Machine Learning Method for TOC Prediction: Taking Wufeng and Longmaxi Shales in the Sichuan Basin, Southwest China as an Example

Abstract

The total organic carbon content (TOC) is a core indicator for shale gas reservoir evaluations. Machine learning-based models can quickly and accurately predict TOC, which is of great significance for the production of shale gas. Based on conventional logs, the measured TOC values, and other data of 9 typical wells in the Jiaoshiba area of the Sichuan Basin, this paper performed a Bayesian linear regression and applied a random forest machine learning model to predict TOC values of the shale from the Wufeng Formation and the lower part of the Longmaxi Formation. The results showed that the TOC value prediction accuracy was improved by more than 50% by using the well-trained machine learning models compared with the traditional Δ Log R method in an overmature and tight shale. Using the halving random search cross-validation method to optimize hyperparameters can greatly improve the speed of building the model. Furthermore, excluding the factors that affect the log value other than the TOC and taking the corrected data as input data for training could improve the prediction accuracy of the random forest model by approximately 5%. Data can be easily updated with machine learning models, which is of primary importance for improving the efficiency of shale gas exploration and development.

Prediction of total organic carbon and E. coli in rivers within the Milwaukee River basin using machine learning methods

“…[17][18][19][20][21][22][23] Several studies explored different ML techniques in predicting TOC to characterize the hydrocarbon potential of source rocks, soil, organic shale, and mudstone. [24][25][26][27][28] However, to the best of the authors' knowledge limited studies have been made for developing TOC prediction models in natural streams. Yeon et al (2008) 29 , Goz et al (2019) 30 , and Kim et al (2021) 31 explored the application of ANN, kernel extreme machine learning, and extreme machine learning models with different activation functions to estimate TOC of rivers.…”

Modeling of Burial History, Source Rock Maturity, and Hydrocarbon Generation of Marine-Continental Transitional Shale of the Permian Shanxi Formation, Southeastern Ordos Basin

“…There may be some discrepancies between the thickness and TOC of source rocks predicted from logging and geological data, which may have affected gas generation. The heterogeneity of source rocks may also be reflected in the differences in the kinetic parameters of hydrocarbon generation. ,, Therefore, the hydrocarbon generation intensity of source rocks at different locations may be different, so that the R o -hydrocarbon generation curve of a single sample leads to uncertainty in gas generation. Otherwise, the TOC content changes during the thermal evolution of the source rocks, and the measured TOC content of the source rock reflects the residual organic carbon content after hydrocarbon generation.…”

“…The maturity of the Shanxi Formation shales is close to that of the major gas-bearing shales in the world, such as the Longmaxi Formation shale, Woodford shale, and Haynesville shale. 44 , 45 According to the gas generation curve of the Upper Paleozoic source rocks, there was a good positive correlation between R o and the gas generation, and the higher the R o value, the higher the gas yield. 46 An approximate calculation revealed that the cumulative gas generation of the Shanxi Formation shale was approximately 44.4 × 10 12 m 3 /km 2 , indicating that the highly mature Shanxi Formation shale in the study area had a high gas production capacity.…”

An Improved Computational Learning-Based Model for Estimating Total Organic Carbon in Unconventional Shale Reservoirs

“…The successful application of computational intelligence (CI) in hydrocarbon exploration and exploitation in recent years, has seen the adoption of intelligence learning models in predicting TOC from well log data. [12][13][14][15][16][17][18][19][20][21][22][23] Computing intelligence is a captivating discipline that combines computational power with human intelligence to develop sophisticated and trustworthy solutions to stunningly nonlinear and complicated problems. The CI models have the advantage of being able to adapt and learn to the dynamic conditions of the reservoir such as depositional and formation environment whilst utilizing the entire suite of well logs for better prediction of TOC.…”