Nuclear Engineering and Design 197 (2000) 197 – 203

www.elsevier.com/locate/nucengdes

Technical note
Effect of specimen thickness on fatigue crack growth rate
Heung-Bae Park a,*, Byong-Whi Lee b
a
Power Engineering Research Institute, Korea Power Engineering Company, Inc. 360 -9, Mabuk-ri, Gusong-myon, Yongin-shi,
Kyunggi-do, 449 -713, South Korea
b
Department of Nuclear Engineering, Korea Ad6anced Institute of Science and Technology 373 -1, Kusong-dong, Yusong-gu,
Taejon, 305 -701, South Korea
Received 25 August 1998; received in revised form 8 June 1999; accepted 13 September 1999

Abstract

Fatigue tests were performed on the compact tension (CT) specimens of Type 304 stainless steel and Inconel 718.
To investigate the effects of specimen thickness on crack tip deformation and fatigue crack growth rate (FCGR),
specimens of different thickness were used. In the analysis, the elastic plastic fracture mechanics (EPFM) parameter
known as the cyclic J-integral, DJ was adopted to observe the local plasticity at the crack tip and compared with the
linear elastic fracture mechanics (LEFM) parameter known as the stress intensity factor range, DK. The results show
that FCGR is a function of specimen thickness, the effect of which is accelerated as specimen thickness increases.
Therefore, it is thought that not only applied stress level but also specimen thickness should be taken into account
in the measurement of FCGR, which is not considered in ASTM E 647 (ASTM E 647, 1995. Standard test method
for measurement of fatigue crack growth rates.). © 2000 Published by Elsevier Science S.A. All rights reserved.

1. Introduction Linear elastic fracture mechanics (LEFM)

methods are valid only as long as there is a linear
Paris and Erdogan (1963) proposed that the use and elastic relationship between the loads applied
of stress intensity factor range, DK could be used to a cracked structure and the resulting displace-
to describe fatigue crack propagation. Their sim- ments. That is, a zone of plastically deformed
ple empirical fit has been applied with remarkable material in the immediate vicinity of the crack
success over the various metallic materials and in must be small compared with the other dimen-
some cases to non-metallic materials. However, sions of the structure. Hence, a much larger re-
the Paris–Erdogan law is short of mechanistic gion of elastically loaded material surrounds the
bases, since it is applying elastic mechanics to a small plastic zone, and the elastic response is
process of crack advance to be dominated by essentially preserved.
near-crack-tip plasticity (Taylor, 1989). In large scale yielding, DK is no longer valid
parameter, and in this case one must appeal to
other parameters to overcome this restriction. In
* Corresponding author. this context, a number of studies have been per-

0029-5493/00/$ - see front matter © 2000 Published by Elsevier Science S.A. All rights reserved.
PII: S 0 0 2 9 - 5 4 9 3 ( 9 9 ) 0 0 2 5 3 - 8
198 H.-B. Park, B.-W. Lee / Nuclear Engineering and Design 197 (2000) 197–203

formed in recent years to provide a description of where sy is the yield strength of materials.
fatigue crack propagation using elastic plastic Furthermore, according to the requirement of
fracture mechanics (EPFM) parameter in the ASTM E 647 (1995), applied load is chosen to
same way as DK, LEFM parameter. Some com-
 
satisfy the following equation:
mon examples are crack tip opening displacement, 4 Kmax 2
CTOD (Brown et al., 1988), cyclic plastic zone (W− a)] , (2)
p sy
size, rpc (Brown et al., 1988; Nicholls and Martin,
1990) and cyclic J-integral, DJ (Brown et al., where a is crack length and W is specimen width.
1988; Chow and Lu, 1989; Nicholls and Martin, It is limiting the relative size of the plastic zone in
1990). a specimen by limiting the minimum size of the
Recently, Wang and Blom (1991) and Wang remaining uncracked ligament (W− a). And it is
(1993) analyzed the plasticity around a fatigue equivalent to restricting the plane stress

 
crack on the basis of the strip yield assumption monotonic plastic zone size, 2rp:
and boundary element treatment. He suggested 1 Kmax 2
2rp = , (3)
that the plastic deformation, especially the cyclic p sy
plastic deformation, might be the main mechani- to about 25% of the uncracked ligament. How-
cal driving force for the fatigue crack growth ever, even if above ASTM criterion, Eq. (2), is
under cyclic loading. satisfied, the case of violating Eq. (1) can occur as
In the present experiment, every test was car- in this study. Furthermore, for relatively low-
ried out in small scale yielding conditions. How- strength materials that exhibit considerable strain
ever, DJ was also calculated to observe the local hardening, the use of flow stress (average of yield
plasticity at the crack tip, and compared with DK. strength and ultimate tensile strength) in lieu of
The present study aims to investigate the effects yield strength in Eq. (2) can lead to violating Eq.
of specimen thickness on crack tip deformation (1) severely.
and fatigue crack growth rate (FCGR). Another assumption that material is uniformly
homogeneous is practically impossible. If material
is homogeneous, FCGR is independent of mi-
2. Theoretical consideration crostructure, and specimens at the same DK level
have the same crack growth rate. However, mate-
The FCGR can be expressed by the Paris–Er- rials have many phases, precipitates, grain
dogan equation (Paris and Erdogan, 1963). This boundaries, and so on. As fracture process zone
equation can be derived with three assumptions has its characteristic size according to inclusion
(Kang and Liu, 1974): (1) small scale yielding, (2) size and inclusion spacing (Wnuk and Mura,
material homogeneity, and (3) independence of 1983), FCGR is strongly dependent on the mi-
crack tip stress and strain from specimen crostructure of the material. Therefore, the as-
thickness. sumption of material homogeneity is unrealistic.
For a number of years, LEFM has been used to In FCGR test using compact tension (CT) spec-
provide a description of fatigue crack propaga- imen, ASTM E 647 recommends the following
tion, which is related to the stress intensity factor, requirement for specimen thickness:
K at the crack tip. However, its applicability is W W
limited by the prerequisite condition for valid 5B 5 , (4)
20 4
LEFM analysis that the stress far from the crack
should be well below yield stress. Thus, applied where B is specimen thickness. Generally, speci-
stress, sapp, should satisfy the following condition men thickness for FCGR test is smaller than that
for small scale yielding assumption: of fracture toughness test. Kang and Liu (1974)
suggested that in order to avoid the effect of crack
sy tip necking, plates satisfying following equation
sapp B , (1)
3 should be used for fatigue test:
 H.-B. Park, B.-W. Lee / Nuclear Engineering and Design 197 (2000) 197–203 199

B] 2.5
 DK 2
. (5)
plane stress condition. Similarly, Rice and
sy Sorensen (1978) extended Wnuk’s concept using
their own critical crack tip opening angle for
It is equivalent to specimen thickness requirement continued crack growth in plane strain condition.
for plane strain fracture toughness measurement
(ASTM E 399, 1990). Recently, Liu et al. (1991)
suggested that FCGR is a function of DK, stress
3. Experiment
ratio and specimen thickness.
Plastic zone size is determined by the applied
stress, yield strength of material, crack tip geome- 3.1. Materials and specimens
try and specimen thickness. Likewise, crack tip
stresses and strains are strongly dependent upon The materials used in the present experiments
specimen thickness, since stress states vary from were Type 304 stainless steel and Inconel 718. 304
plane stress to plane strain or vice versa as the stainless steel is used widely in power plants and
specimen thickness varies. Inconel 718 is used for hold down spring and
For these reasons a number of EPFM parame- anti-debris filter of nuclear fuel assemblies. The
ters such as CTOD, rpc and DJ have been sug- chemical compositions and mechanical properties
gested for characterizing fatigue crack of these materials in as-received conditions are
propagation behaviors. Among these parameters shown in Table 1 and Table 2 respectively.
rpc and DJ have received some attention, and rpc Compact tension specimens of different thick-
has advantages over other parameters such as (1) ness were machined for fatigue tests. Specimens
facilitating correlation of fast growing crack (2) were manufactured according to ASTM E 647,
having an obvious physical interpretation. Brown and knife edge for crack opening displacement
et al. (1988) proposed that if crack closure is (COD) gauge was manufactured according to
taken into account, rpc is the most suitable ASTM E 399 (1990). The dimensions of speci-
parameter for stress up to yield stress. mens are shown in Table 3.
The Dugdale model to estimate the plastic zone
size is most appropriate for plane stress problems, 3.2. Fatigue test procedure
whereas plane strain conditions are generally
more relevant to fatigue crack growth (Kang and Fatigue tests were carried out using electro-
Liu, 1974; Budiansky and Hutchinson, 1978). servo hydraulic dynamic testing machines, Instron
Early, Wnuk (1974) has employed the Dugdale model 8502 with Instron 2490 series application
model to formulate his final stretch criterion for programs. Cyclic loading with sine waveform at

Table 1
Chemical compositions (weight %)

Materials Fe Cr Ni C Ti Mn Si Mo Nb

304 stainless steel Bal. 18.26 8.22 0.027 – 1.81 0.55 0.07 –
Inconel 718 18.5 19.0 Bal. 0.08 0.9 – – 3.0 5.1

Table 2
Mechanical properties

Materials Yield strength (MPa) Tensile strength (MPa) Young’s modulus (GPa) Elongation (%)

304 Stainless steel 270 600 193 50

Inconel 718 1090 1280 205 18
200 H.-B. Park, B.-W. Lee / Nuclear Engineering and Design 197 (2000) 197–203

Table 3 the requirement of ASTM E 647. The closure load

Dimension of CT specimens (unit: mm)
was determined by finding the inflection point on
Materials Thickness Width Notch height the load-displacement curve. The effective stress
(B) (W) (H) intensity factor range, DKeff was calculated from
the crack closure load.
304 Stainless steel 25 100 6
6 50 3
3 50 3
4. Results and discussion
Inconel 718 5 50 3
Constant amplitude loading fatigue tests were
carried out in the steady state FCGR range,
which is called region II or Paris region. The
Paris–Erdogan law for FCGR is:
da
= C(DK)m, (6)
dN
where da/dN is in mm/cycle and DK is in MPam1/2.
The C and m are material constants.
If crack closure phenomena (Elber, 1970) is
taken into account, the FCGR in terms of DKeff
for different stress ratio fall into a single scatter
band which may be expressed as:
da
=Ce(DKeff)me, (7)
dN
where Ce and me are material constants.
Results of fatigue test with these relations are
plotted on logarithmic coordinates in Figs. 1 and
2. And the material constants are summarized in
Table 4 for 304 stainless steel and Inconel 718.

Fig. 1. Fatigue crack growth rate of 304 stainless steel (a)

da/dN versus DK (b) da/dN versus DKeff.

10 Hz was used in fatigue tests. The tests were

performed at room temperature with stress ratios,
R (Pmin/Pmax), of 0.1 and 0.2. Crack length and
crack closure were monitored by COD gauge
using compliance method. Constant loading am-
plitude tests were performed to evaluate the fa-
tigue crack growth properties of specimens. In
every test, the test load was chosen according to Fig. 2. Fatigue crack growth rate of 5 mm-thick Inconel 718.
 H.-B. Park, B.-W. Lee / Nuclear Engineering and Design 197 (2000) 197–203 201

Table 4
Material constants with relation to DK and DKeff

Materials B (mm) R C m Ce me

304 Stainless steel 25 0.1 3.8×10−9 3.1 1.1×10−8 3.0

25 0.2 5.2×10−9 3.1 2.0×10−9 3.5
6 0.2 6.2×10−9 3.0 2.2×10−8 2.7
3 0.2 3.7×10−9 3.1 5.2×10−9 3.1
Inconel 718 5 0.2 5.8×10−9 3.5 5.5×10−8 2.4

Fig. 3. Comparison of normalized crack closure loads (a) 3 mm-thick 304 SS (b) 6 mm-thick 304 SS (c) 25 mm-thick 304 SS (d) 5
mm-thick Inconel 718.

To investigate the effect of specimen thickness thick 304 stainless steel was higher than that of
on FCGR, 304 stainless steel specimen was manu- 3mm-thick specimen, and in case of 25mm-thick
factured with three thicknesses, such as 3mm, specimen, FCGR with stress ratio of 0.2 was
6mm and 25mm. Test results of 304 stainless steel higher than that of stress ratio of 0.1.
with different thickness were compared as a func- The normalized crack closure load levels of 304
tion of DK and DKeff in Fig. 1(a) and (b) respec- stainless steels and Inconel 718 are shown in Fig.
tively. As can be seen in this figure, FCGR of 3. In this figure, crack closure load level of 3 mm
25mm-thick 304 stainless steel was higher than thick 304 stainless steel is lower than that of 6 mm
that of 6mm-thick specimen, and FCGR of 6mm- thick specimen, and 6 mm thick specimen is lower
202 H.-B. Park, B.-W. Lee / Nuclear Engineering and Design 197 (2000) 197–203

than that of 25 mm thick specimen. But, this thin specimens in which plane stress condition is
result disagrees with our expectation because dominant. From above results, it is considered
larger crack closure results in more retardation of that FCGR of 304 stainless steel is a function of
crack growth. specimen thickness, and it increases as the speci-
There is little data in the literature about the men thickness increases.
effect of specimen thickness on FCGR. Three In order to examine the plasticity in the crack
completely contradictory conclusions have been tip region, the analysis of EPFM is applied to the
reported: (1) no effect at all (Hahn et al., 1972; results of the present experiment. In Fig. 4, DKJ
Shahinian, 1976; James, 1981; Kim and Kim, calculated from DJ [DJ = (DKJ )2/E, where E is
1998), (2) accelerated growth rate with thickness Young’s modulus] is compared with DK in terms
(Kang and Liu, 1974; Woodtli et al., 1986), (3) of crack length. The gap between DK and DKJ of
reduced growth rate (Jack and Price, 1972). The 3 mm-thick 304 stainless steel is nearly the same
thickness effect could be a variable of material with that of 6 mm thick specimen, and this gap is
loading conditions, environment and the range of larger than that of 25 mm thick 304 stainless steel
stress intensity factor. In the present experiment, and Inconel 718. These results show that the
the FCGR of a thick specimen was higher than crack tip of thin 304 stainless steel specimen in
that of a thin one with relation to either DK or which plane stress condition is dominant has
DKeff. Generally, thick specimens have more plain more plasticity than those of thick 304 stainless
strain condition in crack front line than that of steel and Inconel 718 in plane strain condition.

Fig. 4. Comparison of DK and DKJ calculated from cyclic J-integral, DJ (a) 3 mm-thick 304 SS (b) 6 mm-thick 304 SS (c) 25
mm-thick 304 SS (d) 5 mm-thick Inconel 718.
 H.-B. Park, B.-W. Lee / Nuclear Engineering and Design 197 (2000) 197–203 203

The crack tip plastic deformation causes crack tip Budiansky, B., Hutchinson, J.W., 1978. Analysis of closure in
bluntness and stress transfer to promote further fatigue crack growth. J. Appl. Mech. 45, 267 – 270.
Chow, C.L., Lu, T.J., 1989. On the cyclic J-integral applied to
plastic deformation in the plastic zone and this
fatigue cracking. Int. J. Fract. 40, R53 – R59.
plastic deformation decreases FCGR. That is, as Elber, W., 1970. Fatigue crack closure under cyclic tension.
the ratio of applied stress to yield strength of Eng. Fract. Mech. 2, 37 – 45.
material increases, the crack tip plastic deforma- Hahn, G.T., Hoagland, R.G., Rosenfield, A.R., 1972. Local
tion increases. In case of high-strain hardening yielding attending fatigue crack growth. Mater. Sci. Eng. 3,
materials as described in ASTM E 647, applied 1189 – 1202.
Jack, A.R., Price, A.T., 1972. Effects of thickness on fatigue
stress can be increased to the extent of violating
crack initiation and growth in notched mild steel speci-
Eq. (2) to reduce test time. In this case the gap mens. Acta Metall. 20, 857 – 866.
between DK and DKJ can be largely increased. So, James, L.A., 1981. Specimen size considerations in fatigue
in large scale yielding or high-strain hardening crack growth rate testing. ASTM STP 738, 45 – 57.
materials, the use of EPFM parameter is more Kang, T.S., Liu, H.W., 1974. Fatigue crack propagation and
appropriate than that of DK in the measurement cyclic deformation at a crack tip. Int. J. Fract. 10, 201–
of FCGR. 222.
E.S. Kim, I.S. Kim, 1998. Effect of specimen thickness on the
near-threshold fatigue crack propagation of nuclear struc-
tural materials. In: Proc. of Workshop on the Integrity of
5. Conclusions
Nuclear Power Plant Equipment, Taejon, Korea, pp. 143–
158.
1. In the Fatigue crack growth rate (FCGR) test Liu, H.W., Chen, Q., Sinha, S.K., 1991. The characterization
of 304 stainless steel, FCGR is a function of of a crack-tip field. Eng. Frac. Mech. 39, 213 – 217.
specimen thickness, and it increases as the Nicholls, D.J., Martin, J.W., 1990. An examination of the
specimen thickness increases. That is, as the reasons for the discrepancy between long and small fatigue
stress state approaches to plane strain condi- cracks in Al-Li alloys. Mater. Sci. Eng. A128, 141 – 145.
Paris, P.C., Erdogan, F., 1963. A critical analysis of crack
tion, FCGR becomes high.
propagation laws. J. Basic Eng. Trans. ASME Ser. D 55,
2. It is thought that there is need to investigate 528 – 534.
the effect of specimen thickness on FCGR Rice, J.R., Sorensen, E.P., 1978. Continuing crack-tip defor-
with several materials, which is not fully con- mation and fracture for plane-strain crack growth in elas-
sidered in ASTM E 647. tic-plastic solids. J. Mech. Phys. Solids 26, 163 – 186.
3. In large scale yielding or high-strain hardening Shahinian, P., 1976. Influence of section thickness on fatigue
crack growth in Type 304 stainless steel. Nucl. Tech. 30,
materials, the use of stress intensity factor
390 – 397.
range, DK in the measurement of FCGR is Taylor, D., 1989. Fatigue Thresholds. Butterworth and Co,
inadequate, and in this case the use of elastic pp. 1 – 7.
plastic fracture mechanics (EPFM) parameter Wang, G.S., 1993. The plasticity aspect of fatigue crack
is more appropriate. growth. Eng. Fract. Mech. 46, 909 – 930.
Wang, G.S., Blom, A.F., 1991. A strip model for fatigue crack
growth predictions under general load conditions. Eng.
References Fract. Mech. 40, 507 – 533.
Wnuk, M.P., 1974. Quasi-static extension of a tensile crack
ASTM E 399, 1990. Test method for plane-strain fracture contained in a viscoelastic-plastic solid. J. Appl. Mech. 41,
toughness of metallic materials. 234 – 242.
ASTM E 647, 1995. Standard test method for measurement of Wnuk, M.P., Mura, T., 1983. Extension of a stable crack at a
fatigue crack growth rates. variable growth step. ASTM STP 791, 96 – 127.
Brown, M.W., de los Rios, E.R., Miller, K.J., 1988. A critical Woodtli, J., Muster, W., Radon, J.C., 1986. Residual stress
comparison of proposed parameters for high-strain fatigue effects in fatigue crack growth. Eng. Fract. Mech. 24,
crack growth. ASTM STP 924, 233–259. 399 – 412.