An adsorbed gas estimation model for shale gas reservoirs via statistical learning

Shale gas plays an important role in reducing pollution and adjusting the structure of world energy. Gas content estimation is particularly significant in shale gas resource evaluation. There exist various estimation methods, such as first principle methods and empirical models. However, resource evaluation presents many challenges, especially the insufficient accuracy of existing models and the high cost resulting from time-consuming adsorption experiments. In this research, a low-cost and high-accuracy model based on geological parameters is constructed through statistical learning methods to estimate adsorbed shale gas content. The new model consists of two components, which are used to estimate Langmuir pressure (PL) and Langmuir volume (VL) based on their quantitative relationships with geological parameters. To increase the accuracy of the model, a “big data” set that consists of 301 data entries was compiled and utilized. Data outliers were detected by the K-Nearest Neighbor (K-NN) algorithm, and the model performance was evaluated by the leave-one-out algorithm. The proposed model was compared with four existing models. The results show that the novel model has better estimation accuracy than the previous ones. Furthermore, because all variables in the new model are not dependent on any time-consuming experimental methods, the new model has low cost and is highly efficient for approximate overall estimation of shale gas reservoirs. Finally, the proposed model was employed to estimate adsorbed gas content for nine shale gas reservoirs in China, Germany, and the U.S.A.