You are on page 1of 9

Connection between the Results of the Low Cycle Fatigue

and the Fatigue Crack Growth Tests



Gyula Nagy
1
and Jnos Lukcs
1

1
University of Miskolc H-3515 Miskolc-Egyetemvros, Hungary

Keywords: low cycle fatigue (LCF), fatigue crack growth (FCG), Manson-Coffin equation, Paris-
Erdogan law, steel, superalloy, aluminium alloy.

Abstract. Material characteristics determined for low cycle fatigue (LCF), high cycle fatigue (HCF)
or fatigue crack growth (FCG) are different. The question arises, whether common mechanical and
microstructural features can be found or not in this field of loading types, based on which
relationship among the materials constants can be assumed. The paper introduces the similarity of
the stress and strain state, furthermore the dislocation structure having been developed during low
cycle fatigue and fatigue crack growth. This can form a basis for the establishment of a connection
between the material characteristics, the exponents of the Manson-Coffin equation and the Paris-
Erdogan law. The validity of the hypothesis has been illustrated here by comparing the test results
determined for different material grades, steels, one superalloy and aluminium alloys.

Introduction

The behaviour and the properties of the different materials are determined by the quality of
materials (chemical composition, type of atomic bonding, spatial arrangement of atoms, micro- and
macrostructure), furthermore the stress state, the strain rate and the temperature. Due to the great
variety of these influencing factors we do not have a general model for describing the behaviour of
materials, thus we have to use several material constant to be able to characterise it.
Even if we consider the fatigue loading of the possible loading conditions we have to use more
equations (e.g. Manson-Coffin, Whler, Paris-Erdogan laws) and we have to determine more
material constants to be able to provide sufficient data for design and control. We determine
different properties in case of low cycle fatigue (LCF), high cycle fatigue (HCF), and fatigue crack
growth (FCG) rate measurements. We can ask if there are common mechanical or microstructural
properties at such a relatively small area that could be taken as a basis in order to build a
connection between the materials constants.
The aim of this paper is to introduce the similarity of the mechanical stress state and the
dislocation structure in case of LCF and FCG, which can be the basis of the connection of the
determined constants. The assumption is supported by the results measured and analysed on the
same steel, superalloy and aluminium alloy grades. Micro-alloyed and thermomechanically treated
steels, one superalloy and aluminium alloys with different microstructure (without and with
hardening) were used for the investigation.

Comparison and Similarity of the Mechanical Stress State and the Dislocation Structure

Comparison of the mechanical stress state. In case of LCF the load on the specimen is so large
that the whole volume of measuring part of the specimen will suffer plastic deformation during the
first half of the loading cycle. At the majority of the experiments the controlled parameter is the
total strain amplitude (e
a
) and its time variation generally follows a sinusoidal, a trapezoidal or a
triangular wave form. Since the connection between the stress and strain is not linear in the region
of plastic deformation, in case of total strain controlled LCF the time dependence of the stress is
influenced by the material, as well.

Materials Science Forum Vols. 414-415 (2003) pp 289-296
Online available since 2003/Jan/15 at www.scientific.net
(2003) Trans Tech Publications, Switzerland
doi:10.4028/www.scientific.net/MSF.414-415.289
All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of TTP,
www.ttp.net. (ID: 141.58.45.15, UB Stuttgart, Stuttgart, Germany-30/07/13,21:00:47)
Fig. 1 shows the time variation of the strain amplitude and the stress amplitude in case of low
cycle fatigue with a saturated strain amplitude of e
a
= 0.5%. The difference is can be seen well. If
the measuring part of the specimen is cylindrical with a uniaxial apprehension the same axial stress
raises in each point of the cross section analysed.

Fig. 1 Time variation of the total strain amplitude and the stress amplitude in case of LCF

Measuring the fatigue crack growth rate a plastic zone is forming at the crack tip during the first
tensile cycle. The maximum size of the plastic zone (w) depends on the yield stress of the material
(s
0
) and the stress intensity factor (K
I
). Supposing an ideally elastic-plastic material and maximal
loading the stress distribution in y direction as a function of the distance from the crack tip is going
to form as shown in Fig. 2 a) [1]. The maximum stress in the plastic zone is equal to the yield
strength (s
y
= s
0
). Decreasing the load the stress distribution will change. When the load is zero
again the stress distribution in y direction ahead of the crack tip is as shown in Fig. 2 b), which
demonstrates the reversed plastic zone, too. This means that a compressive stress arises in the
plastic zone. Repeating the loading up (Fig. 2 c)) and down the stress will be similar as experienced
during LCF.

Fig. 2 Stress distribution at the crack tip of the fatigue specimen [1]
290 Materials Science, Testing and Informatics I
Comparison of the dislocation structure. Mild metals and alloys close to the equilibrium state
can be characterised by a relatively small dislocation density (10
11
10
13
m
-2
) and a homogeneous
dislocation distribution [1]. Under the influence of LCF the dislocation density is increasing and the
dislocation distribution becomes heterogeneous even in the first cycle. At the early stage the edge
dislocations, dislocation loops and dipoles, the labouring dislocation parts form uncondensed,
blurred cell walls. With increasing number of cycles the number of multipoles and the dislocation
density of the walls increases, the cell walls and the matrix will become more and more separated
[1,2]. At the beginning the cell size decreases with increasing the cycle number and after the
dislocation density becomes stable the cell size will not change significantly. The orientation
difference of the matrix crystal planes at the two sides of the cell-walls is increasing continuously
during the entire fatigue process. The size of the cell is inversely proportional to the strain
amplitude [3,4]. Fig. 3 shows a typical cell dislocation structure [2].



Fig. 3 Cell dislocation structure developed during fatigue

In case of fatigue crack growth, a very large local plastic deformation will develop in front of the
crack in the plastic zone, although its measure is different in the different points of the zone. Due to
the large plastic deformation a cell structure will form [5,6,7]. In the crack vicinity the size of the
cells is very small, while increasing the distance away from the crack their size is increasing. As
well, the orientation difference of the planes of the neighbouring cells is the biggest near by the
crack and getting more and more away the orientation difference is decreasing [8].
Based on these facts we can establish that the dislocation structure developing during LCF is
very similar to that which forms around the crack vicinity during FCG.

Investigations and Results

Materials. For the examinations different micro-alloyed steel grades (37C, A106B, A515 Grade70,
KL7D, DX52; own tests and [9,10]) and thermomechanically treated steel grades (X80TM,
QStE690TM/StE690; own tests and [11]), one superalloy (Waspaloy, [12,13]) and aluminium
alloys with different microstructure (AlMg3, AlMg5/AlMg5.1Mn, AlMg4.5Mn without hardening
and 2024-T4/2024-T6, 7075-T6 with hardening; own tests and [14,15,16,17]) were used.
The chemical compositions of the examined steels, superalloy and aluminium alloys are
summarised in Table 1, Table 2 and Table 3, respectively; the mechanical properties of all materials
are listed in Table 4.
Materials Science Forum Vols. 414-415 291
Table 1 Chemical composition of the tested steels [wt %]

Material grade C Si Mn P S Al Nb V Cr Mo
37C 0.15 0.38 0.89 0.029 0.016 0.016 0.021 0.023
A106B 0.3 0.10 1.06 0.048 0.058
A515 Grade70 0.22 1.21 0.030 0.004
KL7D
(1)
0.17 0.24 1.31 0.020 0.036 0.049 0.01 0.11 0.02
DX52
(2)
0.18 0.15-0.20 1.50 0.030 0.035 0.06 0.25
X80TM
(3)
0.077 0.30 1.84 0.012 0.002 0.036 0.046
QStE690TM_FCG 0.08 0.29 1.75 0.011 0.002 0.041 0.04 0.061 0.037 0.32
StE690_LCF 0.15 0.53 0.87 0.011 0.004 0.038 0.63 0.22
(1)
Ni = 0.08%, Ti = 0.0017%, Cu = 0.26%.
(2)
Cu 0.30%.
(3)
Ti = 0.018%, N = 0.005%.


Table 2 Chemical composition of the investigated superalloy [wt %]

Material grade C Cr Ni Mo Ti Al Fe Co B Zr
Waspaloy_FCG 0.07 19.5 57 4.3 3 1.4 2 13.5 0.006 0.09
Waspaloy_LCF 19 59 4.25 3 1.25 13.5

Table 3 Chemical composition of the investigated aluminium alloys [wt %]

Material grade Si Cu Fe Mn Mg Ti Cr Zr Zn
AlMg3
(1)
0.179 0.022 0.310 0.2770 2.950 0.040 0.0430 0.0012 0.044
AlMg5_FCG 0.31 0.01 0.32 0.45 5.20 0.020 0.1 0.01
AlMg5.1Mn_LCF <0.40 <0.10 <0.40 0.75 5.1 0.20 0.12 0.25
AlMg4.5Mn_FCG 0.21 0.01 0.27 0.77 4.53 0.015 0.1 0.03
AlMg4.5Mn_LCF 0.19 0.026 0.30 0.72 4.82 0.01 0.096 0.085
7075-T6_FCG 0.15 1.59 0.35 0.005 2.70 0.19 5.70
7075-T6_LCF <0.4 1.2-2.0 <0.50 <0.30 2.1-2.9 <0.20 0.18-0.35 5.1-6.1
(1)
Pb = 0.009%, Bi = 0.0150%, Be = 0.0018%, Al = 96.1170%.

Low cycle fatigue tests. LCF tests have been executed in air, with total strain amplitude-control.
The tests on steels and aluminium alloys have been performed at room temperature, with a
sinusoidal loading function and the investigations on the superalloy have been executed at elevated
temperature (650C), with trapezoidal loading wave form. The change of deformation was
measured by a diametral extensiometer. The stress ratio was R = -1, the cycle number until failure
was chosen at the 25% decreasing of the maximum tensile load. During the measurement the
maximum values of the total strain and stress amplitudes and the hysteresis loop were recorded.
The results were evaluated by classical methods. The exponent (c) ant the constant (e
f
)

of the
Manson-Coffin equation [18] were determined by the following expression:

,
,
c
ap f t
N e e = (1)

where e
ap
is the plastic strain amplitude, N
t
is the cycle number until failure, e
f


and c are material
constants. The calculated parameters, the material constants (c) for the investigated different steel
grades, the superalloy and the different aluminium alloys are listed in Table 5 (bm = base material,
wm = weld metal).
292 Materials Science, Testing and Informatics I
Table 4 Mechanical properties of the investigated materials

Material grade Yield strength
R
y
[N/mm
2
]
Tensile strength
R
m
[N/mm
2
]
Elongation
A
5
[%]
Reduction of area
Z [%]
Source
37C 269 405 33.5 63.5 own
A106B 317 496 36.0 [9]
A515 Grade70 318 633 [10]
KL7D 392 535 19.0 own
DX52 396 543 25.0 71.0 own
X80TM 540 625 25.1 73.1 own
QStE690TM_FCG 768 854 20.0 own
StE690_LCF 743 885 57 [11]
Waspaloy_LCF 927 (21C) 1310 (21C) [12]
809 (650C) 1127 (650C)
AlMg3 127.7 217.9 27.0 [14]
AlMg5_FCG 185 288 14.5 [14]
AlMg5.1Mn_LCF 235 400 34.6 [15]
AlMg4.5Mn_FCG 230 296 18.0 [14]
AlMg4.5Mn_LCF 226 348 17.0 22.5 [15]
2024-T6_FCG 304 476 35.0 [16]
2024-T4_LCF 332 432 17.3 [15]
7075-T6_FCG 533 594 9.2 [17]
7075-T6_LCF 470 580 33.0 [15]

Fatigue crack growth tests. The fatigue crack growth rate measurements were made on
compact tension (CT) and three point bending (TPB) specimens, in air. Steels and aluminium alloys
were tested at room temperature, with a sinusoidal loading function and the superalloy was
investigated at elevated temperature (650C), with trapezoidal wave form. The stress ratio was R =
0 0,1. The crack size was determined by optical method and using the compliance method. From
the collected data and results of the fatigue crack growth measurements the exponent (n) and the
constant (C) of the Paris-Erdogan equation [19] have been determined:

,
n
da
C K
dN
= D (2)

where da/dN is the fatigue crack growth rate, DK is the stress intensity factor range. In each case
first we have calculated the related part of the kinetic diagram of the fatigue crack growth (da/dN-
DK diagram), then we have determined the two material constants (C and n) with linear regression.
From these results we have calculated the average of the data measured on 3-26 specimens. The
average and standard deviation values of the exponent (n) are also listed in Table 5.
Standard deviation coefficients (standard deviation/average) of fatigue crack growth tests are
generally less than 0.2, which means reliable and reproducible test procedure.

Evaluation of the Results

In our earlier works we showed that there is a correlation between the two parameters of both the
Manson-Coffin and the Paris-Erdogan equation [20]. Fig. 4 shows the correlation between the C
and n parameters of the Paris-Erdogan law for steels (270 data pairs) and their welded joints (111
data pairs), under mode I and mixed mode I+II loading conditions [21].
Materials Science Forum Vols. 414-415 293
Table 5 Parameters of the Manson-Coffin and the Paris-Erdogan equation

Material grade Manson-Coffin equation, c Paris-Erdogan equation (R = 0.1), n
average standard deviation
37C -0.287 3.44 0.311
A106B -0.487 3.52
A515 Grade70_bm -0.553 3.6 (R = 0.03)
A515 Grade70_wm -0.709 3.2 (R = 0.03)
KL7D -0.498 3.36 0.381
DX52 -0.784 3.11 0.140
X80TM -0.478 2.49 0.561
QStE690TM_FCG 2.39 0.498
StE690_LCF -0.659
Waspaloy (650C)_FCG 3.66 (R = 0.05) 0.195
Waspaloy (650C)_LCF -0.721
AlMg3 -0.787 3.37 0.503
AlMg5_FCG 3.71 0.405
AlMg5.1Mn_LCF -0.655
AlMg4.5Mn_FCG 3.57 0.444
AlMg4.5Mn_LCF -0.755
2024-T6_FCG 2.21 (R = 0)
2024-T4_LCF -0.649
7075-T6_FCG 2.13
7075-T6_LCF -0.987
Fig. 4 Connection between the two Paris-Erdogan constants for steels and their welded joints

1
2
3
4
5
6
7
8
9
10
-17 -15 -13 -11 -9 -7 -5
Paris-Erdogan constant, lgC
P
a
r
i
s
-
E
r
d
o
g
a
n

e
x
p
o
n
e
n
t
,

n
Base materials (mode I and I+II)
Welded joints (mode I)
Linear (Base materials (mode I and I+II))
Linear (Welded joints (mode I))
Mode I:
D DD D
K
Mixed mode I+II:
D DD D
K
eff
da/dN: mm/cycle;
D DD D
K,
D DD D
K
eff
: MPa m
1/2
294 Materials Science, Testing and Informatics I
Based on these results it is sufficient to study the relationship of the two exponents (c and n).
Fig. 5 shows the relation of these material constants.
Fig. 5 Connection between the exponents of Manson-Coffin and Paris-Erdogan equations

Despite of the relatively few measurement data we can establish that there is an admissible
connection between the exponents of the Manson-Coffin equation characterising low cycle fatigue
and the Paris-Erdogan law describing fatigue crack growth.
The connection among the steel groups with different microstructures is different, which is
demonstrated by the example of five micro-alloyed steels (37C, A106B, A515 Grade70, KL7D,
DX52) and two grades of thermomechanically treated steels (X80TM, QStE690TM/StE690). Both
the results from base materials and from weld metals can be used for the evaluation of the
connections.
The connection between the aluminium alloy groups of different microstructures is different, too,
which is demonstrated by the example of some alloys without hardening (AlMg3,
AlMg5.1Mn/AlMg5, AlMg4.5Mn) and other alloys with hardening (2024-T6/2024-T4, 7076-T6).
It is probable, that there is a third connection for superalloys. There is only one data point for this
material group, therefore the verification of this assumption requires further investigations.
Comparison of the various material grades (e.g. micro-alloyed steels aluminium alloys, or
micro-alloyed steels stainless steels, or micro-alloyed steels superalloys), and comparison of the
results corresponding to the same material group at various temperatures requires further
investigations.

Summary

Based on the theoretical considerations and the experimental results the following establishments
can be made.
In the plastic zone in front of the crack tip the time dependence of the stress is similar in case of
low cycle fatigue and fatigue crack growth.
2
2.2
2.4
2.6
2.8
3
3.2
3.4
3.6
3.8
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2
Manson-Coffin exponent, c
P
a
r
i
s
-
E
r
d
o
g
a
n

e
x
p
o
n
e
n
t
,

n
37C, A106B, A515Gr70 bm/wm, KL7D, DX52
X80TM, QStE690TM/StE690
AlMg3, AlMg5/AlMg5.1Mn, AlMg4.5Mn
2024-T6/2024-T4, 7075-T6
Waspaloy
Materials Science Forum Vols. 414-415 295
The cyclic plastic deformation results in near the same cell type of dislocation structure in case
of low cycle fatigue and fatigue crack growth.
There is a connection between the exponents of the Manson-Coffin equation characterising low
cycle fatigue and the Paris-Erdogan law describing fatigue crack growth. The connection has been
shown for different groups of steels (micro-alloyed steels, thermomechanically treated steels) and
aluminium alloys (alloys without hardening and with hardening).

Acknowledgements

The authors wish to acknowledge the assistance given by the National Scientific Research
Foundation (Orszgos Tudomnyos Kutatsi Alapprogramok) for supporting the research (OTKA
F4418, OTKA T025428, OTKA T022020 and OTKA T034503).

References

[1] M. Klesnil and P. Lukas: Fatigue of Metallic Materials. (Academia, Prague 1980).
[2] E. Feltner and C. Laird: Trans. Met. Soci. AIME (1980), p. 1253.
[3] E. Feltner and C. Laird: Acta Met. Vol. 15 (1967), p. 1633.
[4] H. Nahm, J. Moteff and D. R. Diercks: Acta Met. (1977), p. 107.
[5] S. Taira and K. Tanaka: Eng. Fract. Mech. Vol. 4 (1972), p. 925.
[6] J. Awatani, K. Katagiri and T. Shiraishi: Met. Trans. (1976), p. 807.
[7] J. Awatani, K. Katagiri and H. Nakai: Met. Trans. (1978), p. 111.
[8] P. Lukas, M. Klesnil and R. Fiedler: Phil. Mag. (1969), p. 799.
[9] P.K. Liaw, A. Saxena and J. Perrin: Eng. Fract. Mech. Vol. 45 (1993), p. 759.
[10] H.A.L Aberasturi, J.M.R. Ibabe and M.F. Perez: ECF9 Reliability and Structural Integrity of
Advanced Materials, Eds.: S. Sedmak, A. Sedmak and D. Ruzis, Vol. I (EMAS, Warley, West
Midlands 1992), p. 397.
[11] Chr. Boller and T. Seeger: Materials Data for Cyclic Loading. Part B: Low-alloy steels.
(Elsevier, Amsterdam 1987).
[12] J. Byrne et al.: Proceedings of the Sixth International Fatigue Congress, Eds.: G. Ltjering
and H. Nowack, Vol II (Pergamon, Elsevier 1996), p. 849.
[13] C. Moura Branco, Lecture notes Preprints of the NATO ASI Mechanical behaviour of
materials at high temperature. Sesimbra, (1995).
[14] I. Trk: Publ. Univ. Miskolc, Series. C, Mech. Engng. Vol 46 (1996), p. 33.
[15] Chr. Boller and T. Seeger: Materials Data for Cyclic Loading. Part B: Aluminium and
titanium alloys. (Elsevier, Amsterdam 1987).
[16] J.B. Lee and D.N. Lee: Proceedings of the Sixth International Conference on Fracture, Eds.:
S.R. Valluri et al., Vol 3 (Pergamon Press 1984), p. 1727.
[17] J.-K. Lim et al.: Proceedings of the Sixth International Fatigue Congress, Eds.: G. Ltjering
and H. Nowack, Vol II (Pergamon, Elsevier 1996), p. 947.
[18] S.S. Manson: Behaviour of materials under conditions of thermal stress, NACA Report 1170.
(Levis Flight Propulsion Laboratory, Cleveland 1954).
[19] P. Paris and F. Erdogan: Journ. Bas. Engng. Trans. ASME (1963), p. 528.
[20] L. Tth, GY. Nagy and P. Romvri: Proceedings of the 7th European Conference on
Fracture, Ed.: E. Czoboly, (EMAS, Warley, West Midlands 1988), p. 649.
[21] J. Lukcs: Habilitation booklet: Reliability of cyclic loaded welded joints and structures
having cracks, (University of Miskolc, Miskolc 1998).

296 Materials Science, Testing and Informatics I
Materials Science, Testing and Informatics I
10.4028/www.scientific.net/MSF.414-415


Connection between the Results of the Low-Cycle Fatigue and the Fatigue Crack Growth Tests
10.4028/www.scientific.net/MSF.414-415.289

DOI References
[3] E. Feltner and C. Laird: Acta Met. Vol. 15 (1967), p. 1633.
doi:10.1016/0001-6160(67)90138-1
[4] H. Nahm, J. Moteff and D. R. Diercks: Acta Met. (1977), p. 107.
doi:10.1016/0001-6160(77)90114-6
[6] J. Awatani, K. Katagiri and T. Shiraishi: Met. Trans. (1976), p. 807.
doi:10.1007/BF02644077
[7] J. Awatani, K. Katagiri and H. Nakai: Met. Trans. (1978), p. 111.
doi:10.1007/BF02647179
[8] P. Lukas, M. Klesnil and R. Fiedler: Phil. Mag. (1969), p. 799.
doi:10.1007/BF01712864
[9] P.K. Liaw, A. Saxena and J. Perrin: Eng. Fract. Mech. Vol. 45 (1993), p. 759.
doi:10.1016/0013-7944(93)90064-Y