Finite Elements in Analysis and Design 31 (1999) 179—191

Three-dimensional dynamic finite element analysis of shot-peening

induced residual stresses
S.A. Meguid *, G. Shagal , J.C. Stranart , J. Daly

Engineering Mechanics & Design Laboratory, Department of Mechanical and Industrial Engineering, University of Toronto,
5 King+s College Road, Toronto, Ontario, M5S 3G8 Canada

Metal Improvement Company, 10 Forest Avenue, Paramus, New Jersey, 07652 USA
Received 27 July 1998; accepted 27 July 1998

Abstract

This investigation is devoted to the modelling and simulation of the residual stress field resulting from the shot-peening
process. In this dynamic elasto-plastic analysis, single and twin spherical indentations were examined using the finite
element method. The contact between the shots and the target was modelled using contact elements of the penalty
function type. Attention was devoted to three related issues. The first is concerned with the effect of the shot velocity, size
and shape upon the plastic zone development and growth, and unloading residual stresses. The second with the effect of
the separation distance between two impinging shots upon the equivalent stress trajectories and unloading residual
stresses. Finally, the study examines the effect of the strain-hardening rate of the target upon the development and
the spread of the plastic zone. The results reveal the important role played by the shot and target characteristics upon the
quality of the treatment, as measured by the mechanically induced residual stresses.  1999 Elsevier Science B.V.
All rights reserved.

Keywords: Finite element analysis; Dynamic; Three dimensional; Nonlinear; Shot-peening; Residual stresses; Plasticity;
Co-indentation

1. Introduction

Shot-peening is a cold-working process used mainly to improve the fatigue life of metallic
components [1—3]. The results are accomplished by bombarding the surface of the component with
small spherical shots made of hardened cast-steel, conditioned cut-wire, glass or ceramic beads at
a relatively high velocity (40—70 m/s). After contact between the target and the shot has ceased,
a small plastic indentation is formed causing stretching of the top layers of the exposed sur-
face. Upon unloading, the elastically stressed sub-surface layers tend to recover their original

* Corresponding author.

0168-874X/99/$ — see front matter  1999 Elsevier Science B.V. All rights reserved
PII: S 0 1 6 8 - 8 7 4 X ( 9 8 ) 0 0 0 5 7 - 2
180 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191

dimensions, but the continuity of the material in both zones, the elastic and the plastic, does not
allow this to occur. Consequently, a compressive residual stress field followed by tensile is trapped
in the bounded solid.
This surface compressive residual stress field is highly effective in preventing premature failure
under conditions of cyclic loading, since fatigue generally propagates from the upper most surface
of the component, and usually starts in a region which is subjected to high tensile stresses. This is
indeed the motivation behind the use of the treatment in the aerospace, automotive and power
generation industries. In the case of the aerospace industry, the treatment leads to a reduction in
structural weight for a specified reliability level. Typically, turbine and compressor discs and the
attached blades, propellers and harmonic drives, main rotor spindles and main and nose landing
gear components are shot-peened. In the automotive industry, it means that relatively small
low-cost components can be upgraded for conservative operation at stress levels that would
represent poor practice without the treatment. Springs, gears of all types and shapes, connecting
rods, cam shafts and torsion bars are examples of components that can be upgraded by the
treatment. The ability to upgrade the mechanical properties of a component by shot-peening offers
obvious opportunities in the correction of undersized components when, e.g., fatigue failure occurs
after a product is standardized or in field service by the power generation industry.
The effectiveness of the shot-peening treatment depends to a large extent on peening intensity and
peening coverage. Peening intensity is a measure of the consistency of the treatment and of the
plastic dissipation of the impinging shots. Peening coverage, on the other hand, is a measure of the
area covered by peening indentations. Both intensity and coverage depend on numerous variables,
including: workpiece characteristics, shot characteristics, flow conditions, jet obliquity, stand-off
distance and exposure time. Peening intensity is measured using the arc-height resulting from
peening standard spring steel strips, known as Almen strips. Coverage, on the other hand, is
typically measured using a magnifying glass. Further details concerning these variables and the
techniques adopted in measuring them can be found, for example, in references: Desvignes et al. [4]
and Meguid [5].
The modelling and simulation of the shot-peening process has received some attention from the
scientific community. This includes the contributions made by Shaw and De Salvo [6], Meguid
et al. [7—9], Khabou et al. [10] and Li et al. [11] of the process using quasistatic analysis. The
dynamic modelling of a single shot was initially conducted by Johnson [12] using a pseudo-
dynamic approach. In his approach, Johnson [12] took into account only the inertial properties of
the shot. As a result, a relationship between the depth of the plastic zone and the shot parameters,
such as radius, mass and velocity, was obtained. This relationship was later validated by Clausen
[13] and Iida [14]. Edberg et al. [15] conducted dynamic three-dimensional finite element analysis
of a single shot impinging viscoplastic and elastoplastic materials. Their results showed that the
two material models give similar residual stress distributions.
Review of most of the existing work in this area, however, reveals a number of shortcomings.
These include: (i) the selection of unrealistic or irrelevant peening parameters, (ii) the focus of
attention to the peculiarities of the specific models rather than upon the relevance of the work to
the process, and (iii) the use of improper and inaccurate modelling techniques. It is believed that the
present study addresses these issues more carefully.
The objective of this study is therefore to evaluate the elasto-plastic dynamic indentation
behaviour of a bounded solid by a single and multiple spherical shots. The effect of the separation
 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191 181

distance and shot velocity as well as the mechanical properties of the bounded solid upon the
resulting residual stress field are examined and discussed. In addition to being of importance to the
shot-peening process, this study is of value to a number of related areas, such as friction and wear,
ball bearing technology, abrasive erosion and surface finish.
This article is divided into 4 sections. Following this introduction, Section 2 describes the finite
element model used. In Section 3, we verify the finite element results and discuss the effect of the
pertinent parameters upon the plastic zone developed and unloading residual stresses. Finally, in
Section 4, we conclude the work.

2. Finite element modelling

Two models are considered using the commercial finite element code ANSYS 5.3. The first is
concerned with a single shot and the second with twin shots, each of radius R, impinging a metallic
target at normal incidence. Due to symmetry, only one-quarter of the single shot model was
discretized, as depicted in Fig. 1a. The following dimensions were selected for the target: width
¼"7R, height H"4R and breadth B"5R. These dimensions were carefully determined as
a result of numerous axisymmetric model analyses to establish the effect of the boundary.
Both eight-noded brick and four-noded tetrahedral finite elements, with large strain and
displacement capabilities, were used to discretize the target. In view of their higher accuracy,
brick elements were used in the impact region around the common normal. The remaining regions
of the target were conveniently discretized using tetrahedral elements. Convergence tests were
conducted using the meshes depicted in Fig. 1b—Fig. 1d, revealing some, but not problematic,
distortion.
To model the shot-target interface, contact element were used on both bodies. The three-
dimensional contact elements adopt a contact node — target segment approach in conjunction with
the penalty function method [16]. The contact nodes were created on the top surface of the peened
target around the common normal. The target segments were generated only on the lower half of
the shot surface (Fig. 1), since it was anticipated that contact would take place only in that area.
Elastic Coulomb’s law, with friction coefficient k"0.25, was used.
A major problem in the implementation of contact elements is the assignment of values to the
normal (K ) and tangential (K ) stiffnesses, which govern the convergence and accuracy of the
 
solution [17]. The use of excessively high values of K and K result in ill-conditioned global
 
stiffness matrix, leading to numerical errors and divergence. On the other hand, the use of smaller
values of K and K results could lead to convergence to the wrong solution, allowing for
 
interpenetration and incorrect estimates of the stick and slip regions. In order to avoid these
problems, several runs were conducted in which the values of K and K were tested. The values
 
assigned to our finite element model were 10 N/m for K and 10 N/m for K .
 
The study was further extended to examine the effect of a twin shot impact upon the residual
stress field. Fig. 2a shows an outline of two spheres each of radius R impinging a target. The
separation distance between these spheres is 2C. Fig. 2b shows the discretized geometry of the
target using the same element types discussed earlier. The boundary conditions imposed on the
above models were as follows: symmetry boundary conditions on the z"0 and x"0 planes, and
fully constrained workpiece bottom surface. For the purpose of this study, the shot was assumed to
182 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191

Fig. 1. Geometry and discretized models used in the single-shot model: (a) one-quarter of geometry, and (b—d) three
discretized geometries of target with different mesh densities.

be rigid. This rigidity stems from the relatively high yield and hardness values of typical steel shots
compared to the target material. It is worth noting that the effect of strain rate and elastic wave
propagation was not considered in the study.

3. Results and discussions

3.1. The single-shot model

A number of preliminary runs were conducted to establish the appropriate mesh design for the
model. The dynamic analysis was carried out using Newmark implicit time-integration scheme
 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191 183

Fig. 2. Geometry and discretized model used in dynamic co-indentation studies: (a) geometry and notation used, and
(b) discretized geometry showing contact elements.

with adjustable time steps. The total integration time was t "10 ls. As a result of these runs, the
mesh given in Fig. 1b was used for the single-shot dynamic analysis.
In order to verify our model, a comparison with the earlier work of Edberg et al. [15] was
conducted. The target and shot characteristics considered for verification were the same as those
selected by Edberg et al. [15]. It should be recalled that these parameters do not represent true
peening parameters, but this case was the only comparable published work available. While not
identical, the results demonstrated substantial agreement.
184 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191

Three aspects of the single shot model are examined in the current work. The first is concerned
with the effect of the shot velocity, the second with the shot size and shape, the third with the strain
hardening coefficient upon the plastic zone development and unloading residual stresses.

3.1.1. Effect of shot velocity

The single-shot finite element analysis was conducted on high-strength steel targets of mass
density o"7800 kg/m and a modulus of elasticity E"200 GPa. The elasto-plastic behaviour
was assumed to be bilinear with an initial yield stress p "600 MPa and a strain-hardening rate

H"800 MPa. The shot used was assumed rigid with radius R and velocity v.
Three different impact velocities were used: 50, 75 and 100 m/s. The model was conducted using
a steel shot with a radius R"0.5 mm. Fig. 3a and Fig. 3b show the variation of the equivalent
plastic strain e and normalized unloading residual stress p /p , beneath the centre line of the
 VV 
shot, with the depth of the treated target for the three impact speeds selected. These figures reveal
that 100% increase in the shot velocity results in: (i) an increase in the maximum equivalent plastic
strain by 65%, (ii) a significant increase in the depth of the compressed layer, and (iii) an
insignificant change in the position and the magnitude of the maximum residual p stress. It is also
VV
interesting to note that the change in the residual stress from compression to tension coincides with
the depth of the plastic zone.

3.1.2. Effect of shot size and shape

Fig. 4 shows the p stress trajectories for the entire target, while Fig. 5 show the corresponding
VV
normalized residual stress p /p beneath the centre line of the shot for different shot radii. The shot
VV 
was assigned a velocity v"75 m/s and the target material was assigned a strain hardening rate
H"800 MPa. Interestingly, the results indicate that in spite of the increase in the volume of the
plastically deformed material with the increase in the shot radius, the magnitude of the surface and

Fig. 3. Effect of shot velocity upon: (a) equivalent plastic strain versus depth, and (b) normalized transverse residual
stress versus depth for a shot radius R"0.5 mm and target strain hardening rate H"800 MPa.
 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191 185

Fig. 4. Effect of shot size upon the transverse residual stress trajectories: (a) shot radius R"0.25 mm, (b) R"0.5 mm
and (c) R"1.0 mm for a shot velocity v"75 m/s and target strain hardening rate H"800 MPa.

maximum subsurface residual stress p remains unchanged. However, the position at which the
VV
maximum subsurface stress takes place increases with the increase in the shot radius.
The study also accounted for the effect of the shape of the shot upon the normalized residual
p /p variation with depth. In this case, the shot was modelled in the form of an ellipsoid with
VV 
major and minor dimensions being a and b, respectively. Mesh refinement was used to ensure the
validity of the results. Fig. 6 shows clearly the effect of the aspect ratio (a/b), for the same mass of
the shot, upon the residual stress field. It is worth noting that the decrease in the aspect ratio results
in an increase in the maximum subsurface residual stress p . This is, however, coupled with an
VV
increase in the surface roughness of the target, as evidenced by Fig. 7a—Fig. 7c, where a pro-
nounced pile-up is apparent for the case when a/b"0.5. Whilst pile-up does occur to peened
components, the extent illustrated here is unusual. In typical components, multiple indentations
resulting from full coverage will reduce these asperities.
186 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191

Fig. 5. Effect of shot size upon normalized transverse residual stress residual stress distribution beneath the centre line of
the shot for a velocity v"75 m/s and target strain hardening rate H"800 MPa.

Fig. 6. Effect of shot shape upon normalized transverse residual stress distribution beneath the centre line of the shot for
a velocity v"75 m/s and target strain hardening rate H"800 MPa.
 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191 187

Fig. 7. Effect of shot shape upon transverse residual stress trajectories: (a) shot aspect ratio a/b"2, (b) a/b"1, and
(c) a/b"0.5, for the same shot mass with a velocity v"75 m/s and target strain hardening rate H"800 MPa.

3.1.3. Effect of target hardening characteristics

The effect of target hardening characteristics was examined by assigning different hardening
coefficients H to the assumed bilinear material model. The shot radius and velocity selected were
0.5 mm and 75 m/s, respectively. Fig. 8 shows the residual stress variation with the depth of the
target for three different values of hardening rates, 50, 800 and 1600 MPa. The following observa-
tions can be made from that figure. Increasing the hardening coefficient of the material results in an
increase in the depth of the compressed layer; little or no change in the magnitude of the maximum
subsurface residual stress, p ; and a decrease in the surface residual stress.
VV
3.2. The dynamic co-indentation model

The work was further extended to account for the effect of two shots impinging the target
simultaneously. Special attention was devoted to examining the effect of the separation distance
188 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191

Fig. 8. Effect of target strain hardening rate upon normalized transverse residual stress distribution beneath the centre
line of the shot for a velocity v"75 m/s and radius R"0.5 mm.

Fig. 9. Effect of separation distance between shots upon the normalized transverse residual stress field: (a) at the centre
line of the target, and (b) beneath the centre line of the shot for a shot radius R"0.5 mm, velocity v"50 m/s,
H"800 MPa.

between the shots, impact velocity and target hardening characteristics upon the resulting residual
stress field. The finite element models corresponding to these dynamic co-indentation studies are
similar to those adopted in the single shot investigation.
Let us now focus our attention to the dynamic spherical co-indentation results. Fig. 9 shows the
effect of the separation distance upon the residual stress, p . The analysis was conducted using the
VV
 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191 189

mechanical properties outlined in the single shot model with separation ratios C/R"1, 1.5 and 2.
The effect of the separation ratio upon the residual stress field is evident at the centre line of the
target, where interactions from the co-indenting spheres take place. However, insignificant interac-
tions take place beneath the centre line of each shot, as evidenced by the distribution of the residual
stress there.
The effect of the kinetic energy of the impinging shots for different separation ratios is shown in
Fig. 10a and Fig. 11a at the centreline of the target and in Fig. 10b and Fig. 11b beneath the
centreline of the shot. The residual stress distributions of Fig. 10b and Fig. 11b resulting from

Fig. 10. Effect of shot velocity upon the normalized transverse residual stress field: (a) at the centre line of the target, and
(b) beneath the centre line of the shot for a shot radius R"0.5 mm, velocity v"50 m/s, H"800 MPa and separation
distance C/R"1.

Fig. 11. Effect of shot velocity upon the normalized transverse residual stress field: (a) at the centre line of the target, and
(b) beneath the centre line of the shot for a shot radius R"0.5 mm, velocity v"50 m/s, H"800 MPa and separation
distance C/R"2.
190 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191

dynamic co-indentation are very similar to those obtained from the single-shot model. However,
the single-shot model is incapable of predicting the relatively large compressive residual stresses
present at the centre line of the target between the co-indenting shots, depicted in Fig. 10a and
Fig. 11a. These figures indicate that the maximum compressive stresses occur at the surface of the
target. The transition from compression to tension occurs at comparable distances to those
observed beneath the shot. The effect of the strain-hardening characteristics of the target was very
similar to that observed in the single-shot model.

4. Conclusions

A three-dimensional finite element model was developed to simulate the shot-peening process.
Dynamic single and twin elasto-plastic spherical indentations were examined using rigid spherical
shots and metallic targets. The effect of shot velocity, size and shape upon the equivalent stress
trajectories, equivalent plastic strains and unloading residual stresses of a target exhibiting bilinear
material behaviour is examined and discussed. The work was further extended to account for the
effect of the separation distance between the co-indenting shots upon the residual stress field. The
results reveal that the depth of the compressed layer, surface and sub-surface residual stresses are
significantly influenced by the shot velocity, shot shape and separation distance between the
co-indenting shots and to a much lesser extent by the strain-hardening rate of the target.
Furthermore, the study shows the large variability in residual stress distribution resulting from
either incomplete coverage or variability of the shape of the shot.
The current work also indicates that the proposed finite element model is capable of capturing
the main features of the induced residual stress field, thus implying its viability as an effective tool
for the control and optimisation of the treatment.

Acknowledgements

This work was supported by the Natural Sciences and Engineering Research Council (NSERC)
of Canada and the Manufacturing Research Corporation (MRCO) of Ontario.

References

[1] H.O. Fuchs, Shot-peening, Mechanical Engineer’s Handbook, Wiley, Cincinnati, OH, 1986, pp. 941—951.
[2] J. Harrison, Controlled shot-peening: cold working to improve fatigue strength, Heat Treatment 19 (1987) 16—18.
[3] S.A. Meguid, Effect of partial coverage upon the fatigue fracture behaviour of peened components, Fatigue
Fracture Eng. Mater. Structures 14 (5) (1991) 515—530.
[4] M. Desvignes, B. Gentil, L. Castex, Fatigue with residual stresses due to shot-peening: effects and evolution. Proc.
2nd Int. Conf. on Residual Stresses, Nancy, 1988, pp. 791—796.
[5] S.A. Meguid (Ed.), Surface Engineering, Elsevier, Amsterdam, 1990.
[6] M.C. Shaw, G. De Salvo, On the plastic flow beneath a blunt axi-symmetric indentor, Trans. ASME J. Eng.
Industry (1970) 480—494.
 S.A. Meguid et al. / Finite Elements in Analysis and Design 31 (1999) 179—191 191

[7] S.A. Meguid, I.F. Collins, W. Johnson, The co-indentation of a layer by two flat plane or spherical-headed rigid
punches, Int. J. Mech. Sci. 19 (1977) 1—9.
[8] S.A. Meguid, M.S. Klair, An examination of the relevance of co-indentation studies to incomplete coverage in
shot-peening using the finite-element method, J. Mech. Working Technol. 11 (1985) 87—104.
[9] S.A. Meguid, M.S. Klair, Elasto-plastic co-indentation analysis of a bounded solid using finite element method, Int.
J. Mech. Sci. 27 (3) (1985) 157—168.
[10] M.T. Khabou, L. Castex, G. Inglebert, Effect of material behaviour law on the theoretical shot peening results,
European J. Mech. A/Solids 9 (6) (1990) 537—549.
[11] A.K. Li, M. Yao, D. Wang, R. Wang, Mechanical approach to the residual stress field induced by shot-peening,
Mater. Sci. Eng. A 147 (1991) 167—173.
[12] W. Johnson, Impact Strength of Materials, Arnold, London, 1972.
[13] R. Clausen, Ermittlung von Einflussgrossen beim Kugelstralen durch Einzelkornversuche, 1st Int. Conf. on
Shot-Peening, Paris, 1981, pp. 279—285.
[14] K. Iida, Dent and affected layer produced by shot peening, 2nd Int. Conf. on Shot Peening, Chicago, USA, 1984,
pp. 217—227.
[15] J. Edberg, L. Lindgren, K. Mori, Shot peening simulated by two different finite element formulations, in: Shen,
Dawson (Eds.), Simulation of Materials Processing: Theory, Methods and Applications, Balkema, Rotterdam,
1995, pp. 425—430.
[16] Z.H. Zhong, Finite Element Procedures for Contact-Impact Problems, Oxford University Press, Oxford, 1993.
[17] M.H. Refaat, S.A. Meguid, A novel finite element approach to frictional contact problems, Int. J. Numer. Methods
Eng. 39 (1996) 3889—3992.