Hydraulic Fracturing with Supercritical CO2 in the Laboratory: Ultrasonic Monitoring of Fracture Propagation at in Situ Stress and Temperature

Abstract

ABSTRACT The paper reports results of hydraulic fracturing experiments. The fracturing was completed under in situ conditions in 15 cm cube of American Black granite and Montney shale samples using supercritical CO2. Different modes of ultrasonic monitoring have been used. The final fractures were also imaged with fluorescent resin. In contrast to previously reported tests, here we use very strong rocks, high confinement, and stress contrast. INTRODUCTION Injection of carbon dioxide (CO2) into subsurface rock formations presents multiple benefits, including reduction in greenhouse gas emissions into the atmosphere. One way of improving the economics of geo-sequestration is to use CO2 as a fracturing fluid to enhance gas productivity in unconventional reservoirs. CO2 can also be used to enhance production of coal seam reservoirs (CO2-ECSB). Understanding fracture initiation/propagation in presence of CO2 is also crucial for the long-term containment of the storage. Monitoring remotely fracture propagation can be achieved using seismic methods. To improve the interpretation of seismic monitoring data in the field we conducted hydraulic fracturing (HF) experiments with supercritical CO2, sc-CO2, in the laboratory under true triaxial stress conditions. We used ultrasonic methods to monitor fracture propagation in cubic rock samples 150 mm in size. For verification purposes the resulting fracture network was visualised after filling it with fluorescent resin. We present here the results of experiments conducted on low permeability American Black granite and shale rock samples. EXPERIMENTAL SETUP Equipment The experiments were conducted in a true triaxial stress cell (TTSC) developed at Curtin University. The maximum compressive load is 1575kN in each spatial direction, while the maximum temperature is about 60°C. The loading stress is transferred to the sample via steel plates (see Figure 1). Further details about the TTSC and the experimental setup can be found in Wang et al. (2019). Each plate accommodates 10 ultrasonic transducers (pink circles in Fig 1). The transducers are spring loaded against the sample's surface. For this study we used commercially available P-wave ultrasonic transducers (Olympus V103-SM) with a resonant frequency of 1 MHz and a bandwidth from 0.5 to 1.5 MHz. The outer diameter of the transducer is 18 mm, and the crystal size is 13 mm. The transducers were connected to a detection/recording system developed in-house at CSIRO. Due to the system limitations at the time of testing, we could only use 32 ultrasonic transducers out of 60 available (see Figure 2).