Available online at www.sciencedirect.

com
Wear 264 (2008) 349358
Deterministic repeated contact of rough surfaces
J. Jamari

, D.J. Schipper
Laboratory for Surface Technology and Tribology, Faculty of Engineering Technology, University of Twente,
Drienerloolaan 5, Postbus 217, 7500 AE Enschede, The Netherlands
Received 14 July 2006; received in revised form 16 February 2007; accepted 22 March 2007
Available online 15 May 2007
Abstract
A theoretical and experimental study to analyze the contact behavior of the repeated stationary deterministic contact of rough surfaces is
presented in this paper. Analysis focuses on the contact of rough surfaces on asperity level without bulk deformation. Contact area and deformation
of isotropic and anisotropic contacting surfaces were simulated for each loading cycle and were plotted along with the experimental results. Very
good agreement is found between the theoretical prediction and the measurement. It was found that the contact behavior becomes elastic soon after
the rst loading has been applied for the same normal load.
2007 Elsevier B.V. All rights reserved.
Keywords: Contact mechanics; Elasticplastic contact; Asperity; Loadingunloading
1. Introduction
For many engineering applications knowledge about contact
between surfaces is of fundamental importance to understand
friction, wear, lubrication, friction-induced vibrations and noise,
thermal and electrical contact resistance, etc. Therefore, the
interest in studying the contact between rough surfaces is very
high. This can be seen in the reviewing papers of Bhushan
[1], Liu et al. [2], Barber and Ciavarella [3] and Adams and
Nosonovsky [4].
Several approaches have been proposed to study the behav-
ior of the contacting surfaces: statistical [511], fractal [12]
and direct or numerical simulation [1316]. The statistical con-
tact model is pioneered by Greenwood and Williamson [5]
where a nominal at surface is assumed to be composed by
hemi-spherical asperities of the same radius and the height of
the asperities is represented by a well-dened statistical dis-
tribution function (i.e., Gaussian). In the fractal theory, the
scale-dependency of the rough surface is diminished; however,
the method can be applied only for fractal related surfaces.
For numerical contact modeling of surfaces, most of the pro-
posed models consider pure elastic deformation; numerical
elasticplastic contact models are rarely reported. Another

Corresponding author. Tel.: +31 53 4892463; fax: +31 53 4894784.

E-mail address: j.jamari@ctw.utwente.nl (J. Jamari).
important consideration to numerical contact modeling is the
computational cost. Karpenko and Akay [17] have introduced
a fast numerical method to calculate the friction force between
two rough surfaces for the elastic contact situation. Contact force
distribution is computed, using the local contact geometry, until
the sum of the local contact forces equals the normal load. This
iterating procedure is similar with the procedure as used in the
present paper. The method of Karpenko and Akay [17] has been
applied to the elasticplastic contact situation for the contact
of rough wavy surfaces [18]. However, in their elasticplastic
contact analysis, a contact point deforms elasto-plastically once
its contact pressure reaches the hardness of the softer material
while it has been widely accepted [1,2] that the elasticplastic
condition starts when the mean contact pressure reach about 0.4
times the hardness of the softer material.
There are many models available to analyze the contact of
rough surfaces in the elastic, elasticplastic and fully plastic
contact regimes, however, the experimental investigations are
rarely published. A theoretically and experimentally study of
the deterministic contact between rough surfaces in the fully
plastic contact regime has been proposed by Jamari et al. [19].
Results showed that the proposed model correlate very well with
the experimental investigation.
Most mechanical contact pairs carry their service load not
just once but for a large number of (repeated) cycles. The
application ranges from micro/nano-systems, such as MEMS
microswitches [20], to normal systems, such as rail-wheel con-
0043-1648/$ see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2007.03.024
350 J. Jamari, D.J. Schipper / Wear 264 (2008) 349358
tact [21]. Several studies have been performed to study the
elasticplastic contact behavior for the loading and unloading
case. By Finite Element Analysis calculation, Vu-Quoc et al.
[22] studied the relation between force and displacement of the
loadingunloading of contacting spherical bodies and Li et al.
[23] analyzed the loadingunloading of the contact between a
rigid at with an elasticplastic sphere or a rigid sphere with
an elasticplastic half-space. Recently, Jones [24] modeled the
loading and unloading of a rough surface based on the statistical
work of Greenwood and Williamson [5].
This paper presents an analysis of the loading and unloading
of a stationary deterministic contact between rough surfaces.
Rough surfaces are modeled based on the measured real sur-
faces. Asperity or roughness deformation without any bulk
deformation is considered. The model introduces a new method
to determine the initial asperity geometry and the remaining or
deformed asperity geometry after unloading. In order to ver-
ify the proposed model, an experimental investigation has been
performed.
2. Model
2.1. Model condition
In the proposed deterministic contact of rough surfaces, the
deformationof the contactingsurfaces is studiedonlyonasperity
level without any surface bulk deformation, therefore, a criterion
as was proposed in [25,26] is used to make sure that the contact
system is operating in the asperity deformation regime only. A
deformation criterion C
m
is dened, according to [25], as:
C
m
=
p
0
H

1 exp

2
3
3/5

< 0.6 (1)

where H is the hardness of the softer material. p
0
and are the
maximum Hertzian contact pressure and the roughness param-
eter, respectively, dened as [27]:
p
0
=

16PE
2

3
R
2

1/3
(2)
=

16RE
2
9P
2

1/3
(3)
where P is the load, R the radius of the curved body and is
the standard deviation of the height distribution. The effective
elastic modulus E is dened as:
1
E
=
1 v
2
1
E
1
+
1 v
2
2
E
2
(4)
where v
1
, v
2
, and E
1
, E
2
, are respectively the Poissons ratio and
the elasticity modulus of surfaces 1 and 2.
2.2. Asperity determination
Schematic representation of the stationary repeated contact
between rough surfaces is presented in Fig. 1(a). Here, the con-
tact loading and unloading takes place at the same position.
Fig. 1. Repeated contact of rough surfaces.
There are several ways to analyze the asperity-based determin-
istic contact of rough surfaces. Greenwood [28], for instance,
introduced a way to calculate the asperity properties using the
summit method. In this method, a summit is dened as a local
surface height higher than its neighboring points whether based
on ve or nine points. Recently, Jamari et al. [19] proposed
a new method to determine the contacting asperity properties
based on volume conservation. Contact areas (Fig. 1(b)) are
determined based on the measured rough surface loaded by a
certain load P or separation given between the means of surface
heights. These areas are converted into elliptical cross-sectional
areas (Fig. 1(c)) by conserving the total area of each asperity
of Fig. 1(b). Other parameters such as asperity radii and asper-
ity heights are determined based on the volume conservation
method, see Figs. 2 and 3. Details of the method can be found
in Appendix A.
Fig. 2 shows the discrepancy of the asperity properties
between the nine-points summit based asperity and the asperity
volume conservation method. An example of surface protru-
sion and its cut-off height is demonstrated in Fig. 2(a). In
Fig. 2(b)(d), at the left-side are the properties obtained by
the volume conservation method and at the right-side are the
properties by using the nine-point summit method. Fig. 2(b)(d)
presenting the asperities property for the indentation by a at
of
1
,
2
and
3
, respectively. As can be seen fromFig. 2(b)(d),
the number of asperity in contact (micro-contacts) changes as the
indentation or separation changes. For the summit method, this
number always increases as the indentation increases. For the
volume conservation method, this number is changing depend-
ing on the surface geometry and the location of the asperities in
contact depending to the indentation. For d =
3
, for instance,
there are 40 asperities in contact according to the summit method
whilst for the volume conservation method there is only 1 large
asperity which represents a real surface in contact much better
than the summit method. It can be concluded that the summit
J. Jamari, D.J. Schipper / Wear 264 (2008) 349358 351
Fig. 2. 3D surface asperities and its properties. (a) Model asperity and its centre prole (b)(d) location micro-contact points (not size) as a function of indentation

1
to
3
, respectively. Left volume conservation method and right the nine-point summit method.
352 J. Jamari, D.J. Schipper / Wear 264 (2008) 349358
Fig. 3. Asperity height determination (a) surface, (b) original asperity represented by paraboloid and (c) height z of paraboloid.
method represents the micro-contacts well only for the very
small interference whilst the volume conservation method the
micro-contacts are represented very well for all the interferences
or cut-off heights. Therefore, in the present model the volume
conservation method is used.
The curvatures of the asperity based on the volume conser-
vation method, see Appendix A, are dened as:

x
= 4
V
A
2
L
y
L
x
(5)

y
=
x
L
2
x
L
2
y
(6)
where V is the volume and A is the contact area. V and A
are determined from the micro-geometry measurement, there-
fore, the elliptic paraboloid representing the actual asperity
has the same volume and contact area as the as measured
asperity.
After determining the asperity curvatures, by Eqs. (5) and
(6), another important parameter in the present deterministic
contact model is the asperity height. Fig. 3 shows the method
how the asperity height is determined. The asperity height, z,
is determined based on the calculated radius in x direction R
x
and the diameter of the cross-sectional elliptical contact area in
x direction L
x
as:
z = d +
(L
x
/2)
2
2R
x
(7)
where d is the contact separation measured from the mean con-
tact of the surface height of the contacting surfaces.
2.3. Elasticplastic elliptical asperity contact model
The elasticplastic to fully plastic contact condition is consid-
ered in the present analysis, however, for subsequence loading
the asperity changes due to plastic deformation, resulting into
a new asperity geometry, and as a result shifts the operating
contact regime to the elastic contact situation.
In the proposed model also an elasticperfectly plastic mate-
rial property without any strain hardening effect is considered.
Therefore, the elasticplastic elliptical asperity contact model
of Jamari and Schipper [29] is employed.
Fig. 4. Flow diagram for the contact calculation of repeated stationary contacts.
J. Jamari, D.J. Schipper / Wear 264 (2008) 349358 353
Fig. 5. Matching procedure: surface before test (a), surface after test (b) and their difference (c).
2.4. Calculation procedure
Principally, the procedure to simulate the repeated stationary
contact situation is similar to the calculation of single contact as
for instance was presented in [19]. The only difference is that
for the repeated stationary contacts the new surfaces, as a result
of the previous loading condition, are used as the input for the
next loading cycle.
Fig. 6. Setup of the experiment.
Fig. 4 shows schematically the iterative procedure for ana-
lyzing the repeated stationary contact of rough surfaces using
the volume conservation method. The 3D surface height data
of the contacting surfaces, z
1
(x, y) and z
2
(x, y) are taken from
a roughness measurement and are used as inputs. The hard-
ness properties of the material H of the surfaces are obtained
from indentation tests. A normal load is applied to bring these
surfaces into contact. In the present study, analysis of the con-
tact is based on the contact interference; therefore, the load is
calculated based on the separation between the surfaces. For
this purpose, a separation is given as an initial guess. At this
separation, there will be a number of asperities in contact (micro-
contacts). Subsequently, all the contact input parameters such as
curvatures, height and location of each asperity in contact are
determined based on the volume conservation method as was
discussed early. Given all the contact input parameters, the con-
tact load, the contact area, etc. can be calculated readily from
the analysis of a single asperity as was discussed in Section
2.3 and formulated in [29]. Summation of all the micro-contact
loads (P
i
) gives the macro-contact load (input load), therefore,
an iterative procedure is applied by changing the separation until
the difference between summation of all the micro-contacts and
the applied load reach a certain criterion, . When the criterion
(=0.01) is satised the iteration loop is terminated and the
contact parameters such as load, contact area, plastic deforma-
tion, etc. for each asperity are taken from the last separation.
Here, the surface topography z

1
(x, y) and z

2
(x, y) are the output
from the contact model simulation. These outputs are used as
the input for the next loading cycle. The condition where there
354 J. Jamari, D.J. Schipper / Wear 264 (2008) 349358
Fig. 7. Isotropic aluminium surface before contact is applied (a) and location (not size) of the corresponding asperities (b).
is no plastic deformation anymore, i.e., the difference between
the subsequent input and output values of the surface topogra-
phy satises a certain running-in surface criterion
r
(=0.01), is
referred as the run-in condition.
3. Experiment
3.1. Specimens
Experiments are performed by loading a hard and smooth
curved surface onto a deformable nominally at rough sur-
face. A hard Silicon Carbide ceramic SiC ball (H=28 GPa,
E=430 GPa and v =0.17) with a diameter of 6.35 mm was used
as hard smooth curved indenter. To comply with the assumption
of perfectly smooth surface as was used in the present anal-
ysis, the r.m.s. roughness of the ceramic ball of 0.01 m was
chosen. An elasticperfectly plastic aluminium (H=0.24 GPa,
E=75.2 GPa and v =0.34), as was used for the single asperity
contact experiment in [30], was used for the rough at surface
specimen. Isotropic and anisotropic at surfaces were used. The
r.m.s. roughness is respectively, 1.3 and 1.4 m for the isotropic
and the anisotropic at surfaces.
3.2. Matching and stitching tool
The plastic deformation of the at surface by a spherical body
after being loaded against each other, for instance, is distinc-
tive but the plastic deformation of the asperities (without bulk
deformation) cannot be observed with the naked eye. Sloetjes
et al. [31] have developed a new robust technique to determine
the changes of the surface topography locally by matching the
measurement images before and after an indentation test.
Procedure of the nal matching process is described in Fig. 5.
Fig. 5(a) shows the 3D measurement result of the surface before
the test. This surface is brought into a contact with a hard spheri-
cal indenter. This surface is measured again after the indentation
test and is presented in Fig. 5(b). In this case, to determine the
change of the surface cannot be done by simply subtracting the
surface before and after the test since the surface after the test
has translated and rotated relatively to the surface taken before
the test, therefore, the matching process is needed. In the match-
ing process the mutual translations and rotations are determined
by aligning or repositioning both surfaces. Once these mutual
translations and rotations are found the surfaces are matched and
the difference image is determined readily by subtracting the
matched surfaces before and after test which results in Fig. 5(c).
Fig. 8. Contact area of the isotropic aluminium surface after the rst load cycle: (a) model and (b) experiment.
J. Jamari, D.J. Schipper / Wear 264 (2008) 349358 355
3.3. Experiment details
Fig. 6 shows the experimental setup used in the present exper-
iments. The setup can bear a maximumload of about 300 N. The
spherical and the at specimens were cleaned with acetone and
dried in air prior to any test. An optical interference microscope
was used for measuring the three-dimensional surface roughness
of the surface before and after a test.
As is shown in Fig. 6, an XY table which is controlled by
stepper motors is used to positioning the at specimen from
the loading position (position A) to the surface measuring posi-
tion (position A

) and the other way around. The at surface

was measured at the statically optical interference microscope
(position A

) before the loading test. After nishing the sur-

face measurement in position A

, the at surface was moved to

the loading position (position A). In this loading position the
statically mounted spherical specimen was moved down by a
loading screw and subsequently loaded by a dead weight load
system. The contact region was lubricated in order to reduce
possible effects of friction on the contact condition. The load
was applied to the contact system for 30 s and then removed
by lifting up the loading arm manually. Next, the at surface is
brought to the surface measuring position (position A

) by using
the XY table. Before measuring the surface after loading with
the optical interference microscope, again the at specimen was
cleaned and dried using the same procedure as was described
earlier. Subsequently, the surface is acquisitioned by the inter-
ference microscope for surface topography data. Until this step,
the rst loading cycle is nished.
To continue the test for the next loading cycle the same pro-
cedure is applied. When position A is reached, the load is gently
reapplied. All the surface topography images from every load-
ing cycle are matched with the initial surface topography image.
This was done separately by a personal computer.
4. Results and discussions
Fig. 7 to Fig. 9 show the deformation results for the isotropic
surface. Elastic perfectly plastic aluminium as was used in [30]
Fig. 9. Prole of the matched and stitched isotropic aluminium surface: (a)
x-prole at y =120 mand (b) y-prole at x =129 m; n =number of load cycles.
has a hardness constant for full plasticity c
h
of 0.71 and the
inception of fully plastic deformation
2
equals 80 times the
inception of the deformation where the rst yield occurs
1
for
circular contact, or c
A
=160. These constants are used in the
Fig. 10. Anisotropic aluminium surface before load is applied (a) and location (not size) of the corresponding asperities (b).
356 J. Jamari, D.J. Schipper / Wear 264 (2008) 349358
Fig. 11. Contact area of the anisotropic aluminium surface after the rst load cycle: (a) model and (b) experiment.
contact simulations. The explanation of these constants can be
found in [30].
Fig. 7(a) presents the initial isotropic surface before brought
in contact with a hard smooth ball. A load of 0.2 N was applied
to the contact. The dashed line represents macroscopic contact
area. The position of the asperities in contact due to the applied
load is presented in Fig. 7(b). The location (not size) of the
corresponding asperities is represented by dot markers in this
gure. 0.2 N load together with the r.m.s. roughness value R
q
of 1.3 m and its material properties yields a C
m
value for the
criterion of Eq. (1) of 0.16 which means that deformation of
the surface on asperity level is expected. As can be seen from
Fig. 9, there is no bulk deformation, all the plastic deformation
occurs at asperity level so that the deformation criterion works
accurately.
The contact area prediction by the model for the cycle n =1
is depicted in Fig. 8(a). Impression of the size of the geometry
of the asperities may be illustrated from this gure. The exper-
imental results of the contact area from the contact system are
presented in Fig. 8(b). As can be seen, the model predicts the
contact area very well. More insight in the model prediction and
the measured results may be seen in the plot of the asperities
in x and y direction as presented in Fig. 9. The model predicts
the change of the surface topography accurately. Experimental
results showthat there is almost nodifference betweenthe prole
for cycle 1, 2, 3, 10 and 20. In the rst loading cycle the asperity
deforms elastic, elasticplastic or fully plastic depending on its
geometry, location and height of the contact indentation. The
contact area for each asperity is developed to support the load.
When the load is removed, plastic deformation will modify the
surface geometry due to elasticplastic and fully plastic deform-
ing asperities which normally increase the degree of conformity
to the counter-surface (indenter). If the same load is reapplied
at exactly the same position (stationary), the same contact area
as for the rst loading will be developed. In this second loading
step the asperities deform elastically as a result of the residual
stresses and geometrical changes induced by the rst loading,
therefore there is no change of the surface topography or of the
contact area. Residual stresses will increase the yielding stress
and the change of the geometry will reduce the level of applied
stresses. In the present model the change of the surface geometry
is the only factor. Since the same loading condition is applied
for cycle 1, 2, 3, 10 and 20, therefore, the same phenomena are
observed.
Fig. 12. Prole of the matched and stitched anisotropic aluminium surface: (a)
x-prole at y =118 mand (b) y-prole at x =126 m; n =number of load cycles.
J. Jamari, D.J. Schipper / Wear 264 (2008) 349358 357
Results of the anisotropic aluminium surface are presented
in Figs. 1012. The initial anisotropic aluminium surface and
its asperity location are shown in Fig. 10. The same load as
was applied to the isotropic aluminium surface is used for the
anisotropic surface. Fig. 11 presents the result of the contact area
of the model prediction and of the experimental investigation for
the rst cycle. From this gure it can be seen that the surface
is dominantly represented by a single elliptic asperity. With an
r.m.s. roughness R
q
value of 1.4 m, it yields a C
m
value of 0.15
so that asperity deformation is expected. This is conrmed by
the experimental results of Fig. 12 where the plot of the x and y
prole are drawn. Results for cycle 1, 2, 3, 10 and 20 are similar
to the isotropic aluminium surface, i.e., there are almost no dif-
ferences in the contact area and the surface topography. Small
deviations between the model prediction and the experimental
results are observed for the y prole in Fig. 12(b). This may be
caused by the assumption of the rigid indenter where, in fact,
there is some elasticity of the ceramic ball indenter.
5. Conclusions
Theoretical and experimental studies have been performed
to analyze the deterministic contact behavior of the repeated
stationary contact of rough surfaces. Bulk deformation of the
surface is excluded in the analysis and concentrated to the con-
tact deformation on asperity level. Contact area and deformation
of the isotropic and anisotropic contacting surfaces were sim-
ulated for every loading cycle and were plotted along with the
experimental ndings. Results show that the theoretical model
predicts the measurement results very well.
In the repeated stationary contact of rough surfaces, it was
found that the contact behavior becomes elastic soon after the
rst loading has been applied for the same normal load. This is
because of the degree of conformity of the contacting surfaces is
reached, so that the developed contact area supports the applied
load elastically.
Acknowledgements
The nancial support of SKF ERC B.V. Nieuwegein, The
Netherlands and The Dutch Technology Foundation (STW) are
gratefully acknowledged.
Appendix A. Volume conservation method for
determining asperity geometry
In this appendix, a method to model the micro-contacts of
real rough surfaces which asperities are represented by elliptical
paraboloids will be described. The size of the asperities is based
on a volume conservation method.
An elliptical paraboloid is dened as a paraboloid having an
elliptical cross-section in the xy-plane and paraboloids in the xz-
and the yz-plane respectively, see Fig. 2. In mathematical form,
this elliptical paraboloid is expressed as:
x
2
a
2
+
y
2
b
2

z
c
= 0 (A.1)
where x, y, and z are the coordinate system and a, b, c are con-
stants. The volume displaced, V, due to contact of this elliptical
paraboloid is the same as the volume above a certain cut-off
height (volume conservation), hence:
V =

x
high
x
low

y
high
y
low

z
high
z
low
dx dy dz (A.2)
The limits of integral in Eq. (A.2) are determined as follows.
For the z-coordinate the upper integration limit is the cut-off
height of the micro-contact as:
z
low
= (A.3)
and the lower limit is determined rearranging Eq. (A.1) into:
z
high
=
c
a
2
b
2
(x
2
b
2
+y
2
a
2
) (A.4)
For the y-coordinate the following equations are valid at the
edge of the contact:
c
a
2
b
2
(x
2
b
2
+y
2
a
2
) = (A.5)
Solving y in Eq. (A.5) gives:
y
low
=
b
ac

c(a
2
cx
2
) (A.6)
y
high
=
b
ac

c(a
2
cx
2
) (A.7)
Now, the integration limits for x are left to consider. These
can be determined by substituting y
low
=y
high
=0 into Eqs. (A.6)
and (A.7) which results:
x
low
= a

c
(A.8)
x
high
= a

c
(A.9)
By substituting Eqs. (A.3)(A.9) into Eq. (A.2) prior to inte-
gration and simplifying yields:
V =

2
ba
c

2
(A.10)
The length of the micro-contact area in x-direction L
x
is
calculated by subtracting Eq. (A.9) by Eq. (A.8) as:
L
x
= 2a

c
(A.11)
and the length of the micro-contact area in y-direction L
y
is
calculated by subtracting Eq. (A.7) by Eq. (A.6) at x =0 as:
L
y
= 2
b
ac

ca
2
(A.12)
Substituting Eqs. (A.11) and (A.12) into Eq. (A.10) results a
new expression for the volume as:
V =
1
8
L
x
L
y
(A.13)
The curvature is dened as the second derivative of the ellip-
tical paraboloid, thus the curvature
x
in x-direction and the
358 J. Jamari, D.J. Schipper / Wear 264 (2008) 349358
curvature
y
in y-direction are found by applying the second
derivative to Eq. (A.1) as:

x
=
2
a
2
c (A.14)

y
=
2
b
2
c (A.15)
A combination is made by rearranging Eqs. (A.10)(A.15):

y
= 2
c
ab
=
8
L
x
L
y
= 4
V
A
2
(A.16)

y
=
L
2
y
L
2
x
(A.17)
Substituting Eq. (A.17) into Eq. (A.16) and rearranging gives
the nal expressions for the elliptic paraboloid that will be used
for tting the real micro-contact region as:

x
= 4
V
A
2
L
y
L
x
(A.18)

y
=
x
L
2
x
L
2
y
(A.19)
References
[1] B. Bhushan, Contact mechanics of rough surfaces in tribology: multiple
asperity contacts, Tribol. Lett. 4 (1998) 135.
[2] G. Liu, Q.J. Wang, C. Lin, A survey of current models for simulating the
contact between rough surfaces, Tribol. Trans. 42 (1999) 581591.
[3] J.R. Barber, M. Ciavarella, Contact mechanics, Int. J. Solids Struct. 37
(2000) 2943.
[4] G.G. Adams, M. Nosonovsky, Contact modellingforces, Tribol. Inter. 33
(2000) 431442.
[5] J.A. Greenwood, J.B.P. Williamson, Contact of nominally at surfaces,
Proc. R. Soc. Lond. A295 (1966) 300319.
[6] W.R. Chang, I. Etsion, D.B. Bogy, An elasticplastic model for the contact
of rough surfaces, ASME J. Tribol. 109 (1987) 257263.
[7] J.H. Horng, An elliptic elasticplastic asperity microcontact model for
rough surface, ASME J. Tribol. 120 (1998) 8288.
[8] Y. Zhao, D.M. Maietta, L. Chang, An asperity microcontact model incor-
porating the transition fromelastic deformation to fully plastic ow, ASME
J. Tribol. 122 (2000) 8693.
[9] Y.R. Jeng, P.Y. Wang, An elliptical microcontact model considering elas-
tic, elastoplastic, and plastic deformation, ASME J. Tribol. 125 (2003)
232240.
[10] L. Kogut, I. Etsion, A nite element based elasticplastic model for the
contact of rough surfaces, Tribol. Trans. 46 (2003) 383390.
[11] R.L. Jackson, I. Green, Astatistical model of elasto-plastic asperity contact
between rough surfaces, Tribol. Inter. 39 (2006) 906914.
[12] A. Majumdar, B. Bhushan, Fractal model of elasticplastic contact between
rough surfaces, ASME J. Tribol. 113 (1991) 111.
[13] M.N. Webster, R. Sayles, A numerical model for the elastic friction-
less contact of real rough surfaces, ASME J. Tribol. 108 (1986) 314
320.
[14] X. Liang, Z. Linqing, A numerical model for the elastic contact of three-
dimensional real rough surfaces, Wear 148 (1991) 91100.
[15] N. Ren, S.C. Lee, Acontact simulation of three-dimensional rough surfaces
using moving grid method, ASME J. Tribol. 115 (1993) 597601.
[16] L. Chang, Y. Gao, Asimple numerical method for contact analysis of rough
surfaces, ASME J. Tribol. 121 (1999) 425432.
[17] Y.A. Karpenko, A. Akay, A numerical model of friction between rough
surfaces, Tribol. Inter. 34 (2001) 531545.
[18] Y.A. Karpenko, A. Akay, A numerical method for analysis of extended
rough wavy surfaces in contact, ASME J. Tribol. 124 (2002) 668679.
[19] J. Jamari, M.B. de Rooij, D.J. Schipper, Plastic deterministic contact of
rough surfaces, ASME J. Tribol., accepted.
[20] S. Majumder, N.E. McGruer, G.G. Adams, A.P. Zavracky, P.M. Morrison,
J. Krim, Study of contacts in an electrostatically actuated microswitch,
Sensor Actuators A 93 (2001) 1926.
[21] A. Kapoor, F.J. Franklin, S.K. Wong, M. Ishida, Surface roughness and
plastic ow in rail wheel contact, Wear 253 (2002) 257264.
[22] L. Vu-Quoc, X. Zhang, L. Lesberg, A normal force-displacement model
for the contacting spheres accounting for plastic deformation: force-driven
formulation, ASME J. Appl. Mech. 67 (2000) 363371.
[23] L.-Y. Li, C.-Y. Wu, C. Thornton, A theoretical model for the contact of
elastoplastic bodies, Proc. Instn. Mech. Engrs. 216C (2002) 421431.
[24] R.E. Jones, Models for contact loading and unloading of a rough surface,
Int. J. Eng. Sci. 42 (2004) 19311947.
[25] J. Jamari, D.J. Schipper, Deformation due to contact between a rough
surface and a smooth ball, Wear 262 (2007) 138145.
[26] J. Jamari, Running-in of Rolling Contacts, Ph.D. Thesis, University of
Twente, The Netherlands.
[27] J.A. Greenwood, K.L. Johnson, E. Matsubara, ASurface roughness param-
eter in hertz contact, Wear 100 (1984) 4757.
[28] J.A. Greenwood, A unied theory of surface roughness, Proc. Roy. Soc.
Lond. A393 (1984) 133157.
[29] J. Jamari, D.J. Schipper, An elasticplastic contact model of ellipsoid bod-
ies, Tribol. Lett. 21 (2006) 262271.
[30] J. Jamari, D.J. Schipper, Experimental investigation of fully plastic contact
of a sphere against a hard at, ASME J. Tribol. 128 (2006) (2006) 230
235.
[31] J.W. Sloetjes, D.J. Schipper, P.M. Lugt, J.H. Tripp, The determination of
changes in surface topography using image processing techniques, in: Pro-
ceeding of the International Tribology Conference, Nagasaki, 2000, pp.
241246.