Documentos de Académico
Documentos de Profesional
Documentos de Cultura
forming
Authors
J.Hola, M.V. Cid Alfarob, M.B. de Rooijc, T. Meindersd
a
Please Note
Care has been taken to ensure that the information herein is accurate, but Tata Steel and its subsidiary companies do not accept
responsibility for errors or for information which is found to be misleading. Suggestions for or descriptions of the end use or
applications of products or methods of working are for information only and Tata Steel and its subsidiaries accept no liability in respect
thereof. Before using products supplied or manufactured by Tata Steel the customer should satisfy themselves of their suitability
All drawings, calculations and advisory services are provided subject to Tata Steel Standard Conditions available on request.
Abstract
The Coulomb friction model is frequently used for sheet metal forming simulations. This model incorporates a constant coefficient
of friction and does not take the influence of important parameters such as contact pressure or deformation of the sheet material
into account. This article presents a more advanced friction model for large-scale forming simulations based on the surface changes
on the micro-scale. When two surfaces are in contact, the surface texture of a material changes due to the combination of normal
loading and stretching. Consequently, shear stresses between contacting surfaces, caused by the adhesion and ploughing effect
between contacting asperities, will change when the surface texture changes. A friction model has been developed which accounts
for the change of the surface texture on the micro-scale and its influence on the friction behavior on the macro-scale. This friction
model has been implemented in a finite element code and applied to a full-scale sheet metal forming simulation. Results showed a
realistic distribution of the coefficient of friction depending on the local process conditions.
Keywords:
friction modeling, friction mechanisms, asperity contact, flattening, real contact area, ploughing, adhesion
1. Introduction
The automotive industry makes extensive use of Finite Element (FE) software for formability analyses to reduce the cost
and lead time of new vehicle programs. In this respect, FE
analysis serves as a stepping stone to optimize manufacturing
processes. However, an accurate forming analysis of an automotive part can only be made if, among others, the material behavior and friction conditions are modeled accurately. For material models, significant improvements have been made over
recent decades. However, in the majority of simulations a simple Coulomb friction model is still used. This model does not
incorporate the influence of important parameters on the contact behavior, such as pressure, punch speed or deformation of
the sheet material. Consequently, even using the latest material models, it is still cumbersome to predict the draw-in and
springback of a blank during the forming process correctly.
To better understand contact and friction conditions during
lubricated sheet metal forming (SMF) processes, experimental and theoretical studies have been performed. At the microscopic level, friction is due to adhesion between contacting asperities [1, 2], the ploughing effect between asperities [1, 2]
and the appearance of hydrodynamic friction stresses [3, 4].
Ploughing effects between asperities and adhesion effects between boundary layers are the main factors causing friction in
the boundary lubrication regime. If the contact pressure is carried by the asperities and lubricant flow - as in the mixed lubri Corresponding
cation regime - or fully carried by the lubricant - as in the hydrodynamic lubrication regime - hydrodynamic shear stresses will
contribute or even predominate. This article will focus on the
two friction mechanisms present in the boundary layer regime:
ploughing and adhesion.
Wilson [1] developed a model which treated the effect of adhesion and ploughing separately. A more advanced model was
developed by Challen & Oxley [2] which takes the combined
effect of ploughing and adhesion on the coefficient of friction
into account. Challen & Oxley performed a slip-line analysis on
the deformation of a soft flat material by a hard wedge-shaped
asperity and derived expressions for the coefficient of friction
and wear rates. Westeneng [5] extended the model of Challen
& Oxley to describe friction conditions between multiple tool
asperities and a flat workpiece material. Their model considers the flattened plateaus of the workpiece asperities as soft and
perfectly flat, and the surface texture of the tool as hard and
rough.
The amount of ploughing and adhesion depends on the real
area of contact. Hence, the coefficient of friction will change
if the real area of contact changes. The real area of contact
depends on the various flattening and roughening mechanisms
of the deforming asperities. The three dominating flattening
mechanisms during SMF processes are: (1) flattening due to
normal loading [6]; (2) flattening due to stretching [7, 8]; and
(3) flattening due to sliding [9]. Flattening increases the real
area of contact, resulting in a higher coefficient of friction.
Roughening of asperities, observed during stretching of the deformed material [10], tends to decrease the real area of contact
November 15, 2010
work-hardening parameters and are able to describe the interaction between asperities.
A further increase of the real area of contact could occur if,
during normal loading, a bulk strain is applied to the material.
The effective hardness of the asperities can be largely reduced
if a bulk strain is present in the underlying material [7]. Wilson & Sheu [7] developed an analytical upperbound model to
describe the flattening behavior using wedge-shaped asperities
with a constant angle, Figure 1a. The length of the asperities
is much greater than the width of the asperities. Therefore, a
plane-strain state transverse to the asperities (x-direction) and a
plane-stress state in the direction of the asperities (y-direction)
is assumed since the stress in this direction might be neglected.
The semi-empirical relation of Wilson & Sheu provides a relation between the effective hardness, the real area of contact
and a non-dimensional strain rate. Sutcliffe [8] extended the
model of Wilson & Sheu to describe a plane-strain situation in
the direction of the asperities (strain in y-direction equals zero,
Figure 1b). A slip-line analysis is performed to describe the
flattening of transverse wedge-shaped asperities. Westeneng
[5] developed a strain model which describes the influence of
strain on a surface geometry using arbitrary shaped asperities.
His method is based on volume and energy conservation laws
and assumes that the crushed asperities cause a constant rise of
the non-contacting asperities. The model is applicable to both
plane-strain and plane-stress situations, depending on the definition of the non-dimensional strain rate [5].
In this article, a friction model is proposed which includes
various friction mechanisms: flattening due to normal loading,
flattening due to straining, ploughing and adhesion. Existing
models have been used to describe these mechanisms:
- To describe the flattening behavior of asperities due to normal loading the contact model of Westeneng [5] has been
used. His contact model include flattening parameters
which are not included in other loading models. Therefore, the model of Westeneng will likely have better predicting capabilities in describing the flattening behavior of
FN
FN
y
y , y = 0
y = 0
y
x
x
x = 0
(a)
(b)
Figure 1: Representation plane-stress model of Wilson and Sheu (a) and plane-strain model of Sutcliffe (b)
2
160
Real distribution
133
107
(z)(m1 )
z(m)
Surface profile
80
-2
53.3
-4
26.7
-6
0.8
2.4
1.6
3.2
-4
-2.4
0.8
-0.8
x(mm)
2.4
z(m)
(b)
(a)
Figure 3: Surface profile (a) and corresponding surface height distribution (b)
Fourier series makes it possible to describe non-smooth asymmetric distribution functions from which the accuracy of the
evaluation depends on the number of expansions used.
The results discussed in this article are obtained by evaluating the surface height distribution functions (z) by a half range
sine Fourier function [16], given by:
(z) =
X
n=1
bn sin
n
z
L
rough surface (z), the uniform rise of the non-contacting surface U (based on volume conservation) and the separation between the tool surface and the mean plane of the asperities of
the rough surface d.
2.3.1. Flattening due to normal loading
Using the normalized surface height distribution (z), the
amount of flattening of the contacting asperities d and the rise
of the non-contacting asperities U can be calculated based on
energy and volume conservation. Contact between a flat hard
smooth surface and a soft rough surface is assumed without
sliding and bulk deformation. Only plastic deformation of asperities is assumed without work-hardening effects.
The amount of external energy must equal the internal energy in order to account for energy conservation. The amount
of external energy is described by the energy needed to indent
contacting asperities. The internal energy is described by the
energy absorbed by the indented asperities and the energy required to lift up the non-contacting asperities. The indentation variables used to describe Westenengs model are depicted
in Figure 5. A distinction is made between asperities in contact with the indenter, asperities which will come into contact
due to the rise of asperities and asperities which will not come
into contact with the indenter. The amount of indentation is
described by the variable z while the rise of asperities is described by the variable u. The number of asperities in contact
with the indenter is indicated by the counter N with corresponding indentation heights of zi (i = 1, 2, ..., N). The number of
asperities coming into contact with the indenter due to a rise of
non-contacting asperities is described by the counter N with
indentation heights of z j ( j = 1, 2, ..., N ). Hence, the total
number of asperities in contact with the indenter after applying
the normal load equals N + N . Asperities which will not come
into contact during the load step are indicated by the counter
N with corresponding rising heights of ul (l = 1, 2, ..., N ).
(1)
with:
2
bn =
L
ZL
0
f (z) sin
n
z
L
(2)
in which n represents the number of expansions, L the evaluation domain and f (z) the discrete form of the surface height
distribution. In Figure 3b, the discrete surface height distribution from the workpiece material (Figure 3a) is evaluated by a
Fourier function using 15 expansions.
2.3. Flattening mechanisms
Two flattening mechanisms have been implemented in the
friction model to calculate the real area of contact of the workpiece: flattening due to normal loading and flattening due to
stretching. The models of Westeneng are used for this purpose
[5]. Westeneng assumed the tool as rigid and perfectly flat,
which indents into a soft and rough workpiece material. This
assumption is valid since the difference in hardness and length
scales between the tool and workpiece material is significant in
the case of sheet metal forming processes. Westeneng modeled
the asperities of the rough surface by bars which can represent
arbitrarily shaped asperities, Figure 4. Westeneng introduced
3 stochastic variables as presented in Figure 4: The normalized surface height distribution function of the asperities of the
4
Rise non-contacting
Indented asperities
Surface U
Tool surface
d
Mean plane
(z)
Workpiece
asperities
U
N
X
F Ni zi +
i=1
N
X
F N j z j =
j=1
N+N
X
F Nk zk
(3)
k=1
dk
Wext = 1 F N z
with
1 =
Wintab =
N+N
X
N+N
X
HAzk
(6)
k=1
F Nk zk
k=1
(4)
dk F N
z
dk the maximum
In which F N represents the total force and z
indentation height. 1 is called the energy factor since it is influenced by the amount of external energy required to indent
the rough surface.
F Ni
(5)
N
N
X
X
ul HA
(8)
Wintri = u j HA +
N
zi
j=1
z j
u j
V
l=1
ul
Wintri = H
N
X
i=1
zi A
(9)
dk
Wintri = 3 H (Ar N A) z
with
3 =
N
X
zi
i=1
dk
N z
(10)
N
X
Anom
3 = 2 = 1
(12)
F N (z) (z) dz
N+N = M
U (1 ) =
zk d
zk d + U
(15)
Zd
Z
Pnom
H
dU
(18)
z1
(z d) (z) dz
(19)
(14)
for zk > d
for d U zk < d
(d z) (z) dz
(13)
(z) dz = M
Zd
dU
and:
(
Zz1
dU
zk =
(z d) (z) dz =
Anom
dU
(17)
u j A
j=1
FN = M
N
X
| {z } | {z }
(11)
ul A +
l=1
| {z }
zi A =
i=1
N
X
d S dS dtS
d S
=
d
dS dtS d
6
(20)
(21)
(22)
US (1 S ) =
dtS
1
=
d
(23)
(24)
l
va + vb
(25)
2 QS
(26)
(27)
It is assumed that the fraction of the real contact area for one
asperity S equals the total fraction of contact area S . Therefore, the stochastic form of the real contact area (Equation 28)
can be used to solve the differential equation in Equation 27:
S =
(z) dz
(28)
dS US
dS
d
=
d (US dS ) d (US dS )
(z) dz
dS US
(29)
= s
(32)
= = arctan p
2
(2t )
(33)
= (dS US )
Substituting Equation 29 into Equation 27 yields:
l
d S
= (dS US )
d
E
(31)
(z dS ) (z) dz
dS US
The second term, which represents the velocity of the indenting asperity, is determined by the downward velocity of the indenting asperity va and the upward velocity of the rising asperities vb , Equation 23. The third term represents the strain in the
bulk material of the asperities and can be written as:
dS
= va + vb
dtS
(30)
7
Tool surface
Mean plane
F Nasp
t
t (s)
r
Workpiece surface
(a)
(b)
Boundary layer
(c)
Fw = t S Anom
(34)
sin
A = 1 + + arccos fC 2 2 arcsin p
2
1 fC
(35)
(40)
(36)
Zsmax
t (s) ds Pnom Anom
0 = HS Anom
(41)
The integral in Equation 41 represents the fraction of the nominal contact area of the tool penetrating into the workpiece material.
If the shear stresses are known from Equation 40 the coefficient of friction can finally be obtained by:
Assuming that only half of the contacting area carries the load
during ploughing and that only small indentations takes place
( << t ), F Nasp can be written as:
F Nasp = t H
Fwasp t (s) ds
with:
F Nasp = AH = 2t 2 H
Zsmax
Fw
FN
(42)
3. Implementation
The friction model described in Section 2.1 has been implemented into the in-house implicit FE code Dieka, developed at
the University of Twente. The friction model is called for every
node in contact during a FE simulation. If a node is in contact,
the nominal contact pressure and strain in the bulk material is
(39)
From which the friction force Fwasp acting on one asperity can
be calculated.
8
Pnom
0 = 1 + (z) dz
H
0=
(z d + U) (z) dz U
dU
, d and U
2
f2 (S )
Hef f
E=
f1 (S )
called from the source code. The discrete surface height distribution is described by a continuous function using the half
range Fourier serie (Section 2.2). Then, the fraction of real
contact area due to normal loading and stretching, shear stresses
due to ploughing and adhesion and the coefficient of friction are
being calculated by the equations described in Section 2. This
section describes the implementation of these equations into the
FE code in more detail.
E=
(43)
z =
=
zd
zd+U
for z > d
for d U z < d
Pnom
S k
(45)
1
0.184 + 1.21 exp (1.47)
(46)
with:
with:
(
Hef f =
with
(44)
Hef f
(1 S )
4
(47)
dS
l
= (dS US )
d
E
U, d and from
load model
Pnom and in
S = S + dS
0=
Z
0=
(z) dz S
dS US
(z dS US ) (z) dz US
dS US
if + d < in
dS and US
Figure 8: Calculation scheme strain step, Pnom and in are coming in from the FE code
if the real area of contact is known from the two flattening
mechanisms. Shear stresses due to ploughing and adhesion effects between asperities and the coefficient of friction can be
calculated if the amount of indentation is known, as presented
in Figure 9.
An expression for the shear factor fC is necessary (Equation
34) in order to solve the system of Equations presented in Figure 9. fC is defined as /k with describing the friction force
in the boundary layer and k the shear strength of the softer material. The model of Timsit & Pelow [5, 18] has been imple-
During ploughing, the contact pressure p equals the effective hardness Hef f of the softer material since ideal plasticity
is assumed. The shear strength
k is related to the hardness H
Zsmax
0 = Hef f S Anom
t (s) ds Pnom Anom
Fwasp t (s) ds
(49)
Fw = t S Anom
= 20.47Hef0.19
f
k
Timsit & Pelow performed experiments to obtain an empirical relation between the normal stress and the shear strength of
a stearic acid film deposited on aluminum. The relation is applicable for contact pressures in between 70MPa and 740MPa,
which are likely to occur on a micro scale during deep drawing
processes.
Zsmax
(48)
Fw
FN
Tool
Tool
Workpiece
Workpiece
(VM stresses)
(VM stresses)
(a)
(b)
will be lower due to work-hardening effects, which in turn result in a lower amount of indentation and real area of contact
(Figure 11).
Combined normal loading and stretching the underlying bulk
material decreases the effective hardness [7]. A lower hardness
results in an increase of the real area of contact. Both the analytical and the FE results of analysis 2, where a rough surface has
been indented by a nominal load and a bulk strain has been applied to the underlying material, are presented in Figure 12. For
the FE solution, only the development of the real area of contact is shown, since the deformation of the asperities is difficult
to separate from the deformation of the underlying bulk material. It can be concluded from Figure 12b that work-hardening
effects have a large influence on the flattening behavior of the
asperities. A difference of 20% in the real area of contact is
obtained at the end of the simulation between the results of the
elastic ideal-plastic and the elastic nonlinear-plastic simulation.
The non-dimensional strain rate E, used in the analytical
model, can be described by the definition given by Wilson &
Sheu (Equation 45) or the definition of Sutcliffe (Equation 46).
The results of the FE simulation presented in Figure 12b are
based on a plane-strain deformation mode in the longitudinal
direction of the asperities. Therefore, the definition proposed
by Sutcliffe has been used. The density of workpiece asperities (in mm2 ) is an unknown parameter in the analytical strain
model. Various methodologies exist to extrude the asperity density from the surface profile. Results obtained by these methodologies are highly dependent on the chosen method and the
resolution of the used roughness measurement device. Future
work is planned to determine this parameter using the most suitable method. Until then, the asperity density will be taken as an
unknown parameter. Calculations have been performed using
realistic values for the asperity density for DC04 to show the
importance of this parameter, see Figure 12. From Figure 12, it
can be concluded that the asperity density of the workpiece has
a significant influence on the development of the real area of
11
2.6
0.5
FEM: elastic ideal-plastic
= 0.5
Indentation (m)
2.2
= 1.0
= 0.0
1.7
1.3
0.87
FEM: elastic ideal-plastic
FEM: elastic nonlinear-plastic
0.43
0.42
0.33
= 0.0
0.25
= 0.5
= 1.0
0.17
0.083
Analytical solution
0
20
40
80
60
100
20
40
80
60
100
(a)
(b)
Figure 11: Results analysis 1: Amount of indentation rough surface (a) and development of real area of contact (b)
5000 asp/mm
2.5
Indentation workpiece (m)
0.6
3000 asp/mm2
9000 asp/mm2
2
1.5
1
0.5
3000 asp/mm2
0.4
5000 asp/mm2
9000 asp/mm2
0.3
0.2
0.1
0.5
Analytical
0
0.02
0.04
0.06
0.08
0.1
0.02
0.04
0.06
0.08
0.1
Strain (-)
Strain (-)
(a)
(b)
Figure 12: Results analysis 2: Amount of indentation rough surface (a) and development of real area of contact (b)
contact. The amount of indentation of the workpiece asperities
will be lower if a higher value of the asperity density is used,
Figure 12a. Consequently, a lower amount of indentation results in a lower amount of real area of contact, Figure 12b. The
trend of the graphs corresponds well to the flattening behavior
obtained by the FE simulations. Using an asperity density of
5000 asp/mm2 it is possible to describe the results of the elastic
ideal-plastic FE solution (which has comparable material characteristics) precisely.
5. Application
The cross-die product is a test piece designed at Renault
which approximates process conditions of complex automotive
parts. The cross-die product is used to test the numerical performance of the developed friction model in a large-scale FE
simulation. In this article, the focus is on the numerical performance and feasibility of the friction model that has been developed. To validate this model, an experimental test procedure
needs to be developed and executed.
Due to symmetry of the cross-die product only a quarter of
the workpiece was modeled. The workpiece was meshed by
9000 triangular discrete Kirchhoff shell elements using 3 integration points in plane and 5 integration points in thickness
12
region D is compressed which causes thickening of the material and region E is stretched over the die radius. On the other
hand, low values of the coefficient of friction can be observed
in low-strain regions. Overall it can be concluded that the distribution of the coefficients of friction lies within the range of
expectation. The increase in calculation time is also promising. An increase of 60% for the first simulation and 200% for
the second simulation was obtained compared to the calculation
time required to perform a Coulomb based FE simulation of the
cross-die product.
During the implementation of the friction model into FE software some important assumptions had to be made:
- Full recovery of asperities is assumed during unloading
of the workpiece rough surface. The amount of recovery
of indented asperities is described by the amount of elastic spring-back in reality. Due to elastic spring-back, the
amount of recovery will be smaller than in case of full recovery. This will result a higher real area of contact at
lower loads. Consequently, the coefficient of friction will
be smaller due to a smaller amount of indentation of tool
asperities into the softer workpiece asperities. A realistic
unloading model is required to describe this effect.
- A definition of the non-dimensional strain rate is required
to calculate the amount of flattening due to bulk straining.
Various definitions exist to describe this variable, but most
of them are based on a plane-strain or a plane-stress assumption. The plane-strain definition of Sutcliffe (Section
3.2), taken the equivalent plastic strain as a strain measure,
has been used for the application discussed in this section.
However, strains and stresses occur in different directions
during sheet metal forming in reality. This gives rise to the
question if Sutcliffes model is still applicable and how the
strains should be accounted for.
- The model of Timsit & Pelow [5, 18] has been implemented in the friction model to describe the shear factor
fC (Section 3.3). Timsit & Pelow performed experiments
to obtain an empirical relation between the normal stress
and the shear strength of a stearic acid film deposited on
aluminum. The applicability of Timsit & Pelows model
0.3
Unit
1400
1
5.0 103
2.0 103
2.0 102
10
Sutcliffe
MPa
0.25
mm
mm2
mm
0.2
0.7
0.15
0.53
0.1
0.35
0.05
0.18
0
-4
0.88
-2.4
-0.8
0.8
2.4
z(m)
(m1 ) tool
Value
(m1 ) workpiece
Roughness parameter
1.1
workpiece
tool
0.145
5.2%
A
0.0%
0.13
(b)
(a)
Figure 14: Development ratio of real to apparent area of contact (a) and coefficient of friction (b) for normal loading only (gray
represents the non-contacting area)
0.19
29.4%
C
0.13
0.0%
D
(b)
(a)
Figure 15: Development ratio of real to apparent area of contact (a) and coefficient of friction (b) for normal loading only (gray
represents the non-contacting area)
must be checked when using other metal-lubricant combinations.
F N (z) (z + U d) (z) dz +
F N (z) (z d) (z) dz
dU
Zd
(z + U d) (z) dz +
(z d) (z) dz
dU
2
Z
(A.3)
F N (z) (z) dz
dU
or:
2 =
(A.4)
with:
Ar
=
=
Anom
7. Acknowledgments
(z) dz
(A.5)
dU
Appendix A.
3 = 2 (N + N )
N
X
zi
i=1
N
X
(A.6)
zk
k=1
(z)dz:
N
X
zk
k=1
3 =
(A.1)
(N + N )2
1 =
dU
(z) dz
Z
(z d) (z) dz
Zd
F Nk zk
k=1
FN
(z) dz
N+N
X
(N + N ) 2
(A.2)
dU
zk
(z + U d) (z) dz +
(A.7)
(z d) (z) dz
or:
k=1
3 = 2
(A.8)
Appendix B.
Value
Unit
210
0.3
0.00243
500
0.2
GPa
MPa
Value
Unit
210
0.3
GPa
45
90
R-value
Uniaxial factor ( fun )
Plane-strain factor ( f ps )
Plane-strain ratio ()
Pure shear factor ( f sh )
Equi-biaxial factor ( fbi )
Equi-biaxial ratio (bi )
2.1537
1.0
1.317
0.5
0.5442
1.169
0.8427
1.3154
1.0531
1.3146
0.5
0.5846
2.1942
0.9979
1.3077
0.5
0.5295
Value
Unit
114.8
250.34
0.25
8.1014
0.75
0.005
600
300
2.2
MPa
MPa
MPa
K
16