International Journal of Mechanical Sciences 77 (2013) 107–112

Contents lists available at ScienceDirect

International Journal of Mechanical Sciences

journal homepage: www.elsevier.com/locate/ijmecsci

Subsurface treatment of a contact subjected to a hertz pressure

A.-C. Palade a,b, G.-P. Pillon a, E. Cicala a,b, D. Grevey a,n, L. Marsavina b
a
Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS-Université de Bourgogne, 1 allée des Granges Forestiers,
F-71100 Chalon sur Saône Cedex, France
b
Universitatea “Politehnica” din Timisoara, Facultatea de Mecanica, 1 Blvd. M. Viteazu, 300222 Timisoara, Romania

art ic l e i nf o a b s t r a c t

Article history: The study proposes an approach allowing the favorable modification of the stresses field of a structure
Received 26 April 2013 subjected to a Hertzian contact. It is based on the study of a classic Hertzian contact in the presence of
Received in revised form subsurface circular defects. The idea is to limit the field of tangential stresses without large increasing of
27 August 2013
the global stress (Von Mises) of the system. The location of the defects is made according to the
Accepted 31 August 2013
mechanical parameters of the contact (Fn, has, z). The results of the numerical calculations (FEM) are
Available online 21 September 2013
compared with photoelastic measurements and highlight an excellent correlation. The location of the
Keywords: defect on the symmetry axis (x ¼ 0) is the dominating parameter for the tangential stress decreasing.
Hertzian contact & 2013 Elsevier Ltd. All rights reserved.
Subsurface treatment
Shearing stress decreasing
Photoelasticity
Numerical simulation

1. Introduction the treatment, and we led a numerical study to be able to

generalize this approach in other materials (metals in particular),
The approach proposed here wants to be original in view to of big industrial interest. The numerical study of the contact was
strengthen the properties of a structure subjected to a Hertzian investigated using ABAQUS/STANDARD.
contact. It is intended to modify the subsurface of a component, The mechanical, optical and dimensional characteristics of used
without altering the integrity of the initial surface. In fact it allows solids are defined in the Table 1. To guarantee a good correlation
modifying the stress field in subsurface allowing the mechanical part between the experimental and numerical approaches, the proper-
to support better and to transmit the efforts which are subjected to it. ties of the polycarbonate were experimentally determined accord-
Our approach is based on numerous works dedicated to the analysis ing to the standard ASTM D 625-02a, and the order of the fringe fr
of the Hertzian contact and to the deformation of solids [1–9] and was obtained from 4 points bending tests.
became refined with the researches of Taraf [10] and Kabo [11–13]. The studied Hertzian contact is a cylinder/plan type. The
They were interested in the influence of the presence of circular and maximal shearing stress is expressed versus the contact pressure
elliptic defects on the initiation of on-surface cracks and in subsur- P0 (τmax ¼0.3  P0) and is located in subsurface, at the depth
face, in the case of the Fatigue of a Rolling Contact. First, the life cycle z¼ 0.78a. We will take the depth z as reference parameter, in
decreases when the size of holes increases. They show that for the order to size the contact parameters and to position the subsurface
rolling contact, the elliptic defects are more dangerous than circulars defects. A normal strength of 714 N is applied to the cylinder. It
defects, because the behavior of plastic accommodation near a leads to a half width of contact a ¼3.122 mm, with a location of the
circular hole is transformed in an effect of “rochet” near the elliptic maximal shearing stress τmax, at z ¼2.5 mm. The maximal corre-
defect when it is located in subsurface. More the most critical sponding pressure for the contact is P0 ¼ 29.716 MPa.
configurations are a vertical or diagonal alignment of several circular
defects. In our approach, the purpose is to modify favorably the
stresses field in subsurface by replacing the zones of strong con- 2. Methodology
centration by circular holes.
We chose to work with a transparent optic active material to be 2.1. Numerical approach (FEM)
able to observe easily, by photoelasticity, the effects produced by
The numerical model implemented to simulate the Hertzian
contact is a 2D model taking into account plane strain conditions.
n
Corresponding author. Tel.: þ 33 385731021; fax: þ 33 385731120. The cylinder was considered perfectly rigid and with an infinitely
E-mail address: dominique.grevey@u-bourgogne.fr (D. Grevey). long length [4,5]. It is modeled as an analytical rigid body [14–16]

0020-7403/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijmecsci.2013.08.013
108 A.-C. Palade et al. / International Journal of Mechanical Sciences 77 (2013) 107–112

Table 1 (Fig. 1a). The prism is considered as a homogeneous and deform-

Mechanical and optical properties of materials. able solid [5]. The normal pressure is elliptically distributed along
the contact region 2a.
Geometrical size Ec ν Rc fs (MPa/
(MPa) (MPa) fringe) The mesh is built with 169,898 linear elements and 170,005
nodes in two dimensions (CPE4R and CPE3) (Fig. 1b). It is divided
100 Cr6 Cylinder – diameter: 210,000 0.29 443 – into two zones: a refined mesh near the contact with 160,000
180 mm, length: 20 mm elements 10 mm² size and an enlargement of the mesh size
Polycarbonate prism – 50  50  49 mm3 1490 0.37 116 1.39
according to the distance of the first zone. The use of elements
of “bias” type allows reducing the duration of the calculation.
The size of the contact zone relatively to the whole sample allows
saying that there will be no visible interaction between the stress
field and the borders of the sample [15,17]. According to the Hertz
theory, the interaction type is standard frictionless, hard “surface to
surface” contact, where the cylinder is defined as “master” surface in
the master–slave contact pair. Being controlled by a reference point,
it acts vertically, pressing on the surface of the planar body. The
constraints of the planar body assumes the locking of the vertical
displacement of the nodes located on the lower edge and the locking
of the horizontal displacement of the nodes located on the symmetry
axis (x¼0). The upper portion is left free.
For our configuration, the comparison of the analytical results
based on the methodology proposed by Nowell and Hills [2] and
the results of the numerical calculation are in a good agreement.
This is visible on the Fig. 2 which presents the stress distribution
under the surface (Fig. 2a) and the contact pressure distribution
(Fig. 2b). The use of the mesh of the refined zone is proven to be
sufficient for giving convergent solutions.

2.2. Photoelasticity approach

A particular care was brought to the realization of the polycarbo-

nate contact area to have a smooth surface, what is imperative by the
theory of Hertz. Indeed, we need to have the flattest surface possible
and a low roughness. The roughness of the “plan” (Ra 0.026 mm and
Rt 0.046 mm) and of the cylinder (Ra0.26 mm and Rt1.90 mm)
are very low in front of the radius of the area of studied contact
a¼ 3.122 mm, what is satisfactory.
The experimental setup consists of a compression machine
(MTS – Adamel Lhomargy-DY36, equipped with a load cell of 5 kN)
and of a transmission polariscope (Jessop–Leech) running with
white circular polarized light.
The polycarbonate sample is loaded and the acquisition of
images is realized by transparency, by using a digital camera. Then,
the images are treated with the Photoshop CS5 software and their
dimensions are calibrated by means of Sigma Scan Pro 5.0.
The polarization of the light transmitted through the loaded
Fig. 1. (a) Configuration of the contact geometry and (b) FE mesh of the sample allows linking the stress with the birefringence of the
polycarbonate part. polycarbonate [18]. So each N isochrome order corresponds to a

Fig. 2. Comparison of the FEM results with Nowell and Hills solutions [2]. (a) Sub-surface stresses and (b) Normalized contact pressure distribution.
 A.-C. Palade et al. / International Journal of Mechanical Sciences 77 (2013) 107–112 109

line of equal value of (s1  s2) that is linked to the maximal Student law and a 95% confidence interval. The experimental
shearing stress with: τmax ¼(s1 s2)/2 where (s1  s2) is the values are detected from a depth of 0.5 mm.
difference between the principal stresses. The plan where the By keeping the previous operating conditions, we validated the
tangential stress is observed is perpendicular to the direction (z) of method by adding a hole with a diameter of 0.82 mm and located
the applied force, namely the polarization plan, and in the 2.5 mm under the surface, location point of τmax, when a load
direction (y) of propagation of the light (Fig. 3). Fn ¼714 N is applied (Fig. 5).
The holes from the polycarbonate samples were obtained by The distribution of the stress field between the numerical and
drilling. The size of these holes (Ø0.82 mm) is obtained with a experimental models is in a good agreement. Thus we assume that
Ø0.76 mm lubricated drill, operated at 0.081 cm s  1 feed rate and the numerical model, showing a good estimation of the shearing
475 tr min  1 rotation speed. These parameters were obtained after stress, allows estimating the distribution of the principal stresses.
an optimization process involving an experimental plan. The main The numerical calculation will thus allow predicting the distribu-
goal was to reduce the residual stress resulting from this process. tion of the shearing and Von Mises's stresses sVM in a pierced solid
or which presents a defect. Samples were mechanically drilled and
the absence of effect is verified by photoelasticity.
2.3. Methodology validation

The results obtained numerically, experimentally and analyti-

3. Résults and discussion
cally are in good agreement for a basic Hertzian contact (Fig. 4).
Each experimental value is the average value of 10 tests and the
3.1. Effect of the presence of a hole in the axis of the contact
measurement uncertainties are statistically obtained using the
The presence of a hole on the symmetry axis allows a 42%
decreasing of the shearing stress, especially near the hole, as
illustrated in Fig. 6. On the other hand, Von Mises's stress
decreases of 8% at the surface (Fig. 7a) and is multiplied by 3 near
the drilling (Fig. 7b).

3.2. Influence of the number and of the location of the drillings

We broadened the digital approach by studying the influence of

various drilling configurations. For more simplicity, we kept the
previous parameters (strength, depth of location of the drilling,
diameter of drilling) and we added holes at the border of the
contact zone, i.e. at ( a) and (þa). So we have 3 new cases (Fig. 8).
Making it, we obtain a system in which the positions of holes are
governed by data which define the contact: Fn, z and a.
The addition of side holes (case 2, 3 and 4) has only a small
Fig. 3. Sub-surface distribution of the stresses within PC. Schematic variation influence on the level of the shearing stress (Fig. 9). It is only the
of τmax.

Fig. 4. Sub-surface distribution of τmax within PC. Comparison of numerical, Fig. 6. Sub-surface distribution of τmax within PC at 714 N. Comparison FEM-
analytical and experimental approaches. experimental.

Fig. 5. (a) Photoelastic fringes of a PC charged. (b) FEM: Tresca distribution.

110 A.-C. Palade et al. / International Journal of Mechanical Sciences 77 (2013) 107–112

Fig. 7. Distribution of Von Mises stress (MPa) along x for Fn ¼714 N: (a) for z ¼0 mm; (b) for z¼ 2.5 mm.

Fig. 8. Tresca distribution within drilled sample, at 714 N. FEM approach.

presence of the hole on the symmetry axis which allows the 60

decreasing of this stress. On the other hand, Von Mises's stress is τmax (MPa) σVM (MPa)
increased by a nearby factor of 3, except in the case 2, where the 50
absence of drilling on the axis minimizes the value.
40
MPa

30
3.3. Influence of the drilling diameter
20
The approach was completed by introducing an additional
factor: the diameter of drilling of the cylindrical hole. The results 10
are presented on Fig. 10 and are expressed according to normal-
ized parameters. On the y-axis, we have the ratio of Von Mises or 0
REF CASE 1 CASE 2 CASE 3 CASE 4
the shearing stress on the maximal contact pressure, Pmax. On the
x-axis, the ratio drilling diameter (ϕ)/half-width of contact (a) has Fig. 9. Distribution of τmax along z for Fn ¼714 N, z(τmax) ¼ 2.5 mm, a ¼3.122 mm.
 A.-C. Palade et al. / International Journal of Mechanical Sciences 77 (2013) 107–112 111

been chosen. The will is that from these curves, it is possible to a big importance for the control of Von Mises's stress. So, and
predict the behavior for various configurations of material. because we want to keep a general vision, we follow the perfor-
For ϕ/a o0.03, we have simultaneously an increase (between mance indicator versus the ratio surface of drilling/surface trans-
50 and 300%, depending on the considered case) of the normalized verse of the contact, where:
Von Mises's stress and a 60% decrease of the normalized tangential
stress.  The transverse surface of the contact is defined as being the
For 0.03 oϕ/ao 0.1, Von Mises's stress continues to increase product 2za;
while the tangential stress remains constant. The interest of the  The surface of the drilling is Nπϕ2⧸4, with N, the number of
drilling as regards the decrease of the tangential stress is elimi- drillings and ϕ, the diameter of the drillings.
nated by the overload of Von Mises's stress.
Beyond a ratio ϕ/a ¼0.1, Von Mises's stress and the tangential Fig. 11, below, shows that the case 2 allows a global decrease of
stress don't vary any more. The tangential stress undergoes a the stresses, only when the drilling surface is lower than 0.2% of
decrease of the order of 40% while Von Mises's stress increased the transverse surface of the contact.
by 300%.

4. Conclusion
3.4. Definition of the most favorable conditions
The decrease of the ecological footprint of the industrial
The study of the results shows that the case 2 is the one which processes brings the researcher to propose new solutions, envir-
leads to the lowest increase of Von Mises's stress. This case onment-friendly, to improve the properties of surface of mechan-
represents the best compromise between the biggest decrease of ical components. Our contribution consisted to show the interest
the tangential stress and the lowest increase of Von Mises's stress. of a subsurface treatment to strengthen the properties of a
Thus the product of the two stresses appears as a good perfor- component subjected to a Hertzian contact. In the case of a roll
mance indicator of the treatment, by accepting the fact that each plan contact, our works bring to light a good correlation between
of the components has the same importance in the result. The the experimental and numerical approach for the study of the
normalization of this indicator allows keeping a classic approach. stresses field. This good agreement is obtained as well in the
Thus it becomes: sVMτmax/P2max. classic Hertzian case as in the presence of circular defects under
The analysis also shows that the size of the drillings has low the surface of the contact. The normalized location of defects, in
influence on the results in terms of tangential stress, but presents the Hertzian sense, allows to define a space of study and allows
showing that the presence of a hole on the symmetry axis (x ¼0) is
the dominating factor when we aim at the decrease of the
tangential stress. On the other hand, the presence of holes
generates an increase of Von Mises's stress we have to take into
account not to plasticize the material. The optimization of this
approach allows to refine the strategy for the limitation of the
stresses field: realization of drillings in border of the contact zone
(case 2), with a drilling surface lower than 0.2% of the transverse
surface of the contact. As a result the performance indicator
(sVMτmax/P2max) shows a global decrease of the stresses level.
From a practical point of view, this imposes to realize cavities of
the order of 100 mm of diameter, by laser drilling for “transparent”
materials or by using the laser sintering technique for the other
materials.

Fig. 10. Normalized stresses's distribution (Von Mises and Tresca) versus the ratio Acknowledgement
ϕ/a.
Authors want to acknowledge the Grant accorded by the
uB-Filiale, subsidiary of Université de Bourgogne, which supports
this study.

References

[1] Johnson KL. Contact mechanics. Cambridge: Cambridge University Press;

84–104 (ISBN 0-521-34796-3).
[2] Hills DA, Nowell D, Sackfield A. Mechanics of elastic contacts. Oxford:
Butterworth-Heinemann Ltd.; 20–4 (ISBN 978-90-481-4409-9).
[3] Kamran Y, Larsson PL. Second-order effects at microindentation of elastic
polymers using sharp indenters. Materials and Design 2011;32(6):3645–53.
[4] Komvopoulos K, Gong Z-Q. Stress analysis of a layered elastic solid in contact
with a rough surface exhibiting fractal behavior. International Journal of Solids
and Structures 2007;44:2109–29.
[5] Yan W, Sun Q, Feng XQ, Qian L. Analysis of spherical indentation of super-
elastic shape memory alloys. International Journal of Solids and Structures
2007;44:1–17.
[6] Terekhina S, Fouvry S, Salvia M, Bulanov I. An indirect method based on
fretting tests to characterize the elastic properties of materials: application to
Fig. 11. Evolution of the performance indicator according to the ratio (surface of an epoxy resin RTM6 under variable temperature conditions. Wear 2010;269:
drilling/transverse surface of the contact). 632–7.
112 A.-C. Palade et al. / International Journal of Mechanical Sciences 77 (2013) 107–112

[7] Frija M, Hassine T, Fathallah R, Bouraoui C, Dogui A. Finite element modelling [13] Kabo E, Ekberg A. Fatigue initiation in railway wheels – a numerical study of
of shot penning process: prediction of the compressive residual stresses, the the influence of defects. Wear 2002;253:26–34.
plastic deformations and the surface integrity. Materials Science and Engi- [14] Wang L, Wang Y, Sun XG, Pan ZY, He JQ, Li CG. Influence of pores on the
neering A 2006;426:173–80. surface microcompression mechanical response of thermal barrier coatings
[8] Dick T, Paulin C, Cailletaud G, Fouvry S. Experimental and numerical analysis fabricated by atmospheric plasma spray–finite element simulation. Applied
of local and global plastic behavior in fretting wear. Tribology International Surface Science 2011;257:2238–49.
2006;39:1036–44. [15] Poon B, Rittel D, Ravichandral G. An analysis nanoindentation in linearly
[9] Mohd Tobi AL, Ding J, Bandak G, Leen SB, Shipway PH. A study on the elastic solids. International Journal of Solids and Structures 2008;45:6018–33.
interaction between fretting wear and cyclic plasticity for Ti–6Al–4V. Wear [16] Liao Y, Zhou Y, Huang Y, Jiang L. Measuring elastic-plastic properties of thin
2009;267:270–82. films on elastic–plastic substrates by sharp indentation. Mechanics of Materi-
[10] Taraf M. Endommagement par fatigue de contact de roulement d’une roue als 2009;41:308–18.
ferroviaire: étude numérique par éléments finis [Thèse doctorale No.: [17] Bagherifard S, Ghelichi R, Guagliano M. A numerical model of severe shot
2008061S]. Metz: Université Paul Verlaine; 2008. penning (SSP) to predict the generation of a nanostructured surface layer of
[11] Kabo E. Material defects in rolling contact fatigue – influence of overloads and material. Surface and Coatings Technology 2010;204:4081–90.
defects clusters. International Journal of Fatigue 2002;24:887–94. [18] Ghita E., Marsavina L. Fotoelasticimetria. Metoda moderna in analiza experi-
[12] Kabo E, Ekberg A. Material defects in rolling contact fatigue of railway wheels - mentala a tensiunilor, Editura Eurostampa, Timisoara, 2002. pp. 63–67. ISBN
the influence of defect size. Wear 2005;258:1194–200. 973-8244-96-X.