Abstract

The energy demand is still increasing across the globe, while environmental concerns about global warming effect and greenhouse gases have augmented recently. CO2 injection into mature oil reservoirs is an interesting operation that could help us supply the uprising demand while saving the environment. Having accurate knowledge about CO2 minimum miscibility pressure (MMP) is of utmost importance in designing a successful operation. This study mainly focuses on proposing several tools based on powerful tree-based and deep learning algorithms for estimating the MMP of CO2-crude oil system based on an extensive databank. The models employed in this study include extreme gradient boosting (XGBoost), categorical boosting (CatBoost), light gradient boosting machine (LGBM), random forest (RF), deep multi-layer neural network (deep MLN), deep belief network (DBN), and convolutional neural network (CNN). The models were trained and verified using 310 data points. Along with intelligent models, seven popular empirical correlations and two computational approaches, which are based on thermodynamics, were utilized to be compared with the proposed models. The outcomes expressed that the CatBoost model could estimate CO2 MMP values, using mole percent of volatile (C1 and N2) and intermediate (CO2, H2S, and C2–C5) fractions of oil, the average critical temperature of injection gas (Tcave), reservoir temperature (Tres), and molecular weight of C5+ fraction of oil (MWc5+) as input variables, with a total AARD of 1.34 %. Moreover, the variable impact examination showed that reservoir temperature greatly affects the MMP predictions. Finally, the Leverage approach verified the reliability of the databank and wide applicability domain of the developed CatBoost model spotting 5 outlier points (out of 310 points) only. The findings of this communication shed light on the high accuracy and reliability of CatBoost model in estimating CO2 MMP in a wide range of operational conditions.

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.