Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 129 (2015) 458 464

International Conference on Industrial Engineering

Prediction of stress-life curve of a material under random or cyclic

loading with asymmetry
Erpalov A.V.a*, Shefer L.A.a
a
South Ural State University, 76, Lenin Avenue, Chelyabinsk, 454080, Russian Federation

Abstract

Estimation of structures fatigue life under dynamic loading remains an unexplored area in the field of mechanical engineering.
Accounting the asymmetry under operational loading is one of features of durability computation for a part of a structure.
Common methods of asymmetry considerations result in significant errors in the lifetime of computed structures. This paper
describes the methodology approach that allows to predict a fatigue curve for any asymmetry coefficients using the results of a
limited number of tests with symmetry and asymmetry under random or cyclic loading. This approach is based on the
experimental results of the herein presented research. The scattering signature of the material is necessary for fatigue probability
calculations. The methodology also presents the recipes for prediction of fatigue curve and scattering signature with the given
asymmetry coefficient and given probability of failure.
2015 The Authors. Published by Elsevier Ltd.
2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the organizing committee of the International Conference on Industrial Engineering (ICIE-
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
2015).
Peer-review under responsibility of the organizing committee of the International Conference on Industrial Engineering (ICIE-2015)

Keywords: S-N curve; Asymmetry loading; Random loading; Cyclic loading; Scattering signature.

1. Introduction

Currently, coefficient is used for the estimation of fatigue life of a structure under asymmetric loading [1-4]. It
is assumed that the Haigh diagram of stress is described by single straight line, and the value of the coefficient
varies within a small range for various materials. According to [5], coefficient varies from 0.1 to 0.2 for carbon
steels, and from 0.2 to 0.3 for alloyed steels. The selection of a value for a material is difficult and it does not

* Corresponding author. Tel.: +7-351-267-90-62

E-mail address: aerpalov@ya.ru

1877-7058 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the International Conference on Industrial Engineering (ICIE-2015)
doi:10.1016/j.proeng.2015.12.151
 A.V. Erpalov and L.A. Shefer / Procedia Engineering 129 (2015) 458 464 459

correlate well with the experimental results in most cases. It is worth noting that the majority of researches study the
cyclic loading [6-9], while random loading with asymmetry remains poorly explored [10, 11].
The purpose of this study is to determine the influence of asymmetry loading on the fatigue curve of a material
under cyclic or random loadings.

Nomenclature

coefficient of Haigh diagram

a stress amplitude value
S standard deviation of stress value
P asymmetry coefficient
m mathematical expectation value of stress
Rm tensile strength of a material
G loading process structure parameter
lgN logarithm of the lifetime
A constant parameter of fatigue curve
B constant parameter of fatigue curve
S0 standard deviation of stress value corresponding to unbounded value fatigue strength
P probability of failure
r correlation coefficient
Ss standard deviation of endurance strength
vs coefficient of variation
S expectation value of endurance strength
SlgN standard deviation of lifetime logarithm
tq inverse normal distribution for the failure probability

2. Fatigue curve equation

Random asymmetrical loading process cannot be characterized by the standard parameters for the cyclic process
without schematization of the original stress history. For example, the standard for the cyclic process stress
amplitude a is not a valid parameter for random process. In this case, the amplitude for the random process is
replaced by the standard deviation (S) of stress value. For cyclic loading with amplitude a, the standard deviation of
stress value is calculated by
Va
S . (1)
2
In this paper, the asymmetry coefficient for cyclic and random loadings is defined as the ratio of the
mathematical expectation value of stress m to tensile strength Rm of the material:
Vm
P (2)
Rm
The processes of loading (cyclic and random) characterize the combination of the following parameters described
in [12]: the mathematical expectation value of stress m, the standard deviation of stress value S and the loading
process structure parameter G. The number of the intersections by the process of the mean level of the mathematical
expectation with the positive sign of the process derivative is taken as the lifetime of the random process. For the
cyclic loading the value N0 is the number of cycles. There are many methods to process the results of fatigue tests
[13-17]. The algorithm described in [12] is used in this work. The expectation of logarithm of the lifetime <lgN> is
described by S-N curve equation:

lg N0i A  B lg  Si  S0  . (3)
460 A.V. Erpalov and L.A. Shefer / Procedia Engineering 129 (2015) 458 464

3. Fatigue tests

The research was performed using random and cyclic types of loading. Flat-shaped specimens made of light
alloys AMg-61 and MA-15 were tested (see Fig. 1). Chemical compositions of the tested alloys are given in Tables
1 and Table 2. Fatigue properties of specimens were tested by the electrodynamic shaker LDS V780 that allows
cyclic or random loading of different bandwidths. Description of the equipment is given in [18].

grabs

R40
30

10 40
100

Fig.1. Geometry of specimen.

Table 1. Chemical composition of AMg-61 alloy (in %)

Mg Al
5,70 94,30

Table 2. Chemical composition of MA-15 alloy (in %)

Mg Al Si Ca Fe Zn Cd
85,74 6,76 0,39 0,41 0,42 1,60 4,68

At each stress level at least six specimens were tested. The specimens were tested until complete failure.

4. Analysis of the results

From the test results of AMg-61 alloy, the fatigue curves under cyclic and random loadings for the symmetry and
asymmetry with the probability of failure P = 0.5 were obtained using Eq. 3. (see. Fig. 2). Each point in these plots
is the expectation of logarithm of the lifetime corresponding to the stress level. From the plots it is hard to assess if
the curves are parallel to each other thus we have to use different variables: the standard deviation of stress value S
and the asymmetry coefficient .
 A.V. Erpalov and L.A. Shefer / Procedia Engineering 129 (2015) 458 464 461

Fig. 2. (a) fatigue curves of AMg-61 alloy under random; (b) cyclic loadings with asymmetry.

According to [19], the range of interest for the high cycle fatigue life corresponds to lgN = 4 ... 7 for calculating
the lifetime of structures. Therefore, the set of lifetimes lgNi = {4; 5; 6; 7} in increments of lgN = 1 was selected.
The values of standard deviation of stress for each lifetime lgNi were found using the plots of fatigue curves (see
Fig. 2) and Eq. 3, accounting for asymmetry. Points in Fig. 3 are well fitted with the linear function with correlation
coefficients are r > 0,993. Each line corresponds to different value of lifetime logarithm.

Fig. 3. (a) standard deviation of stress S vs. asymmetry coefficient for AMg-61 alloy under random; (b) cyclic loadings.

As can be seen from Fig. 3, all lines converge in one point. However, as tests have shown, coordinates of this
point depend on the loading process structure parameter G and cannot be predicted due to limited number of tests. It
should be noted that the point of convergence of all lines has no physical meaning because it lies outside the region
of tensile strength of the material.
In order to verify the tests results, the vibration tests of MA-15 alloy under cyclic loading with asymmetry have
been performed. The plots of fatigue curves and relations of standard deviation of stress value S to the asymmetry
462 A.V. Erpalov and L.A. Shefer / Procedia Engineering 129 (2015) 458 464

coefficient are shown in Figure 4. As seen from the plots, the lines converge in a single point, same as in the case
of AMg-61 alloy.

Fig. 4. (a) for cyclic loading, fatigue curves of MA-15 alloy with asymmetry; (b) standard deviation of stress S vs. asymmetry coefficient .

The analysis of the plots shows the possibility of predicting fatigue curves with asymmetry. It is sufficient to
perform fatigue tests under symmetric loading and just one additional test with an arbitrary value of asymmetry.
Then, one needs to plot linear graphs of standard deviation of stress value S vs. asymmetry coefficient . If the lines
converge in single point, then test results are correct. Thus, for any value of the asymmetry coefficient one can
find the corresponding standard deviation of stress value S for each line. The number of these lifetime lines can be
increased by adding more values lgN. Thus, it is possible to obtain the points (lgN; S) that are processed as in [12]
and obtain the material fatigue curve for a given asymmetry with failure probability P = 0.5.
For the probability calculation of lifetime it is necessary to know the characteristics of scattering signature of a
material under asymmetric loading. Dispersion of endurance strength obeys the normal distribution and may be
estimated from the standard deviation Ss. However, in practice it is convenient to use the relative value a
coefficient of variation vs instead of Ss. According to [20], coefficient of variation vs is a constant value and does not
depend on number of cycles to fracture
Ss
vs const . (4)
S
The coefficient of variation can be found from the standard deviation of lifetime logarithm SlgN [20]:
2.3 Slg Ni  Si  S0 
vsi (5)
B Si
Thus, the standard deviation of stress value S with a failure probability can be calculated as follows:

S S r tq S s (6)

According to Eq. 4:

S S r tq vs S , (7)

Consider the relationship of the standard deviation of stress value S and the asymmetry coefficient for AMg-61
alloy under random loading from Fig. 5, where two solid lines for lgN = 5 and lgN = 7 lifetimes were plotted. For
 A.V. Erpalov and L.A. Shefer / Procedia Engineering 129 (2015) 458 464 463

each experimentally obtained point, the confidential interval with confidence level tq = 3 was obtained using Eq. 7.
The regression values through the points were plotted in Fig. 5 using dashed lines. The smallest correlation
coefficient is r = 0.995. These lines show the scattering signature for the failure probability.

Fig. 5. Standard deviation of stress S vs. asymmetry coefficient for AMg-61 alloy under random loadings, and scattering signature.

5. Conclusions

The methodology has been developed and implemented. It has predicted fatigue curves of a material with
different asymmetry under cyclic and random loadings using limited experimental tests. Furthermore, the
methodology allows accounting for the scatter signature of a fatigue curve and plots this fatigue curve with given
asymmetry and failure probability.

References

[1] J. Schijve, Fatigue of Structures and Materials, CBS Publishers, New Delhi, 2009.
[2] A.V. Zhelubovskiy, A.D. Pogrebnjak, M.N. Regulski, A.T. Sagitov, Y.V. Klyuchnikov, P.V. Kondrashev, Reserve estimate of strength of
machine parts subject laden with asymmetrically, Easter-European Journal of Enterprise Technologies. 66 (2013) 812.
[3] V.P. Golub, E.S. Kochetkova, A.D. Pogrebnyak, Metodika postroeniya diagrammy predel'nykh napryazheniy pri asimmetrichnom
mnogotsiklovom nagruzhenii, Industrial laboratory. Materials diagnostics. 72 (2006) 4248.
[4] M. Nihei, P. Heuler, C. Boller, T. Seeger, Evaluation of mean stress effect on fatigue life by use of damage parameters, International Journal
of Fatigue. 8 (1986) 11926.
[5] L.M. Shkol'nik, Metodika ustalostnykh ispytaniy. Spravochnik, Metallurgy, Moscow, 1978.
[6] Ying Xiong, Yanyao Jiang, Fatigue of ZK60 magnesium alloy under uniaxial loading, International Journal of Fatigue. 64 (2014) 7483.
[7] Johannes Dallmeier, Otto Huber, Holger Saage, Klaus Eigenfeld, Uniaxial cyclic deformation and fatigue behavior of AM50 magnesium
alloy sheet metals under symmetric and asymmetric loadings, Materials and Design. 70 (2015) 1030.
[8] Guozheng Kang, Yujie Liu, Zhao Li, Experimental study on ratchetting-fatigue interaction of SS304 stainless steel in uniaxial cyclic stressing,
Materials Science and Engineering. 435-436 (2006) 396404.
[9] M. Liakat, M.M. Khonsari, An experimental approach to estimate damage and remaining life of metals under uniaxial fatigue loading,
Materials and Design. 57 (2014) 289297.
[10] Tadeusz Lagoda, Ewald Macha, Roland Pawliczek, The influence of the mean stress on fatigue life of 10HNAP steel under random loading,
International Journal of Fatigue. 23 (2001) 283291.
[11] Krzysztof Kluger, Tadeusz Lagoda, Fatigue lifetime under uniaxial random loading with different mean values according to some selected
models, Materials and Design. 28 (2007) 26042610.
[12] L.A. Shefer, A.V. Erpalov, D.Kh. Valeev, Obobshchennaya diagramma ustalosti materialov pri deystvii razlichnykh sluchaynykh,
garmonicheskikh i poligarmonicheskikh protsessov, Industrial laboratory. Materials diagnostics. 82 (2015) 5862.
464 A.V. Erpalov and L.A. Shefer / Procedia Engineering 129 (2015) 458 464

[13] S.A. Faghidian, A. Jozie, M.J. Sheykhloo, A. Shamsi, A novel method for analysis of fatigue life measurements based on modified Shepard
method, International Journal of Fatigue. 68 (2014) 144149.
[14] Yong Xiang Zhao, A fatigue reliability analysis method including super long life regime, International Journal of Fatigue. 35 (2012) 79-90.
[15] Qin Sun, Hong-Na Dui, Xue-Ling Fan, A statistically consistent fatigue damage model based on Miners rule, International Journal of
Fatigue. 69 (2014) 1621.
[16] G. Zonfrillo, A. Rossi, A fuzzy approach for drawing SN curves in the nite life region, Fatigue & Fracture of Engineering Materials &
Structures. 34 (2011) 868876.
[17] Adam Niesony, Chalid el Dsoki, Heinz Kaufmann, Peter Krug, New method for evaluation of the MansonConBasquin and Ramberg
Osgood equations with respect to compatibility, International Journal of Fatigue. 30 (2008) 19671977.
[18] A.V. Erpalov, L.A. Shefer, E.E. Rikhter, P.A. Taranenko, Ustalostnye ispytaniya materialov i konstruktsiy s ispol'zovaniem sovremennogo
oborudovaniya, Bulletin of the SUSU, Series "Mechanical engineering industry". 15 (2015) 7080.
[19] V.P. Kogaev, N.A. Makhutov, A.P. Gusenkov, Raschety detaley mashin i konstruktsiy na prochnost' i dolgovechnost'. Spravochnik,
Mashinostroenie, Moscow, 1985.
[20] L.A. Shefer, Veroyatnostnye metody rascheta resursa i zapasov prochnosti nesushchikh elementov konstruktsiy transportnykh system, SUSU
Published, Chelyabinsk, 2000.