ABSTRACT

Hydraulic fracturing creates highly conductive flow channels by injecting fracturing fluid at high pressures. A complex fracture network is generated in shale due to the contribution of the natural fractures, bedding planes and other zones of weakness. X-ray micro-computed tomography (μCT) has been used to visualize fracture complexity. The visualization accuracy is limited to the scanning resolution and density contrast within the material. An experimental approach is presented in this paper to three-dimensionally (3D) map the fracture network at micron (μm) resolution. A 3D fracture network is reconstructed based on serial-sectioned digital images of a shale cube hydraulically fractured under true-triaxial conditions. The hydraulic fractures are identified as opened natural fractures, activated bedding planes, and newly generated hydraulic fractures. Activated bedding planes represent the majority of hydraulic fractures, which significantly contributes to the lateral growth of the fracture geometry. Rough and crooked natural fracture planes can be observed. On the contrary, the activated bedding planes are smooth and relatively straight. This fracture mapping method can also be applied to quantitatively study fracture aperture to ultimately improve the conductivity prediction of the fracture network.

1. INTRODUCTION

Hydraulic fracturing creates highly conductive flow channels in low-permeability reservoirs by injecting fracturing fluid at high pressures to stimulate reservoir production. The complexity of the fracture network created by hydraulic fracturing in shale is a result of the interference between the hydraulically induced fractures and the existing weak planes, such as bedding planes and natural fractures. Increasing the stimulated reservoir volume (SRV) or enhancing the complexity of the fracture network can be beneficial to improve production. Understanding the fracturing mechanism by studying the fracture complexity is of vital importance to hydraulic fracturing design and to enhance production.

A variety of destructive and non-destructive experimental methods (Ramandi et al., 2017; Tan et al., 2017) have been used for studying fracture complexity in rock mass. The non-destructive observational method μCT has advantages in mapping the fracture complexity over destructive methods. However, visualization of the fracture network in large samples can be problematic. The problem arises from the compromise between the resolution of acquired images using the μCT and capturing a representative field of view (FOV) of the sample. Capturing a large FOV means that micron-scale fractures are represented by only a few voxels in the direction normal to the fractures and in some cases they are sub-resolution. Additionally, existing weak planes that exhibit low density contrast to the rock matrix may be misconstrued as fractures during the process of image segmentation. To overcome these two limitations of relying solely on μCT to capture fracture complexity in large samples, an experimental method is proposed in this paper to three-dimensionally (3D) map the fracture network at micron (μm) resolution utilizing the idea of serial section reconstruction. The method of 3D reconstruction of serial section has been successfully used in medicine and biology for centuries since its first applications in embryology (Levinthal and Ware, 1972). To the authors knowledge, this is the first time this method is being applied to reconstruct the fracture geometry in a hydraulically fractured shale sample.

This content is only available via PDF.
You can access this article if you purchase or spend a download.