Abstract:
Laboratory-scale experiments are an important method for improving understanding of the hydraulic fracturing process under controlled circumstances, even though it is difficult to realistically simulate downhole conditions in a laboratory. Laboratory-scale experiments can also help determine if a numerical model can be used as a predictive tool for hydraulic fracturing analysis. This paper discusses both experimental and numerical modeling performed for a hydraulic fracturing test on a 30×30×45 cm specially designed cement block. The experiment was conducted using a true poly-axial testing system, and was designed to obtain data regarding the behavior of a hydraulically induced fracture and its interaction with a pre-existing fracture. The experimental data can provide insight to better understand the mechanisms of fracture initiation, propagation, and interaction. A computational algorithm was used to perform numerical modeling for the laboratory fracturing test. The computational algorithm tightly couples geomechanics and fluid dynamics models while providing fracture propagation criteria for hydraulically induced fractures. The numerical simulation results are compared in this work to experimental results, illustrating that the numerical model is capable of describing the interaction between pre-existing fractures. It also can help improve the understanding of complex fracture growth in reservoir conditions.
Introduction
Complex fracture geometries are often observed in naturally fractured reservoirs, such as tight gas sandstones and shales. Mineback studies can provide direct evidences of fracture geometries, such as the studies conducted by Warpinski, 1987. But mineback is costly and only viable for shallow depths. Another way to study the complex fracture geometry is through laboratory-scale experimental studies, which can improve understanding of the hydraulic fracturing process under controlled circumstances.
Laboratory-scale experiments have been used to investigate the interaction between hydraulic fractures and pre-existing fractures. Blanton, 1982 shows that the principal stress differential and the approaching angle between hydraulic fractures and natural fractures are key parameters. Beugelsdijk et al., 2000 examined the impact of pre-existing fractures, stress difference, and injection rate on fracture path complexity. These studies mainly focused on the interaction between hydraulically induced fractures and pre-existing natural fractures. The fracture interaction and propagation geometry within pre-existing fractures is still not well studied in laboratory-scale experiments.
