Incorporating subcritical crack growth mechanics into natural fracture characterization for improved reservoir simulation

Access full-text files




Philip, Zeno George

Journal Title

Journal ISSN

Volume Title



In conventional reservoir simulation, grid block permeabilities must frequently be assigned values systematically larger than those observed in core measurements to obtain reasonable history matches. Although part of the discrepancy might be due to improper 3-D scaleup, part of it might be due to substantial flow through natural fractures unaccounted for in the simulation. Here, a method to obtain equivalent permeabilities from fracture networks generated by a geomechanical model, constrained by geologic deformation and mechanical properties of the reservoir rock, is presented. A fracture mechanics based crack growth simulator, using laboratory measured subcritical indices and reservoir rock properties, rather than a purely stochastic method, was used to generate fracture networks with realistic clustering, spacing and fracture length distributions. The effects of subcritical index and bed-thickness on fracture attributes such as mean and total lengths were observed. These fracture networks were represented in a finite-difference simulator explicitly by using enhanced grid cell permeabilities and implicitly by using non-neighbor connections. Coupled fracture-matrix fluid flow simulations of the fracture networks, under uniform pressure boundary conditions, were performed to obtain equivalent permeabilities. Permeability enhancements by factors of between 2 and 10 were observed. The results also indicate that even though the permeability of individual fractures is highly sensitive to the fracture aperture, the computed equivalent permeabilities of a weakly-connected fractured region are not. They are more sensitive to the total and mean fracture lengths, as well as the nature of the fracture pattern, which include connectivity and clustering attributes. Superposing diagenetic effects (mineralization) on a fracture network can also reduce overall equivalent permeabilities. Equivalent permeability estimates made on the fracture networks under uniform flow rate boundary conditions generate lower values than those obtained with uniform pressure boundaries. Comparison of the equivalent permeabilities and flow behavior for fractured networks embedded in a larger homogeneous unfractured matrix indicate that uniform flow rate boundaries yield more reliable estimates of the equivalent permeability than uniform pressure boundaries.