Natural fracture pattern characterization using a mechanically-based model constrained by geologic data—Moving closer to a predictive tool

Abstract

Natural fracture patterns are difficult to characterize because observations are too sparse and often difficult to interpret. Fortunately, fracture pattern characteristics are sensitive to other observable/inferable quantities such as in situ stress state, pore pressure, strain rate, material properties and fractured bed thickness. To take advantage of this, modelers have started to turn toward techniques that are sensitive to the mechanics of fracture propagation. The model presented here is based on a displacement-discontinuity, boundary-element numerical technique. A subcritical crack propagation law is used to better represent fracture growth under geologic conditions (i.e., long term, low strain rate loading). Mixed-mode fracture propagation examples display how in situ stress contrasts and bed thickness can control fracture connectivity. Variations in material properties such as the subcritical growth index affect fracture spacing and length distributions for a given loading history. An approximate correction factor has been incorporated into the model for fracture height (bed thickness) effects. In addition, the ability to load a rock body at various strain rates has been added, allowing detailed observations on the time-dependent factors that control subcritical crack growth. Results from the forward modeling indicate that there is a systematic relationship between boundary conditions and final fracture geometry that can be exploited for the purpose of fracture pattern inversion from observed data.