Development of new models for predicting crude oil bubble point pressure, oil formation volume factor, and solution gas-oil ratio using genetic algorithm | ||||
| Journal of Petroleum and Mining Engineering | ||||
| Article 3, Volume 22, Issue 2, December 2020, Page 17-39 PDF (1011.42 K) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/jpme.2020.31955.1035 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
Ahmed Soliman   ; Attia Attia; Muhamed Abdelwahab
| ||||
| Petroleum Engineering Department, Faculty of Engineering, British University in Egypt (BUE), Elshorouk city, Cairo, Egypt | ||||
| Abstract | ||||
| Bubble point pressure (Pb), oil formation volume factor (BO), and solution gas-oil ratio (Rs) are considered the key parameters required to describe and characterize the crude oil. Accurate determination for crude oil properties are necessary for multi-operation in reservoir evaluation, such as reserve estimation, enhanced oil recovery, oil reservoir performance prediction, designing pipelines and production equipment, and reservoir simulation. Traditional techniques used to calculate PVT data are usually expensive or unavailable, so there are a huge number of empirical correlations developed to estimate PVT properties as a function of production data. But when we used these correlations to predict crude oil properties, big errors are attained. The main target of this study is to find a better and accurate approach for predicting the properties of crude oil. This paper developed new empirical correlations for predicting the properties of reservoir oil as a function of PVT properties such as (P, T, Bo, Rs) using genetic algorithm technique. The simulation model is built using MATLAB software which contains the optimization tool that includes a genetic algorithm tool in it. To validate these correlations, 130 data sets of different crude oils were used. The results obtained showed that the developed empirical correlations from the genetic algorithm model appeared excellent accuracy of predicting crude oil properties compared to their relevant published correlations. The average absolute error for all correlations that the genetic algorithm applied to them is decreased. This technique can be applied to predict crude oil properties with a high level of accuracy. | ||||
| Keywords | ||||
| PVT properties; empirical correlations; Genetic algorithm | ||||
|
Statistics Article View: 1,275 PDF Download: 877 |
||||
