Hydraulic fracturing under waterflooding conditions in unconsolidated sands

Abstract

Rapid injectivity decline is frequently observed during injection in unconsolidated sand reservoirs. Field data suggests that hydraulic fracture processes are directly or indirectly related to this injectivity decline. Conventional fracture theories do not apply to unconsolidated sand since this material has little to no cohesion and tensile strength. The main fracture mechanism hypotheses are shear failure of the zone ahead of the fracture tip and fluidization. For both mechanisms, a fluid pressure high enough to initiate and propagate the fracture is required. The fluid pressure is dependent on the injection rate, fluid viscosity and permeability of the formation. Unconsolidated sands have a high permeability, thus under normal waterflood conditions a high injection pressure is not expected. Three main impairment mechanisms leading to the injectivity decline have been identified based on field evidence and previous work:

-Plugging
-Wellbore fill
-Resorting of grains and finer particles

Plugging results from the infiltration of fines originating from the injection fluid, crossflow or drilling mud. The external and/or internal filter cake can locally reduce the permeability of the formation. During surface shut-ins, backflow and/or crossflow can occur leading to the infiltration of solid particles and fluids into the wellbore. This reduces the leak-off area of the well. Lastly, resorting of grains and finer particles can result in a denser packing of the reservoir. The dynamically mixing of particles can lead to lower permeability regions.

Research goal
The main goal of this research is to develop a better qualitative and quantitative description of the fracturing process and the impairment mechanisms causing the observed injectivity decline. This thesis comprises of the first phase of this research, focusing on the capabilities of the equipment to create and detect fractures under waterflooding conditions. What makes this research unique is the use of low viscosity fluids, to mimic field conditions. Other work often involves the use of efficient fracturing fluids that have a high viscosity and/or good filter cake building capabilities to minimalize the leak-off. Next to that, injection of fluids is performed live in a CT scanner. This allows the visualization of fractures or low-density regions through density distributions in three dimensions over time.
Equipment
Injection takes place in a high-strength aluminium vessel with a sample volume of 3.84 dm3. See images 2.1 and 2.2 of an overview of the setup and pressure vessel. Axial and radial pressure can be controlled independently up to 20 MPa. A pore fluid system records the outflow mass and provides a fluid pressure on the sample. The sample consists of a very fine, very well sorted sand with a permeability around 5 Darcy. The main injection fluids are water and Fluorinert FC-770. This is a high density, low viscosity fluid that is used to visualize the preferential flow path of the injection fluid in the CT scanner.
Results
Fractures have been successfully created using a high viscosity fluid during the first experiment. The goal of that experiment was to test the setup and the equipment. The fractures were created at an injection pressure of 38 MPa and were up to 1 cm long and 2 mm wide. Experiments 2 and 3 were performed in the CT scanner with the use of Fluorinert as the injection fluid. The infiltration zone of this fluid was clearly visible but no fractures were created. Sand infiltration in the injection tube leaded to a number of problems during the experiments. Experiments 4 and 5 added fines to the injection water. Quartz powder was used in experiment 4 and bone meal in experiment 5. The fines leaded to a gradual increase in injection pressure, but did not lead to a higher density in the CT scans. No fractures were observed but low-density regions in front of the perforations were created during both experiments as a result of backflow. Experiment 6 introduced the use of internal pressure sensors in the sample and used a sample created with two sands and Kaolinite, a non-swelling clay. During two high flowrate injection cycles, the clays migrated away from the near wellbore region, leaving behind lower density regions.
Conclusion & future work
Creating fractures with low viscosity fluids in a laboratory environment has proven to be difficult. No fractures have been created throughout the low viscosity experiments. Several impairment mechanisms that were identified in the field have also been observed in the experiments. This thesis forms a solid basis for the next research phase to investigate these impairment mechanisms more closely. By lowering the confining stresses, increasing the flow rates and decreasing the sample permeability, there is a good probability that fractures can be created with this equipment in future work.