Integrating complementarity into the 2D displacement discontinuity boundary element method to model faults and fractures with frictional contact properties
Introduction
Faults and fractures are commonly modeled as cracks in an otherwise homogeneous and isotropic linear elastic material. This approximation allows us to idealize the problem of faulting and fracturing in Earth's brittle crust. Although many analytical solutions for crack problems exist, such solutions are not viable for problems where the cracks have irregular geometry or the boundary conditions cannot be prescribed. Numerical tools allow us to solve new and more realistic crack problems for geoscience applications. We introduce a numerical method that merges two existing computational tools: the displacement discontinuity boundary element method (DDM) and a complementarity algorithm.
The elementary implementation of the two-dimensional DDM (Crouch and Starfield, 1983, Chapter 5) is only accurate for problems for which the contact conditions are known a priori. Incorporating a complementarity algorithm into the DDM is an effective way to numerically model quasi-static slip on cracks and the resulting deformation of the nearby material (De Bremaecker et al., 2004, Griffith et al., 2010, Mutlu and Pollard, 2008, Ritz and Pollard, 2011). Complementarity is a popular optimization method used in engineering, economics, and various scientific disciplines (e.g., Cottle et al., 1992, Harker and Pang, 1990, Murty, 1988, Pang, 1994, Pang and Trinkle, 1996). The aim of this article is to communicate this simple yet powerful tool to the geoscience community, as we and several other researchers have recently used this approach to investigate the mechanical behavior of faults and fractures in Earth's brittle crust (De Bremaecker et al., 2004, Griffith et al., 2010, Mutlu and Pollard, 2008, Ritz and Pollard, 2011). Here, the DDM with complementarity formulation is described, and examples are provided to verify the validity of this method as well as demonstrate its utility.
Section snippets
DDM formulation
DDM is a special type of boundary element method (BEM) (Crouch, 1976, Crouch and Starfield, 1983). The DDM can treat cracks embedded in an infinite homogeneous, isotropic, elastic material as curvilinear boundaries discretized into linear elements of constant displacement discontinuity (Fig. 1). A major advantage of using a BEM is that only discretization of the crack is required, whereas other popular methods, such as the finite element method, require discretization of the surrounding
Frictional contact properties
The shear and normal stress components at the center of a boundary element are related to the specified traction boundary conditions using Cauchy's Formula (Pollard and Fletcher, 2007, p. 213), and also are related to each other by the Coulomb criterion:here μ is the coefficient of friction. Laboratory values for the coefficient of friction of rocks are generally reported to be 0≤μ≤1 (Byerlee, 1967, Hoskins et al., 1968, Jaeger, 1959), and commonly are found to be
Complementarity formulation
The DDM discussed in Section 2 does not include frictional contact properties because interpenetration is allowed, preventing contact between the two crack surfaces. By introducing a complementarity algorithm, we can correct the initial DDM solution and include the frictional contact properties discussed in Section 3.
Since the behavior of each boundary element is not known a priori, the complementarity solver utilizes the basic relationships between sticking, slipping, and opening to correct
Implementation in Matlab
We implement the DDM with complementarity in Matlab, which imports an Excel spreadsheet. All input parameters are managed in a spreadsheet so that problems can be more easily organized, saved, and shared. Alternatively, the input parameters can be prepared using other text formats, readable by Matlab. The user-defined parameters are the elastic moduli, remote Cartesian stress components, the number of fractures, the coordinates for each endpoint of every boundary element, the coefficient of
Validation
In linear elastic fracture mechanics (LEFM) theory, if the gradient of displacement at the model crack tip is infinite, the stresses are singular there. However, this singularity is physically unrealistic because natural materials cannot support infinite stress. One explanation is that real materials limit the stress concentrations at crack tips by locally deforming inelastically (Dugdale, 1960). If the size of the zone of inelastic deformation at a crack tip is negligible in comparison with
Stick or slip conditions
To illustrate the utility of the DDM with complementarity, we explore the deformation of an isolated, planar crack with various frictional contact properties. Throughout this discussion, the elastic moduli are E=50,000 MPa, and ν=0.25. The crack is divided into 500 boundary elements of equal length, lies along the x-axis, and is centered on the y-axis (Fig. 1). A remote principal stress ratio of drives slip on the crack in all cases, with the maximum principal stress and
Discussion
Faults are rarely isolated, continuous, or planar. While CEZs may warp the planar geometry of a fracture (Martel, 1997) the DDM with complementarity also can model segmented, branching, and other types of non-planar fractures. Fig. 8 schematically illustrates various non-planar fractures that have been observed in nature; segmented and stepping, ramped, intersecting and branching, splayed, and curved.
Many numerical studies addressing fracture shapes shown in Fig. 8 have developed from an
Conclusion
The DDM with complementarity accurately computes slip, opening, and stress distributions along cracks, and displacement and stress within the surrounding material, while incorporating the contact properties μ and Sf. This technique can accommodate any number of cracks, each with an arbitrary distribution of friction and frictional strength along its length. Additionally, this numerical tool can model natural fracture and fault geometries if a sufficient number of boundary elements are used to
Acknowledgments
We thank Michael Ferris for permission to use PATH as the exemplar complementarity code. We also thank two anonymous reviewers for comments that greatly improved the manuscript. Funding was provided by DOE Basic Energy Sciences, grant DE-FG02-04ER15588-5.
References (59)
The mathematical theory of equilibrium cracks in brittle fracture
Advances in Applied Mechanics
(1962)- et al.
Slip distributions on faults: effects of stress gradients, inelastic deformation, heterogeneous host-rock stiffness, and fault interaction
Journal of Structural Geology
(1994) - et al.
Physical explanation for the displacement-length relationship of faults using a post-yield fracture mechanics model
Journal of Structural Geology
(1992) - et al.
Including gravity in the displacement discontinuity method when accounting for contact interaction
International Journal of Rock Mechanics and Mining Sciences
(2004) - et al.
A comparison of two algorithms for solving closed crack problems
Engineering Fracture Mechanics
(2000) Yielding of steel sheets containing slits
Journal of the Mechanics and Physics of Solids
(1960)- et al.
A fast iterative boundary element method for solving closed crack problems
Engineering Fracture Mechanics
(1999) - et al.
A medium-scale direct friction experiment
International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts
(1968) - et al.
Slip distributions on intersecting normal faults
Journal of Structural Geology
(1999) Effects of cohesive zones on small faults and implications for secondary fracturing and fault trace geometry
Journal of Structural Geology
(1997)
The initiation and growth of en-echelon veins
Journal of Structural Geology
Theoretical displacements and stresses near fractures in rock
Slip heterogeneity on a corrugated fault
Earth and Planetary Science Letters
The geometry of echelon fractures in rock—implications from laboratory and numerical experiments
Journal of Structural Geology
An iterative algorithm for modeling crack closure and sliding
Engineering Fracture Mechanics
The morphology of strike-slip faults: examples from the San Andreas Fault, California
Journal of Geophysical Research
Friction of rocks
Pure and Applied Geophysics
Frictional characteristics of granite under high confining pressure
Journal of Geophysical Research
Frictional Slip and Fractures Associated with Faults and Folds. PhD dissertation
Fracture localization along faults with spatially varying friction
Journal of Geophysical Research
The Linear Complementarity Problem
Sur une application des règles de maximis & minimus a quelques problèmes de statique, relatifs a l'architecture
Académie Royale des Sciences Mémoires de Mathématique & de Physique
Solution of plane elasticity problems by the displacement discontinuity method. I. Infinite body solution
International Journal for Numerical Methods in Engineering
Boundary Element Methods in Solid Mechanics
The PATH solver: a non-monotone stabilization scheme for mixed complementarity problems
Optimization Methods and Software
Feasible descent algorithms for mixed complementarity problems
Mathematical Programming
Cited by (23)
Investigating stress shadowing effects and fracture propagation patterns: Implications for enhanced geothermal reservoirs
2021, International Journal of Rock Mechanics and Mining SciencesContact between rough rock surfaces using a dual mortar method
2020, International Journal of Rock Mechanics and Mining SciencesCitation Excerpt :Simulation tools that resolve small-scale roughness, while allowing computation of contact stresses and stress variations in the rock mass, are rare. Most analytical or numerical studies at the laboratory or reservoir scale rely on simplifying assumptions and regularizations, such as averaging asperity scale processes with an empirical relationship, reduction to two dimensions, using variants of boundary element methods, not modeling explicitly the mechanical contact between two surfaces, treating the surfaces as two parallel plates, using penalty methods or augmented Lagrangian methods, or only matching triangulations at the surface.4,14–23 In this work, we present a new parallel implementation of an approach that does not require any of the previously mentioned assumptions.
3D scene and geological modeling using integrated multi-source spatial data: Methodology, challenges, and suggestions
2020, Tunnelling and Underground Space TechnologyCitation Excerpt :The model can be used for groundwater seepage and grouting diffusion analysis, as shown in Fig. 17(a). Compared with a traditional plan and section, the 3D geological model can quickly and clearly display the three-dimensional shape and distribution of a geological body (Tertois and Mallet, 2007; Caumon et al., 2009; Ritz et al., 2012). Users can directly see the crosscutting relationship between the fault and the stratum and the coal bed, change the observation angle at will, close the secondary horizon, and highlight the expression of the spatial relationship of the three-dimensional model through various means, such as interface coloring to reflect the elevation changes (Caumon et al., 2009).
Slip on wavy frictional faults: Is the 3rd dimension a sticking point?
2019, Journal of Structural GeologyCitation Excerpt :This is deemed sufficiently accurate in capturing the boundary condition set. In these equations, h is the distance from the mid-point (geometric incenter) of the boundary element to the fracture's tip (Ritz et al., 2012), Dn is the normal displacement of this element, DII is the displacement perpendicular to the crack edge, and DIII is the displacement parallel with the edge. The correction factor c is used to correct for the errors due to the numerical approximation.
The distinct element method (DEM) and the extended finite element method (XFEM) application for analysis of interaction between hydraulic and natural fractures
2018, Journal of Petroleum Science and EngineeringMechanical evolution of transpression zones affected by fault interactions: Insights from 3D elasto-plastic finite element models
2018, Journal of Structural GeologyCitation Excerpt :In this sense, fault steps are zones of slip transfer between discontinuous sub-parallel fault segments in which segments interact through their associated stress field and any hard-linkages. Various non-planar fractures and fault geometries observed in nature and studied both theoretically and numerically (e.g., Rodgers, 1980; Segall and Pollard, 1980; Aydin and Schultz, 1990) are segmented/stepping, ramped, intersecting/branching, splayed, and curved (Ritz et al., 2012, 2015; Ritz, 2013). Fault steps localize either contraction or extension as a function of their manual geometries (right- or left-stepping) and fault kinematics (left- or right-lateral) (Christie-Blick and Biddle, 1985; Woodcock and Fischer, 1986; Ramsay and Huber, 1987; Reches, 1987; Crider, 2001; Storti et al., 2003; Cunningham and Mann, 2007; Mann, 2007).