Unlocking Shale Gas Reserves with Machine Learning

Strategic Insights -

A new study from the School of Geosciences and Info-Physics has utilized machine learning techniques to predict the fracability index of shale gas formations, offering a faster and more cost-effective alternative to traditional experimental methods. The research, focusing on the Upper Ordovician Wufeng to Lower Silurian Longmaxi Formation in the Weiyuan shale gas field, Sichuan Basin, China, employed deep neural networks and metaheuristic algorithms to achieve exceptional predictive accuracy.

Key Takeaways:

  • The study used deep neural networks that integrated two metaheuristic algorithms-genetic algorithm (GA) and particle swarm optimization (PSO)-with the back-propagation technique to predict the fracability index (FI) of shale gas reservoirs.
  • The GABPNN model achieved R2, RMSE, and MAE values of 0.97531, 0.024754, and 0.0042875, respectively, while the PSOBPNN yielded values of 0.97494, 0.024938, and 0.0048962, respectively.
  • The PSOBPNN model attained a R2 value of 0.99848 when predicting FI values for a test well, indicating exceptional predictive accuracy.
  • The GABPNN and PSOBPNN models demonstrated nearly perfect prediction accuracy for FI in the testing dataset, underscored by their high R2 values.
  • The GABPNN model exhibited superior capability in mitigating overfitting, a common challenge in ML applications.
  • The research concluded that the ML-based approaches have the potential to optimize operations in shale gas reservoirs by facilitating the identification of sweet spots for fracturing.
  • The study was funded by the National Key R&D Program of China and the National Natural Science Foundation of China (NSFC).
  • Additional authors include Mbula Ngoy Nadege, Meshac B. Ngungu, Mutangala Emmanuel Arthur, and Kouassi Verena Dominique.

Statistics:

  • R2 value of the GABPNN model: 0.97531
  • RMSE value of the GABPNN model: 0.024754
  • MAE value of the GABPNN model: 0.0042875
  • R2 value of the PSOBPNN model: 0.97494
  • RMSE value of the PSOBPNN model: 0.024938
  • MAE value of the PSOBPNN model: 0.0048962
  • R2 value of the PSOBPNN model for a test well: 0.99848
  • The study used a dataset from the Weiyuan shale gas field, Sichuan Basin, China.

Sources:

  • Application of Ga/pso Metaheuristic Algorithms Coupled With Deep Neural Networks for Predicting the Fracability Index of Shale Gas Formations. Natural Resources Research, 2025.
  • School of Geosciences and Info-Physics, Cent South Univ, 932 South Lushan Rd, Changsha 410083, Hunan, People's Republic of China.
  • Springer. Van Godewijckstraat 30, 3311 Gz Dordrecht, Netherlands.