LAUSR

Abstract Hydraulic fracturing is an effective stimulation technology for the development of tight and shale reservoirs. Because the propagation morphology of hydraulic fractures is closely related to the production performance and the estimated ultimate recovery of unconventional formations, it is essential to accurately characterize the fracture morphology on the reservoir scale. On the basis of the fracture propagation simulations obtained using the cohesive element method, we estimated the fractal dimensions of the hydraulic fractures under different conditions using an improved box-counting method. We analyzed the propagation process of hydraulic fractures and its dependence on natural fractures. To obtain the exact fractal dimensions, we utilized the box flex method to increase the fitted data points and analyzed the effects of different factors. Reasonable selections of the self-similar interval and the image resolution play significant roles in the estimation of fractal dimension. We found that the fractal dimension increases with time and the existence of natural fractures increases the fractal dimension. We obtained the correlations of different fracture parameters against the fractal dimension to quantitatively characterize the propagation of hydraulic fractures. The fractal dimension mainly depends on the injection rate of fracturing fluid, whereas the number of perforation clusters and fluid viscosity show only negligible influences on the fracture morphology. This study can be useful for the characterization of hydraulic fracture networks and the production performance prediction in tight reservoirs.