Understanding proppant transport plays a critical role in estimating propped fracture dimensions and performance. Existing models generally assume a vertical planar geometry, whereas the reality in the subsurface may be much more complex. In this study, we use the discrete element method to simulate proppant transport in a 3D fracture system. The system geometry involves a hydraulic fracture intersecting preexisting natural fractures. In the numerical investigations for this study, we consider different natural fracture apertures and intersection, as well as proppant of various sizes and concentrations.
By analyzing the proppant distribution in both the hydraulic fractures (HF) and natural fractures (NF), we identified the conditions under which continuous flow of proppants can be achieved. Our results show that large apertures of the natural fractures are critical in achieving continuous proppant transport. Narrow natural fractures increase the particle-particle and particle-wall interaction, thus causing blockage at the intersection of HFs and NFs, and, consequently, limiting the proppant transport efficiency and the fracture effectiveness. However, our results show that the proppant concentration can influence its transport efficiency by alleviating the blockage at these HF and NF intersections. Thus, injecting fluid with low proppant concentration can wash out the proppant pack accumulated at these points and relieve the blockage, leading to improved proppant placement efficiency in the fractures.
The success of oil and gas production from shale reservoirs depends greatly on the effectiveness of hydraulic fracturing treatment. To achieve an optimal propped (effective) fracture area, understanding proppant transport plays a critical role. Model development for proppant transport in low viscosity fluids (slick water) has been a fastgrowing field of research. In general, the proposed models can be categorized into three groups: the Eulerian two-fluids model (Dontsov and Peirce, 2014), the Lagrangian particle model (Tsai et al. 2012) and the hybrid model (Tong et al., 2016). The Eulerian model treats proppants as a continuous phase. This approach is computational efficient, and is thus capable of simulating large-scale transport processes in a timely manner. However, the assumption of the particles being a continuous phase is questionable, especially when a low viscosity carrier fluid is used in the stimulation. An additional weakness of the Eulerian model is that it does not consider particle-particle collisions and particle-boundary collisions. This may result in non-physical transport behavior and over-estimation of the proppant distribution.
