See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/229201137

A numerical model of severe shot peening (SSP)

to predict the generation of a nanostructured
surface layer of material

Article in Surface and Coatings Technology September 2010

DOI: 10.1016/j.surfcoat.2010.05.035

CITATIONS READS

62 291

3 authors:

Sara Bagherifard Ramin Ghelichi

Politecnico di Milano Politecnico di Milano
53 PUBLICATIONS 639 CITATIONS 28 PUBLICATIONS 377 CITATIONS

SEE PROFILE SEE PROFILE

M. Guagliano
Politecnico di Milano
176 PUBLICATIONS 1,668 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Creation of nanograined surface layer for fatigue life increase. View project

All content following this page was uploaded by Sara Bagherifard on 22 October 2014.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
 Surface & Coatings Technology 204 (2010) 40814090

Contents lists available at ScienceDirect

Surface & Coatings Technology

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s u r f c o a t

A numerical model of severe shot peening (SSP) to predict the generation of a

nanostructured surface layer of material
S. Bagherifard, R. Ghelichi, M. Guagliano
Politecnico di Milano, Department of Mechanical Engineering, Via La Masa, 34 20156 Milano, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Generation of a surface layer of material characterized by grains with dimensions up to 100 nm by means of
Received 21 October 2009 severe plastic deformation is one of the most interesting methods to improve the mechanical behaviour of
Accepted in revised form 20 May 2010 materials and structural elements. Among the ways to obtain a surface layer with this characteristic, shot
Available online 4 June 2010
peening is one of the most promising processes, since it is applicable to very general geometries and to all
metals and metal alloys without high-tech equipments. Notwithstanding the fact that the ability of shot
Keywords:
Severe shot peening (SSP)
peening to obtain nanostructured surfaces by using particular process parameters (mainly high impact
Nanocrystalline structure energy and long exposure time) is proved, deep knowledge of the correct choice of quantitative values of
Finite element process parameters and their relation to the grain size and the thickness and uniformity of the
Residual stresses nanostructured layer is still lacking.
Equivalent plastic strain In this paper a nite element model of severe shot peening (SSP) is developed with the aim of predicting the
Severe plastic deformation treatment conditions that lead to surface nanocrystallization. After having assessed the accuracy of the
model as regards mesh parameters and constitutive law of the material, the results are discussed and
interpreted in terms of induced residual stresses and surface work hardening. A method to assess the
formation of nanostructured layer of materials based on the value of the equivalent plastic strain is
developed.
The comparison with experimental results allow to afrm that the model is a useful tool to predict the
generation of a nanostructured surface layer by shot peening and to relate the peening parameters with the
treated surface layer in terms of residual stresses, work hardening, and depth of the nanostructured layer.
2010 Elsevier B.V. All rights reserved.

1. Introduction
Shot peening (SP) is a well-known mechanical surface treatment
generally applied to improve fatigue behaviour of metallic compo-
nents. During the process small spherical peening media which are
accelerated in peening devices of various kinds impact the surface of
work piece with energy able to cause surface plastic deformation. The
process is aimed to create compressive residual stresses and work
harden the near surface layer of material. These effects are very useful
in order to totally prevent or greatly delay failure of the part by fatigue
[16].
In recent years special methods of SP have been recognized to be
benecial in creating ne grained layers of material on surface of
treated components in order to improve service life time and global
behaviour. This improvement is due to the fact that most cases of
material failure such as fatigue fracture, fretting fatigue, wear and
corrosion originate from the exterior layers of the work piece. These
phenomena are all extremely sensitive to the structure and properties
Corresponding author. Tel.: + 39 02 23998206; fax: + 39 02 23998202.
E-mail address: mario.guagliano@polimi.it (M. Guagliano).
of the surface material. Since most of fatigue cracks initiate from the
surface and propagate to the interior, a component with a nanos-
tructured surface layer and coarse-grained interior is expected to have
highly improved fatigue properties. Since both fatigue-crack initiation
and propagation are inhibited by ne grains near the surface and
coarse grains in the interior, respectively. Moreover, the residual
compressive stresses introduced during the process can effectively
stop or delay the initiation and propagation of fatigue cracks [7].
Among the alternative SP methods aimed at surface nanocrystalliza-
tion, we can recall ultrasonic shot peening (USSP) [8], surface
nanocrystallization and hardening (SNH) [8] and high-energy shot
peening (HESP) [8]. The mentioned processes are somehow different
from the conventional SP process both for the needed technological
facilities and for the mechanics of the treatment. The common aspect
is to use special combinations of peening parameters to multiply the
kinetic energy of the shot impacts in order to generate a large number
of defects, dislocations and interfaces (grain boundaries) on the
surface layer of treated part and consequently transform its
microstructure into nanograin size [8]. Some experimental research
has been developed to nd the relationship between the process
mechanisms and the characteristics of the generated ne grained
layer but there are still many aspects to be studied.

0257-8972/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.surfcoat.2010.05.035
4082 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090
It is well-known that numerical simulation can be considerably
helpful to limit the costly and time consuming experiments. Their
efcacy is more appreciable in cases of attempt to obtain nanocrystals
with SP. Since these processes necessitate performing many experi-
ments with different combinations of peening parameters in order to
increase the kinetic energy of process for generation of nanocrystals.
In the present paper, SSP refers to an SP process applied by
conventional air blast shot peening device but using severe para-
meters which include high shot velocity and long exposure time.
These parameters are basically different from those generally applied
in industry to shot peen the same material. The aim of intensifying the
peening parameters is to increase the kinetic energy of the process
and eventually obtain surface nanocrystallization.
Several approaches have been suggested in the literature for
numerical simulations of SP but there are very few studies dealing
with simulation of SSP that leads to surface nanocrystallization. Dai et
al. performed impact simulation of surface nanocrystallization and
hardening process (SNH) and compared the results in terms of
roughness, residual stress and effective plastic strain [9,10]. Ma et al.
developed a routine using nite element method to perform a
molecular dynamic simulation and calculate the dislocation density
generated by ultrasonic SP [11].
In this study, nite element simulation of air blast shot peening
(ABSP) with unconventional and severe parameters is performed to
provide quantitative description of effect of peening parameters in
order to describe the distribution and magnitude of residual stresses
and also the thickness of the work-hardened layer. After having
described the development of the model in terms of geometry, mesh
parameters and the material mechanical behaviour schematization,
the attention is focused on the critical assessment of the mesh
convergence and the way the number of impacts can be related to the
real treatment time and consequently to the resulting coverage.
In the nal part of the paper the results are critically discussed in
terms of residual stresses and surface work hardening.
Studies performed to distinguish the essential parameters for
structural evolution and generation of nanograins have acknowledged
large strains as the most important condition favorable to produce
nanocrystals [12,13]. There is also a criterion set on the accumulated
equivalent plastic strain (PEEQ) value proposed to assess the
formation of nanograins [12,13]. In this paper a method based on
the mentioned criterion is proposed for the estimation of the
treatment parameters needed to obtain nanostructured surface
layer of material and to assess its depth.
The comparison of the obtained numerical results with experi-
mental data regarding the favorable condition for formation of
nanograins allows validating the presented SSP numerical model.
Some guidelines for future works aimed to rene the model are nally
drawn.
2. Finite element simulation development
2.1. FE model geometry
A 3D model developed using commercial nite element code
Abaqus/Explicit 6.7 is utilized to investigate single and multiple
impact effects. Due to symmetry only one quarter of the specimen is
modelled for studying single impact while in case of multiple impacts
the full model is analyzed to maintain the arbitrary aspect of the
process.
The target is modelled as a rectangular body (3 3 1.5 mm3),
sufciently large to avoid the effects of boundary conditions on the
residual stress state in the impact area. The impact area of (1 1 mm2)
is located in the centre of the rectangular face. Target mesh is set up by
513604 C3D8R 8-node linear brick elements with reduced integration
and hourglass control. All side faces including target's base are
surrounded by so called half innite elements that provide quiet
boundaries by minimizing the reection of dilatational and shear
waves back in to the region of interest [14]. The shape and orientation
of this element are similar to C3D8R element excluding that the
element must be attached such that the innite end faces away from
the model. Innite elements are allowed only with linear elastic
behaviour, so they must be positioned sufciently distant from the
non-linear interaction region to ensure accuracy [15].
Steel shots with a diameter of 0.6 mm similar to the shots used in
experiment (commercial grade S230) are modelled as spherical
bodies consisting of tetrahedral C3D4 elements with an isotropic
elastic behaviour. Velocity in the z-direction was dened as initial
condition to all the shots, regarding an impact angle = 90 as it is
typical for ABSP.
General contact was used as the criteria of contact with an
isotropic Coulomb friction coefcient equal to = 0.2 [16]. The
scheme of full model is presented in Fig. 1.
It is to be underlined that thermal effects have been neglected in
the developed simulation.
2.2. Material model
Since strain rate dependency of target material will have notable
effects on stress prole and the extent of surface hardening, it is
recommended to use a material model able to describe the near
surface high strain rate response due to the high-energy successive
impacts [17]. In this study non-linear kinematic Chaboche hardening
model [18] and also JohnsonCook equation [19], which are both
constitutive models representing effect of strain rate on material
hardening, were initially chosen to describe target material's
behaviour. The results for both material models were compared
with experimental results and nally combined isotropic kinematic
model was chosen for the rest of analysis due to better consistency
with experimental results. JohnsonCook model highly overestimated
surface residual stress and also the maximum residual stress below
the surface.
The material used in this study is steel (39NiCrMo3, according to
the Italian nomenclature) with the chemical composition presented in
Table 1. Static tension tests were carried out according to ASTM E8M
[20] in order to obtain monotonic properties of the material. The
specimens were prepared following the instructions of ASTM A370-05
[21]. Obtained results are reported in Table 2. For characterization of
plastic behaviour of material a set of cyclic tension compression
experiments were performed to obtain the properties needed for the
Chaboche hardening model. The adopted model consists of a
combined isotropic kinematic hardening formulation rstly intro-
duced by Armstrong and Fredrick [22], and subsequently modied by
Chaboche [18,2325]. The detailed description of its formulation is
provided in [26].
Low-cycle fatigue tests up to the stabilized cycle were performed
at different deformation intervals ( = 0.012, = 0.014,
= 0.016), all symmetric with respect to zero, using cylindrical
specimens following instructions of ASTM E606 [27]. Fig. 2 shows the
monotonic tension and stabilized cyclic tension compression curves
obtained from tests. The results represent that the material is
subjected to cyclic softening.
The results obtained for different deformation intervals were
elaborated to dene parameters of hardening model [18]. For
simplication it was supposed that the isotropic component can be
excluded from hardening model taking in to account that shot
peening process provides small number of loading unloading cycles
and consequently the inuence of cycling softening of material after
small number of cycles can be negligible in this case.
Determination of combined hardening parameters was performed
by Abaqus/CAE which provides a special function for determination of
hardening parameters based on experimental data of a stabilized
 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090 4083

Table 2
Monotonic mechanical characteristic of the material.

y 734 MPa
u 908 MPa
E 210.522 GPa
A% 14.8

of the values obtained in monotonic tension tests. The obtained

kinematic parameter values and the nal chosen properties are
reported in Table 3.
In order to assess the correspondence of the obtained data from
Abaqus and the experimental tests, a low cycle pushpull fatigue test
was simulated on a simple cylindrical model ( = 0.016). A
sinusoidal deformation as shown in Fig. 3a was applied to one end
of the cylinder. The model was comparatively coarsely meshed by
8 node brick elements as shown in Fig. 3b.
Comparison of the numerical and experimental results as shown in
Fig. 4 represent a good level of correspondence in the reconstruction
of simulated load cycle if the values in the rst plastic loading phase
are excluded. This difference is due to the fact that the numerical
model is not able to describe the plateau for yield stress in the rst
branch of plastic load as already demonstrated by Broggiato et al.
[26,28].
Moreover, regarding the fact that it is not possible to predict the
level of strain that the material will undergo during SP process, a set of
simulations are performed to check model transferability. The
simulation has been repeated applying different symmetric strain
cycles (max = 0.006 and max = 0.007) while in both cases the
hardening parameters obtained from max = 0.008 test were used for
characterization of material. In both simulations the comparison with
numerical data was in agreement with the experimental results. The
numerical results compared with experimental data as shown in Fig. 5
for max = 0.007, show the independence of the results from the input
data that is max = 0.008 experimental results.

2.3. Assessment of mesh convergence

For a 3D model with multiple impacts, the most challenging aspect

Fig. 1. 3D model used for multiple impact simulation.
cycle obtained in a deformation control condition with a symmetric
strain interval [15].
The stressstrain data was extracted from the test performed at
= 0.016. However instead of the stabilized cycle, the hysteresis
cycle obtained in the beginning of the test (after 3 cycles) was used in
order to make a more realistic model by simulating low number of
cycles. Also cyclic yield stress and elastic modulus were used instead
of simulation is the solution time; since in explicit FE simulation size
of smallest element determines total solution time, setting the
minimum element size would be of great importance.
In all SP simulations available in literature ne mesh is used in the
impact area and coarser elements in the area far from impact region.
Whereas the size of elements utilized in these simulations are
different and the authors did not manage to make out a clear
justication about the chosen element sizes. Moreover all the few
mesh convergence studies have focused on obtaining good resolution
just in terms of residual stress distribution under impact area. Frija et
al. simplied the case and carried out a sensitivity study to optimize
dimensions of the elements in rened zone, comparing stress results
with the elastic Hertz contact problem [29].
Zimmerman et al. used an element size equal to 1/15th of the
dimple diameter produced by a single shot impact for the modelled
shot diameter and velocity for which convergence was obtained in
terms of stresses [30]. Klemenz et al. also used Hertz analytical
solution for a purely elastic material behaviour to examine the
accuracy of FE mesh in the cases of single and double impact model

Table 1
Chemical composition of steel in mass density.

C (wt.%) Si (wt.%) max Mn (wt.%) P wt(wt.%) max S (wt.%) max Cr (wt.%) Mo (wt.%) Ni (wt.%)

0.350.43 0.02 0.40 0.03 0.50.8 0.04 0.025 + 0.005 .035 0.005 0.601.00 0.05 0.150.25 0.03 0.701.00 0.05
4084 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090

Fig. 2. Monotonic and cyclic behaviour of the material.

and for multiple impacts they chose the size of elements equal to 1/
10th of dimple size [31].
In this study the effect of element size not only on stress state but
also on the strains, particularly in terms of PEEQ is examined, since the
accumulated strain is a key parameter to be applied in assessment of
favourable conditions for formation of nanograins.
A single shot impact was simulated in order to estimate the
dimensions of plastic indentation generated on target surface. The
indentation diameter i.e. the diameter of the concaved surface due to
the single shot impact was measured to be 0.2439 mm. Then
convergence evaluation was performed changing element size in
impact zone of the target as a ratio of this dimple diameter. Results
indicate that with ne mesh convergence can be reached in terms of
residual stresses in the impact zone but even with very minute size of
elements it is not possible to nd absolute convergent results in the
case of PEEQ.
ALE (Arbitrary LagrangianEulerian) adaptive meshing was also
used to take benet from adaptive meshing options due to severe
deformations generated in impact zone. ALE adaptive meshing can
often maintain a sufciently rened mesh under severe material
deformation by allowing the mesh to move independently of the
underlying material; and it can be used as a continuous adaptive
meshing tool for problems undergoing large deformations [15].
However even with ALE meshing it was not possible to obtain entirely Fig. 3. a. Applied deformation function. b. Meshing of the simple model.
stabilized results in terms of PEEQ by decreasing element size even in
the order of 1/30th of dimple diameter.
As a solution to describe element size effect, the real maximum the value of maximum PEEQ for element size of zero is calculated to be
PEEQ was worked out by linear extrapolation to zero element size, a 0.611(mm/mm).
method used in similar situations in simulation of cold spray process SP simulations available in literature are usually performed using
which has many aspects in common with SP [32]. Considering the fact rigid shots for numerical simulations [17,3436]. There are few
that while the meshing is excessively ne, it is difcult to conduct the studies that have modelled deformable shots. In all these studies the
calculation owing to the limited system capability and time, Assadi et
al. and Li et al. concluded that the extrapolation of instable results to a
meshing size of zero could be used to stand for the real one while as
the mesh size is decreased variations of instable parameter are almost
linear [32,33]. In this study a similar extrapolation was performed on
values of PEEQ in order to assess the inuence of mesh size on
variation of PEEQ values. Variation of maximum PEEQ in the single
shot model with respect to element size is shown in Fig. 6. As it is
observed, the variations show a clear linear trend. By extrapolating,

Table 3
Cyclic mechanical characteristic of the material.

y0 = 359.26 MPa
E = 190 GPa
= 0.3
C = 169823
Fig. 4. Comparison of experimental and numerical results obtained for max = 0.008
= 501.87
(1st ve cycles).
 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090 4085

Fig. 5. Comparison of experimental and numerical results obtained for max = 0.007
(stabilized cycles).

size of shot element has often been chosen much coarser than the size
of elements in the impact region [10,14,29,31,3740]. In the present
work, for mesh convergence studies the same size of elements of
impact area was used for the shots.
Due to computational costs as a nal solution in the multiple
impacts simulation, the size of the elements on the shots was chosen
as 1/10th of dimple diameter and the size of elements in the impact
zone was selected 1/20th of dimple diameter. These choices result in
acceptable convergence in terms of stress and PEEQ in impact zone
and also avoid extremely long solution times common for ner size of
elements. It is also to be mentioned that in the nal stage the
discussed extrapolating process is performed on results in order to
increase the accuracy of denitive PEEQ values and eliminate effect of
element size.
Top view of elements in the target and close view of mesh density
in impact zone are shown respectively in Fig. 7a and b.

2.4. Coverage determination

To achieve a realistic model of SP process, the developed FE model

shall consist of a large number of identical shots impinging the target
with an impact angle of 90 at random locations and in random
sequences. Fig. 7. a. Top view of elements in the target model. b. Close view of elements in impact
zone.

Surface coverage is dened as the ratio of the area covered by

Fig. 6. Variation of maximum PEEQ just under impact by changing element size of
impact area.
plastic indentation to the whole surface area treated by SP expressed
in percentage. It is one of the most important input variables in SP
simulation and specifying a required level of coverage is a necessary
starting point. However no comprehensive investigations on how to
accurately model coverage have been carried out yet. In fact most of
the 3D multiple impact simulation models developed in recent times
did not focus on coverage but on the general understanding of how
the stress state develops during successive impacts [41].
Miao et al. suggested a method based on distribution of PEEQ to
evaluate percentage of coverage [42]. In their denition coverage is
approximated as the ratio of the number of nodes with PEEQ larger
than the PEEQ at the boundary of the indentation, to the total number
of nodes on the representative surface. The sensitivity of peening
coverage with respect to this denition is not investigated and is let
4086 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090
for future study [42]. In the case of the present study due to much ner
element size used for target model and also the higher impact energy
of the shots, the distribution of PEEQ in the impacted zone is
essentially different from the distribution that Miao et al. obtained in
their simulation, consequently their proposed denition is not
applicable anymore unless the criteria is modied by regarding not
only the PEEQ but also the displacement of the nodes in impact area.
Generally an efcient quantitative procedure shall be applied at
several locations in order to predict and control the coverage.
Theoretical and practical evidence point to an exponential approach
to 100% coverage as the amount of peening is increased [43].
According to these evidences, the established relationship between
coverage (Co%) and the ratio of total indent area to the target area (Ar)
is given in Eq. 1:
Ar
Co % = 1001e  1
SAE J2277 [44] sets a realistic target for coverage and denes full
coverage as being equivalent to 98% actual coverage. In our case

d 2
N =2
Ar = d 2 = 4, where N is the number of shots for full coverage
=2
Fig. 8. Arrangement of impacts on the area over which residual stress measurements
are conducted a) impact centre b) impact dimple.
(98.168%), d = 0.2439 mm and D is assumed to be 1 mm as shown in
Fig. 8. Accordingly full coverage is achieved when an indent ratio of 4
is reached. Regarding dimensions of target surface and single impact
indent area, 67 impacts are needed to obtain full coverage. Due to the
exponential nature of the formula it cannot describe coverage more
than 100% thus in order to increase the coverage percentage up to
200% the peening time and consequently shot numbers were doubled
and the simulation was performed considering 134 shots.
There is another denition put forward by Umemoto et al. [12,45]
that estimates coverage higher than 50% by multiplying SP time
required at the state of 50% coverage. The coverage used in this FEM
simulation corresponds to 575% the coverage of according to
description of Umemoto, for comparison.
Schwarzer showed that shot impacts one after the other has a
different inuence on the development of residual stresses with
respect to impacting simultaneously [14]. Due to this phenomenon
many researchers have avoided simultaneous impacts. Nevertheless
since considering simultaneous impact (not in the same position)
seems a realistic assumption, it was not avoided in this series of
simulations taking in to account that it would be also helpful to
decrease analysis time.
In order to obtain a realistic simulation, the impact sequence and
arrangement of these 134 shots was generated by using the random
function. The impact sequence which provides a dimple pattern of full
coverage on the surface is shown in Fig. 8. The arrangement is
repeated so as to simulate 134 shot impacts to the target surface. The
inner circular area was used for calculation of residual stress prole in
depth.
3. Experimental procedure
Steel (39NiCrMo3, according to the Italian nomenclature) discs
(80 mm radius and 8 mm thickness) were subjected to three different
approaches of ABSP. Using cast steel shots, specimens were treated
with particular combinations of SP parameters (alternative methods
(1) and (2)) which as presented in Table 4 are basically different from
typical parameters.
After SP process X-ray diffraction (XRD) analysis of surface layer of
the as-treated samples was performed using an AST X-Stress 3000 X-
ray diffractometer (radiation Cr K, irradiated area 1 mm2, sin2
method, 11 diffraction angles (2) scanned between 45 and 45).
To obtain the trend of residual stresses, measurements were carried
out in depth step by step removing a very thin layer of material using
an electro-polishing device. The results of the in-depth residual stress
measurements were corrected by using the method described in [46].
Microstructure observations were carried out through optical
microscopy (OM), scanning electron microscopy (SEM) and Trans-
mission electron microscopy (TEM). For SEM observations the
specimens were etched by 2% Nital. Generation of nanostructured
layers over specimens' surface was conrmed by microscopy
observations. Fig. 9 represents a cross-sectional SEM observation of
specimen A. From the overall observation of Fig. 9a distinct region
separated with sharp boundaries from the underlying work-hardened
layer is easily recognized on the top surface. This layer representing a
very dense structure near the surface as stated by Saitoh et al. is
Table 4
Aspects of the SSP processes.
Specimen Type of Shot Shot Almen Treatment
label treatment diameter speed intensity time [s]
[mm] [m/s] [0.0001 in.]
A Alternative method (1) 0.6 90 7C 450
B Alternative method (2) 0.6 90 7C 900
C Typical SP 0.42 30 15A 1030
 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090
Fig. 9. Cross-sectional SEM observation of specimen A.
considered to be the ne grained layer. It is reported that the
repeatedly folding of the deformed regions leads to the formation of
such ne grained regions at the surface of the specimens [47]. Fig. 10
shows TEM bright-eld image and the corresponding selected area
diffraction (SAD) pattern obtained at impacted surface of specimen A.
The bright-eld image represents irregularly shaped grains the
average size of which is measured to be 50 nm. As it is shown in the
upper left corner of Fig. 10, SAD pattern is composed of partially
continuous diffraction rings, which conrm that the as-received large
crystalline grains have been broken down to nanograins at this region.
Theses diffraction patterns also clearly indicate that nanograins at
impacted surface have completely random orientations with high-
angle grain boundaries. More details of surface characterization of
specimens are available in authors' other paper [48]. Results obtained
from X-ray measurements are presented in the following sections in
terms of residual stresses and FWHM.
4. Results and discussion
4.1. Analysis of residual stress prole
Due to random sequence of multiple impacts the target surface
doesn't show an absolutely uniform distribution of residual stresses
Fig. 10. TEM image of the impacted surface of specimen A showing the formation of
nanograins, insert is the correspondent SAD pattern.
4087
and deformation. A difference (even if in some cases negligible)
always exists between the results obtained in dissimilar points on the
treated surface. Both experiments [49] and numerical simulations
have conrmed this irregularity.
XRD measurements imply that a considerable depth of material is
characterized with signicant compressive residual stresses in both
specimens. This is veried by the comparison shown in Fig. 11
between alternatively treated specimens and a similar specimen
treated with typical peening parameters.
Stresses measured by X-ray diffraction are usually an average
value of stresses in the area covered by X-ray. Comparison of multiple
impact simulation results with the XRD measurements was per-
formed using a method suggested by Schwarzer [14]. In this method
mean of residual stresses in an area equal to the area affected by X-ray
is calculated at each depth. The chosen area in the nite element
model is a circle at the centre of impact zone shown in Fig. 8. Within
this area an average of stress components at Gauss points of adjacent
elements parallel to the surface is calculated for each element layer in
order to simulate the residual stress prole also after layer removal
where the residual stresses were determined at new surfaces.
Numerical results are compared with XRD measurements in
Fig. 12. FEM results have satisfactory consistence with XRD measure-
ments. The slight differences observed between XRD measurements
and numerical results can be partially related to the signicant
roughness over the surface of specimens due to severe SP. The high
points in the rough surface contribute more to the diffraction pattern
and consequently they introduce some errors to the XRD measure-
ments especially on the rst layers [50]. In Fig. 12 it is also interesting
to note that increasing treatment time has not resulted in notable
changes neither in experimental approach nor in the numerical
simulation. It can be concluded that changing process time at least in
this scale is not notably affecting residual stresses in terms of on-
surface and maximum value. The main discernible difference (Fig. 11)
is the increment of the depth characterized by high compressive
residual stresses in case of the specimen B which has a longer
treatment time.
4.2. Estimation of work-hardened layer thickness
Another parameter measured by X-ray diffraction is FWHM. This
parameter which is shown in Fig. 13 represents the full width of the
diffraction peak at half of the maximum value of diffraction. FWHM,
assumed as an index of hardening of the material, is related to the
grain distortion, dislocation density and the so called type II micro
residual stresses [51].
As shown in Fig. 13 maximum FWHM is observed on the surface,
declining down to the depth. Also in terms of FWHM in the surface
region there is no considerable difference between the results for
specimen A and B.
Fig. 11. Distribution of residual stresses obtained by XRD as a function of position
measured from impacted surface in steel specimens.
4088 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090
Fig. 12. Distribution of residual stresses obtained by XRD and FEM as a function of
position measured from impacted surface.
Fig. 14 shows the effective plastic strain as a function of the
location measured from the impacted surface obtained from the FE
analysis after 67 and 134 impacts, calculated with the same averaging
method employed for stresses. The equivalent plastic strain in Abaqus
pl
is dened as pl j 0 + t0 dt, where pl j 0 is the initial equivalent
pl
plastic strain. The denition of
depends on the material model. For
q produce nanocrystals [12].
pl pl pl
classical metals (Von-Mises) plasticity is used: = 23 : [15].
As it is observed in Fig. 14 the amount of PEEQ is considerably growing
with increasing the number of impacts, even if the thickness of work-
hardened layer does not change signicantly. It is also interesting to
note that the thickness of the work-hardened layer obtained from
FEM that is the layer characterized with non-zero PEEQ values (almost
0.35 mm) is nearly the same as the depth which shows considerably
increased FWHM values in comparison with core material in
experimental results (approximately 0.30.35 mm in Fig. 13).
4.3. Assessment of nanocrystalline structure formation and thickness of
NC layer
To optimize the known processes or develop new processes for
production of nanocrystals, it is of great importance to identify the
parameters necessary for structural evolution and generation of
nanograins.
High-pressure torsion [5255], ball milling [5661], sliding wear
[62,63], drilling [6467], ball drop test [68,69], shot blasting and
annealing [7073], ultrasonic shot peening [7478] and air blast shot
Fig. 13. FWHM prole obtained by XRD as a function of position measured from
impacted surface.
Fig. 14. Equivalent plastic strain prole within the target measured from impacted
surface.
peening [12,65,77,7982] are demonstrated to be possible processes
to generate NC. The deformation conditions common to these
processes are considered to be large strain and high strain rate.
Reported amount of strain applied to obtain NC varies depending on
the deformation technique and material. However, imposing large
strains seems to be the most important condition favourable to
Studying particle impact and ABSP, Umemoto et al. proposed the
minimum necessary amount of PEEQ as the key parameter for
formation of nanocrystalline structure, to be around 78 (mm/mm).
This theory was tried on PEEQ results extracted from numerical
simulation performed in this study. Two analyses of multiple impacts
were performed considering different element sizes in impact region:
1/10th and 1/20th of dimple diameter. The same averaging method
used for calculation of stresses was used to estimate the average value
of PEEQ on impact surface.
According to the fact that as the mesh size is decreased, variations
of PEEQ are almost linear, performing the zero element size linear
extrapolation, the amount of PEEQ independent from element size
was estimated as shown in Fig. 15. The measured amount of PEEQ on
the surface layer was found to be 8.49 mm/mm. Four other
simulations have been performed developing different random shot
sequences. The values of the surface PEEQ extrapolated for zero
element size was found to be always in the range 7.88.5 which
satises the criteria suggested by Umemoto et al. as the condition
favourable for formation of nanocrystals.
Fig. 15. Estimation of average value of PEEQ on the impact surface independent from
element size.
 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090
Coverage obtained in experiment according to optical measure-
ments was reported to be 100% after 30 s treatment; consequently
after 450 s and 900 s the obtained coverage can be decoded
respectively as 1500% and 3000% according to the experimental
coverage interpretation of Umemoto.
Due to computational costs this PEEQ value (8.49 mm/mm), which
satises Umemoto limit for formation of NC, is calculated just after
134 impacts (i.e. 200% coverage according to the exponential
approach, equivalent to 575% coverage according to Umemoto
denition). In view of the fact that the coverage in experiment has
been much more, and taking into account that the values of PEEQ rise
by increasing the coverage percentage, the average values of PEEQ in
the surface layer of the experimentally treated specimens would be
even more than 8.49 mm/mm. According to the obtained results, it
can be inferred that coverage up to 500% (Umemoto denition) can be
enough to reach the limit value of PEEQ that is around 78 (mm/mm).
Thickness of the nanocrystallized layer after 134 impacts was also
estimated as the thickness of layer with PEEQ value over 7. The
thickness was measured to be 1 m on the specimen surface.
There are studies that consider grain renement process to be
accelerated by other parameters such as high strain rate, low
temperature, multidirectional deformation, existence of impurities,
alloying elements and [13,45,55,83]. Role of such other parameters
are the subjects of future studies.
5. Conclusions
A comprehensive multiple impact model of severe ABSP based on
nite element simulation has been developed. The model is aimed at
the description and the characterization of the surface layer of
material after SSP. Experimental analyses and tests were developed to
assess the accuracy of the numerical model.
On the basis of the obtained results the following conclusions can
be drawn:
1. A FE based procedure is proposed for predicting the effect of SSP on
metal surfaces: after a ne convergence study of the parameters
dening the mesh, a relation between the number of simulated
impacts and the real treatment time and coverage is established
and supported by the comparison with the experimental results;
2. The numerical model of SSP is able to predict the induced residual
stresses and the depth of the surface layer affected by SSP: the
experimental XRD measurements are in good agreement with the
numerical results;
3. A criterion to assess the formation of a nanostructured layer of
material from FE results is presented: it is based on the
accumulated equivalent plastic strain (PEEQ), determined by
using the so called zero element size linear extrapolation.
Comparison with the experimental data afrms that the criterion
is able to evaluate the combination of treatment parameters for
obtaining nanostructured layer of material and also to estimate the
depth of this latter with good approximation. PEEQ, recognized as
common favourable mechanism for generation of nanograins, has
been extracted from simulations;
4. Experimental surface characterization of steel specimens, shot
peened with similar set of peening parameters used in numerical
simulation, conrmed the numerical results: on this basis, the
proposed numerical model can be considered as a useful tool to
predict the formation of nanostructured layer of material.
5. Further studies on other parameters such as strain rate, temper-
ature, and deformation type, that seem to inuence the nanograin
formation process are required to complete the present study.
Taking all effective parameters into account, may provide a more
accurate correlation between the SSP parameters combination and
the expected dimension of the grain size obtained after SSP.
4089
Acknowledgment
We would like to thank Eng. Michele Bandini for the execution of
shot peening in Peen Service srl and his useful suggestions for peening
parameters.
References
[1] J.O. Almen, P.H. Black, Residual Stresses and Fatigue in Metals, McGraw-Hill Publ.
Company, 1963.
[2] K.J. Marsh, Shot Peening: Techniques and Applications, EMAS, London, 1993.
[3] A. Blarasin, M. Guagliano, L. Vergani, Fatigue Fract. Eng. Mater. Struct. 20 (1997)
1171.
[4] L. Wagner, Mater. Sci. Eng. A 263 (1999) 210.
[5] I. Altenberger, B. Scholtes, U. Martin, H. Oettel, Mater. Sci. Eng. A 264 (1999) 1.
[6] V. Schulze, Modern Mechanical Surface Treatment, States, Stability, Effects,
WILEY-VCH, 2006.
[7] Y. Ochi, K. Masaki, T. Matsumura, T. Sekino, Int. J. Fatigue 23 (2001) 441.
[8] S. Bagheri, M. Guagliano, Surf. Eng. 25 (2009) 1.
[9] K. Dai, J. Villegas, Z. Stone, L. Shaw, Acta Mater. 52 (2004) 5771.
[10] K. Dai, L. Shaw, Mater. Sci. Eng. A 463 (2007) 46.
[11] X. Ma, W. Wang, W. Yang, Acta Mech. Sin. 19-2 (2003) 172.
[12] M. Umemoto, Y. Todaka, K. Tsuchiya, Mater. Trans. 44-7 (2003) 1488.
[13] M. Umemoto, Mater. Trans. 44-10 (2003) 1900.
[14] J. Schwarzer, V. Schulze, O. Vohringer, Proc. of 8th International Conference of
Shot Peening (ICSP8), 2002, p. 507, Munich (Germany).
[15] Abaqus Analysis User's Manual (2007), version 6.7.
[16] T. Kim, J.H. Lee, H. Lee, S.K. Cheong, Mater. Des. 31 (2010) 50.
[17] S.T.S. Al-Hassani, Proc. of 7th International Conference of Shot Peening (ICSP7),
1999, p. 217, Warsaw (Poland).
[18] J. Lemaitre, J.L. Caboche, Mcanique des matriaux solides, DUNOD, Paris, 1998.
[19] G.R. Johnson, W.H. Cook, Proc. of the 7th International Symposium on Ballistic,
1973, p. 541, Huge (Netherlands).
[20] ASTM E8M. Standard Test Methods for Tension Testing of Metallic Materials.
[21] ASTM A370-05, Standard Test Methods and Denitions for Mechanical Testing of
Steel Products.
[22] P.J. Armstrong, C.O. Fredrick, CEGB Report, RD/B/N731, Berkeley Nuclear
Laboratories, 1996.
[23] J.L. Caboche, Int. J. Plast. 2 (1986) 149.
[24] J.L. Caboche, Int. J. Plast. 5 (1989) 247.
[25] J.L. Caboche, Int. J. Plast. 24 (2008) 1642.
[26] G.B. Broggiato, F. Campana, L. Cortese, Meccanica 43 (2008) 115.
[27] ASTM E606. Standard Practice for Strain-Controlled Fatigue Testing.
[28] G.B. Broggiato, F. Campana, L. Cortese, Proceeding of XXXV AIAS National
Conference, 2006 Ancona (Italy).
[29] M. Firja, T. Hassine, R. Fathallah, C. Bouraoui, A. Dogui, Mater. Sci. Eng. A 426
(2006) 173.
[30] M. Zimmerman, V. Schulze, H.U. Baron, D. Lohe, Proc. of the 10th International
Conference on Shot Peening (ICSP10), 2008, p. 63, Tokyo (Japan).
[31] M. Klemenz, V. Schulze, O. Vohringer, D. Lohe, Mater. Sci. Forum 524525 (2006)
349.
[32] W.Y. Li, H. Liao, C. Li, G. Li, C. Coddet, X. Wang, Appl. Surf. Sci. 253 (2006) 2852.
[33] H. Assadi, F. Gartner, T. Stoltenhoff, H. Kreye, Acta Mater. 51 (2003) 4379.
[34] S.A. Meguid, G. Shagal, J.C. Stranart, J. Daly, Finite Elem. Anal. Des. 31 (1999) 179.
[35] M. Guagliano, L. Vergani, M. Bandini, F. Gili, Proceeding of 7th International
Conference of Shot Peening (ICSP7), 1999, p. 274, Warsaw (Poland).
[36] T. Hong, J.Y. Ooi, J. Favier, B. Shaw, Proceeding of 9th International Conference of
Shot Peening (ICSP9), 2005, Paris (France).
[37] S.A. Meguid, G. Shagal, J.C. Stranart, Int. J. Impact Eng. 27 (2002) 119.
[38] G.H. Majzoobi, R. Azizi, A. Alavinia, J. Mater. Process. Technol. 164165 (2005)
1226.
[39] J. Schwarzer, V. Schulze, O. Vohringer, Mater. Sci. Forum 426432 (2003) 3951.
[40] M. Meo, R. Vignjevic, Adv. Eng. Softw. 34 (2003) 569.
[41] V. Schulze, M. Klemenz, M. Zimmerman, Proceedings of the 10th International
Conference on Shot Peening (ICSP10), 2008, p. 53, Tokyo (Japan).
[42] H.Y. Miao, S. Larose, C. Perron, M. Lvesque, Adv. Eng. Softw. 40-10 (2009) 1023.
[43] D. Kirk, The Shot Peener 23-2 (2009) 24.
[44] SAE Standard J2277 Shot Peening Coverage, Developed by Surface Enhancement
Committee, Publication Date: Jan 1, 2003.
[45] Y. Todaka, M. Umemoto, Y. Watanabe, A. Yamazaki, C. Wang, K. Tsuchiya, ISIJ Int.
47-1 (2007) 157.
[46] M.G. Moore, W.P. Evans, SAE Trans. 66 (1958) 340.
[47] H. Saitoh, T. Ochi, M. Kubota, Proceedings of the 10th International Conference on
Shot Peening (ICSP10), 2008, p. 488, Tokyo (Japan).
[48] S. Bagherifard, I. Fernndez Pariente, M. Guagliano, 12th International Conference
on Fracture ICF (proceeding on CD-Rom), 2009, Ottawa, Ontario (Canada).
[49] B. Scholtes, Eigenspannungen in mechanisch raiidschichtverformten Werkstoff-
zust3nden, Ursachen-Ermittlung-Bewertung, DGM-Informationsgesellschaft,
Oberursel, 1990.
[50] B.D. Cullity, S.R. Stock, Elements of X-ray Diffraction, Prentice Hall Inc., 2001
[51] I.C. Noyan, J.B. Cohen, Residual Stress-Measurement by Diffraction and Interpre-
tation, Springer-Verlag, New York, 1987.
[52] A.V. Korznikov, I.M. Safarov, V.P. Pilyugin, R.Z. Valiev, Nanostruct. Mater. 4 (1994)
156.
4090 S. Bagherifard et al. / Surface & Coatings Technology 204 (2010) 40814090

[53] R.Z. Valiev, Yu.V. Ivanisenko, E.F. Rauch, B. Baudelet, Acta Mater. 44 (1996) 4705. [69] M. Umemoto, X.J. Hao, T. Yasuda, K. Tsuchiya, Mater. Trans. 43 (2002) 2536.
[54] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000) 103. [70] X.Y. Wang, D.Y. Li, Electrochim. Acta 47 (2002) 3939.
[55] Y. Todaka, M. Umemoto, J. Yin, Zh. Liu, K. Tsuchiya, Mater. Sci. Eng. A 462 (2007) [71] L. Wang, D.Y. Li, Surf. Coat. Technol. 167 (2003) 188.
264. [72] L. Wang, D.Y. Li, Wear 255 (2003) 836.
[56] H.J. Fecht, E. Hellstern, Z. Fu, W.L. Johnson, Metall. Mater. Trans. A 21 (1990) 2333. [73] X.P. Jiang, X.Y. Wang, J.X. Li, D.Y. Li, C.S. Manc, M.J. Shepard, T. Zhai, Mater. Sci. Eng.
[57] C.H. Moelle, H.J. Fecht, Therm. Nanostruct. Mater. 6 (1995) 421. A 429 (2006) 30.
[58] J. Yin, M. Umemoto, Z.G. Liu, K. Tsuchiya, ISIJ Int. 41 (2001) 1389. [74] N.R. Tao, M.L. Sui, J. Lu, K. Lua, Nanostruct. Mater. 11 (1999) 433.
[59] Y. Todaka, M. Umemoto, K. Tsuchiya, ISIJ Int. 42 (2002) 1429. [75] G. Liu, J. Lu, K. Lu, Mater. Sci. Eng. A 286 (2000) 91.
[60] Y. Todaka, P.G. McCormick, K. Tsuchiya, M. Umemoto, Mater. Trans. 43 (2002) [76] Z.G. Liu, H.J. Fecht, M. Umemoto, Mater. Sci. Eng. A 375377 (2004) 839.
667. [77] Y. Todaka, M. Umemoto, K. Tsuchiya, Mater. Trans. 45 (2004) 376.
[61] L. Shaw, H. Lou, J. Mater. Sci. 42 (2007) 1415. [78] F.A. Guo, N. Trannoy, J. Lu, Mater. Sci. Eng. A 369 (2004) 36.
[62] P. Heilmann, W.A.T. Clark, D.A. Rigney, Acta Metall. 31 (1983) 1293. [79] Y. Todaka, M. Umemoto, Y. Watanabe, K. Tsuchiya, Trans. Mater. Res. Soc. Jpn. 29
[63] D.A. Hughes, D.B. Dawson, J.S. Korellis, L.I. Weingarten, Wear 181183 (1995) 458. (2004) 3523.
[64] Y. Todaka, M. Umemoto, S. Tanaka, K. Tsuchiya, Mater. Trans. 45 (2004) 2209. [80] K.S. Raja, S.A. Namjoshi, M. Misra, Mater. Lett. 59 (2005) 570.
[65] Y. Todaka, M. Umemoto, J. Li, K. Tsuchiya, Rev. Adv. Mater. Sci. 10 (2005) 409. [81] M. Umemoto, Y. Todaka, J. Metastable Nanocryst. Mater. 2425 (2005) 571.
[66] M. Umemoto, Y. Todaka, J. Li, K. Tsuchiya, Mater. Sci. Forum 539543 (2007) 2787. [82] T. Wang, J. Yu, B. Dong, Surf. Coat. Technol. 200 (2006) 4777.
[67] J.G. Li, M. Umemoto, Y. Todaka, K. Tsuchiya, J. Alloy. Comp. 434435 (2007) 290. [83] R. Pippan, F. Wetscher, M. Hafok, A. Vorhauer, I. Sabirov, Adv. Eng. Mater. 811
[68] M. Umemoto, B. Huang, K. Tsuchiya, N. Suzuki, Scr. Mater. 46 (2002) 383. (2006) 1046.

View publication stats