Role of fault gouge during Interaction between hydraulic fracture and a preexisting fracture

Abstract

Enhanced reservoir connectivity generally requires maximizing the intersection between hydraulic fracture (HF) and preexisting underground natural fractures (NF), while having the hydraulic fracture continue to propagate across the natural fractures. Observations of downhole core samples suggest that these natural fractures are in fact veins filled with minerals such as calcite (Mighani et al., 2016). We study this interaction during the approach of a hydraulic fracture to a smooth saw-cut fracture under triaxial stress conditions. The specimen is Solnhofen limestone, a fine-grained (<5 µm grain), low permeability (<10 nD) carbonate. The differential stress (1-20 MPa) and inclination of the fault which determines the approach angle, θ (30, 60°) are the experimental variables. We conduct the experiments on both bare surface and gouge-filled fault surfaces. The gouge is a 1 mm thick crushed powder of Solnhofen limestone with <106 µm grain size. During the hydraulic fracture, acoustic emissions (AE), inferred slip velocity, axial stress and pore pressure are recorded at a 5 MHz sampling rate. The hydraulic fracture was able to cross the bare surface fault with small induced fault slip. The fault gouge increased the coefficient of friction significantly from 0.12 (bare, polished surface) to > 0.44 (gouged layer). However, the gouge-filled fault arrested the hydraulic fracture and generated a slip event with different characteristics: 1- The stress drop was larger while the generated AE signals had lower magnitude. 2- Slip velocity recorded by the vibrometer was of the same order of magnitude for the bare and gouge-filled faults, but the slip duration increased from 29 µsec for bare surface to ~2.5 msec (~90 times longer rise time) for the gouge-filled fault. The experiments suggest that the gouge-filled fault can accommodate much larger displacement while promoting slow slip on the fault which is harder to detect as AE signals. The observed long duration slip events are similar to the field observations of the long period and long duration (LPLD) events during the stimulation of clay-rich shale formations (Zoback et al., 2012). While the intrinsic low strength, high ductility, and unfavorably oriented natural fractures in shale formations are expected to reduce the occurrence of induced seismicity, our experiments suggest an additional mechanism for the observed LPLD events, i.e. the role of fault gouge. They also suggest that the microseismic detection techniques may under-predict the stimulated volume as the activation of natural gouge-filled fractures may proceed aseismically.