- Conference Article
2
- 10.3997/2214-4609.202011960
Tuning of Peng-Robinson Equation of State For Simulation of Oil Compositional Change During EOR Processes
- Jan 01, 2021
- M Majidi + 2 more +2
Summary In most of the EOR processes, including CO2 injection, miscible gas flooding and WAG recovery, oil production is normally occurred along with rapid changes in reservoir fluid phase, hence a robust PVT model is required to predict the fluid properties in the reservoir. Application of the Equation of State (EOS) is the most common method to predict the fluid phase behavior in dynamic reservoir simulation, however, because of the oil mixture complexities, EOS initial predictions always have some degree of error. The most common approach to encounter this problem is to calibrate or tune an EOS model against experimental data, before applying in the simulator. In this work, Peng-Robinson (PR) EOS was used to predict the results of a Constant Composition Expansion (CCE) test for a real crude oil. Using sensitivity analysis, dominant regression parameters were picked to match the CCE test results. Direct Search and Genetic Algorithm were used to optimize the EOS predictions and it was shown that Direct Search can tune the PR EOS more effectively. While Direct Search could reach the minimum AAD% of 2.16% in 207 runs, Genetic Algorithm required 329 runs to reach 4.45% of the minimum AAD% to match the experimental results.
Read more