Abstract:Abstract Shale gas reservoirs have become prominent contributors to the world's hydrocarbon resources and production. They exhibit multiple storage mechanisms, two of which are linked to the free and adsorbed gas phase. Since the adsorbed gas may be stored as a denser phase than the free gas, the contribution of the adsorbed phase can be significant. The adsorbed volume is related to the total organic carbon (TOC) and thus, higher TOC can indicate higher hydrocarbon inplace. Furthermore, productivity can be linked to TOC through the potential for overpressure and conversion of kerogen to pore space. However, estimation of the TOC is not a trivial problem, as it depends on geological factors such as depositional environment. In this study, we propose an integrated workflow using concepts of machine learning to estimate TOC. The workflow is divided into 3 sections which are area selection, sub-region categorization, and prediction modeling. Firstly, 3 active exploration and development areas (Kaybob, Pembina, and East shale basin) of the Duvernay Formation are highlighted and the geology of each specific area is analyzed. Thereafter, using the available core data and average properties of the attributes (Gamma Ray, resistivity, density, and distance from mean vitrinite reflectance line), each area is clustered into sub-regions using SVM, logistic regression, and k-means clustering. Finally, using Random Forest prediction, models for each sub-region are developed and ranked with average mean square errors and standard deviations. It is observed that the Kaybob area can be clustered into 2 regions. This observation is supported by the principal component plot (PC1 vs PC2), which shows a dual cloud structure. This is further supported through clustering analysis, which also revealed the same observation. Results of the prediction modeling found random forest as the best predictor, achieving a match wiht the core data with a error less than 10% and in some cases only a 1% deviation. Shale reservoir characterization requires estimation of the key parameters such as TOC. However, it is difficult to estimate TOC with purely physics-based or purely statistical models, as it requires limited specialized data and is impacted by subtle variations in the reservoir. This study suggests that TOC can be accurately estimated by combining geological interpretation and machine learning based algorithms without bearing cost of the specialized data.

Prediction of Shale-Gas Production at Duvernay Formation Using Deep-Learning Algorithm

Application of Machine Learning Method of Data-Driven Deep Learning Model to Predict Well Production Rate in the Shale Gas Reservoirs

Time Series Modeling for Production Prediction of Shale Gas Wells

Prediction of shale gas horizontal wells productivity after volume fracturing using machine learning – an LSTM approach

A deep-learning-based prediction method of the estimated ultimate recovery (EUR) of shale gas wells

A physical constraint-based machine learning model for shale oil production prediction

Prediction of gas production potential based on machine learning in shale gas field: a case study

A Deep Learning-Based Approach for Predicting Oil Production: A Case Study in the United States

Optimized long short-term memory (LSTM) network for performance prediction in unconventional reservoirs

A Combined Neural Network Forecasting Approach for CO2-Enhanced Shale Gas Recovery

Shale gas well flowback rate prediction for Weiyuan field based on a deep learning algorithm

Machine Learning Based Integrated Approach to Estimate Total Organic Carbon in Shale Reservoirs – A Case Study from Duvernay Formation, Alberta Canada

Deep Learning and Data Assimilation for Real-Time Production Prediction in Natural Gas Wells

A Gaussian Decomposition Method and Its Applications to the Prediction of Shale Gas Production

A framework for timely and accessible long-term forecasting of shale gas production based on time series pattern matching

Application of Machine Learning for Shale Oil and Gas "Sweet Spots" Prediction

A novel governing equation for shale gas production prediction via physics-informed neural networks

Predicting production-rate using wellhead pressure for shale gas well based on Temporal Convolutional Network

Enhancing pressure gradient prediction in multi-phase flow through diverse well geometries of North American shale gas fields using deep learning

Gas well performance prediction using deep learning jointly driven by decline curve analysis model and production data

Production Forecast of Deep-Coalbed-Methane Wells Based on Long Short-Term Memory and Bayesian Optimization