GEOMECHANICAL MODELING OF HYDRAULIC FRACTURING: IMPORTANCE OF MECHANICAL STRATIGRAPHY, STRESS STATE, AND PRE-EXISTING STRUCTURES

The increasing exploration and production in unconventional resource plays in the past decade has been accompanied by a greater need for understanding the effectiveness of multi-stage hydraulic fracturing programs, particularly in long (>1500 m or 5000 ft) sub-horizontal boreholes (“laterals”). Traditional (analytical) analysis techniques for estimating the size and orientation of fractures induced by fluid injection typically result in predictions of relatively long and planar extension (mode I) fractures that may not be representative of natural systems. While these traditional approaches offer the advantage of rapid analysis, the lack of inclusion of key features of the natural system (e.g., realistic mechanical stratigraphy, pre-existing natural faults and fractures, and heterogeneity of in situ stresses) may render results unrealistic for planning, executing, and interpreting multimillion dollar hydraulic stimulation programs. Numerical geomechanical modeling provides a means of including key aspects of natural complexity in simulations of hydraulic fracturing. In this study, we present results of finite-element-based geomechanical modeling of fluid-injection-induced rock deformation that combines a coupled stress-pore pressure analysis with a continuum damage-mechanics-based constitutive relationship. The models include both the natural mechanical stratigraphic variability as well as the in situ stress state anisotropy, and permit tracking of the temporal and spatial development of shear and tensile permanent strains that develop in response to fluid injection. Our results show that simple, long planar fractures are unlikely to be induced in most mechanically layered natural systems under typical in situ stress conditions. Analyses that assume this type of fracture geometry may significantly overestimate the reach of hydraulically induced fractures and/or effectively stimulated rock volume.