Use of Well Testing in Discrete Fracture Network Modeling

Over the past thirty years, the research efforts of radioactive waste programs, results from groundwater contamination research sites, and advances in petroleum technology have developed tools for characterizing flow in fracture networks.  A modern workflow includes the following:

• Locating conductive fractures using flow logging or other forms of detailed hydraulic testing,
• Characterizing the geology and geometry of conducting features using image logs, and
• Isolating and monitoring significant conductors using multi-zone monitoring systems.

These methods provide a basis for building practical discrete fracture network models that simulate the major, deterministic features having known locations and properties along with stochastic fractures that represent background connectors. 

Well test data provide an important means of defining network geometry and testing the validity of network models.  The pressure derivative method, which has revolutionized well-test interpretation, is fundamental to evaluating the hydraulic geometry of the fracture network.  Examples using simple fracture-network models illustrate how networks control the shape of the pressure derivative.   Specific cases show the influence of fracture connectivity in a two-dimensional region, the effect of a single, dominant fracture orientation, and the influence of faults or other features that provide vertical connectivity across stratified aquifers. 

Hydraulic diffusivity is a key parameter for understanding flow in fracture networks. Diffusivity is the ratio of a network’s transmissive properties to its storage properties.  High diffusivities indicate the likelihood that system is fracture-controlled, and variations in point-to-point diffusivities provide perhaps the most reliable means for mapping network connectivity.  Simulations of pumping well behaviors in fracture network models are non-unique with respect to the scale of pressure responses.  Diffusivity values from observation-well responses constrain the appropriate range of parameter combinations.