Full Program »

Image-based modeling of flow into and through proppant-filled fractures under variable stress

In the past decade, new technologies and advances in horizontal hydraulic fracturing have improved oil and gas production rates. However, there are still a number of fundamental questions regarding Darcy and non-Darcy flow processes in proppant-filled fractures under typical reservoir stresses. Improved understanding of the flow regimes and head losses as the fracture width changes and the proppants rearrange and/or break should lead to improved well production estimates. Here, we use synchrotron-based x-ray computed tomography (XCT) to nondestructively image, at 6 micron voxel resolution, a bulk proppant packed between Berea walls and a single-layer proppant located between two shale walls. Both systems were exposed to a series of loading stresses ranging from zero and 20,000 psi (137.9 MPa). The resulting XCT image datasets were segmented and quantitatively analyzed for packing and pore structure changes. Image-based flow modeling, using a Finite Element Method, was used to look at the flow patterns from the rock matrix into the fracture and at low- and moderate-Reynolds number flow through the proppant-filled fracture.

Both the images and quantitative grain analysis revealed changes as stress increased: changes in the packing structure, corresponding reduction in porosity, and some embedding into the rock of <0.5 proppant diameter. The shale system exhibited more embedding. At the highest stress in the Berea system (20000 psi or 137.9 MPa), individual proppant particles failed and the broken particles caused significant loss of permeability. Simulation results for each of the loadings showed that permeability is less sensitive to loading than experimental (vendor-reported) permeability values, but also show reasonable agreement at 8000 psi (55.16 MPa) for both systems. For the shale system, the embedding had a significant impact on the simulated permeability/fracture conductivity. Visualization of the flow and pressure distributions will be used to emphasize some of the important insights.

Author(s):

Clinton Willson    
Department of Civil & Environmental Engineering, Louisiana State University
United States

Paula Sanematsu    
Department of Civil & Environmental Engineering, Louisiana State University
United States

Yijie Shen    
Cain Department of Chemical Engineering, Louisiana State University
United States

Karsten Thompson    
Craft and Hawkins Department of Petroleum Engineering, Louisiana State University
United States