Tiny micrometer-scale fractures, or “cleats,” permeate coal beds and create networks that often harbor a form of natural gas known as coal seam gas. Coal seam gas consists primarily of methane and is extracted for energy production in a growing number of countries, including the United States.
Coal seam gas has generated controversy in certain regions, with some raising questions about its potential environmental effects. This has created interest in applying advanced characterization methods to better understand gas production from these resources and control its environmental impacts. A new paper by Jing et al. validates a new strategy to advance understanding of coal cleat networks, which is key to modeling the flow of gas through the tiny cracks.
The new strategy builds on an existing method in which X-ray microcomputed tomography (micro-CT) imaging is used to analyze rock properties. A shortcoming of this method is that “noisy” micro-CT data of coal can create the appearance of extra cleats and connections between cleats where there are none.
To address this problem, the authors of the new study recently developed a new discrete fracture network (DFN) model—a computational tool for analyzing fractures in rocks. The new model, published in 2016, incorporates certain coal cleat characteristics (shape, length, and orientation) gleaned from micro-CT images, but it was unclear how well it captured connections between the tiny fractures. These connections control how gas will be produced from coal beds.
The authors have now put their new DFN model to the test. First, they used a micro-CT instrument to scan a coal sample from the Moura coal mine in Australia. Then they constructed three different representations of the sample’s internal cleat network: one using the original micro-CT images, one using micro-CT images that were filtered to remove noisy data, and one using the DFN model, which incorporated some of the micro-CT data.
The researchers then developed and applied a series of mathematical tests to compare the accuracy of the three representations. They found that the unfiltered micro-CT images showed extensive connections between individual cleats, especially at smaller scales. Meanwhile, the filtered images and the DFN model showed lower levels of connectivity at both small and whole-sample scales.
The authors concluded that the developed DFN model more realistically captures cleat connections than do typical segmented micro-CT images. It also requires less computing power and avoids the trade-offs between sample size and image resolution, which often hamper the micro-CT method. Thus, the DFN model could improve simulations of gas flow through coal, with potential applications for commercial extraction. (Journal of Geophysical Research: Solid Earth, https://doi.org/10.1002/2017JB014667, 2017)
—Sarah Stanley, Freelance Writer