absolute Deviations (AAD) of 0.03%. Thirteen mixing rules of viscosity are examined by comparing the predicted values with the experimental viscosity of the recombined live oil + CO2 mixture. The performance of some rules using compositional fractions (molar, volume and weight) is also evaluated. Thus, a total of 28 different ways to calculate the mixture viscosities were tested in this study. The worst result was obtained with Bingham’s method, leading to 148.6% AAD. The best result was obtained from Lederer’s method with 2% AAD and a maximum deviation of 5.8% using volume fractions and the fitting parameter α. In addition, deviations presented by the predictive methods of Chevron, Double log, and Kendall did not exceed 9% AAD, using weight fractions (Chevron and Double log) and molar fractions (Kendall and Monroe).
ISSN:1294-4475
1953-8189