Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Phase field modelling of crack propagation, branching and coalescence in rocks

Phase field modelling of crack propagation, branching and coalescence in rocks We present a phase eld model (PFM) for simulating complex crack patterns including crack propagation, branching and coalescence in rock. The phase eld model is imple- mented in COMSOL and is based on the strain decomposition for the elastic energy, which drives the evolution of the phase eld. Then, numerical simulations of notched semi-circular bend (NSCB) tests and Brazil splitting tests are performed. Subsequently, crack propaga- tion and coalescence in rock plates with multiple echelon aws and twenty parallel aws are studied. Finally, complex crack patterns are presented for a plate subjected to increasing internal pressure, the (3D) Pertersson beam and a 3D NSCB test. All results are in good agreement with previous experimental and numerical results. Keywords: Phase eld, Rock, COMSOL, Crack propagation, Crack branching arXiv:1902.06686v1 [cond-mat.soft] 14 Feb 2019 1 Introduction Fracture-induced failure has gained extensive concern in engineering because of the huge threat to engineering safety (Anderson 2005). The prediction of fracture in rock is challenging. Rock masses have many pre-existing aws, such as micro cracks, voids and soft minerals. Many e orts have been made to study crack propagation in rock, see for instance the contributions in Bobet and Einstein (1998), Wong et al. (2001), Sagong and Bobet (2002), Wong and Einstein (2009), Park and Bobet (2009), Park and Bobet (2010), Lee and Jeon (2011), and Zhou et al. (2014). However, many studies focus on uniaxial compressive loads since tensile loads or more complicated load cases, which are more dicult to perform in practical tests. Numerical methods are a good alternative to study fracture problems. They are less expen- sive than experimental tests and can provide physical insight dicult to gain through 'pure' experimental testing. Computational methods for fracture can be classi ed in discrete and continuous approaches. Ecient remeshing techniques (Areias and Rabczuk 2017, Areias et al. 2013, Areias and Rabczuk 2013), multiscale method (Budarapu et al. 2014a,b, Yang et al. 2015), strain-softening element (Areias et al. 2014), the extended nite element method (Nanthaku- mar et al. 2014, Mo es and Belytschko 2002), the phantom node method (Rabczuk et al. 2008a, Chau-Dinh et al. 2012, Vu-Bac et al. 2013) and speci c meshfree methods (Rabczuk et al. 2007a, Rabczuk and Zi 2007, Rabczuk et al. 2007b, 2008a, Rabczuk and Samaniego 2008, Rabczuk et al. 2008b, Amiri et al. 2014a) are classical representatitves of the rst class. The cracking parti- cles method (CPM) (Rabczuk and Belytschko 2004, 2007, Rabczuk et al. 2010), Peridynamics (Rabczuk and Ren 2017) and dual-horizon peridynamics (Ren et al. 2016, 2017) are also dis- crete crack approaches but they share the simplicity of continuous approaches to fracture as they also do not require any explicit representation of the crack surface and any crack tracking algorithms. Element-erosion (Belytschko and Lin 1987, Johnson and Stryk 1987) directly sets the stresses of the elements to zero when the elements ful ll the fracture criterion. However, the element-erosion method cannot simulate crack branching correctly (Song et al. 2008). Gradient models (Thai et al. 2016), non-local models (Pijaudier-Cabot et al. 2004), models based on the screend-poisson equation (Areias et al. 2016a) and also phase eld models are typical continuous approaches to fracture. 2 In this paper, we pursue the phase eld model (PFM) (Bourdin et al. 2008, Miehe et al. 2010a,b, Hesch and Weinberg 2014, Borden et al. 2012) to model crack propagation, branching and coalescence in rock. The origins of the PFM can be traced back to Bourdin et al. (2008), but a thermodynamic consistent framework was rst presented by Miehe et al. (2010a). Considerable attention has been paid to PFMs due to their ease in implementation and applicability to multi- physics problems. The PFM does not treat the crack as a physical discontinuity but uses a scalar eld (the phase eld) to smoothly transit the intact material to the broken one. Thus, the sharp crack is represented by a 'damage-like' zone. The shape of the crack is controlled by a length scale parameter and propagation of the crack is obtained through the solution of a di erential equation. Thus, the PFM does not require any external criterion for fracture and additional work to track the fracture surface algorithmically (Borden et al. 2012). It is believed that for this reason, the phase eld is therefore has some advantage over other approaches in modeling branching and merging of multiple cracks. Phase eld models have been discretized in the context of the nite element method (Areias et al. 2016b), meshfree methods (Amiri et al. 2014b) and isogeometric analysis (Borden et al. 2012); the latter two approaches use a fourth-order di erential equation for the phase eld exploiting the higher continuity of the meshfree and isogeometric approximation. The PFM for brittle cracks has also been implemented in commercial software such as ABAQUS (Msekh et al. 2015, Liu et al. 2016). However, the extension of the implementation in ABAQUS to problems with more elds { as hydraulic fracturing { is dicult. Hence, we present an implementation of the phase eld model in COMSOL Multiphysics, a software particularly dedicated to multi- eld modeling. This paper is organized as follows. The phase eld model for brittle fractures is presented in Section 2. Subsequently, the numerical implementation of the phase eld model in COMSOL is described in Section 3. Then, simulations of initiation, propagation, branching, and coalescence of cracks in rock are shown in Section 4 before Section 5 concludes our manuscript. 3 2 Theory of phase eld modeling 2.1 Theory of brittle fracture Consider an elastic body R (d 2 f1; 2; 3g) as shown in Figure 1, whose external boundary and internal discontinuity boundary are denoted as @ and , respectively; x is the position vector and u(x; t)  R the displacement vector at time t. In Fig. 1, the body satis es the time-dependent Dirichlet boundary conditions (u (x; t) = g (x; t) on @ ), and also the i i g time-dependent Neumann conditions on @ ; b(x; t)  R is the body force and f (x; t) the traction on boundary @ Figure 1: Phase eld approximation of the crack surface Given that the stored elastic energy can be transformed into dissipative forms of energy, the classical Grith's theory (Anderson 2005) for brittle fracture states that the crack starts to propagate when the stored energy is sucient to overcome the fracture resistance of the material. Therefore, the crack propagation is regarded as a process to minimize a free energy L that consists of the kinetic energy (u _ ), elastic energy , fracture energy and external kin " f work W : ext Z Z Z Z L = (u _ ) (")d G dS + b ud + f  udS (1) kin " c | {z } | {z } | {z } " f ext 4 @u where u _ = , is the elastic energy density, and G is the critical energy release rate. The " c @t linear strain tensor " = "(u) is given by " = ru + (ru) (2) The kinetic energy is given by (u _ ) = u _ d (3) kin where  indicates the density. 2.2 Phase led approximation for the fracture energy The phase eld method (Miehe et al. 2010a,b, Borden et al. 2012) uses a scalar eld, i.e. the phase eld, to smear out the crack surface (see Fig. 1) over the domain . The phase eld (x; t) 2 [0; 1] has to satisfy the following conditions: 0; if material is intact = (4) 1; if material is cracked A typical one dimensional phase eld approximated by the exponential function is given by (Miehe et al. 2010a) jxj=l (x) = e (5) l denoting the length scale parameter, which controls the transition region of the phase eld and thereby re ects the width of the crack. The distribution of the one dimensional phase eld is shown in Fig. 2. The crack region will have a larger width as l increases and the phase eld will represent a sharp crack when l tends to zero. It can be shown that the crack surface density per unit volume of the solid is given by (Miehe et al. 2010a) (;5) = + jrj (6) 2l 2 5 Figure 2: Distribution of the one dimensional phase eld across a crack Thus, the fracture energy is approximated by Z Z G dS = G + jrj d (7) c c 2l 2 The variational approach (Bourdin et al. 2000) states that the crack surface energy is trans- formed from the elastic energy, which drives the evolution of the phase eld. To capture cracks only under tension, the elastic energy is decomposed into tensile and compressive parts (Miehe et al. 2010b): " = h" i n n (8) a  a a a=1 where " and " are the tensile and compressive strain tensors, respectively. In addition, " + a is the principal strain and n is the direction of the principal strain. The operators hi are de ned as : hi = (j j)=2. Consequently, the positive and negative elastic energy densities are expressed as 2 2 (") = htr(")i + tr " (9) where  > 0 and  > 0 are the Lam e constants. The Lam e constants are related to the Young's modulus E and Poisson's ratio  of the solid through the well known relation: (1 + )(1 2) (10) :  = 2(1 + ) The phase eld is assumed to a ect only the positive elastic energy density, which introduces 6 a sti ness reduction as (Borden et al. 2012) 2 + (") = (1 k)(1 ) + k (") + (") (11) " " where 0 < k  1 is a stability parameter for avoiding numerical singularities because the positive elastic energy density disappears as the phase eld  tends to 1. 2.3 Governing equations By substituting Eqs. (3), (7), and (11) into Eq. (1), the energy functional L is rewritten as Z Z 2 2 + L = u _ d (1 k)(1 ) + k (") + (") d " " Z Z Z G + jrj d + b ud + f  udS (12) 2l 2 Employ the rst variation of the functional L = 0, it can be shown that the strong form of the governing equations are given by (Borden et al. 2012) Div() + b = u  (13) + + 2l (1 k) 2l (1 k) 0 0 > " 2 2 " + 1  l r  = G G c c @ u where u  = and  is Cauchy stress tensor given by @t = @ " " (14) = (1 k)(1 ) + k [htr(")i I + 2" ] + htr(")i I + 2" + + dd with unit tensor I 2 R . The phase eld requires the irreversibility condition (x; s) 2 (x; t)(s < t) during compres- sion or unloading, i.e. the crack cannot be healed. Therefore, we introduce a strain-history eld H (x; t) (Miehe et al. 2010a,b) to ensure a monotonically increasing phase eld: H (x; t) = max ("(x; s)) (15) x2[0;t] 7 The history eld H satis es the Kuhn-Tucker conditions for loading and unloading (Miehe et al. 2010a): + + _ _ H  0; H  0; H ( H ) = 0 (16) " " By replacing by H (x; t) in Eq. (13), the strong form is rewritten as Div() + b = u  (17) 2l (1 k)H 2l (1 k)H 0 0 > 2 2 + 1  l r  = G G c c We denote m as the outward-pointing normal vector to the boundaries, and the governing equations are subjected to the Dirichlet and Neumann boundary conditions u = g on @ > g (18) m = f on @ r m = 0 on @ along with the initial conditions u(x; 0) = u (x) x 2 (19) v(x; 0) = v (x) x 2 (x; 0) =  (x) x 2 2.4 The choice of l Borden et al. (2012) and Zhang et al. (2017) proposed an analytical solution for the critical tensile stress  that a one-dimensional bar can sustain: cr 9 EG = (20) cr 16 3l There is an apparent singularity when l tends to zero, i.e. in case of a sharp crack, which is phyiscally meaningless. However, assuming all other parameters except of l are known, eq. 8 (20) can be solved for l : 27EG l = (21) cr In Eq. (21), the critical energy release rate G and Young's modulus E can be obtained by conducting regular experimental tests, while the critical stress  can be approximated by the cr tensile strength  through the standard tensile test. Hence, Eq. (21) can estimate the length scale. Note that the accuracy is unknown for more complex cases. 3 Numerical implementation 3.1 Finite element method We use the nite element method to solve the governing equations (17) given in weak form by Z Z Z (u   u  : ") d + b ud + f  udS = 0 (22) and Z Z 2(1 k)H (1 )d + G l rr +  d = 0 (23) c 0 We use the standard vector-matrix notation and denote the nodal values of the displacement and phase eld as u and  . The discretization is thereby given by i i u = N d;  = N  (24) where d and  are the vectors consisting of node values u and  . N and N are shape i i u function matrices given by 2 3 N 0 0 : : : N 0 0 1 n 6 7 h i 6 7 N = 6 7 ; N = (25) u 0 N 0 : : : 0 N 0  N N : : : N 1 n 1 2 n 4 5 0 0 N : : : 0 0 N 1 n where n is the node number in one element and N is the shape function of node i. The same 9 discretization is applied to the test functions and we obtain u = N d;  = N  (26) where d and  are the vectors consisting of node values of the test functions. The gradients are thereby calculated by ^ ^ " = B d; r = B ; " = B d; r = B  (27) u  u where B and B are the derivatives of the shape functions de ned by 2 3 N 0 0 : : : N 0 0 1;x n;x 6 7 6 7 2 3 6 7 0 N 0 : : : 0 N 0 1;y n;y 6 7 N N : : : N 1;x 2;x n;x 6 7 6 7 6 0 0 N : : : 0 0 N 7 1;z n;z 6 7 6 7 B = ; B = 6 7 (28) u  N N : : : N 1;y 2;y n;y 6 7 4 5 6 7 N N 0 : : : N N 0 1;y 1;x n;y n;x 6 7 N N : : : N 1;z 2;z n;z 6 7 6 7 0 N N : : : 0 N N 1;z 1;y n;z n;y 4 5 N 0 N : : : N 0 N 1;z 1;x n;z n;x By applying the nite element approximation, the equations of weak form (22) and (23) are then written as " # Z Z Z Z T T T T T T (d) N N d d + B D B d d + (d) N bd + N f dS = 0 (29) u e u u u u u Z Z T T T T T ^ ^ ^ () B G l B + N + 2(1 k)H N d + () 2(1 k)HN d = 0 c 0 (30) where D is the degraded sti ness matrix. D can be calculated from the fourth order elasticity e e tensor D: D = @" 2 2 + =  (1 k)(1 ) + k H (tr(")) + H (tr(")) J + 2 (1 k)(1 ) + k P + P " " (31) 10 P P 3 3 where J =   ,  being the Kronecker and P = H (" ) n n n n + ijkl ij kl ij " a ab ai aj bk bl ijkl a=1 b=1 P P 3 3 h" i h" i 1 a n n (n n + n n ) with n the i-th component of vector n ; H hxi ai bj ak bl bk al ai a " a=1 b6=a 2 " " is the Heaviside function: 1; x > 0 H hxi = (32) 0; x  0 Since P cannot be computed when " = " , we apply a \perturbation" technology for the a b ijkl principal strains (Miehe 1993) with an unchanged " : " = " (1 + ) if " = " 1 1 1 2 (33) " = " (1 ) if " = " 3 3 2 3 with the perturbation  = 1 10 . For admissible arbitrary test functions, Eqs. (29) and (30) always hold, thereby producing the discretized weak form as Z Z Z Z T T T T N N d d B D B d d + N bd + N f dS = 0 (34) e u u u u u | {z } | {z } | {z } int ine  F =K d F =Md u ext u F Z     Z T T T B G l B + N + 2(1 k)H N d + 2(1 k)HN d = 0 (35) c 0 | {z } | {z } ext int ^ F F =K ine int ext where F , F , and F are the inertial, internal, and external forces for the displacement u u u int ext eld and F and F are the internal and external force terms of the phase eld. In addition, the mass and sti ness matrices follow M = N N d > Z K = B D B d (36) u e u > Z > T T K = B G l B + N + 2(1 k)H N d c 0 In this paper, we use a staggered scheme to solve the displacement and phase elds. Thus, the Newton-Raphson approach is adopted to obtain the residual of the discrete equations R = 11 ext ine int ext int F F F = 0 and R = F F = 0, respectively. u u u 3.2 COMSOL implementation We implemented the phase eld approach into COMSOL Multiphysics, which can simulate mathematical and physical problems easily by adding application-speci c modules. Therefore, it is suitable for multi- eld modeling. In this paper, we construct three main modules, namely, the Solid Mechanics, History-strain and Phase Field Modules, which employ the standard nite element discretization in space as described in Subsection 3.1. In addition, a pre-set Storage Module is employed to evaluate and store the intermediate eld variables in a time step, such as the positive elastic energy and principal strains. Based on a linear elastic material library, the Solid Mechanics Module is used for the dis- placement u. The boundary and initial conditions shown in Section 2 are added to the Solid Mechanics Module, while a non-linear stress-strain relationship is considered. The sti ness ma- trix D is modi ed in a time step as follows 2 3 D D D D D D 1111 1122 1133 1112 1123 1113 6 7 6 7 6 D D D D D D 7 2211 2222 2233 2212 2223 2213 6 7 6 7 6 7 D D D D D D 3311 3322 3333 3312 3323 3313 6 7 D = (37) 6 7 6 D D D D D D 7 1211 1222 1233 1212 1223 1213 6 7 6 7 6 7 D D D D D D 2311 2322 2333 2312 2323 2313 4 5 D D D D D D 1311 1322 1333 1312 1323 1313 The governing equation (17) is presented for dynamic crack problems. However, for a quasi- static problem, the inertia term must be neglected in the Solid Mechanics Module. The Phase Field Module is established for the phase eld  by revising a pre-de ned module, which is governed by the Helmholtz equation. The boundary condition Eq. (18) and initial condition (19) are also implemented in this module. For the history strain eld H , the Distributed ODEs and DAEs Interfaces are used to construct the History-strain Module, where the history strain eld is not solved directly. We use a \previous solution" solver to record the results in the 12 previous time steps and obtain the eld H by applying the following format in COMSOL: + + H = 0; if > H " " (38) H H = 0; if  H Additionally, the initial condition H (x) = 0 is used for the History-strain Module. Figure 3 shows the relationship among all the established modules. The mechanical re- sponses, such as the principal strains, the directions of principal strain, and the elastic energy, are naturally exported from the Solid Mechanics Module and stored in the Storage Module. The History-strain Module then call the positive elastic energy and update the local history strain eld H . The Phase Field Module employs the updated H to solve the phase eld. In a time step, the updated phase eld and the stored principal strains as well as the directions are used to modify the sti ness matrix of the Solid Mechanics Module and subsequently to obtain the mechanical responses. The detailed procedure of the staggered (segregated) scheme is depicted in Fig. 4. The equations of displacement, history strain and phase- eld are solved independently. An implicit Generalized- method (Borden et al. 2012) is used for time integration. The Generalized- method is unconditionally stable and requires a prediction step and a correction step. When a new time step starts, an initial guess from the linear extrapolation of the previous solution is used in the prediction stage. Newton-Raphson iterations are used to solve the residuals for each module in the correction stage. That is, in the iteration step j + 1 of a given time step i, the dis- j+1 j j j placement u is rst solved by using the results (u , H and  ) of the previous iteration step i i i i j+1 j. Following the updated displacements u , the history strain is updated. Subsequently, an- j+1 other Newton-Raphson iteration is applied to solve the phase- eld  . The total relative error is estimated and the iterations continue until the tolerance requirement is met. The maximum number of iteration in one time step is set as 50 in our simulations. We accelerate the convergence by using the Anderson acceleration (Toth and Kelley 2015) where the dimension of the iteration space eld is chosen as more than 50. A ow chart of our implementation is shown in Fig. 5. The source code can be found in \https://sourceforge.net/projects/phase eldmodelingcomsol/". 13 Figure 3: Relationship between all the modules established Figure 4: Segregated scheme for the coupled calculation in phase eld modeling 14 Figure 5: COMSOL implementation of the phase eld modeling 4 Numerical examples of crack propagation, branching and coalescence in rocks 4.1 Simulation of notched semi-circular bend (NSCB) tests Let us consider the notched semi-circular bend (NSCB) test rst. Geometry and boundary condition of the rock specimen are shown in Fig.6. The mechanical properties of the rock are taken from the Laurentian granite (LG) from Grenville province of Canada (Gao et al. 2015). The rock density  is 2630 kg/m , while the Young's modulus E and Poisson's ratio  are 92 GPa and 0.21, respectively. We follow the 1D solution of Borden et al. (2012) for the critical 2 4 stress of material softening, and choose G = 7.6 J/m and the length scale l = 4:5 10 m. c 0 This produces a critical stress close to the quasi-static tensile strength of the specimen (12.8 MPa) (Gao et al. 2015). The phase eld modeling is performed by using 93587 6-node quadratic triangular elements and the maximum element size is h = 2:25 10 m. We apply a vertical displacement on the 15 top of the specimen to drive crack propagation from the tip of the notch. During the simulation, we apply the displacement increment u = 5 10 mm in each time step. Figure 7 shows the crack initiation and propagation in the rock specimen by using the phase eld model. When the applied displacement u reaches to 6:72  10 mm, the crack initiates from the tip of the notch. Subsequently, the crack propagates straightly in the vertical direction 3 3 when u accumulates to 6:74 10 mm and 6:77 10 mm. When the displacement reaches to 6:86 10 mm, the tip of the propagating crack is close to the upper boundary of the specimen and failure of the semi-circular rock specimen occurs. The crack patterns obtained by the phase eld simulation are in good agreement with the experimental results in Gao et al. (2015). Figure 6: Geometry and boundary condition of the notched semi-circular bend (NSCB) tests We test the crack patterns for G = 5:6, 6.6, 8.6, and 9.6 J/m . The results show that G has c c little in uence on the crack patterns. Di erent G have the same crack paths. Figure 8 gives the curves of the reaction force on the upper boundary of the specimen versus the displacement for di erent G . A sudden drop of the load is observed after the crack initiation. In addition, all the curves have the same slope before the crack initiation under di erent G , while the maximum load increases as G increases . We also test the in uence of the maximum mesh size h under a xed G = 7.6 J/m and the 4 4 4 5 length scale l = 4:5 10 m. We choose h = 2:25 10 m, 1:13 10 m, and 5:63 10 m in the tests. The resulting load-displacement curve is shown in Fig. 9. As expected, the load-displacement curve converges with the mesh re nement. To test the in uence of the length 4 2 scale l , we x the mesh size h = 2:25  10 m and G = 7.6 J/m , and then present the 0 c 16 (a) (b) (c) (d) Figure 7: Crack propagation of the notched semi-circular bend (NSCB) tests at a displacement 3 3 3 of (a) u = 6:72  10 mm, (b) u = 6:77  10 mm, (c) u = 6:74  10 mm, and (d) 3 2 u = 6:86 10 mm for G = 7:6 J/m Figure 8: Load-displacement curves of the notched semi-circular bend (NSCB) tests 17 load-displacement curve for di erent l in Fig. 10. Figure 10 shows a decreasing peak load and displacement range when the length scale increases. Figure 9: Load-displacement curves of the NSCB tests under di erent mesh size Figure 10: Load-displacement curves of the NSCB tests under di erent length scale l 4.2 Simulation of Brazil splitting tests Brazil splitting tests are commonly used to obtain the tensile strength of rocks and many re- searchers simulated crack propagation in a Brazilian disc under compression, such as Cai (2013) 18 and Zhou and Wang (2016). Figure 11 presents the geometry of the Brazilian disc along with the boundary conditions. Figure 11: Geometry and boundary condition of the Brazil splitting tests These material parameters are adopted:  = 2630 kg/m , E = 31.5 GPa, and  = 0.25. The length scale parameter l is xed to 1 mm. 26700 linear triangular elements (base mesh) are used to discretize the disc with the maximum element size h = 0:5 mm, k is set to 1  10 . Finally, we conduct the simulation by using ve di erent G : 50, 75, 100, 125, and 150 J/m . Figure 12 shows the crack initiation and propagation in the Brazil splitting tests for G = 100 J/m . When the displacement u approaches a value of 0.476 mm, the crack occurs in the center of the disc where the maximum tensile stress occurs which is in good agreement with the experimental and analytical results (Atkinson et al. 1982, Entacher et al. 2015). When u reaches a value of 0.477 mm, the crack propagates with an increasing width. Then, the crack continues to propagate and the crack tips move close to the upper and bottom ends of the disc when u reaches to 0.478 mm. The crack branching is observed when u is close to 0.480 mm. In addition, the crack cannot penetrate deeply into the ends of the disc because of the locally compressed area around both ends. We compare the curve of the reaction force on the upper end of the Brazilian disc versus the displacement u with the experimental result in Fig. 13. The experimental curve is originated from the work of Erarslan and Williams (2012). Figure 13 shows that the phase eld model can reproduce the experimental results well. The main reason for the di erence in Fig. 13 is 19 (a) (b) (c) (d) Figure 12: Crack propagation of the Brazil splitting tests at a displacement of (a) u = 0:476 mm, (b) u = 0:477 mm, (c) u = 0:478 mm, and (d) u = 0:480 mm for G = 100 J/m 20 that there are contact issues in the actual Brazilian tests and a compaction stage is commonly observed before the elastic stage. Figure 13: Comparison of the load-displacement curves obtained by the experimental test and phase eld modeling Figure 14 presents the curves of the reaction force on the upper end of the Brazilian disc versus the displacement u for di erent G . Similar to the NSCB tests, a sudden drop of the load is observed after the phase eld increases to 1. In addition, the peak load increases with the increase in G . We then test the in uence of the length scale l under a xed G = 100 J/m c 0 c and mesh size h = 0:5 mm. The tested length scale parameters are l = 0.5, 1, 2, and 3 mm, respectively. The load-displacement curves under di erent length scale l are shown in Fig. 15. As shown in Fig. 15, the peak load of the specimen decreases with the increase in the length scale l . We also test the in uence of mesh size h under a xed G = 100 J/m and l = 1 mm. c 0 The maximum mesh size h is set as 1, 0.5, 0.25, and 0.125 m, respectively. The resulting load- displacement curve is shown in Fig. 16. The load-displacement curve converges with mesh re nement as expected. In addition, in the Brazil splitting test, the tensile strength  of the rock specimen is given by 2P peak = (39) DL where P is the peak load, and D and L are the diameter and length of the rock specimen. peak Thus, we compare the tensile strength by the phase eld simulation and the critical stress for 21 Figure 14: Load-displacement curves for the Brazil splitting tests for di erent G Figure 15: Load-displacement curves for the Brazil splitting tests for di erent l 22 1D by Borden et al. (2012) in Fig. 17. The tensile strength increases at a decreasing rate with the increase in G . The critical stress has the same trend as the tensile strength. However, the critical stress is far larger than the tensile strength. This observation indicates that the critical stress for 1D analysis cannot be applied directly to 2D or 3D. Figure 16: Load-displacement curves for the Brazil splitting tests for di erent h Figure 17: Comparison of tensile strength by Brazil splitting tests and critical stress for 1D 23 Figure 18: Geometry and boundary condition of three pre-existing aws 4.3 Propagation of multiple echelon aws We consider a 50 mm  50 mm square rock sample subjected to tension. The rock sample has three pre-existing aws, whose position and geometry are shown in Fig. 18. All the aws have the same length, spacing, and inclination angle of 45 . These parameters are adopted: the rock density  = 2500 kg/m , the Young's modulus E = 30 GPa, the Poisson's ratio  = 0:333, 2 5 G = 3 J/m , k = 110 , and the length scale l = 0:25 mm. Vertical displacements are applied c 0 on the top and bottom boundaries of the rock sample as shown in Fig. 18. The simulation is performed by using 106852 6-node quadratic triangular elements where the maximum element size h is 0.25 mm. In each time step, the displacement increment is u = 5 10 mm. Figure 19(a)-(f ) shows the propagation and coalescence of the three pre-existing aws. We also calculate the reaction force on the upper boundary of the rock sample, and present the load-displacement curves in Fig. 20. As the displacement u increases, both the load and the phase eld around the tips of the aws increase. When u reaches to 2:6 10 mm, the phase eld increases close to 1 and the load approaches to the maximum value. When u reaches to 2:63  10 mm, the rst tensile cracks occurs around the left and right tips of the aw . These two cracks are perpendicular to the direction of the applied displacement. In addition, the load reaches to the maximum as shown in Fig. 20. When the displacement u reaches to 2:64 10 mm, the cracks from the tips of the aw propagates perpendicular to u while the 24 load starts to drop after the peak load. Additionally, new cracks occur from the right tip of the aw 1 and the left tip of the aw 3 . 1 2 3 Figure 19(d) shows the coalescences of the cracks from the tips of the aws , , and when u = 2:65 10 mm. Thus, the aw 2 links the aws 1 and 3 , which is accompanied by a sudden drop of the load after the maximum value in Fig. 20. At the same time, new 1 3 cracks from the left tip of the aw and the right tip of the aw are observed. When u = 2:66 10 mm, the cracks from the aws 1 and 3 continue to propagate and the load decreases. When the displacement increases to 2:69 10 mm, the cracks initiating from the 1 3 left tip of the aw and the right tip of the aw reach the boundaries of the rock sample. This indicates the rock loses its load-bearing capacity, which can also be veri ed by Fig. 20. We now consider the same square rock sample subjected to tension with nine pre-existing aws in Fig. 21. These aws have the same inclination angle of 45 , while the lengths and spacing are not xed. The position and geometry of the aws are shown in Fig. 21. The parameters are the same as those in the example of three aws. A total of 107982 6-node quadratic elements are used to discretize the rock sample and the maximum element size h is 0.25 mm. The displacement increment u = 1 10 mm is applied in each time step. Figure 22(a)-(f ) shows the propagation and coalescence process of the nine pre-existing aws, and the reaction force on the upper boundary of the rock sample is depicted in Fig. 23. As the displacement u increases, the phase eld around the tips of the aws and the load both increase. The rst tensile cracks initiate from the left tips of the aws 8 and 9 when the displacement u reaches to 3:02  10 mm. At this time, the load achieves the maximum value. As the displacement increases to 3:04 10 mm, the crack from the left tip of the aw 8 propagates and links up the aw 7 . The crack from the left tip of the aw 9 continues to propagate 7 8 while new cracks initiate from the left tip of the aw and the right tips of the aws and 9 . The load then has a drop after the peak pint (a). When u reaches to 3:06 10 mm and 3:08 10 mm, the cracks initiating from the left tip of the aw 9 and the right tip of the aw continue to propagate at a decreasing rate and at a small angle with the horizontal. However, the cracks from the left tip of the aw 7 and the right tip of the aw 9 propagate at relatively large velocity nearly along the horizontal direction. The load decreases sharply as the applied displacement increases. 25 (a) (b) (c) (d) (e) (f ) Figure 19: Propagation and coalescence of the three pre-existing aws in a square rock sample 3 3 3 at a displacement of (a) u = 2:6 10 mm, (b) u = 2:63 10 mm, (c) u = 2:64 10 mm, 3 3 3 (d) u = 2:65 10 mm, (e) u = 2:66 10 mm, and (f ) u = 2:69 10 mm 26 Figure 20: Load-displacement curves for the square rock sample with three pre-existing aws Figure 21: Geometry and boundary condition of nine pre-existing aws 27 Figure 22(e) shows that the crack from the left tip of the aw 7 propagates close to the left boundary of the rock sample when u = 3:12 10 mm. The crack initiating from the right tip of the aw propagates at a relatively small rate because of a larger distance from the boundary where the applied displacement is applied. When u = 3:14 10 mm, the cracks from the left tip of the aw 7 and the right tip of the aw 9 both reaches the left and right boundaries of the rock sample, indicating the failure of rock sample. The rock loses its load-bearing capacity and the load drops to near 0 in Fig. 23. In addition, Fig. 22 shows no cracks initiation from the tips of the aws 1 - 6 . The reason is the stress shielding and ampli cation e ects due to the interaction of the aws Zhou et al. (2015). 4.4 Propagation and coalescence of twenty parallel aws A 2D square rock sample with twenty parallel pre-existing aws subjected to tension is tested. All aws have the same length, spacing, and inclination angle of 0 . We consider the aws in doubly periodic rectangular and diamond-shaped arrays, respectively. The arrangement and geometry of the aws are depicted in Fig. 24. The rock sample are 50 mm  50 mm. These parameters are adopted: the rock density  = 2450 kg/m , the Young's modulus E = 30 GPa, 2 9 the Poisson's ratio  = 0:3, G = 100 J/m , k = 1  10 , and the length scale l = 0:4 mm. c 0 The rock sample is discretized by using uniform 8-node quadratic elements with the element size h = 0:2 mm. We adopt the displacement increment u = 5 10 mm for each time step. Figure 25 presents the load-displacement curves for the square rock sample with twenty parallel aws. A sudden drop of the load is also observed after the maximum value is obtained. Figure 26 presents the propagation and coalescence of the doubly periodic rectangular array of twenty pre-existing aws. As the displacement u increases, the phase eld  concentrates at the tip of each aw and the load achieves the maximum when u = 2:185  10 mm. When the displacement u reaches to 2:19 10 mm, the rst tensile cracks initiate from the left and right tips of the aws 1 , 5 , 17 , and 20 . In addition, the load has a small drop from the peak. When u = 2:195  10 mm, the rst cracks continue to propagate perpendicular to the direction of the applied displacement and the load decreases. However, when u = 2:2 10 mm, the cracks fully connect the aws 1 , 5 , 17 , and 20 with the aws 2 , 3 , 18 , and 19 . Moreover, some new 2 18 3 cracks initiate from the right tips of the aws and as well as the left tips of the aws and 28 (a) (b) (c) (d) (e) (f ) Figure 22: Propagation and coalescence of the nine pre-existing aws in a square rock sample 3 3 3 at a displacement of (a) u = 3:02 10 mm, (b) u = 3:04 10 mm, (c) u = 3:06 10 mm, 3 3 3 (d) u = 3:08 10 mm, (e) u = 3:12 10 mm, and (f ) u = 3:14 10 mm 29 Figure 23: Load-displacement curves for the square rock sample with nine pre-existing aws (a) Doubly periodic rectangular array (b) Diamond-shaped array Figure 24: Arrangement and geometry of the twenty pre-existing aws and the boundary con- dition 30 2 19 . The load then drops to approximately 60% of the maximum load. When u = 2:205 10 mm, the cracks initiating from the right tips of the aws 2 and 18 and the cracks from the left 3 19 1 tips of the aws and coalesce. Finally, the cracks from the left tips of the aws and 17 as well as the cracks from the right tips of the aws 4 and 20 reach the left and right sides of the rock sample when u = 2:21  10 mm, indicating that the rock loses the load-bearing 5 16 capacity. In addition, no cracks initiate from the tips of the aws - . (a) Doubly periodic rectangular array (b) Diamond-shaped array Figure 25: Load-displacement curves for the square rock sample with twenty parallel pre-existing aws Figure 27 presents the propagation and coalescence of the twenty pre-existing aws in the 31 (a) (b) (c) (d) (e) (f ) Figure 26: Propagation and coalescence of the doubly periodic rectangular array of twenty pre- 2 2 existing aws at a displacement of (a) u = 2:185  10 mm, (b) u = 2:19  10 mm, (c) 2 2 2 2 u = 2:195 10 mm, (d) u = 2:2 10 mm, (e) u = 2:205 10 mm, and (f ) u = 2:21 10 mm 32 diamond-shaped array. The crack patterns for the diamond-shaped array are di erent from those for the doubly periodic rectangular array. As the displacement u increases, the phase eld rapidly increases around the tips closest to the left and right boundaries of the rock sample. When the displacement u reaches to 2:11 10 mm, the load increases to the maximum. When u = 2:115 10 mm, the tensile cracks rst emanate from the left tips of the aws 5 and 13 4 20 as well as the right tips of the aws and . When u = 2:12 10 mm, another two cracks initiate from the left tips of the aws 4 and 20 . The rst four cracks continue to propagate along 2 2 the horizontal direction, while the load at u = 2:12 10 mm is close to that at u = 2:11 10 mm and u = 2:115 10 mm. The aws close to the top and bottom edges of the sample are placed perpendicular to the direction of loading. The arrangement of the internal aws has little e ect on the propagation of the edge aws. Therefore, cracks propagating from those edge aws will show typical features of Mode-I fracture. As shown in Fig. 27, when u = 2:130 10 mm, the originally initiating cracks propagate as expected, while new cracks emanate from the right tips of the aws 3 , 5 , 13 19 , and . The load then starts to drop from the peak region. However, when u = 2:135 10 mm, the cracks fully connect the aws 3 and 19 with the aws 4 and 20 . In addition, two new cracks initiate from the left tips of the aws 3 and 19 . At this time, the load decreases to approximately half of the peak load. As the displacement u increases to 2:145 10 mm, the propagating cracks and the aws 2 , 3 , 5 , 6 , 13 , 14 , 18 , and 19 coalesce. At the same time, the cracks from the left tips of the aws 5 and 13 as well as the cracks from the right tips of 4 20 the aws and reach the left and right sides of the rock sample. The load at u = 2:145 mm is less than zero followed by new cracks emanating from the left tip of the aw 2 and the right tip of the aw 6 . 2 2 When the displacement u reaches to 2:195 10 mm and 2:3 10 mm, the crack initiating from the right tip of the aw 14 slowly propagates; however, the cracks from the left tip of the aw 2 and the right tip of the aw 6 continue to propagate obliquely at a relatively large velocity. The di erence in the crack patterns from Figs. 26 and 27 implies that initiation, propagation and coalescence of the cracks are signi cantly a ected by the arrangement of the aws. 33 (a) (b) (c) (d) (e) (f ) (g) (h) Figure 27: Propagation and coalescence of the diamond-shaped array of twenty pre-existing aws 2 2 2 at a displacement of (a) u = 2:11 10 mm, (b) u = 2:115 10 mm, (c) u = 2:120 10 mm, 2 2 2 2 (d) u = 2:13010 mm, (e) u = 2:13510 mm, (f ) u = 2:14510 mm, (f ) u = 2:19510 mm, and (g) u = 2:3 10 mm 34 4.5 Crack branching in a plate subjected to internal pressure This example is a square plate subjected to internal pressure with geometry and boundary conditions shown in Fig. 28. The internal pressure is applied on the upper and lower boundaries of the notch with p  = 1 MPa/s, while the outer boundaries of the plate are traction-free. In this example, dynamic cracks are considered, and these parameters are adopted: the rock density 3 2 9 = 2450 kg/m , the Poisson's ratio  = 0:3, G = 1 J/m , k = 1 10 , and the length scale l = 0:4 mm. The plate is discretized by using uniform Q4 elements with the element size h = l , 0 0 and we adopt the time step size t = 0:01 s. Figure 28: Geometry and boundary conditions of the plate subjected to internal pressure We consider the plate as a heterogeneous material and apply a Weibull distribution to the Young's modulus: m1 m m E E '(E) = exp (40) E E E 0 0 0 where ' is the probability density function and coecient m determines the shape of '. m also re ects the homogeneity of the material. As m increases, the material becomes more homogeneous and vice versa. In this paper, we consider m = 1, 3, 5, 7, and 9 along with m = 1 representing a homogeneous plate. We use E = 30 GPa and Fig. 29 shows the distribution of Young's modulus for di erent m. Figure 30 presents the crack patterns of the plate subjected to internal pressure for di erent m. Complex crack patterns, such as many crack branching, are observed. These observations are di erent from the previous examples. When the time t reaches 20 s, cracks initiate from the left and right tips of the notch and then start to branch. When the time t increases to 35 (a) m = 1 (b) m = 3 (c) m = 5 (d) m = 7 (e) m = 9 (f ) Homogeneous Figure 29: Young's modulus of the plate 36 30 s, the branching cracks propagate. When t = 50s, large \damage" region are observed around the notch where  is large. At t = 60 s, some cracks initiate from the upper and lower boundaries of the plate, while the cracks from the notch keep propagating. Finally, when t = 70 s, many new cracks occur inside the plate and the cracks from the upper and lower boundaries of the plate start to branch. For a smaller m, more crack branching is observed during the crack propagation. Therefore, a smaller m produces more complex crack patterns. 4.6 3D Petersson beam In this example, the phase eld model is applied to a single edge notched beam subjected to three-point bending (the so-called Petersson beam). A 3D simulation is conducted in this example. The geometry and boundary conditions are depicted in Fig. 31 according to Petersson (1981). The thickness of the beam is 50 mm. The following mechanical properties of the beam are chosen Petersson (1981): Young's modulus E = 27 GPa, Poisson's ratio  = 0.21, critical energy release rate G = 56 J/m and length scale l = 1 mm. We choose the maximum element c 0 size h = 5 mm in most of the beam but h = 1 mm in the region where the crack is expected to propagate. A displacement increment of u = 2:5 10 mm is used. Figure 32 compares the load-displacement curves obtained of the phase eld model with the experimental test (Petersson 1981). The results obtained by the phase eld model are in good agreement with the experimental test. Figure 33 presents the crack propagation in the beam at the displacements u = 0:34 mm, 0:355 mm, and 0:46 mm. 4.7 3D NSCB tests In the last example, we simulate the crack propagation in a 3D NSCB specimen. The geometry and boundary conditions are similar to the 2D tests except a thickness of 16 mm in 3D. The same parameters are used as those in the 2D tests and G = 7:6 J/m . We re ne the elements with the maximum size h = 4:5 10 m in the region where the crack is expected to propagate, 3 7 while in the rest region h = 1:8 10 m. In addition, the displacement increment u = 5 10 mm is applied. 3 3 Figure 34 presents the crack propagation at the displacements u = 6:6710 mm, 6:7110 37 t = 20 s t = 30 s t = 50 s t = 60 s t = 70 s m = 1 m = 3 m = 5 m = 7 m = 9 m = inf Figure 30: Crack patterns of the plate subjected to internal pressure 38 Figure 31: Geometry and boundary condition of Petersson beam Figure 32: Load-displacement curves of 3D petersson beam 39 (a) (b) (c) Figure 33: Crack propagation in the 3D petersson beam at a displacement of (a) u = 0:34 mm, (b) u = 0:355 mm, and (c) u = 0:46 mm 40 3 mm, and 6:745 10 mm. Only the domain with  > 0:95 are displayed for the crack shape. The crack patterns are the same as those in the 2D simulation, and also in good agreement with the results of the experimental tests (Gao et al. 2015). The example of 3D NSCB test shows the ability and practicability of the phase eld method in modeling crack propagation of rocks in 3D. (a) (b) (c) Figure 34: Crack propagation of the 3D NSCB test at a displacement of (a) u = 6:67  10 3 3 2 mm, (b) u = 6:71 10 mm, and (c) u = 6:745 10 mm for G = 7:6 J/m 41 5 Conclusions The phase eld theory for fracture is applied to study the crack propagation, branching and coalescence in rocks. Implementation details of the phase eld modeling in COMSOL are pre- sented with the consideration of cracks only due to tension. The numerical simulations of the 2D notched semi-circular bend (NSCB) tests and Brazil splitting tests are then performed. The presented results are in good agreement with those of the previous experimental tests. Subse- quently, the crack propagation and coalescence in plates with multiple echelon aws and twenty parallel aws are studied. We also present the complex crack patterns in a plate subjected to internal pressure, the increase of which produces crack propagation and branching. Finally, the simulation of a 3D Petersson beam and a 3D NSCB test are performed to show the practicability of phase eld modeling in 3D rocks. All the numerical examples presented by this work show that the initiation, propagation, coalescence, and branching of cracks are autonomous, while the phase eld modeling does not require external criterion for fracture and setting propagation path in advance. These obser- vations highlight the advantages of the phase eld method over other numerical methods in modeling complex crack propagation in rocks. Therefore, the phase eld modeling approach will be useful and practicable for other crack problems in rock engineering in future research. In addition, the presented phase eld model cannot predict the shear cracks when a rock reaches its shear strength. The reason is that the shear strength is not involved in the formulation of the phase eld method and the crack propagation is only driven by the elastic energy. In this sense, a modi ed phase eld model coupled with the shear model of rocks will be also attractive in the future. Acknowledgement The nancial support provided by the Sino-German (CSC-DAAD) Postdoc Scholarship Program 2016, the Natural Science Foundation of China (51474157), and RISE-project BESTOFRAC (734370) is gratefully acknowledged. 42 References Ted L Anderson. Fracture mechanics: fundamentals and applications. CRC press, 2005. Antonio Bobet and HH Einstein. Fracture coalescence in rock-type materials under uniaxial and biaxial compression. International Journal of Rock Mechanics and Mining Sciences, 35(7): 863{888, 1998. RHC Wong, KT Chau, CA Tang, and P Lin. Analysis of crack coalescence in rock-like materials containing three awspart i: experimental approach. International Journal of Rock Mechanics and Mining Sciences, 38(7):909{924, 2001. M Sagong and A Bobet. Coalescence of multiple aws in a rock-model material in uniaxial compression. International Journal of Rock Mechanics and Mining Sciences, 39(2):229{241, LNY Wong and HH Einstein. Crack coalescence in molded gypsum and carrara marble: part 1. macroscopic observations and interpretation. Rock Mechanics and Rock Engineering, 42(3): 475{511, 2009. CH Park and A Bobet. Crack coalescence in specimens with open and closed aws: a comparison. International Journal of Rock Mechanics and Mining Sciences, 46(5):819{829, 2009. CH Park and A Bobet. Crack initiation, propagation and coalescence from frictional aws in uniaxial compression. Engineering Fracture Mechanics, 77(14):2727{2748, 2010. Heekwang Lee and Seokwon Jeon. An experimental and numerical study of fracture coalescence in pre-cracked specimens under uniaxial compression. International Journal of Solids and Structures, 48(6):979{999, 2011. XP Zhou, H Cheng, and YF Feng. An experimental study of crack coalescence behaviour in rock-like materials containing multiple aws under uniaxial compression. Rock mechanics and rock engineering, 47(6):1961{1986, 2014. 43 P. Areias and T. Rabczuk. Steiner-point free edge cutting of tetrahedral meshes with applications in fracture. Finite Elements in Analysis and Design, 132:27{41, 2017. doi: 10.1016/j. nel. 2017.05.001. P. Areias, T. Rabczuk, and D. Dias-da Costa. Element-wise fracture algorithm based on rotation of edges. Engineering Fracture Mechanics, 110:113{137, 2013. doi: 10.1016/j.engfracmech. 2013.06.006. P. Areias and T. Rabczuk. Finite strain fracture of plates and shells with con gurational forces and edge rotations. International Journal for Numerical Methods in Engineering, 94(12): 1099{1122, 2013. doi: 10.1002/nme.4477. Pattabhi R Budarapu, Robert Gracie, St ephane PA Bordas, and Timon Rabczuk. An adaptive multiscale method for quasi-static crack growth. Computational Mechanics, 53(6):1129{1148, 2014a. Pattabhi R Budarapu, Robert Gracie, Shih-Wei Yang, Xiaoying Zhuang, and Timon Rabczuk. Ecient coarse graining in multiscale modeling of fracture. Theoretical and Applied Fracture Mechanics, 69:126{143, 2014b. Shih-Wei Yang, Pattabhi R Budarapu, D Roy Mahapatra, St ephane PA Bordas, Goangseup Zi, and Timon Rabczuk. A meshless adaptive multiscale method for fracture. Computational Materials Science, 96:382{395, 2015. P. Areias, T. Rabczuk, and P.P. Camanho. Finite strain fracture of 2d problems with injected anisotropic softening elements. Theoretical and Applied Fracture Mechanics, 72(1):50{63, 2014. doi: 10.1016/j.tafmec.2014.06.006. SS Nanthakumar, Tom Lahmer, and Timon Rabczuk. Detection of multiple aws in piezo- electric structures using xfem and level sets. Computer Methods in Applied Mechanics and Engineering, 275:98{112, 2014. Nicolas Mo es and Ted Belytschko. Extended nite element method for cohesive crack growth. Engineering fracture mechanics, 69(7):813{833, 2002. 44 Timon Rabczuk, Goangseup Zi, Axel Gerstenberger, and Wolfgang A Wall. A new crack tip element for the phantom-node method with arbitrary cohesive cracks. International Journal for Numerical Methods in Engineering, 75(5):577{599, 2008a. Thanh Chau-Dinh, Goangseup Zi, Phill-Seung Lee, Timon Rabczuk, and Jeong-Hoon Song. Phantom-node method for shell models with arbitrary cracks. Computers & Structures, 92: 242{256, 2012. Nam Vu-Bac, Hung Nguyen-Xuan, Lei Chen, Chang-Kye Lee, Goangseup Zi, X Zhuang, Gui Rong Liu, and Timon Rabczuk. A phantom-node method with edge-based strain smooth- ing for linear elastic fracture mechanics. Journal of Applied Mathematics, 2013, 2013. Timon Rabczuk, PMA Areias, and Ted Belytschko. A simpli ed mesh-free method for shear bands with cohesive surfaces. International Journal for Numerical Methods in Engineering, 69(5):993{1021, 2007a. Timon Rabczuk and Goangseup Zi. A meshfree method based on the local partition of unity for cohesive cracks. Computational Mechanics, 39(6):743{760, 2007. Timon Rabczuk, St ephane Bordas, and Goangseup Zi. A three-dimensional meshfree method for continuous multiple-crack initiation, propagation and junction in statics and dynamics. Computational Mechanics, 40(3):473{495, 2007b. T Rabczuk and E Samaniego. Discontinuous modelling of shear bands using adaptive meshfree methods. Computer Methods in Applied Mechanics and Engineering, 197(6):641{658, 2008. Timon Rabczuk, Goangseup Zi, St ephane Bordas, and Hung Nguyen-Xuan. A geometrically non- linear three-dimensional cohesive crack method for reinforced concrete structures. Engineering Fracture Mechanics, 75(16):4740{4758, 2008b. F. Amiri, C. Anitescu, M. Arroyo, S.P.A. Bordas, and T. Rabczuk. Xlme interpolants, a seamless bridge between xfem and enriched meshless methods. Computational Mechanics, 53(1):45{57, 2014a. doi: 10.1007/s00466-013-0891-2. 45 T. Rabczuk and T. Belytschko. Cracking particles: A simpli ed meshfree method for arbitrary evolving cracks. International Journal for Numerical Methods in Engineering, 61(13):2316{ 2343, 2004. doi: 10.1002/nme.1151. T. Rabczuk and T. Belytschko. A three-dimensional large deformation meshfree method for arbitrary evolving cracks. Computer Methods in Applied Mechanics and Engineering, 196 (29-30):2777{2799, 2007. doi: 10.1016/j.cma.2006.06.020. Timon Rabczuk, Goangseup Zi, Stephane Bordas, and Hung Nguyen-Xuan. A simple and robust three-dimensional cracking-particle method without enrichment. Computer Methods in Applied Mechanics and Engineering, 199(37):2437{2455, 2010. Timon Rabczuk and Huilong Ren. A peridynamics formulation for quasi-static fracture and contact in rock. Engineering Geology, 2017. Huilong Ren, Xiaoying Zhuang, Yongchang Cai, and Timon Rabczuk. Dual-horizon peridynam- ics. International Journal for Numerical Methods in Engineering, 108(12):1451{1476, 2016. Huilong Ren, Xiaoying Zhuang, and Timon Rabczuk. Dual-horizon peridynamics: A stable solution to varying horizons. Computer Methods in Applied Mechanics and Engineering, 318: 762{782, 2017. Ted Belytschko and Jerry I Lin. A three-dimensional impact-penetration algorithm with erosion. International Journal of Impact Engineering, 5(1-4):111{127, 1987. Gordon R Johnson and Robert A Stryk. Eroding interface and improved tetrahedral element algorithms for high-velocity impact computations in three dimensions. International Journal of Impact Engineering, 5(1-4):411{421, 1987. Jeong-Hoon Song, Hongwu Wang, and Ted Belytschko. A comparative study on nite element methods for dynamic fracture. Computational Mechanics, 42(2):239{250, 2008. Tran Quoc Thai, Timon Rabczuk, Yuri Bazilevs, and Gun  ther Meschke. A higher-order stress- based gradient-enhanced damage model based on isogeometric analysis. Computer Methods in Applied Mechanics and Engineering, 304:584{604, 2016. 46 Gilles Pijaudier-Cabot, Khalil Haidar, and Jean-Fran cois Dub e. Non-local damage model with evolving internal length. International journal for numerical and analytical methods in ge- omechanics, 28(7-8):633{652, 2004. P. Areias, M.A. Msekh, and T. Rabczuk. Damage and fracture algorithm using the screened poisson equation and local remeshing. Engineering Fracture Mechanics, 158:116{143, 2016a. doi: 10.1016/j.engfracmech.2015.10.042. Blaise Bourdin, Gilles A Francfort, and Jean-Jacques Marigo. The variational approach to fracture. Journal of elasticity, 91(1):5{148, 2008. Christian Miehe, Martina Hofacker, and Fabian Welschinger. A phase eld model for rate- independent crack propagation: Robust algorithmic implementation based on operator splits. Computer Methods in Applied Mechanics and Engineering, 199(45):2765{2778, 2010a. C Miehe, F Welschinger, and M Hofacker. Thermodynamically consistent phase- eld models of fracture: Variational principles and multi- eld fe implementations. International Journal for Numerical Methods in Engineering, 83(10):1273{1311, 2010b. C Hesch and K Weinberg. Thermodynamically consistent algorithms for a nite-deformation phase- eld approach to fracture. International Journal for Numerical Methods in Engineering, 99(12):906{924, 2014. Michael J Borden, Clemens V Verhoosel, Michael A Scott, Thomas JR Hughes, and Chad M Landis. A phase- eld description of dynamic brittle fracture. Computer Methods in Applied Mechanics and Engineering, 217:77{95, 2012. P. Areias, T. Rabczuk, and M.A. Msekh. Phase- eld analysis of nite-strain plates and shells including element subdivision. Computer Methods in Applied Mechanics and Engineering, 312:322{350, 2016b. doi: 10.1016/j.cma.2016.01.020. Fatemeh Amiri, Daniel Mill an, Yongxing Shen, Timon Rabczuk, and M Arroyo. Phase- eld modeling of fracture in linear thin shells. Theoretical and Applied Fracture Mechanics, 69: 102{109, 2014b. doi: 10.1016/j.tafmec.2013.12.002. 47 Mohammed A Msekh, Juan Michael Sargado, Mostafa Jamshidian, Pedro Miguel Areias, and Timon Rabczuk. Abaqus implementation of phase- eld model for brittle fracture. Computa- tional Materials Science, 96:472{484, 2015. Guowei Liu, Qingbin Li, Mohammed A Msekh, and Zheng Zuo. Abaqus implementation of monolithic and staggered schemes for quasi-static and dynamic fracture phase- eld model. Computational Materials Science, 121:35{47, 2016. Blaise Bourdin, Gilles A Francfort, and Jean-Jacques Marigo. Numerical experiments in revisited brittle fracture. Journal of the Mechanics and Physics of Solids, 48(4):797{826, 2000. Xue Zhang, Chet Vignes, Scott W Sloan, and Daichao Sheng. Numerical evaluation of the phase- eld model for brittle fracture with emphasis on the length scale. Computational Mechanics, 59(5):737{752, 2017. C Miehe. Computation of isotropic tensor functions. International Journal for Numerical Meth- ods in Biomedical Engineering, 9(11):889{896, 1993. Alex Toth and CT Kelley. Convergence analysis for anderson acceleration. SIAM Journal on Numerical Analysis, 53(2):805{819, 2015. G Gao, S Huang, K Xia, and Z Li. Application of digital image correlation (DIC) in dynamic notched semi-circular bend (NSCB) tests. Experimental Mechanics, 55(1):95{104, 2015. M Cai. Fracture initiation and propagation in a brazilian disc with a plane interface: a numerical study. Rock mechanics and rock engineering, 46(2):289{302, 2013. Xiao-Ping Zhou and Yun-Teng Wang. Numerical simulation of crack propagation and coales- cence in pre-cracked rock-like brazilian disks using the non-ordinary state-based peridynamics. International Journal of Rock Mechanics and Mining Sciences, 89:235{249, 2016. C Atkinson, RE Smelser, and Jo Sanchez. Combined mode fracture via the cracked brazilian disk test. International Journal of Fracture, 18(4):279{291, 1982. Martin Entacher, E Schuller, and R Galler. Rock failure and crack propagation beneath disc cutters. Rock Mechanics and Rock Engineering, 48(4):1559{1572, 2015. 48 N Erarslan and DJ Williams. Investigating the e ect of cyclic loading on the indirect tensile strength of rocks. Rock mechanics and rock engineering, 45(3):327{340, 2012. Xiao-Ping Zhou, Xin-Bao Gu, and Yun-Teng Wang. Numerical simulations of propagation, bifurcation and coalescence of cracks in rocks. international journal of Rock Mechanics and Mining Sciences, 80:241{254, 2015. Per-Erik Petersson. Crack growth and development of fracture zones in plain concrete and similar materials. Report TVBM 1006, 1981. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Phase field modelling of crack propagation, branching and coalescence in rocks

Zhou, Shuwei; Zhuang, Xiaoying; Zhu, Hehua; Rabczuk, Timon
Condensed Matter , Volume 2019 (1902) – Feb 14, 2019
Download PDF
49 pages

Loading next page...
 
/lp/arxiv-cornell-university/phase-field-modelling-of-crack-propagation-branching-and-coalescence-RMHUEYCsJJ

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

ISSN
0167-8442
eISSN
ARCH-3331
DOI
10.1016/j.tafmec.2018.04.011
Publisher site
See Article on Publisher Site

Abstract

We present a phase eld model (PFM) for simulating complex crack patterns including crack propagation, branching and coalescence in rock. The phase eld model is imple- mented in COMSOL and is based on the strain decomposition for the elastic energy, which drives the evolution of the phase eld. Then, numerical simulations of notched semi-circular bend (NSCB) tests and Brazil splitting tests are performed. Subsequently, crack propaga- tion and coalescence in rock plates with multiple echelon aws and twenty parallel aws are studied. Finally, complex crack patterns are presented for a plate subjected to increasing internal pressure, the (3D) Pertersson beam and a 3D NSCB test. All results are in good agreement with previous experimental and numerical results. Keywords: Phase eld, Rock, COMSOL, Crack propagation, Crack branching arXiv:1902.06686v1 [cond-mat.soft] 14 Feb 2019 1 Introduction Fracture-induced failure has gained extensive concern in engineering because of the huge threat to engineering safety (Anderson 2005). The prediction of fracture in rock is challenging. Rock masses have many pre-existing aws, such as micro cracks, voids and soft minerals. Many e orts have been made to study crack propagation in rock, see for instance the contributions in Bobet and Einstein (1998), Wong et al. (2001), Sagong and Bobet (2002), Wong and Einstein (2009), Park and Bobet (2009), Park and Bobet (2010), Lee and Jeon (2011), and Zhou et al. (2014). However, many studies focus on uniaxial compressive loads since tensile loads or more complicated load cases, which are more dicult to perform in practical tests. Numerical methods are a good alternative to study fracture problems. They are less expen- sive than experimental tests and can provide physical insight dicult to gain through 'pure' experimental testing. Computational methods for fracture can be classi ed in discrete and continuous approaches. Ecient remeshing techniques (Areias and Rabczuk 2017, Areias et al. 2013, Areias and Rabczuk 2013), multiscale method (Budarapu et al. 2014a,b, Yang et al. 2015), strain-softening element (Areias et al. 2014), the extended nite element method (Nanthaku- mar et al. 2014, Mo es and Belytschko 2002), the phantom node method (Rabczuk et al. 2008a, Chau-Dinh et al. 2012, Vu-Bac et al. 2013) and speci c meshfree methods (Rabczuk et al. 2007a, Rabczuk and Zi 2007, Rabczuk et al. 2007b, 2008a, Rabczuk and Samaniego 2008, Rabczuk et al. 2008b, Amiri et al. 2014a) are classical representatitves of the rst class. The cracking parti- cles method (CPM) (Rabczuk and Belytschko 2004, 2007, Rabczuk et al. 2010), Peridynamics (Rabczuk and Ren 2017) and dual-horizon peridynamics (Ren et al. 2016, 2017) are also dis- crete crack approaches but they share the simplicity of continuous approaches to fracture as they also do not require any explicit representation of the crack surface and any crack tracking algorithms. Element-erosion (Belytschko and Lin 1987, Johnson and Stryk 1987) directly sets the stresses of the elements to zero when the elements ful ll the fracture criterion. However, the element-erosion method cannot simulate crack branching correctly (Song et al. 2008). Gradient models (Thai et al. 2016), non-local models (Pijaudier-Cabot et al. 2004), models based on the screend-poisson equation (Areias et al. 2016a) and also phase eld models are typical continuous approaches to fracture. 2 In this paper, we pursue the phase eld model (PFM) (Bourdin et al. 2008, Miehe et al. 2010a,b, Hesch and Weinberg 2014, Borden et al. 2012) to model crack propagation, branching and coalescence in rock. The origins of the PFM can be traced back to Bourdin et al. (2008), but a thermodynamic consistent framework was rst presented by Miehe et al. (2010a). Considerable attention has been paid to PFMs due to their ease in implementation and applicability to multi- physics problems. The PFM does not treat the crack as a physical discontinuity but uses a scalar eld (the phase eld) to smoothly transit the intact material to the broken one. Thus, the sharp crack is represented by a 'damage-like' zone. The shape of the crack is controlled by a length scale parameter and propagation of the crack is obtained through the solution of a di erential equation. Thus, the PFM does not require any external criterion for fracture and additional work to track the fracture surface algorithmically (Borden et al. 2012). It is believed that for this reason, the phase eld is therefore has some advantage over other approaches in modeling branching and merging of multiple cracks. Phase eld models have been discretized in the context of the nite element method (Areias et al. 2016b), meshfree methods (Amiri et al. 2014b) and isogeometric analysis (Borden et al. 2012); the latter two approaches use a fourth-order di erential equation for the phase eld exploiting the higher continuity of the meshfree and isogeometric approximation. The PFM for brittle cracks has also been implemented in commercial software such as ABAQUS (Msekh et al. 2015, Liu et al. 2016). However, the extension of the implementation in ABAQUS to problems with more elds { as hydraulic fracturing { is dicult. Hence, we present an implementation of the phase eld model in COMSOL Multiphysics, a software particularly dedicated to multi- eld modeling. This paper is organized as follows. The phase eld model for brittle fractures is presented in Section 2. Subsequently, the numerical implementation of the phase eld model in COMSOL is described in Section 3. Then, simulations of initiation, propagation, branching, and coalescence of cracks in rock are shown in Section 4 before Section 5 concludes our manuscript. 3 2 Theory of phase eld modeling 2.1 Theory of brittle fracture Consider an elastic body R (d 2 f1; 2; 3g) as shown in Figure 1, whose external boundary and internal discontinuity boundary are denoted as @ and , respectively; x is the position vector and u(x; t)  R the displacement vector at time t. In Fig. 1, the body satis es the time-dependent Dirichlet boundary conditions (u (x; t) = g (x; t) on @ ), and also the i i g time-dependent Neumann conditions on @ ; b(x; t)  R is the body force and f (x; t) the traction on boundary @ Figure 1: Phase eld approximation of the crack surface Given that the stored elastic energy can be transformed into dissipative forms of energy, the classical Grith's theory (Anderson 2005) for brittle fracture states that the crack starts to propagate when the stored energy is sucient to overcome the fracture resistance of the material. Therefore, the crack propagation is regarded as a process to minimize a free energy L that consists of the kinetic energy (u _ ), elastic energy , fracture energy and external kin " f work W : ext Z Z Z Z L = (u _ ) (")d G dS + b ud + f  udS (1) kin " c | {z } | {z } | {z } " f ext 4 @u where u _ = , is the elastic energy density, and G is the critical energy release rate. The " c @t linear strain tensor " = "(u) is given by " = ru + (ru) (2) The kinetic energy is given by (u _ ) = u _ d (3) kin where  indicates the density. 2.2 Phase led approximation for the fracture energy The phase eld method (Miehe et al. 2010a,b, Borden et al. 2012) uses a scalar eld, i.e. the phase eld, to smear out the crack surface (see Fig. 1) over the domain . The phase eld (x; t) 2 [0; 1] has to satisfy the following conditions: 0; if material is intact = (4) 1; if material is cracked A typical one dimensional phase eld approximated by the exponential function is given by (Miehe et al. 2010a) jxj=l (x) = e (5) l denoting the length scale parameter, which controls the transition region of the phase eld and thereby re ects the width of the crack. The distribution of the one dimensional phase eld is shown in Fig. 2. The crack region will have a larger width as l increases and the phase eld will represent a sharp crack when l tends to zero. It can be shown that the crack surface density per unit volume of the solid is given by (Miehe et al. 2010a) (;5) = + jrj (6) 2l 2 5 Figure 2: Distribution of the one dimensional phase eld across a crack Thus, the fracture energy is approximated by Z Z G dS = G + jrj d (7) c c 2l 2 The variational approach (Bourdin et al. 2000) states that the crack surface energy is trans- formed from the elastic energy, which drives the evolution of the phase eld. To capture cracks only under tension, the elastic energy is decomposed into tensile and compressive parts (Miehe et al. 2010b): " = h" i n n (8) a  a a a=1 where " and " are the tensile and compressive strain tensors, respectively. In addition, " + a is the principal strain and n is the direction of the principal strain. The operators hi are de ned as : hi = (j j)=2. Consequently, the positive and negative elastic energy densities are expressed as 2 2 (") = htr(")i + tr " (9) where  > 0 and  > 0 are the Lam e constants. The Lam e constants are related to the Young's modulus E and Poisson's ratio  of the solid through the well known relation: (1 + )(1 2) (10) :  = 2(1 + ) The phase eld is assumed to a ect only the positive elastic energy density, which introduces 6 a sti ness reduction as (Borden et al. 2012) 2 + (") = (1 k)(1 ) + k (") + (") (11) " " where 0 < k  1 is a stability parameter for avoiding numerical singularities because the positive elastic energy density disappears as the phase eld  tends to 1. 2.3 Governing equations By substituting Eqs. (3), (7), and (11) into Eq. (1), the energy functional L is rewritten as Z Z 2 2 + L = u _ d (1 k)(1 ) + k (") + (") d " " Z Z Z G + jrj d + b ud + f  udS (12) 2l 2 Employ the rst variation of the functional L = 0, it can be shown that the strong form of the governing equations are given by (Borden et al. 2012) Div() + b = u  (13) + + 2l (1 k) 2l (1 k) 0 0 > " 2 2 " + 1  l r  = G G c c @ u where u  = and  is Cauchy stress tensor given by @t = @ " " (14) = (1 k)(1 ) + k [htr(")i I + 2" ] + htr(")i I + 2" + + dd with unit tensor I 2 R . The phase eld requires the irreversibility condition (x; s) 2 (x; t)(s < t) during compres- sion or unloading, i.e. the crack cannot be healed. Therefore, we introduce a strain-history eld H (x; t) (Miehe et al. 2010a,b) to ensure a monotonically increasing phase eld: H (x; t) = max ("(x; s)) (15) x2[0;t] 7 The history eld H satis es the Kuhn-Tucker conditions for loading and unloading (Miehe et al. 2010a): + + _ _ H  0; H  0; H ( H ) = 0 (16) " " By replacing by H (x; t) in Eq. (13), the strong form is rewritten as Div() + b = u  (17) 2l (1 k)H 2l (1 k)H 0 0 > 2 2 + 1  l r  = G G c c We denote m as the outward-pointing normal vector to the boundaries, and the governing equations are subjected to the Dirichlet and Neumann boundary conditions u = g on @ > g (18) m = f on @ r m = 0 on @ along with the initial conditions u(x; 0) = u (x) x 2 (19) v(x; 0) = v (x) x 2 (x; 0) =  (x) x 2 2.4 The choice of l Borden et al. (2012) and Zhang et al. (2017) proposed an analytical solution for the critical tensile stress  that a one-dimensional bar can sustain: cr 9 EG = (20) cr 16 3l There is an apparent singularity when l tends to zero, i.e. in case of a sharp crack, which is phyiscally meaningless. However, assuming all other parameters except of l are known, eq. 8 (20) can be solved for l : 27EG l = (21) cr In Eq. (21), the critical energy release rate G and Young's modulus E can be obtained by conducting regular experimental tests, while the critical stress  can be approximated by the cr tensile strength  through the standard tensile test. Hence, Eq. (21) can estimate the length scale. Note that the accuracy is unknown for more complex cases. 3 Numerical implementation 3.1 Finite element method We use the nite element method to solve the governing equations (17) given in weak form by Z Z Z (u   u  : ") d + b ud + f  udS = 0 (22) and Z Z 2(1 k)H (1 )d + G l rr +  d = 0 (23) c 0 We use the standard vector-matrix notation and denote the nodal values of the displacement and phase eld as u and  . The discretization is thereby given by i i u = N d;  = N  (24) where d and  are the vectors consisting of node values u and  . N and N are shape i i u function matrices given by 2 3 N 0 0 : : : N 0 0 1 n 6 7 h i 6 7 N = 6 7 ; N = (25) u 0 N 0 : : : 0 N 0  N N : : : N 1 n 1 2 n 4 5 0 0 N : : : 0 0 N 1 n where n is the node number in one element and N is the shape function of node i. The same 9 discretization is applied to the test functions and we obtain u = N d;  = N  (26) where d and  are the vectors consisting of node values of the test functions. The gradients are thereby calculated by ^ ^ " = B d; r = B ; " = B d; r = B  (27) u  u where B and B are the derivatives of the shape functions de ned by 2 3 N 0 0 : : : N 0 0 1;x n;x 6 7 6 7 2 3 6 7 0 N 0 : : : 0 N 0 1;y n;y 6 7 N N : : : N 1;x 2;x n;x 6 7 6 7 6 0 0 N : : : 0 0 N 7 1;z n;z 6 7 6 7 B = ; B = 6 7 (28) u  N N : : : N 1;y 2;y n;y 6 7 4 5 6 7 N N 0 : : : N N 0 1;y 1;x n;y n;x 6 7 N N : : : N 1;z 2;z n;z 6 7 6 7 0 N N : : : 0 N N 1;z 1;y n;z n;y 4 5 N 0 N : : : N 0 N 1;z 1;x n;z n;x By applying the nite element approximation, the equations of weak form (22) and (23) are then written as " # Z Z Z Z T T T T T T (d) N N d d + B D B d d + (d) N bd + N f dS = 0 (29) u e u u u u u Z Z T T T T T ^ ^ ^ () B G l B + N + 2(1 k)H N d + () 2(1 k)HN d = 0 c 0 (30) where D is the degraded sti ness matrix. D can be calculated from the fourth order elasticity e e tensor D: D = @" 2 2 + =  (1 k)(1 ) + k H (tr(")) + H (tr(")) J + 2 (1 k)(1 ) + k P + P " " (31) 10 P P 3 3 where J =   ,  being the Kronecker and P = H (" ) n n n n + ijkl ij kl ij " a ab ai aj bk bl ijkl a=1 b=1 P P 3 3 h" i h" i 1 a n n (n n + n n ) with n the i-th component of vector n ; H hxi ai bj ak bl bk al ai a " a=1 b6=a 2 " " is the Heaviside function: 1; x > 0 H hxi = (32) 0; x  0 Since P cannot be computed when " = " , we apply a \perturbation" technology for the a b ijkl principal strains (Miehe 1993) with an unchanged " : " = " (1 + ) if " = " 1 1 1 2 (33) " = " (1 ) if " = " 3 3 2 3 with the perturbation  = 1 10 . For admissible arbitrary test functions, Eqs. (29) and (30) always hold, thereby producing the discretized weak form as Z Z Z Z T T T T N N d d B D B d d + N bd + N f dS = 0 (34) e u u u u u | {z } | {z } | {z } int ine  F =K d F =Md u ext u F Z     Z T T T B G l B + N + 2(1 k)H N d + 2(1 k)HN d = 0 (35) c 0 | {z } | {z } ext int ^ F F =K ine int ext where F , F , and F are the inertial, internal, and external forces for the displacement u u u int ext eld and F and F are the internal and external force terms of the phase eld. In addition, the mass and sti ness matrices follow M = N N d > Z K = B D B d (36) u e u > Z > T T K = B G l B + N + 2(1 k)H N d c 0 In this paper, we use a staggered scheme to solve the displacement and phase elds. Thus, the Newton-Raphson approach is adopted to obtain the residual of the discrete equations R = 11 ext ine int ext int F F F = 0 and R = F F = 0, respectively. u u u 3.2 COMSOL implementation We implemented the phase eld approach into COMSOL Multiphysics, which can simulate mathematical and physical problems easily by adding application-speci c modules. Therefore, it is suitable for multi- eld modeling. In this paper, we construct three main modules, namely, the Solid Mechanics, History-strain and Phase Field Modules, which employ the standard nite element discretization in space as described in Subsection 3.1. In addition, a pre-set Storage Module is employed to evaluate and store the intermediate eld variables in a time step, such as the positive elastic energy and principal strains. Based on a linear elastic material library, the Solid Mechanics Module is used for the dis- placement u. The boundary and initial conditions shown in Section 2 are added to the Solid Mechanics Module, while a non-linear stress-strain relationship is considered. The sti ness ma- trix D is modi ed in a time step as follows 2 3 D D D D D D 1111 1122 1133 1112 1123 1113 6 7 6 7 6 D D D D D D 7 2211 2222 2233 2212 2223 2213 6 7 6 7 6 7 D D D D D D 3311 3322 3333 3312 3323 3313 6 7 D = (37) 6 7 6 D D D D D D 7 1211 1222 1233 1212 1223 1213 6 7 6 7 6 7 D D D D D D 2311 2322 2333 2312 2323 2313 4 5 D D D D D D 1311 1322 1333 1312 1323 1313 The governing equation (17) is presented for dynamic crack problems. However, for a quasi- static problem, the inertia term must be neglected in the Solid Mechanics Module. The Phase Field Module is established for the phase eld  by revising a pre-de ned module, which is governed by the Helmholtz equation. The boundary condition Eq. (18) and initial condition (19) are also implemented in this module. For the history strain eld H , the Distributed ODEs and DAEs Interfaces are used to construct the History-strain Module, where the history strain eld is not solved directly. We use a \previous solution" solver to record the results in the 12 previous time steps and obtain the eld H by applying the following format in COMSOL: + + H = 0; if > H " " (38) H H = 0; if  H Additionally, the initial condition H (x) = 0 is used for the History-strain Module. Figure 3 shows the relationship among all the established modules. The mechanical re- sponses, such as the principal strains, the directions of principal strain, and the elastic energy, are naturally exported from the Solid Mechanics Module and stored in the Storage Module. The History-strain Module then call the positive elastic energy and update the local history strain eld H . The Phase Field Module employs the updated H to solve the phase eld. In a time step, the updated phase eld and the stored principal strains as well as the directions are used to modify the sti ness matrix of the Solid Mechanics Module and subsequently to obtain the mechanical responses. The detailed procedure of the staggered (segregated) scheme is depicted in Fig. 4. The equations of displacement, history strain and phase- eld are solved independently. An implicit Generalized- method (Borden et al. 2012) is used for time integration. The Generalized- method is unconditionally stable and requires a prediction step and a correction step. When a new time step starts, an initial guess from the linear extrapolation of the previous solution is used in the prediction stage. Newton-Raphson iterations are used to solve the residuals for each module in the correction stage. That is, in the iteration step j + 1 of a given time step i, the dis- j+1 j j j placement u is rst solved by using the results (u , H and  ) of the previous iteration step i i i i j+1 j. Following the updated displacements u , the history strain is updated. Subsequently, an- j+1 other Newton-Raphson iteration is applied to solve the phase- eld  . The total relative error is estimated and the iterations continue until the tolerance requirement is met. The maximum number of iteration in one time step is set as 50 in our simulations. We accelerate the convergence by using the Anderson acceleration (Toth and Kelley 2015) where the dimension of the iteration space eld is chosen as more than 50. A ow chart of our implementation is shown in Fig. 5. The source code can be found in \https://sourceforge.net/projects/phase eldmodelingcomsol/". 13 Figure 3: Relationship between all the modules established Figure 4: Segregated scheme for the coupled calculation in phase eld modeling 14 Figure 5: COMSOL implementation of the phase eld modeling 4 Numerical examples of crack propagation, branching and coalescence in rocks 4.1 Simulation of notched semi-circular bend (NSCB) tests Let us consider the notched semi-circular bend (NSCB) test rst. Geometry and boundary condition of the rock specimen are shown in Fig.6. The mechanical properties of the rock are taken from the Laurentian granite (LG) from Grenville province of Canada (Gao et al. 2015). The rock density  is 2630 kg/m , while the Young's modulus E and Poisson's ratio  are 92 GPa and 0.21, respectively. We follow the 1D solution of Borden et al. (2012) for the critical 2 4 stress of material softening, and choose G = 7.6 J/m and the length scale l = 4:5 10 m. c 0 This produces a critical stress close to the quasi-static tensile strength of the specimen (12.8 MPa) (Gao et al. 2015). The phase eld modeling is performed by using 93587 6-node quadratic triangular elements and the maximum element size is h = 2:25 10 m. We apply a vertical displacement on the 15 top of the specimen to drive crack propagation from the tip of the notch. During the simulation, we apply the displacement increment u = 5 10 mm in each time step. Figure 7 shows the crack initiation and propagation in the rock specimen by using the phase eld model. When the applied displacement u reaches to 6:72  10 mm, the crack initiates from the tip of the notch. Subsequently, the crack propagates straightly in the vertical direction 3 3 when u accumulates to 6:74 10 mm and 6:77 10 mm. When the displacement reaches to 6:86 10 mm, the tip of the propagating crack is close to the upper boundary of the specimen and failure of the semi-circular rock specimen occurs. The crack patterns obtained by the phase eld simulation are in good agreement with the experimental results in Gao et al. (2015). Figure 6: Geometry and boundary condition of the notched semi-circular bend (NSCB) tests We test the crack patterns for G = 5:6, 6.6, 8.6, and 9.6 J/m . The results show that G has c c little in uence on the crack patterns. Di erent G have the same crack paths. Figure 8 gives the curves of the reaction force on the upper boundary of the specimen versus the displacement for di erent G . A sudden drop of the load is observed after the crack initiation. In addition, all the curves have the same slope before the crack initiation under di erent G , while the maximum load increases as G increases . We also test the in uence of the maximum mesh size h under a xed G = 7.6 J/m and the 4 4 4 5 length scale l = 4:5 10 m. We choose h = 2:25 10 m, 1:13 10 m, and 5:63 10 m in the tests. The resulting load-displacement curve is shown in Fig. 9. As expected, the load-displacement curve converges with the mesh re nement. To test the in uence of the length 4 2 scale l , we x the mesh size h = 2:25  10 m and G = 7.6 J/m , and then present the 0 c 16 (a) (b) (c) (d) Figure 7: Crack propagation of the notched semi-circular bend (NSCB) tests at a displacement 3 3 3 of (a) u = 6:72  10 mm, (b) u = 6:77  10 mm, (c) u = 6:74  10 mm, and (d) 3 2 u = 6:86 10 mm for G = 7:6 J/m Figure 8: Load-displacement curves of the notched semi-circular bend (NSCB) tests 17 load-displacement curve for di erent l in Fig. 10. Figure 10 shows a decreasing peak load and displacement range when the length scale increases. Figure 9: Load-displacement curves of the NSCB tests under di erent mesh size Figure 10: Load-displacement curves of the NSCB tests under di erent length scale l 4.2 Simulation of Brazil splitting tests Brazil splitting tests are commonly used to obtain the tensile strength of rocks and many re- searchers simulated crack propagation in a Brazilian disc under compression, such as Cai (2013) 18 and Zhou and Wang (2016). Figure 11 presents the geometry of the Brazilian disc along with the boundary conditions. Figure 11: Geometry and boundary condition of the Brazil splitting tests These material parameters are adopted:  = 2630 kg/m , E = 31.5 GPa, and  = 0.25. The length scale parameter l is xed to 1 mm. 26700 linear triangular elements (base mesh) are used to discretize the disc with the maximum element size h = 0:5 mm, k is set to 1  10 . Finally, we conduct the simulation by using ve di erent G : 50, 75, 100, 125, and 150 J/m . Figure 12 shows the crack initiation and propagation in the Brazil splitting tests for G = 100 J/m . When the displacement u approaches a value of 0.476 mm, the crack occurs in the center of the disc where the maximum tensile stress occurs which is in good agreement with the experimental and analytical results (Atkinson et al. 1982, Entacher et al. 2015). When u reaches a value of 0.477 mm, the crack propagates with an increasing width. Then, the crack continues to propagate and the crack tips move close to the upper and bottom ends of the disc when u reaches to 0.478 mm. The crack branching is observed when u is close to 0.480 mm. In addition, the crack cannot penetrate deeply into the ends of the disc because of the locally compressed area around both ends. We compare the curve of the reaction force on the upper end of the Brazilian disc versus the displacement u with the experimental result in Fig. 13. The experimental curve is originated from the work of Erarslan and Williams (2012). Figure 13 shows that the phase eld model can reproduce the experimental results well. The main reason for the di erence in Fig. 13 is 19 (a) (b) (c) (d) Figure 12: Crack propagation of the Brazil splitting tests at a displacement of (a) u = 0:476 mm, (b) u = 0:477 mm, (c) u = 0:478 mm, and (d) u = 0:480 mm for G = 100 J/m 20 that there are contact issues in the actual Brazilian tests and a compaction stage is commonly observed before the elastic stage. Figure 13: Comparison of the load-displacement curves obtained by the experimental test and phase eld modeling Figure 14 presents the curves of the reaction force on the upper end of the Brazilian disc versus the displacement u for di erent G . Similar to the NSCB tests, a sudden drop of the load is observed after the phase eld increases to 1. In addition, the peak load increases with the increase in G . We then test the in uence of the length scale l under a xed G = 100 J/m c 0 c and mesh size h = 0:5 mm. The tested length scale parameters are l = 0.5, 1, 2, and 3 mm, respectively. The load-displacement curves under di erent length scale l are shown in Fig. 15. As shown in Fig. 15, the peak load of the specimen decreases with the increase in the length scale l . We also test the in uence of mesh size h under a xed G = 100 J/m and l = 1 mm. c 0 The maximum mesh size h is set as 1, 0.5, 0.25, and 0.125 m, respectively. The resulting load- displacement curve is shown in Fig. 16. The load-displacement curve converges with mesh re nement as expected. In addition, in the Brazil splitting test, the tensile strength  of the rock specimen is given by 2P peak = (39) DL where P is the peak load, and D and L are the diameter and length of the rock specimen. peak Thus, we compare the tensile strength by the phase eld simulation and the critical stress for 21 Figure 14: Load-displacement curves for the Brazil splitting tests for di erent G Figure 15: Load-displacement curves for the Brazil splitting tests for di erent l 22 1D by Borden et al. (2012) in Fig. 17. The tensile strength increases at a decreasing rate with the increase in G . The critical stress has the same trend as the tensile strength. However, the critical stress is far larger than the tensile strength. This observation indicates that the critical stress for 1D analysis cannot be applied directly to 2D or 3D. Figure 16: Load-displacement curves for the Brazil splitting tests for di erent h Figure 17: Comparison of tensile strength by Brazil splitting tests and critical stress for 1D 23 Figure 18: Geometry and boundary condition of three pre-existing aws 4.3 Propagation of multiple echelon aws We consider a 50 mm  50 mm square rock sample subjected to tension. The rock sample has three pre-existing aws, whose position and geometry are shown in Fig. 18. All the aws have the same length, spacing, and inclination angle of 45 . These parameters are adopted: the rock density  = 2500 kg/m , the Young's modulus E = 30 GPa, the Poisson's ratio  = 0:333, 2 5 G = 3 J/m , k = 110 , and the length scale l = 0:25 mm. Vertical displacements are applied c 0 on the top and bottom boundaries of the rock sample as shown in Fig. 18. The simulation is performed by using 106852 6-node quadratic triangular elements where the maximum element size h is 0.25 mm. In each time step, the displacement increment is u = 5 10 mm. Figure 19(a)-(f ) shows the propagation and coalescence of the three pre-existing aws. We also calculate the reaction force on the upper boundary of the rock sample, and present the load-displacement curves in Fig. 20. As the displacement u increases, both the load and the phase eld around the tips of the aws increase. When u reaches to 2:6 10 mm, the phase eld increases close to 1 and the load approaches to the maximum value. When u reaches to 2:63  10 mm, the rst tensile cracks occurs around the left and right tips of the aw . These two cracks are perpendicular to the direction of the applied displacement. In addition, the load reaches to the maximum as shown in Fig. 20. When the displacement u reaches to 2:64 10 mm, the cracks from the tips of the aw propagates perpendicular to u while the 24 load starts to drop after the peak load. Additionally, new cracks occur from the right tip of the aw 1 and the left tip of the aw 3 . 1 2 3 Figure 19(d) shows the coalescences of the cracks from the tips of the aws , , and when u = 2:65 10 mm. Thus, the aw 2 links the aws 1 and 3 , which is accompanied by a sudden drop of the load after the maximum value in Fig. 20. At the same time, new 1 3 cracks from the left tip of the aw and the right tip of the aw are observed. When u = 2:66 10 mm, the cracks from the aws 1 and 3 continue to propagate and the load decreases. When the displacement increases to 2:69 10 mm, the cracks initiating from the 1 3 left tip of the aw and the right tip of the aw reach the boundaries of the rock sample. This indicates the rock loses its load-bearing capacity, which can also be veri ed by Fig. 20. We now consider the same square rock sample subjected to tension with nine pre-existing aws in Fig. 21. These aws have the same inclination angle of 45 , while the lengths and spacing are not xed. The position and geometry of the aws are shown in Fig. 21. The parameters are the same as those in the example of three aws. A total of 107982 6-node quadratic elements are used to discretize the rock sample and the maximum element size h is 0.25 mm. The displacement increment u = 1 10 mm is applied in each time step. Figure 22(a)-(f ) shows the propagation and coalescence process of the nine pre-existing aws, and the reaction force on the upper boundary of the rock sample is depicted in Fig. 23. As the displacement u increases, the phase eld around the tips of the aws and the load both increase. The rst tensile cracks initiate from the left tips of the aws 8 and 9 when the displacement u reaches to 3:02  10 mm. At this time, the load achieves the maximum value. As the displacement increases to 3:04 10 mm, the crack from the left tip of the aw 8 propagates and links up the aw 7 . The crack from the left tip of the aw 9 continues to propagate 7 8 while new cracks initiate from the left tip of the aw and the right tips of the aws and 9 . The load then has a drop after the peak pint (a). When u reaches to 3:06 10 mm and 3:08 10 mm, the cracks initiating from the left tip of the aw 9 and the right tip of the aw continue to propagate at a decreasing rate and at a small angle with the horizontal. However, the cracks from the left tip of the aw 7 and the right tip of the aw 9 propagate at relatively large velocity nearly along the horizontal direction. The load decreases sharply as the applied displacement increases. 25 (a) (b) (c) (d) (e) (f ) Figure 19: Propagation and coalescence of the three pre-existing aws in a square rock sample 3 3 3 at a displacement of (a) u = 2:6 10 mm, (b) u = 2:63 10 mm, (c) u = 2:64 10 mm, 3 3 3 (d) u = 2:65 10 mm, (e) u = 2:66 10 mm, and (f ) u = 2:69 10 mm 26 Figure 20: Load-displacement curves for the square rock sample with three pre-existing aws Figure 21: Geometry and boundary condition of nine pre-existing aws 27 Figure 22(e) shows that the crack from the left tip of the aw 7 propagates close to the left boundary of the rock sample when u = 3:12 10 mm. The crack initiating from the right tip of the aw propagates at a relatively small rate because of a larger distance from the boundary where the applied displacement is applied. When u = 3:14 10 mm, the cracks from the left tip of the aw 7 and the right tip of the aw 9 both reaches the left and right boundaries of the rock sample, indicating the failure of rock sample. The rock loses its load-bearing capacity and the load drops to near 0 in Fig. 23. In addition, Fig. 22 shows no cracks initiation from the tips of the aws 1 - 6 . The reason is the stress shielding and ampli cation e ects due to the interaction of the aws Zhou et al. (2015). 4.4 Propagation and coalescence of twenty parallel aws A 2D square rock sample with twenty parallel pre-existing aws subjected to tension is tested. All aws have the same length, spacing, and inclination angle of 0 . We consider the aws in doubly periodic rectangular and diamond-shaped arrays, respectively. The arrangement and geometry of the aws are depicted in Fig. 24. The rock sample are 50 mm  50 mm. These parameters are adopted: the rock density  = 2450 kg/m , the Young's modulus E = 30 GPa, 2 9 the Poisson's ratio  = 0:3, G = 100 J/m , k = 1  10 , and the length scale l = 0:4 mm. c 0 The rock sample is discretized by using uniform 8-node quadratic elements with the element size h = 0:2 mm. We adopt the displacement increment u = 5 10 mm for each time step. Figure 25 presents the load-displacement curves for the square rock sample with twenty parallel aws. A sudden drop of the load is also observed after the maximum value is obtained. Figure 26 presents the propagation and coalescence of the doubly periodic rectangular array of twenty pre-existing aws. As the displacement u increases, the phase eld  concentrates at the tip of each aw and the load achieves the maximum when u = 2:185  10 mm. When the displacement u reaches to 2:19 10 mm, the rst tensile cracks initiate from the left and right tips of the aws 1 , 5 , 17 , and 20 . In addition, the load has a small drop from the peak. When u = 2:195  10 mm, the rst cracks continue to propagate perpendicular to the direction of the applied displacement and the load decreases. However, when u = 2:2 10 mm, the cracks fully connect the aws 1 , 5 , 17 , and 20 with the aws 2 , 3 , 18 , and 19 . Moreover, some new 2 18 3 cracks initiate from the right tips of the aws and as well as the left tips of the aws and 28 (a) (b) (c) (d) (e) (f ) Figure 22: Propagation and coalescence of the nine pre-existing aws in a square rock sample 3 3 3 at a displacement of (a) u = 3:02 10 mm, (b) u = 3:04 10 mm, (c) u = 3:06 10 mm, 3 3 3 (d) u = 3:08 10 mm, (e) u = 3:12 10 mm, and (f ) u = 3:14 10 mm 29 Figure 23: Load-displacement curves for the square rock sample with nine pre-existing aws (a) Doubly periodic rectangular array (b) Diamond-shaped array Figure 24: Arrangement and geometry of the twenty pre-existing aws and the boundary con- dition 30 2 19 . The load then drops to approximately 60% of the maximum load. When u = 2:205 10 mm, the cracks initiating from the right tips of the aws 2 and 18 and the cracks from the left 3 19 1 tips of the aws and coalesce. Finally, the cracks from the left tips of the aws and 17 as well as the cracks from the right tips of the aws 4 and 20 reach the left and right sides of the rock sample when u = 2:21  10 mm, indicating that the rock loses the load-bearing 5 16 capacity. In addition, no cracks initiate from the tips of the aws - . (a) Doubly periodic rectangular array (b) Diamond-shaped array Figure 25: Load-displacement curves for the square rock sample with twenty parallel pre-existing aws Figure 27 presents the propagation and coalescence of the twenty pre-existing aws in the 31 (a) (b) (c) (d) (e) (f ) Figure 26: Propagation and coalescence of the doubly periodic rectangular array of twenty pre- 2 2 existing aws at a displacement of (a) u = 2:185  10 mm, (b) u = 2:19  10 mm, (c) 2 2 2 2 u = 2:195 10 mm, (d) u = 2:2 10 mm, (e) u = 2:205 10 mm, and (f ) u = 2:21 10 mm 32 diamond-shaped array. The crack patterns for the diamond-shaped array are di erent from those for the doubly periodic rectangular array. As the displacement u increases, the phase eld rapidly increases around the tips closest to the left and right boundaries of the rock sample. When the displacement u reaches to 2:11 10 mm, the load increases to the maximum. When u = 2:115 10 mm, the tensile cracks rst emanate from the left tips of the aws 5 and 13 4 20 as well as the right tips of the aws and . When u = 2:12 10 mm, another two cracks initiate from the left tips of the aws 4 and 20 . The rst four cracks continue to propagate along 2 2 the horizontal direction, while the load at u = 2:12 10 mm is close to that at u = 2:11 10 mm and u = 2:115 10 mm. The aws close to the top and bottom edges of the sample are placed perpendicular to the direction of loading. The arrangement of the internal aws has little e ect on the propagation of the edge aws. Therefore, cracks propagating from those edge aws will show typical features of Mode-I fracture. As shown in Fig. 27, when u = 2:130 10 mm, the originally initiating cracks propagate as expected, while new cracks emanate from the right tips of the aws 3 , 5 , 13 19 , and . The load then starts to drop from the peak region. However, when u = 2:135 10 mm, the cracks fully connect the aws 3 and 19 with the aws 4 and 20 . In addition, two new cracks initiate from the left tips of the aws 3 and 19 . At this time, the load decreases to approximately half of the peak load. As the displacement u increases to 2:145 10 mm, the propagating cracks and the aws 2 , 3 , 5 , 6 , 13 , 14 , 18 , and 19 coalesce. At the same time, the cracks from the left tips of the aws 5 and 13 as well as the cracks from the right tips of 4 20 the aws and reach the left and right sides of the rock sample. The load at u = 2:145 mm is less than zero followed by new cracks emanating from the left tip of the aw 2 and the right tip of the aw 6 . 2 2 When the displacement u reaches to 2:195 10 mm and 2:3 10 mm, the crack initiating from the right tip of the aw 14 slowly propagates; however, the cracks from the left tip of the aw 2 and the right tip of the aw 6 continue to propagate obliquely at a relatively large velocity. The di erence in the crack patterns from Figs. 26 and 27 implies that initiation, propagation and coalescence of the cracks are signi cantly a ected by the arrangement of the aws. 33 (a) (b) (c) (d) (e) (f ) (g) (h) Figure 27: Propagation and coalescence of the diamond-shaped array of twenty pre-existing aws 2 2 2 at a displacement of (a) u = 2:11 10 mm, (b) u = 2:115 10 mm, (c) u = 2:120 10 mm, 2 2 2 2 (d) u = 2:13010 mm, (e) u = 2:13510 mm, (f ) u = 2:14510 mm, (f ) u = 2:19510 mm, and (g) u = 2:3 10 mm 34 4.5 Crack branching in a plate subjected to internal pressure This example is a square plate subjected to internal pressure with geometry and boundary conditions shown in Fig. 28. The internal pressure is applied on the upper and lower boundaries of the notch with p  = 1 MPa/s, while the outer boundaries of the plate are traction-free. In this example, dynamic cracks are considered, and these parameters are adopted: the rock density 3 2 9 = 2450 kg/m , the Poisson's ratio  = 0:3, G = 1 J/m , k = 1 10 , and the length scale l = 0:4 mm. The plate is discretized by using uniform Q4 elements with the element size h = l , 0 0 and we adopt the time step size t = 0:01 s. Figure 28: Geometry and boundary conditions of the plate subjected to internal pressure We consider the plate as a heterogeneous material and apply a Weibull distribution to the Young's modulus: m1 m m E E '(E) = exp (40) E E E 0 0 0 where ' is the probability density function and coecient m determines the shape of '. m also re ects the homogeneity of the material. As m increases, the material becomes more homogeneous and vice versa. In this paper, we consider m = 1, 3, 5, 7, and 9 along with m = 1 representing a homogeneous plate. We use E = 30 GPa and Fig. 29 shows the distribution of Young's modulus for di erent m. Figure 30 presents the crack patterns of the plate subjected to internal pressure for di erent m. Complex crack patterns, such as many crack branching, are observed. These observations are di erent from the previous examples. When the time t reaches 20 s, cracks initiate from the left and right tips of the notch and then start to branch. When the time t increases to 35 (a) m = 1 (b) m = 3 (c) m = 5 (d) m = 7 (e) m = 9 (f ) Homogeneous Figure 29: Young's modulus of the plate 36 30 s, the branching cracks propagate. When t = 50s, large \damage" region are observed around the notch where  is large. At t = 60 s, some cracks initiate from the upper and lower boundaries of the plate, while the cracks from the notch keep propagating. Finally, when t = 70 s, many new cracks occur inside the plate and the cracks from the upper and lower boundaries of the plate start to branch. For a smaller m, more crack branching is observed during the crack propagation. Therefore, a smaller m produces more complex crack patterns. 4.6 3D Petersson beam In this example, the phase eld model is applied to a single edge notched beam subjected to three-point bending (the so-called Petersson beam). A 3D simulation is conducted in this example. The geometry and boundary conditions are depicted in Fig. 31 according to Petersson (1981). The thickness of the beam is 50 mm. The following mechanical properties of the beam are chosen Petersson (1981): Young's modulus E = 27 GPa, Poisson's ratio  = 0.21, critical energy release rate G = 56 J/m and length scale l = 1 mm. We choose the maximum element c 0 size h = 5 mm in most of the beam but h = 1 mm in the region where the crack is expected to propagate. A displacement increment of u = 2:5 10 mm is used. Figure 32 compares the load-displacement curves obtained of the phase eld model with the experimental test (Petersson 1981). The results obtained by the phase eld model are in good agreement with the experimental test. Figure 33 presents the crack propagation in the beam at the displacements u = 0:34 mm, 0:355 mm, and 0:46 mm. 4.7 3D NSCB tests In the last example, we simulate the crack propagation in a 3D NSCB specimen. The geometry and boundary conditions are similar to the 2D tests except a thickness of 16 mm in 3D. The same parameters are used as those in the 2D tests and G = 7:6 J/m . We re ne the elements with the maximum size h = 4:5 10 m in the region where the crack is expected to propagate, 3 7 while in the rest region h = 1:8 10 m. In addition, the displacement increment u = 5 10 mm is applied. 3 3 Figure 34 presents the crack propagation at the displacements u = 6:6710 mm, 6:7110 37 t = 20 s t = 30 s t = 50 s t = 60 s t = 70 s m = 1 m = 3 m = 5 m = 7 m = 9 m = inf Figure 30: Crack patterns of the plate subjected to internal pressure 38 Figure 31: Geometry and boundary condition of Petersson beam Figure 32: Load-displacement curves of 3D petersson beam 39 (a) (b) (c) Figure 33: Crack propagation in the 3D petersson beam at a displacement of (a) u = 0:34 mm, (b) u = 0:355 mm, and (c) u = 0:46 mm 40 3 mm, and 6:745 10 mm. Only the domain with  > 0:95 are displayed for the crack shape. The crack patterns are the same as those in the 2D simulation, and also in good agreement with the results of the experimental tests (Gao et al. 2015). The example of 3D NSCB test shows the ability and practicability of the phase eld method in modeling crack propagation of rocks in 3D. (a) (b) (c) Figure 34: Crack propagation of the 3D NSCB test at a displacement of (a) u = 6:67  10 3 3 2 mm, (b) u = 6:71 10 mm, and (c) u = 6:745 10 mm for G = 7:6 J/m 41 5 Conclusions The phase eld theory for fracture is applied to study the crack propagation, branching and coalescence in rocks. Implementation details of the phase eld modeling in COMSOL are pre- sented with the consideration of cracks only due to tension. The numerical simulations of the 2D notched semi-circular bend (NSCB) tests and Brazil splitting tests are then performed. The presented results are in good agreement with those of the previous experimental tests. Subse- quently, the crack propagation and coalescence in plates with multiple echelon aws and twenty parallel aws are studied. We also present the complex crack patterns in a plate subjected to internal pressure, the increase of which produces crack propagation and branching. Finally, the simulation of a 3D Petersson beam and a 3D NSCB test are performed to show the practicability of phase eld modeling in 3D rocks. All the numerical examples presented by this work show that the initiation, propagation, coalescence, and branching of cracks are autonomous, while the phase eld modeling does not require external criterion for fracture and setting propagation path in advance. These obser- vations highlight the advantages of the phase eld method over other numerical methods in modeling complex crack propagation in rocks. Therefore, the phase eld modeling approach will be useful and practicable for other crack problems in rock engineering in future research. In addition, the presented phase eld model cannot predict the shear cracks when a rock reaches its shear strength. The reason is that the shear strength is not involved in the formulation of the phase eld method and the crack propagation is only driven by the elastic energy. In this sense, a modi ed phase eld model coupled with the shear model of rocks will be also attractive in the future. Acknowledgement The nancial support provided by the Sino-German (CSC-DAAD) Postdoc Scholarship Program 2016, the Natural Science Foundation of China (51474157), and RISE-project BESTOFRAC (734370) is gratefully acknowledged. 42 References Ted L Anderson. Fracture mechanics: fundamentals and applications. CRC press, 2005. Antonio Bobet and HH Einstein. Fracture coalescence in rock-type materials under uniaxial and biaxial compression. International Journal of Rock Mechanics and Mining Sciences, 35(7): 863{888, 1998. RHC Wong, KT Chau, CA Tang, and P Lin. Analysis of crack coalescence in rock-like materials containing three awspart i: experimental approach. International Journal of Rock Mechanics and Mining Sciences, 38(7):909{924, 2001. M Sagong and A Bobet. Coalescence of multiple aws in a rock-model material in uniaxial compression. International Journal of Rock Mechanics and Mining Sciences, 39(2):229{241, LNY Wong and HH Einstein. Crack coalescence in molded gypsum and carrara marble: part 1. macroscopic observations and interpretation. Rock Mechanics and Rock Engineering, 42(3): 475{511, 2009. CH Park and A Bobet. Crack coalescence in specimens with open and closed aws: a comparison. International Journal of Rock Mechanics and Mining Sciences, 46(5):819{829, 2009. CH Park and A Bobet. Crack initiation, propagation and coalescence from frictional aws in uniaxial compression. Engineering Fracture Mechanics, 77(14):2727{2748, 2010. Heekwang Lee and Seokwon Jeon. An experimental and numerical study of fracture coalescence in pre-cracked specimens under uniaxial compression. International Journal of Solids and Structures, 48(6):979{999, 2011. XP Zhou, H Cheng, and YF Feng. An experimental study of crack coalescence behaviour in rock-like materials containing multiple aws under uniaxial compression. Rock mechanics and rock engineering, 47(6):1961{1986, 2014. 43 P. Areias and T. Rabczuk. Steiner-point free edge cutting of tetrahedral meshes with applications in fracture. Finite Elements in Analysis and Design, 132:27{41, 2017. doi: 10.1016/j. nel. 2017.05.001. P. Areias, T. Rabczuk, and D. Dias-da Costa. Element-wise fracture algorithm based on rotation of edges. Engineering Fracture Mechanics, 110:113{137, 2013. doi: 10.1016/j.engfracmech. 2013.06.006. P. Areias and T. Rabczuk. Finite strain fracture of plates and shells with con gurational forces and edge rotations. International Journal for Numerical Methods in Engineering, 94(12): 1099{1122, 2013. doi: 10.1002/nme.4477. Pattabhi R Budarapu, Robert Gracie, St ephane PA Bordas, and Timon Rabczuk. An adaptive multiscale method for quasi-static crack growth. Computational Mechanics, 53(6):1129{1148, 2014a. Pattabhi R Budarapu, Robert Gracie, Shih-Wei Yang, Xiaoying Zhuang, and Timon Rabczuk. Ecient coarse graining in multiscale modeling of fracture. Theoretical and Applied Fracture Mechanics, 69:126{143, 2014b. Shih-Wei Yang, Pattabhi R Budarapu, D Roy Mahapatra, St ephane PA Bordas, Goangseup Zi, and Timon Rabczuk. A meshless adaptive multiscale method for fracture. Computational Materials Science, 96:382{395, 2015. P. Areias, T. Rabczuk, and P.P. Camanho. Finite strain fracture of 2d problems with injected anisotropic softening elements. Theoretical and Applied Fracture Mechanics, 72(1):50{63, 2014. doi: 10.1016/j.tafmec.2014.06.006. SS Nanthakumar, Tom Lahmer, and Timon Rabczuk. Detection of multiple aws in piezo- electric structures using xfem and level sets. Computer Methods in Applied Mechanics and Engineering, 275:98{112, 2014. Nicolas Mo es and Ted Belytschko. Extended nite element method for cohesive crack growth. Engineering fracture mechanics, 69(7):813{833, 2002. 44 Timon Rabczuk, Goangseup Zi, Axel Gerstenberger, and Wolfgang A Wall. A new crack tip element for the phantom-node method with arbitrary cohesive cracks. International Journal for Numerical Methods in Engineering, 75(5):577{599, 2008a. Thanh Chau-Dinh, Goangseup Zi, Phill-Seung Lee, Timon Rabczuk, and Jeong-Hoon Song. Phantom-node method for shell models with arbitrary cracks. Computers & Structures, 92: 242{256, 2012. Nam Vu-Bac, Hung Nguyen-Xuan, Lei Chen, Chang-Kye Lee, Goangseup Zi, X Zhuang, Gui Rong Liu, and Timon Rabczuk. A phantom-node method with edge-based strain smooth- ing for linear elastic fracture mechanics. Journal of Applied Mathematics, 2013, 2013. Timon Rabczuk, PMA Areias, and Ted Belytschko. A simpli ed mesh-free method for shear bands with cohesive surfaces. International Journal for Numerical Methods in Engineering, 69(5):993{1021, 2007a. Timon Rabczuk and Goangseup Zi. A meshfree method based on the local partition of unity for cohesive cracks. Computational Mechanics, 39(6):743{760, 2007. Timon Rabczuk, St ephane Bordas, and Goangseup Zi. A three-dimensional meshfree method for continuous multiple-crack initiation, propagation and junction in statics and dynamics. Computational Mechanics, 40(3):473{495, 2007b. T Rabczuk and E Samaniego. Discontinuous modelling of shear bands using adaptive meshfree methods. Computer Methods in Applied Mechanics and Engineering, 197(6):641{658, 2008. Timon Rabczuk, Goangseup Zi, St ephane Bordas, and Hung Nguyen-Xuan. A geometrically non- linear three-dimensional cohesive crack method for reinforced concrete structures. Engineering Fracture Mechanics, 75(16):4740{4758, 2008b. F. Amiri, C. Anitescu, M. Arroyo, S.P.A. Bordas, and T. Rabczuk. Xlme interpolants, a seamless bridge between xfem and enriched meshless methods. Computational Mechanics, 53(1):45{57, 2014a. doi: 10.1007/s00466-013-0891-2. 45 T. Rabczuk and T. Belytschko. Cracking particles: A simpli ed meshfree method for arbitrary evolving cracks. International Journal for Numerical Methods in Engineering, 61(13):2316{ 2343, 2004. doi: 10.1002/nme.1151. T. Rabczuk and T. Belytschko. A three-dimensional large deformation meshfree method for arbitrary evolving cracks. Computer Methods in Applied Mechanics and Engineering, 196 (29-30):2777{2799, 2007. doi: 10.1016/j.cma.2006.06.020. Timon Rabczuk, Goangseup Zi, Stephane Bordas, and Hung Nguyen-Xuan. A simple and robust three-dimensional cracking-particle method without enrichment. Computer Methods in Applied Mechanics and Engineering, 199(37):2437{2455, 2010. Timon Rabczuk and Huilong Ren. A peridynamics formulation for quasi-static fracture and contact in rock. Engineering Geology, 2017. Huilong Ren, Xiaoying Zhuang, Yongchang Cai, and Timon Rabczuk. Dual-horizon peridynam- ics. International Journal for Numerical Methods in Engineering, 108(12):1451{1476, 2016. Huilong Ren, Xiaoying Zhuang, and Timon Rabczuk. Dual-horizon peridynamics: A stable solution to varying horizons. Computer Methods in Applied Mechanics and Engineering, 318: 762{782, 2017. Ted Belytschko and Jerry I Lin. A three-dimensional impact-penetration algorithm with erosion. International Journal of Impact Engineering, 5(1-4):111{127, 1987. Gordon R Johnson and Robert A Stryk. Eroding interface and improved tetrahedral element algorithms for high-velocity impact computations in three dimensions. International Journal of Impact Engineering, 5(1-4):411{421, 1987. Jeong-Hoon Song, Hongwu Wang, and Ted Belytschko. A comparative study on nite element methods for dynamic fracture. Computational Mechanics, 42(2):239{250, 2008. Tran Quoc Thai, Timon Rabczuk, Yuri Bazilevs, and Gun  ther Meschke. A higher-order stress- based gradient-enhanced damage model based on isogeometric analysis. Computer Methods in Applied Mechanics and Engineering, 304:584{604, 2016. 46 Gilles Pijaudier-Cabot, Khalil Haidar, and Jean-Fran cois Dub e. Non-local damage model with evolving internal length. International journal for numerical and analytical methods in ge- omechanics, 28(7-8):633{652, 2004. P. Areias, M.A. Msekh, and T. Rabczuk. Damage and fracture algorithm using the screened poisson equation and local remeshing. Engineering Fracture Mechanics, 158:116{143, 2016a. doi: 10.1016/j.engfracmech.2015.10.042. Blaise Bourdin, Gilles A Francfort, and Jean-Jacques Marigo. The variational approach to fracture. Journal of elasticity, 91(1):5{148, 2008. Christian Miehe, Martina Hofacker, and Fabian Welschinger. A phase eld model for rate- independent crack propagation: Robust algorithmic implementation based on operator splits. Computer Methods in Applied Mechanics and Engineering, 199(45):2765{2778, 2010a. C Miehe, F Welschinger, and M Hofacker. Thermodynamically consistent phase- eld models of fracture: Variational principles and multi- eld fe implementations. International Journal for Numerical Methods in Engineering, 83(10):1273{1311, 2010b. C Hesch and K Weinberg. Thermodynamically consistent algorithms for a nite-deformation phase- eld approach to fracture. International Journal for Numerical Methods in Engineering, 99(12):906{924, 2014. Michael J Borden, Clemens V Verhoosel, Michael A Scott, Thomas JR Hughes, and Chad M Landis. A phase- eld description of dynamic brittle fracture. Computer Methods in Applied Mechanics and Engineering, 217:77{95, 2012. P. Areias, T. Rabczuk, and M.A. Msekh. Phase- eld analysis of nite-strain plates and shells including element subdivision. Computer Methods in Applied Mechanics and Engineering, 312:322{350, 2016b. doi: 10.1016/j.cma.2016.01.020. Fatemeh Amiri, Daniel Mill an, Yongxing Shen, Timon Rabczuk, and M Arroyo. Phase- eld modeling of fracture in linear thin shells. Theoretical and Applied Fracture Mechanics, 69: 102{109, 2014b. doi: 10.1016/j.tafmec.2013.12.002. 47 Mohammed A Msekh, Juan Michael Sargado, Mostafa Jamshidian, Pedro Miguel Areias, and Timon Rabczuk. Abaqus implementation of phase- eld model for brittle fracture. Computa- tional Materials Science, 96:472{484, 2015. Guowei Liu, Qingbin Li, Mohammed A Msekh, and Zheng Zuo. Abaqus implementation of monolithic and staggered schemes for quasi-static and dynamic fracture phase- eld model. Computational Materials Science, 121:35{47, 2016. Blaise Bourdin, Gilles A Francfort, and Jean-Jacques Marigo. Numerical experiments in revisited brittle fracture. Journal of the Mechanics and Physics of Solids, 48(4):797{826, 2000. Xue Zhang, Chet Vignes, Scott W Sloan, and Daichao Sheng. Numerical evaluation of the phase- eld model for brittle fracture with emphasis on the length scale. Computational Mechanics, 59(5):737{752, 2017. C Miehe. Computation of isotropic tensor functions. International Journal for Numerical Meth- ods in Biomedical Engineering, 9(11):889{896, 1993. Alex Toth and CT Kelley. Convergence analysis for anderson acceleration. SIAM Journal on Numerical Analysis, 53(2):805{819, 2015. G Gao, S Huang, K Xia, and Z Li. Application of digital image correlation (DIC) in dynamic notched semi-circular bend (NSCB) tests. Experimental Mechanics, 55(1):95{104, 2015. M Cai. Fracture initiation and propagation in a brazilian disc with a plane interface: a numerical study. Rock mechanics and rock engineering, 46(2):289{302, 2013. Xiao-Ping Zhou and Yun-Teng Wang. Numerical simulation of crack propagation and coales- cence in pre-cracked rock-like brazilian disks using the non-ordinary state-based peridynamics. International Journal of Rock Mechanics and Mining Sciences, 89:235{249, 2016. C Atkinson, RE Smelser, and Jo Sanchez. Combined mode fracture via the cracked brazilian disk test. International Journal of Fracture, 18(4):279{291, 1982. Martin Entacher, E Schuller, and R Galler. Rock failure and crack propagation beneath disc cutters. Rock Mechanics and Rock Engineering, 48(4):1559{1572, 2015. 48 N Erarslan and DJ Williams. Investigating the e ect of cyclic loading on the indirect tensile strength of rocks. Rock mechanics and rock engineering, 45(3):327{340, 2012. Xiao-Ping Zhou, Xin-Bao Gu, and Yun-Teng Wang. Numerical simulations of propagation, bifurcation and coalescence of cracks in rocks. international journal of Rock Mechanics and Mining Sciences, 80:241{254, 2015. Per-Erik Petersson. Crack growth and development of fracture zones in plain concrete and similar materials. Report TVBM 1006, 1981.

Journal

Condensed Matter – arXiv (Cornell University)

Published: Feb 14, 2019

References

  • Xlme interpolants, a seamless bridge between xfem and enriched meshless methods
    Amiri
  • Phase-field modeling of fracture in linear thin shells
    Amiri