Understanding the Controlling Factors for CO₂ Sequestration in Depleted Shale Reservoirs Using Data Analytics and Machine Learning

Abstract

Carbon capture and sequestration is the process of capturing carbon dioxide (CO 2 ) from refineries, industrial facilities, and major point sources such as power plants and storing the CO 2 in subsurface formations. Carbon capture and sequestration has the potential to generate an industry comparable to, if not greater than, the existing oil and gas sector. Subsurface formations such as unconventional oil and gas reservoirs can store significant quantities of CO 2 . Despite their importance in the oil and gas industry, our understanding of CO 2 sequestration in unconventional reservoirs still needs to be developed. The objective of this paper was to use an extensive data set of numerical simulation results combined with data analytics and machine learning to identify the key parameters that affect CO 2 sequestration in depleted shale reservoirs. Machine learning-based predictive models based on multiple linear regression, regression tree, bagging, random forest, and gradient boosting were built to predict the cumulative CO 2 injected. Variable importance was carried out to identify and rank important reservoir and operational parameters. The results showed that random forest provided the best predictive ability among the machine learning techniques and that regression tree had the worst predictive ability, mainly because of overfitting. The most significant variable for predicting cumulative CO 2 sequestration was stimulated reservoir volume fracture permeability. The workflows, machine learning models, and results reported in this study provide insights for exploration and production companies interested in quantifying CO 2 sequestration performance in shale reservoirs.

Development of Empirical and Artificial Neural Network Model for the Prediction of Sorption Time to Assess the Potential of CO2 Sequestration in Coal

“…Machine learning techniques such as ANN have been utilized in solving complex nonlinear problems for CBM gas content prediction; reservoir analysis and prediction of production from well; – prediction of key reservoir properties; prediction of coal properties; and analysis of factors affecting productivity of CBM and factors controlling CO 2 sequestration …”

“…Machine learning techniques such as ANN have been utilized in solving complex nonlinear problems for CBM gas content prediction; 36 reservoir analysis and prediction of production from well; 37−40 prediction of key reservoir properties; 41 prediction of coal properties; 42 and analysis of factors affecting productivity of CBM 43 and factors controlling CO 2 sequestration. 44 Therefore, attempts are further made to explore the utilization of artificial intelligence tools such as ANN, which may provide a better prediction of sorption time as compared to a linear regression model. ANN is capable of identifying nonlinear relationships between the inputs and target which may further reduce the errors.…”