Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Evaluation of oil density estimation methods in reservoir conditions using the composition of the mixture

Bampili Fani

Full record

---

Summary

Density in reservoir conditions is a fundamental property of petroleum fluids with many applications in the oil industry. By linking mass to volume density has a direct role in calculating oil reserves and choosing the right production plan. It defines, among other things, the location of fluids in the porous media, the modeling of the flow system and the distribution of pressures with depth in the reservoir. Finally, it finds application in determining the parameters needed in reservoir simulation.In this diploma thesis methods of estimating the density of the oil in reservoir conditions, which are based both on the composition of the reservoir fluid and on production data, have been studied and evaluated. In the first category are included semi-theoretical models such as the Peng-Robinson equation of state and semi-empirical correlations such as the Alani-Kennedy and Standing-Katz. The second category refers to the semi-empirical correlation of Katz. In order to achieve the goals of the thesis a database of 484 petroleum fluids from different reservoirs around the world was used. For each petroleum fluid of the database there has been among others information including the composition, pressure and temperature of the reservoir.The assessment of densities predictions that were provided by the semi-theoretical and semi-empirical methods was performed by comparing the density predictions against the experimental density measurements that were available in the database in the form of regression equations.The best approach for matching the database densities was found to be the Katz method which uses only production data without the need for composition, exhibiting lowest mean relative (0,30%) and average absolute relative error (2,60%). The Standing-Katz and Alani-Kennedy methods followed with almost similar accurate estimates. The former provided better results than the second one with respect to the average relative error (0,35% for Standing-Katz and 0,80% for Alani-Kennedy). The opposite occurred with respect to the mean absolute relative error (2,75% for Alani-Kennedy and 3,14% for Standing-Katz). Less accurate was found to be the Peng-Robinson (without Tuning) method which exhibited the highest average relative (1,09%) and average absolute relative error (3,80%). The best standard deviation of the mean relative error was achieved by the Alani-Kennedy method (3,89%) followed by the Katz method (4,06%), and then the Peng-Robinson (4,83%) and the Standing-Katz (5,36%) ones.Finally, an analysis was made of the dependence of the mean absolute relative errors on the gas / oil ratio, API density of the data base oils as well as on the conditions of pressure and temperature of the formation. The mean absolute relative errors of the Katz, Standing-Katz and Alani-Kennedy methods were found in general to increase with the volatility of the fluids (GOR) and with pressure and temperature. The exception was the Peng-Robinson method, the accuracy of the densities predictions of which, did not show to depend on GOR and temperature, while its mean absolute relative error showed a decrease with pressure and API density.

Available Files

Services