--> ABSTRACT: Strain Localization Finite Element Models for Fracture Prediction, by Nigel G. Higgs; #90906(2001)

Datapages, Inc.Print this page

Nigel G. Higgs1

(1) Higgs Technologies, L.L.C, Tulsa, OK

ABSTRACT: Strain Localization Finite Element Models for Fracture Prediction

An advanced elastoplastic material model is utilized within a large-strain, nonlinear finite element code to predict the onset and development of complex rock fracture patterns. The technique uses a strain-softening constitutive model from the concrete literature, calibrated so as to simulate triaxial test stress-strain curves and rock failure behavior. The material model captures both the brittle, strain-softening response at low mean stress and the ductile, strain-hardening response at high mean stress, typical of the mechanical behavior of sedimentary rocks at depths of petroleum interest. Strain localization in the brittle response leads to the spontaneous development of fractures and fracture patterns in the numerical models. The geometries of the numerically-generated fractures bear an uncanny resemblance to the patterns of natural fractures observed both within geologic structures and around subsurface wellbores.

An application of the technique is described in which a horizontal wellbore is introduced into a finite element model of a reservoir sandstone stressed so as to simulate in situ conditions. The reservoir is then subjected to pressure depletion, simulating production. Shear fractures initiate around the hole upon drilling, controlled partly by the boundary stresses imposed on the model. Upon reservoir depletion, the fracture network expands as individual fractures propagate, ultimately leading to the collapse of the horizontal hole. In this example the model provides a means of predicting quantitatively the nature and extent of wellbore damage, and the timing of failure of the well. The technique has general applicability to geomechanical problems at any scale.

AAPG Search and Discovery Article #90906©2001 AAPG Annual Convention, Denver, Colorado