RESIDUAL STRESS RELAXATION DUE TO FRETTING FATIGUE
IN SHOT PEENED SURFACES OF T1-6A1-4V
S.A. Martinez1, S. Sathish2, M.P. Blodgett1, S. Namjoshi3, S. Mall1
Air Force Research Laboratory, Wright Patterson Air Force Base, OH 45433
2
University of Dayton Research Institute, Dayton, OH 45469
3 Air Force Institute of Technology, Wright Patterson Air Force Base, OH 45433
ABSTRACT. Fretting fatigue occurs at locations where the materials are sliding against each other
under load. In order to enhance the fatigue life under fretting conditions the surface of the
component is shot peened. In general, the shot peening process produces a compressive stress on the
surface of the material, thereby increasing the resistance of the material to crack initiation. This
paper presents the relaxation of residual stress caused during fretting fatigue. X-ray diffraction has
been utilized as the method to measure residual stress in fretting fatigued samples of Ti-6Al-4V.
INTRODUCTION
The most frequently used method for surface enhancement is shot peening. In shot
peening, the surface of a part is bombarded with high velocity spherical media called shot
[1]. Each shot striking the material acts as a tiny peening hammer, imparting to the surface
a small indentation. This produces a layer of compressive stresses at the surface of the
material that is resistant to fatigue loading and stress corrosion failure. The improvement
in fatigue behavior of the component can be a consequence of the compressive stresses on
the surface induced by shot peening, surface hardening and final surface finish quality
(poor surface finish is detrimental for fatigue and corrosion). The effects of these factors
will depend on the geometry of the structure, applied stress, and hardness of the material.
Residual stresses generated by shot peening are compressive in nature. The major
benefit of shot peening in improving fatigue and fretting fatigue strength of various alloys
is in the residual compressive stresses that are generated in the surface. These residual
compressive stresses retard the propagation of fatigue cracks, since it is known that cracks
will not initiate nor propagate in a compressive stress zone. The compressive stresses near
the surface are balanced by tensile stresses which decrease fatigue strength in the interior
and the stresses are usually lower [2]. Shot peening will result in a large increase of fatigue
strength when the yield strength of the material in question is high and the service induced
stress is low [3]. If the local yield strength of the material in the critical section is exceeded
CP657, Review of Quantitative Nondestructive Evaluation Vol. 22, ed. by D. O. Thompson and D. E. Chimenti
© 2003 American Institute of Physics 0-7354-0117-9/03/S20.00
1531
by service-induced stress the residual compressive stress are relaxed, therefore reducing
fatigue strength.
The effect of shot peening surfaces under fretting wear is of importance in this
research. Fretting is a contact loading that modifies the surface of the component and can
lead to relaxation of the compressive residual stresses and loss of matter. The residual
compressive stresses introduced by shot peening have slight effects on crack nucleation,
but can retard crack propagation. These stresses are beneficial in fretting fatigue as well as
in plain fatigue. In fretting, there are two further advantages: shot peening causes a
reduction in the coefficient of friction and also the rough surface gives a further increase in
fatigue strength in comparison with plain fatigue [4]. If these residual stresses relax under
cyclic loading the fatigue life of the material will not be improved. The decay of residual
compressive stresses early in the cyclic loading will prevent internal fatigue crack
nucleation.
In this investigation, the effects of fretting fatigue on the residual stress distribution
and fatigue life of Ti-6Al-4V shot peened samples have been studied. X-ray diffraction has
been used as a method for measuring surface residual stress.
THEORY
Residual stresses can arise in a number of different ways. In a body, the residual
stresses are those which are not necessary to maintain equilibrium between the body and
the environment that surrounds it [5]. They can be the result of the fabrication method, for
example wire drawing, hot and cold forging, thermal treatment, quenching, milling,
grinding, or processes such as shot peening or surface rolling.
Shot peening affects surface tensile stress by developing residual compressive
stresses, therefore increasing the endurance of parts. The total depth of the compressive
stresses for Ti-6Al-4V can be about 100 Jim as reported by Fridici and co-workers [6]. The
shot peening process belongs to the group of mechanical surface treatments that includes
polishing, surface, rolling, etc. All of these surface treatments improve the fatigue strength
of components by preventing dislocation movement on the surface inducing a layer of
compressive residual stresses at the surface.
Farrahi and co-workers [7] have reported the results of their investigation on the
effect of shot peening variables and the resulting residual stress and fatigue life of spring
steel. It appears that there is a correlation between the fatigue strength and the area under
the residual stress distribution curve. Fatigue life improvement can be attributed to the
increase in surface roughness due to shot peening since it is known that a rougher surface
suffer less fretting wear and a compressive surface stress results in less wear than a tensile
stress [8].
The effects of shot peening in specimens of Ti-6Al-4V undergoing plain fatigue
and fretting fatigue was studied by Namjoshi et al [9]. During this investigation, it was
found that plain fatigue causes relaxation of the residual stresses by as much as 80 percent.
These stresses relax more rapidly under conditions of fretting fatigue.
Many references describe the beneficial effects of compressive residual stresses
induced by shot peening [10]. Residual stresses can relax by processes of shake down in
conditions of alternating stress and by applications of mechanical or thermal energy. The
higher the dislocation density and the internal energy of the material, the faster a given
level of residual stress will relax at a fixed temperature. In our experiments we have
utilized the non-destructive method of X-ray diffraction for surface residual stress
measurements.
1532
X-ray diffraction is one of the most popular techniques for residual stress
measurements. In this technique, monochromatic x-rays impinge on the surface of the
material and crystallographic planes diffract whenever the Bragg's Law: nX=2dsin9is
satisfied. Here n is an integer representing the order of diffraction (typically 1), A,, is the x-ray wavelength, d is the lattice spacing of crystal planes (hkl), and 6 is the diffraction
angle. This spacing depends on the angle of tilt, \j/ and the stress being measured. The
strain in the crystal is calculated from the lattice spacing of a specific set of crystal planes
(hkl).
For conditions where the stress tensor is biaxial (isotropic material), a condition of
plane stress is assumed to exist in the diffracting surface layer [11]. The stress distribution
is described by the principal stresses an and a22 with no stress acting perpendicular to the
free surface. This is described by Eqn. (1).
-d
1+ v
. 2
0
—=———n^ sin iff
d,
E
7
t = 0"ll
COS 2
0 + ^22
Sil1 2
0
(1)
Where a^ is the biaxial stress tensor and d^-do/do is the strain obtained from the
position of the diffraction peak for a given reflection, hkl, acting perpendicular to the free
surface, do, is the unstressed lattice spacing at \|/=0 and d^y, is the lattice spacing obtained
from the position of the diffraction peak for a crystallographic plane (hkl) in the direction
(j),\|/ tilt. E is the Young's modulus and v Poisson's ratio of the material. The principal
stress tensors are an, a22, and 033, To determine the residual stress experimentally, the
strain is measured at different angles (at least two) of \|/ and a plot of d vs. sin2\|/ is
obtained. Using the slope of the curve and the x-ray elastic constant in Eq. (1) the residual
stress can be evaluated. The x-ray elastic constants needed for the evaluation of the
residual stress components are V* 82 = 5.64 x 10~4 and Si = 6.89 MPa"1. Similarly, Young's
Modulus (E=l 19 GPa) and Poisson's ratio (v=.321).
EXPERIMENTS
Sample Description
Three shot peened samples of Ti-6Al-4V with similar microstructure were
examined during this investigation. This material was chosen due to its extensive use in
aerospace applications. The dimensions of the samples are 60 mm x 6.35 mm x 3.81mm.
The sample was peened according to SAE Aerospace Materials Specification (AMS) 2432.
These specimens originated from plate stock. The plate was forged and later it was solution
heat treated at 935°C for 1 hour, fan cooled, then vacuum annealed at 705°C for 2 hours,
slow cooled. The resultant microstructure after heat treatment included primary alpha with
transformed beta structure (Figure 1). The average grain size is 30 Jim.
The first shot peened sample was not exposed to the fretting fatigue test. This
sample was utilized to determine the uniformity and orientation dependence of the residual
stress in shot peened samples. The remaining samples were subjected to fretting fatigue and
hence have a fretting scar. The description of them is as follows: sample 2 with a fretting
scar had no fracture, the maximum and minimum loads applied were 910 and 1 19
1533
FIGURE 1. Microstructure of the T1-6A1-4V Alloy.
in MPa, respectively. This sample was used to determine the uniformity of the stress
distribution across the fretting scar. Sample 3 has a small crack at the fretting scar and the
loads applied were 382 (maximum) and 193 (minimum). This sample was used to
investigate relaxation of residual stress ahead of the crack and directional dependence of
the residual stress inside the fretting scar. Residual stress measurements were performed
using x-ray diffraction. The x-ray diffraction residual stress analyzer has a horizontal
goniometer fixture. Copper K-oc radiation collimated to a 2 mm diameter was used as the
source. Nickel filters were utilized to reduce the intensity of K-(3 radiation from the x-ray
source. The diffraction peaks from the (302) crystallographic planes of the hexagonal
phase of the Ti-6Al-4V were used for residual stress measurements. After correcting the
data for Lorentz-Polarization and absorption, Gaussian peak-fitting algorithm was used to
fit the diffraction peaks. The stress was evaluated using the sin2\j/ technique described
previously.
Fretting Fixture
The fretting fatigue tests were conducted in a servo-hydraulic uniaxial test machine,
at ambient temperature, in a laboratory environment. An overview of the fretting fixture is
shown in Fig. 2. This set-up was described elsewhere [9,12]. The normal fretting load was
applied to the dogbone specimen by using the spring loaded bolts. Alternating torque was
applied to the end of each of the spring bolts until the load cell located in front of the
fretting fixture read a desired applied load. The longitudinal springs provide for the support
between the fixture and the load frame, this minimizes the effects of fixture misalignment.
The contact between the specimen and the fretting pads served as a third grip, this results in
different cyclic loadings felt by the specimen above and below the contact area. A cylinderon-flat configuration was used to introduce the fretting effect. Two fretting pads with
cylindrical radius of 67.3 mm were pressed against the surface of the fatigue specimen
using the fretting fixture. Both specimen and fretting pad were made from Ti-6Al-4V.
1534
•LongitiKJiiial Springs Q
IP
Lateral Springs
FIGURE 2. Schematic of the fretting fatigue test frame. Not to scale.
RESULTS AND DISCUSSIONS
Residual Stress
Shot peened samples were examined before and after subjecting to fretting and the
residual stress at a given location was determined by utilizing seven \j/-tilts. Thirteen peak
positions were obtained from the seven i|/-tilts and fitted to a line using a least squares
linear regression algorithm. The stress was obtained from the slope of this line along with
x-ray elastic constants. According to Eqn. 1, for an isotropic material, a plot of d vs.
sin2\|/ should be linear and the slope of the line represents the residual stress. The residual
stress distribution was examined by scanning in both x and y directions in different regions
of the sample.
The d vs. sin2\|/ plot for sample 1 is shown in Figure 3. The average residual stress
measured for this sample with no fretting scar was -827+7-34 MPa. No departure from the
straight-line was observed when measuring at different regions and orientations of the
sample. In investigations performed by Prevey [13] similar results are obtained showing the
uniformity of the residual stresses for shot peened surfaces.
|
£ 0.803 c
TJ" 0.801 jf
——B
1 0.799 S
| 0-797
0.05
0.1
0.15
0.2
FIGURE 3. Linearity of d vs. sin2\j/ for a shot peened sample 1.
1535
0.25
0.3
0.35
-
2
0
2
Distance (mm)
FIGURE 4. Residual stress measurements across a fretting scar in sample 2.
The uniformity of the stress distribution across a fretting scar (no failure) was
examined in sample 2. Residual stress was measured before and after fretting for this
sample and the values are -793 +/-55 MPa and -406 +/-6S MPa, respectively. This
indicates a relaxation of about 48.8%. This relaxation is expected since it is well
documented that residual stresses can relax by processes of shake down in conditions of
alternating stress, also by applications of mechanical and/or thermal energy (Prevey). The
higher the dislocation density and the internal energy of the material, the faster a given
level of residual stress will relax at a fixed temperature.
A mapping procedure was set up covering a total distance of 12 mm (6 mm on
either side of the scar) with a step size of 1 mm. The measurements for this sample and
others were taken at the center of the fretting scar approximately 3mm from one edge and
2.5 mm from the other edge. Figure 4 shows the residual stress measurements along this
line. A minimum compressive stress of -435 MPa and a maximum compressive stress of844 MPa are observed. Lower compressive stresses correspond to the region surrounding
the fretting scar. Based on the observations we can see that the presence of a fretting scar
produces a non-uniform residual stress distribution in Ti-6Al-4V specimens.
Sample 3 has 2 mm long crack that started from one of the edges. During fretting
fatigue, the relative motion of the sample and the pads is confined along the direction of
fatigue loading. This is expected to give rise to anisotropic relaxation of residual stress
inside the fretting scar. In order to test anisotropy in residual stress measurements were
performed along (phi=0) and perpendicular (phi^O) to the direction of fatigue loading.
Measurements were done at the center of the sample and care was taken to set the x-ray
spot at least 1 mm away from the crack tip. The results of the measurements along a line
parallel to the length of the sample are shown in figure 5. The residual stress variation
1536
-300
-800
-900
-
2
0
2
4
Distance (mm)
FIGURE 5. Residual stress along (phi=0) and perpendicular (phi=90) to fatigue loading directions in
Sample 3.
outside the fretting scar for both directions is similar. On the other hand inside the fretting
scar a significant difference is observed. At the center of the scar the residual stress along
phi = 0 direction is -385 MPa compared to phi = 90 which is -640 MPa. The stress along
the fatigue loading direction has relaxed by 50%, while perpendicular to this direction it
relaxes by 25%. The measurements show that the fretting fatigue produces a biaxial stress
state on an otherwise isotropic residual stress specimen.
CONCLUSIONS
Shot peening induced compressive stresses on Ti-6Al-4V samples have been found
to be uniform. Directional dependence residual stress was not observed in the samples.
Relaxation of residual stress has been observed in fretting fatigued shot peened samples.
The relaxation of the residual stresses has been found to extend beyond the area of the
fretting scar. The residual stress along and perpendicular to the fatigue loading directions
outside the fretting zone are similar. Inside the fretting scar the residual stress relaxation
along the fatigue loading direction is twice that of the perpendicular direction.
ACKNOWLEDGEMENTS
This research was sponsored by, and performed on-site at the NDE Branch,
AFRL/Material and Manufacturing Directorate, Wright-Patterson Air Force Base, Ohio
under the Contract # F33615-98-C-5217.
1537
REFERENCES
1. Metal Improvement Company, Inc., Shot Peening Applications, Seventh Ed., 1986.
2. Fuchs, H.O., "Optimum Peening Intensities", Seventh International Conference on
Shot Peening, 1999, pp. 63 9-645.
3. Schiitz, W., "Shot Peening of Components to Improve Fatigue Strength", Seventh
International Conference on Shot Peening, 1999, pp. 501-508.
4. Waterhouse, R.B., "Effect of material and surface condition on Fretting Fatigue",
Fretting Fatigue, Mechanical Engineering Publications, London, pp. 339-349, 1994
5. Withers, P.J. and Bhadeshia, H.K.D.H., "Residual Stress Part I: Measurement
Techniques", MTS, 4640A, 2001, pp. 355-365.
6. Fridrici, V., Fouvry, S., and Kapsa, P., "Effect of Shot Peening in the wear of T1-6A14V", Wear, 2001, Vol. 250, pp. 642-649.
7. Farrahi, G.H., Lebrun, J.L., and Couratin, D., "Effect of Shot Peening on Residual
Stress and Fatigue Life of a Spring Steel", Fatigue and Fracture of Engineering
Materials and Structures, 1995, Vol. 18, No. 2, pp. 211-220.
8. Waterhouse, R.B. and Trowsdale, A.J., "Residual Stress and surface roughness in
fretting fatigue", Journal of Applied Physics, 1992, Vol.25, pp. A236-A239.
9. Namjoshi, S.A., Jain, V.K., and Mall, S., "Effects of Shot Peening on Fretting Fatigue
Behavior of Ti-6Al-4V", Transactions of the American Society of Mechanical
Engineers, April 2002, Vol. 124, pp. 222-228.
10. Prevey, S., Hornbach, D., and Mason, P., "Thermal Residual Stress Relaxation and
Distortion in Surface Enhanced Gas Turbine Engine Components", Proceeding of the
17th Heat Treating Society Conference and Exposition and the 1st International
Induction Heat Treating Symposium", American Society of Metals, Materials Park,
Ohio,1998,pp.3-12.
11. Noyan, I.C. and Cohen, J.B., "Residual Stresses and sliding wear", Wear,1983, Vol.
84, pp. 183-202.
12. Mall, S., Jin, O., Yuksel, H., and Calcaterra, J.R., "Investigation into Fretting Fatigue
Behavior of Shot Peened Ti-6Al-4V", submitted for publication.
13. Prevey, S., "The Uniformity of Shot Peening Induced Residual Stress," Residual Stress
for Designers and Metallurgists, American Society for Metals, Metals Park, OH, 1981,
pp.151-168.
1538