Beruflich Dokumente
Kultur Dokumente
and M. Nikbakht
Abstract. In this paper, the extended finite element method is employed to model the presence of
discontinuities caused by frictional contact. The method is used in modeling strong discontinuity
within a standard finite element framework. In XFEM technique, the special functions are included
in standard FE method to simulate discontinuity without considering the boundary conditions in
meshing the domain. In this study, the classical finite element approximation is enriched by
applying additional terms to simulate the frictional behavior of contact between two bodies. These
terms, which are included for enrichment of nodal displacements, depend on the contact condition
between two surfaces. The partition of unity method is applied to discretize the contact area with
triangular sub-elements whose Gauss points are used for integration of the domain of elements.
Finally, numerical examples are presented to demonstrate the applicability of the XFEM in
modeling of frictional contact behavior.
1. Introduction
The numerical modeling of engineering contact problems is one of the most difficult and
demanding tasks in computational mechanics. Frictional contact can be observed in many problems;
such as: crack propagation, metal forming operation, drilling pile etc. In metal forming operations
the required shape changes are obtained by either of forming process, such as pressing, hammering,
rolling or extruding the material between the tools which are much stiffer than shaped material.
Because of large difference between deformability of the tool and material, relative movements
occur in contact area. These relative movements produce the normal and tangential stresses, which
have important role on metal flow and may cause serious inhomogeneities in works products.
*
Therefore, it is not surprising that much attention has been given to both the experimental and
numerical research aspects of this complex problem.
From the finite element point of view the modeling of interface friction has been categorized into
three routes; the traction boundary condition, contact node algorithm, and interface element
technique. In the method of 'traction boundary condition', the frictional forces are appended to the
external force vector as a traction boundary force. In this technique, the contact requirements appear
in the variational formulation as constraints. The contact node algorithm is an iterative
implementation of the boundary conditions on velocity or displacement depending on the nodal
forces at the interface friction boundary. This method is a pointwise algorithm, which uses the
contact nodes as indications of the contact conditions. The interface element technique is an
alternative to contact node algorithm, which can be applied by implementation of thin elements
having very high aspect ratio. Early studies on contact problems were largely related to a linear
geometry and often involved node-to-node contacts when two boundaries come into contact
(Fracavilla and Zienkiewicz [1], Hughes et al. [2] and Beer [3]). Such a node-to-node model can
only be applied to problems in which relative sliding displacements of the two contact boundaries
are sufficiently small. Once significant non-linear contact deformations were introduced, methods
tended to switch to node-on-segment contact (Wriggers and Simo [4], Parisch [5], Papadopoulos
and Taylor [6] and Goncalves et al. [7]).
There are important links between the finite element contact problem and mathematical
programming techniques. Indeed, because the contact problems usually involve inequality
constraints (varying contact areas), the mathematics can be related to the method of variational
inequalities. Basically, two main constraint methods of solution have been employed in the finite
element solution of contact problems; the method of Lagrangian multipliers and the penalty
approach. In Lagrangian multipliers approach, the contact forces are taken as primary unknowns
and the non-penetration condition is enforced (Simo et al. [8], Chaudaray and Bathe [9] and Gallego
and Anza [10]). In penalty method, the penetration between two contacting boundaries is introduced
and the normal contact force is related to the penetration by a penalty parameter (Curnier and Alart
[11] and Peric and Owen [12]). The aim of this study is to present a model for simulation of
frictional contact problem using the extended finite element method based on the penalty approach.
The eXtended Finite Element Method (XFEM) is a numerical approach used to simulate the
discontinuity within the standard finite element framework. In this approach, the standard
displacement based approximation is enriched by incorporating discontinuous fields through a
partition of unity method (Melenk and Babuska [13]). The method was developed by Dolbow et al.
[14, 15] to model cracks, voids and inhomogeneities. This allows for the entire crack geometry to be
modeled independently of the mesh, and completely avoids the need to remesh as the crack grows.
A methodology that constructs the enriched approximation based on the interaction of the
discontinuous geometric features with the mesh was developed by Daux et al. [16] in modeling
crack discontinuities. A technique for modeling arbitrary discontinuities in finite elements was
presented by Belytschko et al. [17]. In this method, the discontinuous approximation was
constructed in terms of a signed distance functions and the level sets were used to update the
position of the discontinuities. An algorithm which couples the level set method with the extended
finite element method was proposed by Stolarska et al. [18]. They applied a discontinuous function
based on the Heaviside step function in modeling two-dimensional linear elastic crack-tip
displacement fields. Recently, a numerical technique was developed by Sukumar et al. [19] for
three-dimensional fatigue crack growth simulations. This technique couples the extended finite
element method to the fast marching method using the partition of unity method.
In the present paper, an extended finite element method (XFEM) is developed to simulate the
frictional behavior of contact problems. The classical finite element approximation is enriched by
employing additional terms based on the Heaviside step function. These terms depend on the
contact conditions between two surfaces and model the stress-strain relationship in contact area. As
the contact region has discontinuities in domain, special techniques are implemented in the XFEM.
The partition of unity method (PUM) is applied to discretize the contact area with triangular subelements whose Gauss points are used for integration of the domain of elements. The conditions
that describe frictional contact are formulated as a non-smooth constitutive law on the interface of
contact and the iterative scheme is implemented to solve the nonlinear boundary value problem.
Finally, numerical examples are presented to demonstrate the applicability of the extended finite
element method in modeling of frictional contact behavior.
on
( C ) = ( S ) I (T )
(1)
where g N is the gap between the two bodies, (x) and (y ) are the configuration mapping of the
slave and target bodies, ( S ) and (T ) are the slave and target body surfaces respectively,
( C ) denotes a surface where contact between the two bodies occurs and N is the unit outward
normal vector on the target surface. During the contact and sliding of the bodies, we define p N and
pT as the normal and tangential load acting on the point x , respectively. The contact conditions
may be expressed in the standard Kuhn-Tucker form as
gN 0
pN 0
pN g N = 0
(2)
which is best suited for a variational formulation. Consider that there is no gap between the two
bodies in the sliding contact problem (Figure 1), the normal displacement is assumed to be zero and
a tangential displacement is only considered, which consists of stick and slip decompositions and is
in principle the same as the decomposition of elastic and plastic behavior. Thus, the decomposition
of the tangential displacement at the contact surface can be given as
uT = uTe + uTp
(3)
where u T is the tangential part of the displacement described by uT = (I N N) u and, uTe and u Tp
are the elastic and plastic components of tangential displacement.
In relation (2), the kinetic constraint of impenetrability of two bodies can be satisfied as well as
the static condition of compressive normal load. To resolve the resulting unilateral contact problem
the Lagrange multiplier method, or the penalty method, are typically used. In the case of the
Lagrange multiplier method, however, a large number of additional unknown variables need to be
included owing to incorporation of p N as new variables. On the contrary, the penalty method needs
no additional variable, because the impenetrability condition is approximately satisfied (by
constraint via embedding very stiff springs on the contact surface). Consequently, the normal load
p N can be obtained from multiplication of the penalty factor k N and the displacement in the normal
direction u N . Similarly, the stick (or elastic) component of the tangential load may be obtained by
multiplying the penalty factor kT and elastic part of the displacement in the tangential direction uTe .
The penalty factors k N and kT can be considered as being the normal stiffness constant and shear
stiffness constant, respectively. Constitutive laws for the contact loads can now be summarized as
( )u
= (D ) u
p N = Def
pT
e
f T
(4)
(5)
where (D ef )N and (Def )T are the normal and tangential parts of the elastic modulus tensor for friction
defined as
(D )
(D )
e
f N
= k N (N N )
(6)
e
f T
= kT (I N N )
(7)
In order to perform the additive decomposition of displacement into adherence and slip, a slip
criterion must be introduced. To this end the slip surface F f is postulated in the contact stress space
on which slip will occur. The slip criterion is expressed based on the Coulomb law as
F f (p, w) = p T f ( p N , w) p N c f
= 0
< 0
(8)
where c f denotes the cohesion between two bodies and the Coulomb friction coefficient is defined
as f = tan f , with f denoting the tool friction angle.
The direction of slip is governed by an appropriate slip rule, which can be derived from the
gradient of a convex potential Z . If the potential Z is replaced by the slip criterion F f , the slip rule
becomes associated. Although in plasticity the flow rules associated with the standard criteria prove
realistic for relatively large classes of materials, the slip rule associated to the usual friction criterion
(8) is not acceptable. Indeed, the relative movement at the interface derived from the associated
potential Z = F f results in the creation of gaps (separation). Thus, in order to avoid the slave body
separation from the contacting surface, a non-associated flow rule is typically adopted (Curnier
[20]). Hence, the slip potential Z is introduced as a cylinder with radius PT for isotropic frictional
contact and the slip direction is defined as the outward normal to the slip potential Z .
Consequently, the plastic part of the tangential displacement u Tp , in equation (3), can be defined by
the definition of the slip rule as
duTp = d
Z
= dT
pT
(9)
where d is a constant expressing the collinearity of the slip increment with the outward normal to
the potential Z , and T =
pT
pT
The constitutive model for the contact problem is conveniently described by equations (29).
Following the standard arguments of elasto-plasticity, the continuum tangent tensor for the contact
problem with non-linear frictional evolution can be derived using the consistency condition as
(Khoei [22])
F (p, w)
F f (p, w)
F (p, w)
d p N + f
dw = 0
dpT + f
w
pT
p N
(10)
substituting d into the constitutive law, a linearized equation is obtained in the incremental form
as
dp = D epf du
(11)
D epf = k N (N N ) kT (I N N T T )
kT
1 f ( p N , w)
pN
kT
kN
where = 1 +
f ( p N , w)
kT
f
w
(T T )
f
1
f ( p N , w) + p N
pN
2 f
p N
w
(12)
(T N )
In equation (12), the first term of the continuum tangent tensor D epf indicates the stiffness in the
normal direction to the contact surface. The second term denotes the adhesion stiffness
perpendicular to the sliding direction on the contact surface. The third and forth terms indicate the
adhesion and slip stiffness with hardening, or softening, in the sliding direction, respectively. In the
case of frictional slip without hardening, or softening, equation (12) can be simplified as
D epf = k N (N N ) kT (I N N T T ) f k N (T N )
(13)
Evidently, the non-associative slip rule (9) results in non-symmetry of the slip modulus tensor,
which is defined by (13) under the conditions of frictional slip.
N ( x) u + N
i
j ( x)
f ( x) a j
for ni n and n j n g
(14)
The first term of above equation denotes the classical finite element approximation and the second
term indicates the enrichment function considered in XFEM. In this equation, u i is the classical
nodal displacement, a j the nodal degrees of freedom corresponding to the enrichment functions,
f ( x ) the enrichment function, and N ( x ) the standard shape function. In equation (14), n is the set
of all nodal points of domain, and n g the set of nodes of elements located on discontinuity, i.e.
n g = n j : n j n , j I c 0
(15)
In the above equation, j = supp(n j ) is the support of the nodal shape function N j ( x ) , which
consists of the union of all elements with n j as one of its vertices, or in other words the union of
elements in which N j ( x ) is non-zero.
It must be noted that the enrichment varies from node to node and many nodes require no
enrichment, which is an application of the partition of unity concept. Different techniques may be
used for the enrichment function; these functions are related to the type of discontinuity and its
influences on the form of solution. These techniques are based on the signed distance function,
branch function, Heaviside jump function, level set function, and etc.
The signed distance function is applicable to crack problem, which is discontinuous across the
crack line (Belytschko et al. [17]). The function can be viewed as an enrichment with a windowed
step function, where N ( x ) is the window function. The window function localizes the enrichment
so that the discrete equations will be sparse. For cracks which are not straight, a mapping is required
to align the near-tip discontinuities with the crack edges. In this case, a near-tip function, or branch
function, can be constructed in terms of the distance function, which enables the discontinuity to be
curved or piecewise linear (Dolbow et al. [15]). This function spans the near-tip asymptotic solution
for a crack, and gives very good accuracy for these problems. The level set method is a numerical
scheme developed by Sethian [23] for tracking the motion of interfaces. In this technique, the
interface is represented as the zero level set of a function of one higher dimension. Recently, the
technique of fast marching method, which was first introduced by Sethian [24], was coupled with
XFEM to model crack growth (Sukumar et al. [19] and Chopp and Sukumar [25]). The method
computes the crossing time map for a monotonically advancing front in an arbitrary number of
spatial dimensions.
: (u) d =
b u d +
t u d +
t u d
(16)
in which the last term represents the concept of energy, dissipated in the relative motion of the
contact surface. It is important that the displacement field of domain and the displacement on
contact surface be kinematically admissible. The interfacial constitutive law on the contact surface
is expressed by equation (8) that indicates the stress level at which relative slip motion occurs. The
goal is to obtain the stress and displacement fields on the contact surface which satisfy the
equilibrium and consistency conditions.
For an arbitrary contact displacement field, equation (14) can be rewritten as
u h ( x) =
N ( x) u + N
i
j ( x)H ( x) a j
for ni n and n j n g
(17)
where H ( x ) is the Heaviside jump function. In above relation, the contact surface is considered to
be a curve parameterized by the curvilinear coordinate s , as shown in Figure 3. Considering a point
x in the domain, we denote x * the closest point to x on the contact surface. At point x * , we
construct the tangential and normal vectors to the curve, e s and en , with the orientation of e n taken
such that e s e n = e z . The Heaviside jump function H ( x ) is then given by the sign of the scalar
product ( x x * ) en , in which the function H ( x ) takes the value of +1 above the contact surface,
and 1 below the contact surface, i.e.
+ 1
H ( x) =
1
if ( x x * ) e n 0
(18)
otherwise
On substituting the trial function of equation (17) into the weak form of equilibrium equation of
elasto-plasticity (16), and using the arbitrariness of nodal variations, the discrete system of
equations can be obtained as K d = f , where d is the vector of unknowns of u i and a j at the nodal
points, and K and f are the global stiffness matrix and external force vector, defined as
K uu
K ij = ijau
K ij
K ua
ij
,
aa
K ij
f i = {f iu
f ia }T
(19)
where
K
ij =
f i =
(Bi )T Dep (B j ) d
N i t d +
e
N i b d
e
( , = u , a)
( = u , a )
(20)
where Dep is the elasto-plastic constitutive matrix. In equation (20), N i N i is for a finite element
displacement degree of freedom, and N i N i H for an enriched degree of freedom. The matrices
Bi and B j include the shape function derivatives defined as
B ui
B ia
Ni, x
= 0
N i, y
N i, y
N i , x
( N i H ), x
= 0
( N i H ) , y
(21)
( N i H ), y
( N i H ) , x
For the elements cut by the contact surface, the standard Gauss quadrature points are insufficient
for numerical integration, and may not adequately integrate the discontinuous field. If the
integration of the discontinuous enrichment is indistinguishable from that of a constant function, the
system of equations may be rank deficient. Thus, it is necessary to modify the element quadrature
points to accurately evaluate the contribution to the weak form for both sides of the contact surface.
In what follows, we present the modifications made to the numerical integration scheme for
elements cut by a contact surface, which was applied by Daux et al. [16] in modeling of cracks with
multiple branches.
The discrete weak form is normally constructed with a loop over all elements, as the domain is
approximated by
= U e
(22)
e =1
where m is the number of elements and e is the element sub-domain. For elements located on
contact surface, an appropriate procedure is performed. For those elements, we divide the element
into triangular sub-domains s with boundaries aligned with the contact surface geometry, i.e.,
ms
e = U s
(23)
s =1
where ms denotes the number of sub-polygons of the element. The Gauss points of sub-triangles are
used for numerical integration across the contact surface, as shown in Figure 4. Different algorithms
may be applied to generate these sub-polygons, based on sub-triangles and sub-quadratics. In this
study, sub-triangles are implemented for numerical integration. It is essential to mention that these
sub-polygons only generated for numerical integration and no new degrees of freedom are added to
system. In the construction of the matrix equations, the element loop is replaced by a loop over the
sub-triangles for those elements cut by the contact surface.
D12
D22
(24)
In order to preserve the symmetry of the numerical formulation the off-diagonal term in D epf , which
represents the coupling between the normal and tangential stresses at the interface, is neglected, so
that its effect is brought into the formulation via residual 'pseudo loads'. In this way, the problem is
artificially decomposed into a pure contact in the normal direction and frictional resistance in the
tangential direction, which are linearized separately as
G f
Depf =
0
E f
(25)
In an incremental manner, the stress components are related to the strains through the material
property matrix D epf by = Depf . The stress vector at each Gauss point located in specified
distance from both sides of contact surface, i.e. the contact bond, is assumed to have only two
components; the normal stress n and shear stress , where = { , n }T and = { , n }T .
The material property matrix D epf needs to capture the details of the physical processes taking place
such as asperity contact, adhesion and the consequent 'stick-slip' behavior. Since there is no volume
change due to shearing strains, the shear and normal components of deformation are therefore
uncoupled.
The normal strains are measured only to monitor the normal stress by means of the linear
equation, i.e. n = E f n , where E f is chosen as an arbitrary large number for numerical
convenience. Note that only compressive normal stresses are allowed, i.e. n 0 . Since the
tangential micro-shifts due to adherence are negligible in comparison to the micro-slips due to
sliding, the adherence does not require a very precise numerical treatment. The incremental form of
the shear stress-strain relationship under adherence is defined as = G A , where G A is
proportional to the stick shear modulus of contact surface. The frictional shear force, however, is
limited by the slip criterion. The shear stress-strain relationship is presented in Figure 5, where
T = f n is considered to be a known state variable [22]. The frictional non-linearity is modeled
by an appropriate variation of G f which can be obtained from the shear stress-strain ( )
relationship for the stick and slip region (Figure 5). It must be noted that the shear modulus G f is
divided into two components, the first where the body has not moved but there is a rapid buildup in
load. Under this condition, a stick shear modulus G f = G A is derived directly from the slope of the
curve. On commencement of movement, the slip shear modulus condition G f = GB is now
appropriate and this is derived similarly.
Finally, the stiffness matrix of equation (20) can be evaluated at each Gauss point located on
contact bond by
(K f )
ij =
(Bi )T D epf (B j ) d
( , = u , a )
(26)
and
G f = GA
(27)
Gf =
(28)
which the nodal points of relevant elements are enriched with the Heaviside jump function. For
those elements which are intersected by the contact surface, the concept of the partition of unity is
used to generate the sub-triangles. The Gauss points of sub-triangles are then employed to evaluate
the tangent stiffness matrix of friction. The numerical simulation is performed using 200 six-noded
triangular elements in FEM mesh and 361 four-noded rectangular elements in XFEM mesh. The
convergence tolerance is set to 10 4 and the analysis was performed using 20 increments.
In Figures 8(a)-(d), the distribution of shear stress contours at two different values of shear
modulus, i.e. G f = 1 108 and 1 1010 Kg m 2 are presented for both XFEM and FEM techniques.
Remarkable agreements can be observed between two different methods. Figure 9 presents the
variation of horizontal displacement of top edge with the shear modulus of contact surface using
two different approaches. A convergence study is conducted in Figure 10 using various values of
contact bond width, i.e. the region whose Gauss points have similar property as contact surface. In
this figure, the variation of uniform shear loading with horizontal displacement of top edge is
plotted at three values of contact bond width, i.e. 0.5, 0.7, 1.0. It can be observed from the figure
that the contact bond width of 1.0 gives a remarkable improvement in accuracy of horizontal
sliding.
the contact surface due to pull-out force is presented in Figure 13(a). This result can be compared
with those obtained by Lei [26] using the interface element technique in Figure 13(b).
powder-die friction. Due to large movements of punches, the updated Lagrangian description is
incorporated into the XFEM formulation.
The initial geometry of uncompacted powder in its position before compaction is presented in
Figure 18, where h1 = 100 mm , h2 = 50 mm , w1 = 300 mm and w2 = 30 mm . The material properties
chosen for powder are E = 2 105 Kg m 2 and = 0.3 . The rigid die-wall is assumed to be elastic
with the Youngs modulus of 2 1010 Kg m 2 and Poissons coefficient of 0.3. The powder-tool
friction is modeled by the Coulomb law with c f = 0 and f = 0.3 . On the virtue of symmetry, the
process is modeled for half of specimen, as shown in Figure 19. In order to present the accuracy of
the XFEM technique, a comparison is performed with the finite element solution for a mesh with
approximately the same number of unknowns. The numerical simulation is performed using 240
six-noded triangular elements in FEM mesh and 400 four-noded rectangular elements in XFEM
mesh. The convergence tolerance is set to 10 4 and the analysis was carried out using 20
increments.
The final deformed mesh for both techniques is presented in Figure 20 at the top punch
movement of 34 mm . In Figures 21(a)-(d), the horizontal and vertical displacements contours are
presented for both XFEM and FEM approaches. Also plotted in Figures 22(a)-(f) are the stress
distribution contours of x , y and t xy for two different techniques. Remarkable agreements can be
observed between two different methods. The performance of XFEM in the case of large sliding is
presented in Figure 23 by a detailed plot of traction on the contact surface at the volume reduction
of 34 percents. This figure clearly presents the variation of the shear and normal stresses on the
surface of rigid die-wall with the distance from bottom on the contact surface between the die and
powder. As can be observed, the values of shear and normal stresses are approximately zero on the
portion of the boundary that is free at the top punch movement of 34 mm , in which the two bodies
have already slide past each other. The evolution of top punch vertical reaction force with its
vertical displacement is depicted in Figure 24(a). Complete agreements can be observed between
two methods. It must be noted that on using the finite element simulation, no further advance of the
punch movement is obtained, however the volume reduction of 90 percents can be observed with
the same computational algorithm in XFEM technique, as plotted in Figure 24(b). This example
adequately presents the applicability of XFEM technique in modeling of powder-tool friction in
compaction pressing of powder.
6. Conclusion
In the present paper, a method was developed based on the extended finite element method in
modeling the discontinuity caused by frictional contact. The classical finite element approximation
was enriched by employing additional terms based on the Heaviside step function. The partition of
unity method was applied to discretize the contact area with triangular sub-elements whose Gauss
points were used for integration of the domain of elements. For the elements cut by contact surface,
the integration of stiffness matrix was performed by employing the material property matrix of
contact surface at the integration points located on contact bond. The frictional contact behavior was
formulated by a non-smooth constitutive law on the interface of contact and the iterative scheme
was implemented to resolve the nonlinear boundary value problem. Finally, numerical examples
were presented to demonstrate the accuracy and capability of the XFEM in modeling the frictional
contact behavior. It is shown how the XFEM technique can be effectively used to model 2D contact
problems. In a later work, the implementation of XFEM technique will be presented in modeling 3D
contact problems.
Acknowledgement
Parts of this research were conducted by the first author on his sabbatical leave at the Laboratorie de
Mecanique des Solides, Ecole Polytechnique, France (2002). The author would like to appreciate the
collaboration and guidance of Professor Claude Stolz.
References
1. A. Fracavilla and O.C. Zienkiewicz, A note on the numerical computation of elastic contact problems,
Int. J. Num. Meth. Eng., 9 (1975) 913-924.
2. T.J.R. Hughes, R.L. Taylor, J.L. Sackman, A. Curnier and W. Kanoknukulchai, A finite element method
for a class of contact-impact problems, Comp. Mech. Appl. Meth. Eng., 8 (1976) 249-276.
3. G. Beer, An isoparametric joint/interface element for finite element analysis, Int. J. Num. Meth. Eng, 21
(1985) 585-600.
4. P. Wriggers and J.C. Simo, A note on tangent stiffness for fully nonlinear contact problems, Comm.
Appl. Num. Meth., 1 (1985) 199-203.
5. H. Parisch, A consistent tangent stiffness matrix for three-dimensional contact analysis, Int. J. Num.
Meth. Eng., 28 (1989) 1803-1812.
6. P. Papadopoulos and R.L. Taylor, A mixed formulation for the finite element solution of contact
problems, Comp. Mech. Appl. Meth. Eng., 94 (1992) 373-389.
7. J.P.M. Goncalves, M.F.S.F. de Moura, P.M.S.T. de Castro and A.T. Marques, Interface element
including point-to-surface constraints for three-dimensional problems with damage propagation, Eng.
Comput., 17 (2000) 28-47.
8. J.C. Simo, P. Wriggers and R.L. Taylor, A perturbed Lagrangian formulation for the finite element
solution of contact problems, Comp. Meth. Appl. Mech. Eng., 51 (1985) 163-180.
9. A.B. Chaudaray and K.J. Bathe, A solution method for static and dynamic analysis of contact problems
with friction, Comput. Struct., 24 (1986) 855-873.
10. F.J. Gallego and J.J. Anza, A mixed finite element for the elastic contact problem, Int. J. Num. Meth.
Eng., 28 (1989) 1249-1264.
11. A. Curnier and P. Alart, Generalisation of Newton type methods to contact problems with friction, J. de
Mecanique Theorique et Appliquee, 7 (1988) 67-82.
12. D. Peric and D.R.J. Owen, Computational model for 3D contact problems with friction based on the
penalty method, Int. J. Num. Meth. Eng., 35 (1992) 1289-1309.
13. J.M. Melenk and I. Babuska, The partition of unity finite element method: basic theory and applications,
Comp. Meth. Appl. Mech. Eng., 139 (1996) 289-314.
14. J. Dolbow, An extended finite element method with discontinuous enrichment for applied mechanics,
Ph.D. Thesis, Northwestern University, 1999.
15. J. Dolbow, N. Moes and T. Belytschko, Discontinuous enrichment in finite elements with a partition of
unity method, Finite Elem. Anal. Des., 36 (2000) 235260.
16. C. Daux, N. Moes, J. Dolbow, N. Sukumar and T. Belytschko, Arbitrary branched and intersecting
cracks with the extended finite element method, Int. J. Num. Meth. Eng., 48 (2000) 17411760.
17. T. Belytschko, N. Moes, S. Usui and C. Parimi, Arbitrary discontinuities in finite elements, Int. J. Num.
Meth. Eng., 50 (2001) 9931013.
18. M. Stolarska, D.L. Chopp, N. Moes and T. Belytschko, Modeling crack growth by level sets in the
extended finite element method, Int. J. Num. Meth. Eng., 51 (2001) 943960.
19. N. Sukumar, D.L. Chopp and B. Moran, Extended finite element method and fast marching method for
three-dimensional fatigue crack propagation, Eng. Fracture Mech., 70 (2003) 29-48.
20. A. Curnier, A theory of friction, Int. J. Solids Struct., 20 (1984) 637-647.
21. T. Rodic and D.R.J. Owen, A plasticity theory of friction and joint elements, Computational Plasticity II:
Models, Software and Applications, D.R.J. Owen et al. (Eds.), Pineridge Press, 1043-1062, 1989.
22. A.R. Khoei, Computational Plasticity in Powder Forming Processes, Elsevier (UK), 2005.
23. J.A. Sethian, A marching level set method for monotonically advancing fronts, Proc. Nat. Acad. Sci., 93
(1996) 1591-1595.
24. J.A. Sethian, Level Set Methods and Fast Marching Methods: Evolving Interfaces in Computational
Geometry, Fluid Mechanics, Computer Vision, and Material Science, Cambridge University Press, 1999.
25. D.L. Chopp and N. Sukumar, Fatigue crack propagation of multiple coplanar cracks with the coupled
extended finite element/fast marching method, Int. J. Eng. Science, 41 (2003) 845-869.
26. X. Lei, Contact friction analysis with a simple interface element, Comp. Meth. Appl. Mech. Eng., 190
(2001) 1955-1965.
27. A.R. Khoei and R.W. Lewis, Adaptive finite element remeshing in a large deformation analysis of metal
powder forming, Int. J. Num. Meth. Eng., 45 (1999) 801-820.
Figure Legend:
Figure 1.
Figure 2.
Figure 3.
Illustration of normal and tangential coordinates for a contact surface in the case of
jump function H ( x ) = 1 ; x * is the closest point to x on the contact surface
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Sliding of two bodies; The shear stress contours corresponding to; a) The FEM mesh
with G f = 1 1010 , b) The XFEM mesh with G f = 1 1010 , c) The FEM mesh with
G f = 1 108 , d) The XFEM mesh with G f = 1 108
Figure 9.
Sliding of two bodies; The variation of horizontal displacement of top edge with the
shear modulus of contact surface using the FEM and XFEM techniques
Figure 10.
Sliding of two bodies; The variation of uniform shear loading with horizontal
displacement of top edge at three values of contact bond width
Figure 11.
Figure 12.
Figure 13.
The pull-out of pile; The shear stress distribution of pile along the contact surface; a)
The XFEM model, b) Lei [24]
Figure 14.
Figure 15.
Torsion of FRP rod with concrete; a) The FEM mesh, b) The XFEM mesh
Figure 16.
Torsion of FRP rod with concrete; a) The stress x contour using FEM analysis, b)
The stress x contour using XFEM analysis, c) The stress y contour using FEM
analysis, d) The stress y contour using XFEM analysis, e) The shear stress xy
contour using FEM analysis, f) The shear stress xy contour using XFEM analysis
Figure 17.
Torsion of FRP rod with concrete; a) The horizontal displacement contour using
FEM analysis, b) The horizontal displacement contour using XFEM analysis, c) The
vertical displacement contour using FEM analysis, d) The vertical displacement
contour using XFEM analysis
Figure 18.
Figure 19.
Figure 20.
Figure 21.
Figure 22.
The compaction process; a) The stress x contour using FEM analysis, b) The stress
x contour using XFEM analysis, c) The stress y contour using FEM analysis, d)
The stress y contour using XFEM analysis, e) The shear stress xy contour using
FEM analysis, f) The shear stress xy contour using XFEM analysis
Figure 23.
The compaction process; The variation of the shear and normal stresses on the
surface of rigid die-wall with the distance from bottom of the contact surface at the
top punch movement of 34 mm using the XFEM technique
Figure 24.
The compaction process; The evolution of top punch vertical reaction force with its
vertical displacement; a) A comparison between two different approaches at the top
punch movement of 34 mm , b) The XFEM technique at the top punch movement of
90 mm
(a)
(b)
(c)
Figure 2. Modeling of contact surface between two bodies a) Problem definition, b) The FE mesh
which conforms to the geometry of contact together with employed contact elements, c) A uniform
mesh in which the circled nodes have additional degrees of freedom and enrichment functions
Figure 3. Illustration of normal and tangential coordinates for a contact surface in the case of jump
function H ( x ) = 1 ; x * is the closest point to x on the contact surface
Figure 4. The sub-triangles associated with elements cut by contact surface in XFEM
(a)
(b)
Figure 7. Sliding of two bodies, a) The FEM mesh, b) The XFEM mesh
(a)
(b)
(c)
(d)
Figure 8. Sliding of two bodies; The shear stress contours corresponding to; a) The FEM mesh with
G f = 1 1010 , b) The XFEM mesh with G f = 1 1010 , c) The FEM mesh with G f = 1 108 , d) The
XFEM mesh with G f = 1 108
Displacement (Cm)
1.8
1.6
1.4
1.2
1
0.8
1.0E+10
1e6'
1.0E+09
1e5'
1.0E+08
1e4'
Figure 9. Sliding of two bodies; The variation of horizontal displacement of top edge with the shear
modulus of contact surface using the FEM and XFEM techniques
2500
XFEM (CBW = 0.5)
XFEM (CBW = 0.7)
XFEM (CBW = 1.0)
FEM
2000
1500
1000
500
0.2
0.4
0.6
0.8
1.2
Figure 10. Sliding of two bodies; The variation of uniform shear loading with horizontal
displacement of top edge at three values of contact bond width
(a)
(b)
Figure 13. The pull-out of pile; The shear stress distribution of pile along the contact surface;
a) The XFEM model, b) Lei [24]
(a)
(b)
Figure 15. Torsion of FRP rod with concrete; a) The FEM mesh, b) The XFEM mesh
(a)
(b)
(c)
(d)
(e)
(f)
Figure 16. Torsion of FRP rod with concrete; a) The stress x contour using FEM analysis, b) The
stress x contour using XFEM analysis, c) The stress y contour using FEM analysis, d) The
stress y contour using XFEM analysis, e) The shear stress xy contour using FEM analysis, f) The
shear stress xy contour using XFEM analysis
(a)
(b)
(c)
(d)
Figure 17. Torsion of FRP rod with concrete; a) The horizontal displacement contour using FEM
analysis, b) The horizontal displacement contour using XFEM analysis, c) The vertical
displacement contour using FEM analysis, d) The vertical displacement contour using XFEM
analysis
(a)
(b)
Figure 19. The compaction process; a) Initial FEM mesh, b) Initial XFEM mesh
(a)
(b)
Figure 20. The compaction process; a) Deformed FEM mesh, b) Deformed XFEM mesh
(a)
(b)
(c)
(d)
Figure 21. The compaction process; a) The horizontal displacement contour using FEM analysis, b)
The horizontal displacement contour using XFEM analysis, c) The vertical displacement contour
using FEM analysis, d) The vertical displacement contour using XFEM analysis
(a)
(b)
(c)
(d)
(e)
(f)
Figure 22. The compaction process; a) The stress x contour using FEM analysis, b) The stress x
contour using XFEM analysis, c) The stress y contour using FEM analysis, d) The stress y
contour using XFEM analysis, e) The shear stress xy contour using FEM analysis, f) The shear
stress xy contour using XFEM analysis
100
-5000
80
Normal stress
60
40
Shear stress
20
5000
15000
25000
35000
45000
55000
Figure 23. The compaction process; The variation of the shear and normal stresses on the surface of
rigid die-wall with the distance from bottom of the contact surface at the top punch movement of
34 mm using the XFEM technique
12
8
6
4
2
0
0.00
0.0
(a)
10
0.30
0.6
0.60
1.2
0.90
1.8
1.20
2.4
1.50
3.0
1.80
3.6
45
40
35
30
25
20
15
10
5
0
(b)
0
0.0
0.01
2.0
0.02
4.0
0.03
6.0
0.04
8.0
0.05
10.0