Experimental and Theoretical Investigation Of Multiple Hydraulic Fractures Propagation Behavior In Tight Reservoir

Abstract

Multi-stage fracturing is widely used in low-permeability reservoirs to enhance hydrocarbon recovery. However, stress disturbances (or stress shadow effects) between fractures can cause fluctuations in breakdown pressures and variations in propagation paths. This complexity complicates the accurate characterization of fracture morphology and spatial distribution, posing challenges for fracturing design and field operations. Understanding fracture initiation modes, propagation paths, and distribution patterns in multi-stage fracturing-along with the mechanisms and effects of stress disturbances-is essential for designing and optimizing fracturing schemes to enhance tight oil and gas recovery. However, due to the complexity of this issue and the limitations of existing technologies, fracturing design and field operations remain largely empirical, often resulting in suboptimal reservoir stimulation. This study performed true triaxial hydraulic fracturing experiments on artificial single-well and dual-well models to investigate hydraulic fracture propagation and spatial distribution in multi-stage fracturing. To facilitate comparison and clarify the differences between single-well and dual-well multi-stage fracturing while minimizing the influence of natural rock heterogeneity, this study focused on the impact of fracturing methods on fracture behavior. Tracers, CT imaging, and the U-net convolutional neural network algorithm were developed to extract and quantify the spatial distribution differences of hydraulic fractures. Using the weight function stress intensity factor analytical solution, a solid-phase equilibrium model incorporating parallel fractures was developed by integrating the weight function into the induced stress model. This model validated the experimental results and explored the stress shadow mechanisms induced by variations in fracture spacing and angles. Additionally, a stress disturbance factor was introduced to quantify the mechanical interactions that enhance or inhibit the growth of parallel hydraulic fractures. The results show that in single-well fracturing, hydraulic fracture propagation is primarily governed by reservoir properties and in-situ stresses, with fractures predominantly propagating along the maximum horizontal principal stress (σH). These fractures exhibit smooth surfaces and relatively regular shapes. In contrast, dual-well sequential fracturing reduces stress disturbances, promoting the development of more complex fracture structures. The peak injection pressure in the second stage of sequential fracturing is generally higher than in the first stage. Tracer analysis reveals that fractures generated by sequential fracturing adopt a penny-shaped distribution, with inter-fracture disturbances influencing both fracture radius and propagation paths. A stress disturbance factor was introduced to quantitatively determine the optimal spacing between parallel fractures and elucidate the mechanisms governing complex fracture network formation. This factor enables the quantitative analysis and prediction of fracture diversion behavior during hydraulic fracturing.

Keywords: Hydraulic fracturing, stress shadow, dual-well fracturing, sequential fracturing, weight function-induced stress model