文摘
Proppant is a vital component of hydraulic stimulation operations, improving conductivity by maintaining fracture aperture. While correct placement is a necessary part of ensuring that proppant performs efficiently, the transport behavior of proppant in natural rock fractures is poorly understood. In particular, as companies pursue new propping strategies involving new types of proppant, more accurate models of proppant behavior are needed to help guide their deployment. A major difficulty with simulating reservoir-scale proppant behavior is that continuum models traditionally used to represent large-scale slurry behavior loose applicability in fracture geometries. Particle transport models are often based on representative volumes that are at the same scale or larger than fractures found in hydraulic fracturing operations, making them inappropriate for modeling these types of flows. In the absence of a first-principles approach, empirical closure relations are needed. However, even such empirical closure relationships are difficult to derive without an accurate understanding of proppant behavior on the particle level. Thus, there is a need for experiments and simulations capable of probing phenomena at the sub-fracture scale. In this paper, we present results from experimental and numerical studies investigating proppant behavior at the sub-fracture level, in particular, the role of particle dispersion during proppant settling. In the experimental study, three-dimensional printing techniques are used to accurately reproduce the topology of a fractured Marcellus shale sample inside a particle-flow cell. By recreating the surface in clear plastic resin, proppant movement within the fracture can be tracked directly in real time without the need for X-ray imaging. Particle tracking is further enhanced through the use of mixtures of transparent and opaque proppant analogues. The accompanying numerical studies employ a high-fidelity three-dimensional particle-flow model, capable of explicitly representing the particles, the fracture surface and the interstitial fluid flow. Both studies reveal large-scale vortex motion during particle settling. For the most part, this behavior is independent of the fracture topology, instead driven by interactions between the sinking particles and the upwelling interstitial fluid. This motion results in large amounts of particle dispersion, significantly greater than might be expected from traditional slurry models. The competition between the particles and the fluid also results in a redistribution of particles toward the fracture walls, which has significant implications for the transport of proppant along the fracture.
