This study focused on developing machine learning models to detect leak size and location in transient state conditions. The model was designed for an onshore methane–hydrogen blending gas pipeline in Canada. Base case simulations revealed significant effects on mass flow and pressure due to leaks, with the system taking approximately 6 h to reach a steady state from transient conditions. This made it essential to analyze the flow characteristics during the transient state. Trend data from the pipeline’s inlet and outlet were examined, considering the leak size and location. To better represent the data over time, a method was used to create two-dimensional images, which were then fed into a CNN (convolutional neural network) for training. The model’s accuracy was assessed using classification accuracy and a confusion matrix. By refining the data acquisition process and implementing targeted screening procedures, the model’s classification accuracy increased to over 80%. In conclusion, this study demonstrates that machine learning can enable rapid and accurate leak detection in transient state conditions. The findings are expected to complement existing leak detection methods and support operators in making faster, more informed decisions.

Coalbed methane (CBM) content is generally estimated using the isotherm theory between pressure and adsorbed amounts of methane. It usually determines the maximum content of adsorbed methane or storage capacity. However, CBM content obtained via laboratory experiment is not consistent with that in the in-situ state because samples are usually ground, which changes the specific surface area. In this study, the effect of the specific surface area relative to CBM content was investigated, and diffusion coefficients were estimated using equilibrium time analysis. The differences in adsorbed content with sample particle size allowed the determination of a specific surface area where gases can adsorb. Also, there was an equilibrium time difference between fine and lump coal, because more time is needed for the gas to diffuse through the coal matrix and adsorb onto the surface in lump coal. Based on this, we constructed a laboratory-scale simulation model, which matched with experimental results. Consequently, the diffusion coefficient, which is usually calculated through canister testing, can be easily obtained. These results stress that lump coal experiments and associated simulations are necessary for more reliable CBM production analysis.

Most shale gas reservoirs have extremely low permeability. Predicting their fluid transport characteristics is extremely difficult due to complex flow mechanisms between hydraulic fractures and the adjacent rock matrix. Recently, studies adopting the dynamic modeling approach have been proposed to investigate the shape of the flow regime between induced and natural fractures. In this study, a production history matching was performed on a shale gas reservoir in Canada’s Horn River basin. Hypocenters and densities of the microseismic signals were used to identify the hydraulic fracture distributions and the stimulated reservoir volume. In addition, the fracture width decreased because of fluid pressure reduction during production, which was integrated with the dynamic permeability change of the hydraulic fractures. We also incorporated the geometric change of hydraulic fractures to the 3D reservoir simulation model and established a new shale gas modeling procedure. Results demonstrate that the accuracy of the predictions for shale gas flow improved. We believe that this technique will enrich the community’s understanding of fluid flows in shale gas reservoirs.