See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/312523314

Unification proposals for fatigue crack propagation laws

Article  in  Multidiscipline Modeling in Materials and Structures · August 2017

DOI: 10.1108/MMMS-10-2016-0052

CITATION READS

1 200

1 author:

Vladimir Kobelev
Universität Siegen
196 PUBLICATIONS   654 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Effect of anisotropy on static and dynamic aeroelasticity View project

Exactly solvable optimization problems in structural mechanics View project

All content following this page was uploaded by Vladimir Kobelev on 12 October 2017.

The user has requested enhancement of the downloaded file.

MMMS
13,2
262
Received 17 October 2016
Revised 9 December 2016
4 January 2017
Accepted 19 January 2017
Multidiscipline Modeling in
Materials and Structures
Vol. 13 No. 2, 2017
pp. 262-283
© Emerald Publishing Limited
1573-6105
DOI 10.1108/MMMS-10-2016-0052
Unification proposals for fatigue
crack propagation laws
Vladimir Kobelev
Department of Natural Sciences, University of Siegen, Siegen, Germany
Abstract
Purpose – The purpose of this paper is to propose the new dependences of cycles to failure for a given
initial crack length upon the stress amplitude in the linear fracture approach. The anticipated unified
propagation function describes the infinitesimal crack-length growths per increasing number of load
cycles, supposing that the load ratio remains constant over the load history. Two unification functions with
different number of fitting parameters are proposed. On one hand, the closed-form analytical solutions
facilitate the universal fitting of the constants of the fatigue law over all stages of fatigue. On the other
hand, the closed-form solution eases the application of the fatigue law, because the solution of nonlinear
differential equation turns out to be dispensable. The main advantage of the proposed functions is the
possibility of having closed-form analytical solutions for the unified crack growth law. Moreover, the mean
stress dependence is the immediate consequence of the proposed law. The corresponding formulas for
crack length over the number of cycles are derived.
Design/methodology/approach – In this paper, the method of representation of crack propagation functions
through appropriate elementary functions is employed. The choice of the elementary functions is motivated by
the phenomenological data and covers a broad region of possible parameters. With the introduced crack
propagation functions, differential equations describing the crack propagation are solved rigorously.
Findings – The resulting closed-form solutions allow the evaluation of crack propagation histories on one
hand, and the effects of stress ratio on crack propagation on the other hand. The explicit formulas for crack
length over the number of cycles are derived.
Research limitations/implications – In this paper, linear fracture mechanics approach is assumed.
Practical implications – Shortening of evaluation time for fatigue crack growth. Simplification of the
computer codes due to the elimination of solution of differential equation. Standardization of experiments for
crack growth.
Originality/value – This paper introduces the closed-form analytical expression for crack length over
number of cycles. The new function that expresses the damage growth per cycle is also introduced.
This function allows closed-form analytical solution for crack length. The solution expresses the number of
cycles to failure as the function of the initial size of the crack and eliminates the solution of the nonlinear
ordinary differential equation of the first order. The different common expressions, which account for the
influence of the stress ratio, are immediately applicable.
Keywords Generalized Paris-Erdogan law, Linear fracture mechanics, Stress ratio influence
Paper type Technical paper
Nomenclature
Kmax maximum stress intensity R=Kmin/Kmax o 1 stress ratio of cyclic load
factor per cycle Km=(Kmax+Kmin)/2 mean value of stress
Kmin minimum stress intensity intensity factor
factor per cycle UI(K) unified propagation
K range of stress intensity function of type I
factor UII(K) unified propagation
σ = σmax−σmin stress range function of type II
σmax maximum stress per cycle p>1 fatigue exponent
σmin minimum stress per cycle m2>1 exponent at short-term
Y dimensionless geometry limit
parameter m1>1 endurance limit exponent
c = c(R) material constant for a K2 short-term threshold limit
(K2>K1)
given stress ratio R
 Unification proposals for fatigue crack propagation laws

Unification proposals for fatigue crack

propagation laws
In the present paper, the new dependences of cycles to failure for a given initial crack length upon the
stress amplitude in the linear fracture approach are proposed. The anticipated unified propagation
function describes the infinitesimal crack length growths per increasing number of load cycles, suppos-
ing that the load ratio remains constant over the load history. Two unification functions with different
number of fitting parameters are proposed. The main advantage of the proposed functions is the pos-
sibility of closed form analytical solutions for the unified crack growths law. On one side, the closed
form analytical solutions facilitate the universal fitting of the constants of the fatigue law over all stages
of fatigue. On the other side, the closed-form solution eases the application of the fatigue law, because
the solution of nonlinear differential equation turns to be dispensable. Moreover, the mean stress
dependence is the immediate consequence from the proposed law. The corresponding formulas for
crack length over the number of cycles are derived.

Keywords: generalized Paris-Erdogan law; linear fracture mechanics; stress ratio influence

1
 Unification proposals for fatigue crack propagation laws

Contents
1 The crack growth rate ..................................................................................................................... 2
2 The cyclic crack propagation laws ................................................................................................... 6
3 The unification of Paris law type I ................................................................................................... 8
4 Limit cases for cycles to fracture for the type I ............................................................................. 10
5 The unification of the fatigue law type II ...................................................................................... 11
6 Limit cases for cycles to fracture for the type II ............................................................................ 13
7 Influence of stress ratio ................................................................................................................. 14
8 Conclusion ..................................................................................................................................... 17

1 The crack growth rate

The functional relation between the loading properties, stress gradient and the physical time-depend-
ing characteristics of materials are essential for the statistical fatigue analysis and analytical evaluation
number of the cycles to failure (Carpinteri, 1994, Suresh, 1998, Fatemi and Yang, 1998). The traditional
evaluations concepts based on the Palmgren-Miner’s rule of damage accumulation, rain-flow counting
of time-dependent loads, Wöhler curve and Basquin equation, Paris-Erdogan law and diverse general-
izations of the fracture and damage mechanics approaches (Richard, Sander, 2012). The book (Krupp,
2007) reviews the numerical treatment of fatigue microscopic crack propagation together with their
implementation in fatigue-life prediction models.
There are three stages of fatigue fracture commonly distinguished: initiation, propagation, and final
rupture (Totten, 2008), (Leese, 1989).
The first stage (stage I) of fatigue is referred to as initiation. Initiation is probably the most complex
stage of fatigue fracture. The most significant factor about the initiation stage of fatigue fracture is
that the irreversible alterations in the metal are caused by cyclic shear stresses. The accumulation of
microscopic faults over a large number of load applications, leads to cumulative damage. At the loca-
tion of a severe stress concentration, the number depends on the geometry of the part as well as on
environmental, stress, metallurgical, and strength conditions, as will become apparent. During the
stage I, the spread of fatigue crack per unit cycle decelerates smoothly with number of cycles and the
process approaches its second stage.
The second stage of fatigue is known as crack propagation (stage II). The propagation stage of fatigue
causes the microscopic crack to change direction and grow perpendicular to the tensile stress. The
second, or propagation, stage of fatigue is typically the most readily recognisable area of a fatigue
fracture. The Paris law describes the stage II propagation law (Paris and Erdogan, 1963). The traditional
form of Paris law pronounces the spread of fatigue crack per unit cycle as a power function of the
range of stress intensity factor. Towards the end of the stage II, the spread of fatigue crack per unit
cycle begins to accelerate smoothly with number of cycles and the process turns to the third stage.
The final, third stage of fatigue is the final rupture (stage III). As the propagation of the fatigue crack
endures, progressively sinking the cross-sectional area of the test specimen, it eventually deteriorates
the part so significantly that final, broad fracture occurs with a couple of load cycles. The fracture mode
may be either ductile (with a dimpled fracture surface) or brittle (with a cleavage, or intergranular,
2
 Unification proposals for fatigue crack propagation laws

fracture surface). Occasionally occur the combinations of these modes, depending upon the metal
concerned, the stress level and the environment. In the course of the stage III, the spread of fatigue
crack per unit cycle progressively accelerates with each cycle.
Numerous load damage models, which extend the celebrated Paris and Erdogan propagation law, were
proposed since the 1970s. The article (Beden, Abdullah, Ariffin, 2009) reviews the articles focused on
the prediction of fatigue properties of structures under variable amplitude loading. The reliability and
accuracy of prediction models and the physical concept of fatigue damage was discussed. The reviewed
fatigue life models were based on the scientific and engineering knowledge about fatigue of material
and structures under constant and variable amplitude loading.
The near-threshold deviation of the common propagation law was suggested by (Donahue et al. 1972).
The observed rapidly increasing growth towards ductile tearing was accounted. The brittle fracture
correction of the power law was proposed in the cited article in the analytical form.
In the work (Kanninen, Popelar, 1985) several combinations of high and low amplitude of stress inten-
sity factors values were studied. A similar approach of for transition region between high and low am-
plitude of stress intensity factors was proposed in (McEvily, Groeger, 1977).
A more general ‘‘unified law’’ accounts certain deviations from the power-law regime (Pugno et al
2006). The implemented extension of the Paris law for crack propagation results in the generalized
Wöhler fatigue curves.
Under low cycle fatigue condition the Manson–Coffin law is commonly applied. Well known, that un-
der extremely low cycle fatigue condition the usually applied Manson–Coffin law tends to over-predict
the cyclic life. A different expression is proposed for the correction extremely low cycle fatigue behav-
iour of in the paper (Xue, 2008). The proposed formulas describe the cyclic life by introducing an ex-
ponential function, making possible to cover the entire span of extremely low cycle fatigue to low cycle
fatigue. The curves are fitted by means of an additional material parameter.
The methodology to the modelling of material fatigue in cyclic loading has been suggested on the basis
of the kinetic theory of strength in (Mishnaevsky, Brøndsted, 2007) . According to this concept, the
acting stresses instead of stress changes yield the damage growth in materials under fatigue condi-
tions. In order to model the material degradation the kinetic theory of strength was applied. The model
allows the determination of the stiffness reduction in composite with a definite load history. Moreo-
ver, the model resolves the effect of the loading frequency on the lifetime and the damage growths
per cycle. It was shown that the number of cycles to failure increases almost linearly. With increasing
the loading frequency the damage growth per cycle decreases.
By applying the quantized fracture mechanics the unification and extension of the traditional Paris’
and Wöhler’s laws for fatigue crack propagation were derived in (Pugno, Cornetti, Carpinteri, 2007). In
the cited article the generalized Paris’, Wöhler’s or unified laws was suggested. The performed analysis
demonstrated the applicability of proposed laws for predicting the life time of structures containing
fatigue cracks. The sizes of fatigue cracks was varied from small (the Wöhler’s regime) to large (the
Paris’ regime).
The revision of the fatigue crack growth model was proposed by (Castillo, Fernández-Canteli, Siegele,
2014), CCS model. The crack growth model for derivation of the S–N curves from the crack growth rate
curves was presented. The CCS model complements the common stress based and fracture mechanics
methodologies in fatigue lifetime through the enhanced cracks propagation law. Firstly, the crack

3
 Unification proposals for fatigue crack propagation laws

growth rate curve is defined over all its range as a cumulative distribution function based on a normal-
ized dimensionless stress intensity factor range. The CCS model is based on the assumption that fatigue
crack growth takes the form of a Gumbel cumulative distribution function (Gumbel, 1954). For this
prediction of crack growth rate the Buckingham theorem (Buckingham, 1915) was incorporated. Sec-
ondly, the proprietary theorem was derived. The derived theorem provides an alternative to self-sim-
ilarity and assures the significant reduction of experimental planning. Correspondingly the different
crack growth curves for different stress ranges and initial crack lengths are obtained from a specific
crack growth curve. The S–N field results from the crack growth curves. The close relation between
the fracture mechanics and stress approaches was demonstrated. Thirdly, the model was applied to a
certain set of experimental data. The crack growth rate curve and the S–N curves of a certain material
were derived.
The substantial modification of CSS model was proposed in paper (Blasón et al, 2016). This modifica-
tion accounts the crack opening and closing effects as well as the influence of the stress ratio. The
theoretical model incorporates both effects of the average stress as well crack closure and opening
action. The model provided an analytical expression of the crack propagation rate curve by matching
experimental data by means of the least squares method. The identification of new variables was in-
troduced for the fatigue cracking model. The auxiliary variables include the cracking and opening ef-
fects using the influence of the stress ratio. The normalized variables allows the values in the range 0
to 1. Thus the use of "S-shaped" cumulative distribution functions reproduces the relation between
the variation of stress intensity factor and number of cycles. The integration of several models that
come from the statistical domain to solve the problem of crack growth is possible. The a-N and S-N
curves describe the crack closure effects for the materials on different initial crack size values and
loads.
The common fatigue crack propagation laws must also evaluate opening and closure effects for fatigue
crack. For this purpose a corresponding modification of the CCS fatigue crack growth law was proposed
in the paper (Correia et al, 2016a). The modification incorporated the crack opening and closure effects
together with the stress ratio effect. The fatigue crack opening and closure effects were explained with
the plasticity-induced crack-closure model. With the introduction of the phenomenological auxiliary
parameter, an alternative fatigue crack closure models was implemented. The corresponding modified
CCS crack propagation model was based on the effective stress intensity factor range. The version
model was supported by mathematical and physical considerations.
The model (de Castro, Meggiolaro, Miranda, 2009) predicts crack growth by means of crack initiation
properties and critical damage concepts. The crack was modelled as a sharp notch with a very small
but finite tip radius. The finite radius removes the singularity of elastic fields at the tip vicinity. Accord-
ingly the damage caused by each load cycle and the effects of residual stresses in the vicinity of the
crack tip evaluates immediately through the hysteresis loops caused by the loading. The calculation of
cycle-bicycle crack growth was based on the introduced method. An agreement between the crack
growth forecasts and experiments was achieved in the citing paper for constant and for variable am-
plitude loadings.
A number of prediction models are known for fatigue crack growth under variable amplitude loading
histories. However, no of these model predicts adequately considering loading sequence effects. An
improved crack growth rate model has been proposed in (Chen, Wang, Cui, 2011). The cited model
deals the constant amplitude loading demonstrate its validity for experimental data. The applicability
of the improved crack growth rate model was extended to variable amplitude loading. The extension
accounts crack closure level based on the concept of partial crack closure. The concept was based on
4
 Unification proposals for fatigue crack propagation laws

the crack-tip plasticity. To denote the variation in the affected zone of under- or overload, a modified
coefficient was introduced. The identification of the coefficient was based on Wheeler model
(Wheeler, 1972).
The fatigue crack growth model (Noroozi, Glinka, Lambert, 2005) was based on the elastic–plastic crack
tip stress–strain history. In the FCG model the fatigue crack growth simulates the stress–strain re-
sponse in the vicinity of the crack tip and estimating the accumulated fatigue damage. The fatigue
crack growth was treated as sequential crack re-initiation near the crack tip. The model predicted the
effect of the mean stress and the influence of the compressive stress. A fatigue crack growth was an-
alysed for the plane strain and for plane stress states. The fatigue crack growth was demonstrated to
be controlled by a two parameter driving force. The driving force itself evaluates on the basis of the
local stresses and strains at the crack tip using the Smith–Watson–Topper fatigue damage parameter
(Smith, Watson, Topper, 1970). The residual (internal) stresses are induced by the reversed cyclic plas-
ticity. The effect of the residual stress could be also accounted with the proposed model. The applied
stress intensity factors were modified to the total stress intensity factors to be accounted for the effect
of the local crack tip stresses and strains on fatigue crack growth. The fatigue crack growth was fore-
casted by the stress–strain response in the vicinity of the crack tip on one side and estimating the
accumulated fatigue damage on the other side. The fatigue crack growth is the process of successive
crack re-initiations in the crack tip region. The model estimates the effect of the mean and residual
stresses induced by the cyclic loading. The proposed model evaluates the effect of variable amplitude
loadings on fatigue crack growth. The influence of the internal stress induced by the reversed cyclic
plasticity alternates the resultant stress intensity factors controlling the fatigue crack growth.
Two singularity fields near the crack tip are known. Namely, the first type of tip field is the Hutchinson-
Rice-Rosengren field (Hutchinson, 1968), (Rice, Rosengren, 1968). The second type is the Kujawski-
Ellyin field (Kujawski, Ellyin, 1984) and (Ellyin, 1986). The effect of these singularity fields was examined
in the paper (Shi, Cai, Bao, 2014). The study (Shi, Cai, Bao, 2014) evaluates the common fatigue crack
growth model, which is based on energy balance during growth of the crack. The new parameter was
introduced. With this parameter the effect of different types of singularity fields was incorporated in
the crack propagation model. The model predicted the stage-II fatigue crack growth behaviour from
the basic low cycle fatigue properties.
Commonly accepted, that the fatigue crack growth is influenced by the stress intensity factor range
and the maximum stress intensity factor. These parameters are usually combined into one formula
which is known as the driving force. Alternative expressions for driving forces were suggested in (No-
roozi, Glinka, Lambert, 2008). The driving force concepts could be successfully introduced in the fol-
lowing case. The stress intensity factors are properly linked to the actual elastic-plastic crack tip stress–
strain field. However, the correlation between the stress intensity factors and the crack tip stress–
strain fields are prejudiced by internal or residual stresses.
In the paper (Correia, 2016b) the extended procedure for fatigue life prediction of structural details
was presented. The procedure is based on fracture mechanics approach. The model generalizes the
normalized fatigue crack growth CSS model. Instead of the stress intensity factor range as reference
parameter, the generalization of the CCS model implements the cyclic J-integral range. This allows the
comprehensive elastic-plastic conditions. The proposed approach was applied to a notched plate, us-
ing the equivalent initial flaw size concept.

5
 Unification proposals for fatigue crack propagation laws

2 The cyclic crack propagation laws

The crack propagation is quantified as a function of the range of the stress intensity factor:

K  K max  K min .

The factors in this equation are

K max the maximum stress intensity factor and

K min the minimum stress intensity factor per cycle.

The common form of Paris' law quantifies the fatigue life of a specimen for a given particular crack size
a . The range of stress intensity factor reads:
K  Y a .
The range of stress intensity factor depends on:

   max   min the stress range,

 max the maximum stress,

 min the minimum stress per cycle

Y a dimensionless parameter that reflects the geometry.

The parameter Y possesses the value 1 for a center crack in an infinite sheet. Hereafter, this value is
assumed for briefness: Y  1.
In the present paper the common form for propagation law is implemented:

(1) c
dn
da
 U K  or c
dn
da

 U  a . 
The coefficients in Eq. (1) are:

c  cR  the material constant for a given stress ratio R ,

K min
R 1 the stress ratio of cyclic load,
K max
K max  K min 1  R K
Km   the mean value of stress intensity factor.
2 1 R 2

The unified propagation function U K  describes the infinitesimal crack length growths per increas-
ing number of load cycles n , supposing that the load ratio R remains constant over the load history.
Thus, the range K and the mean value K m of stress intensity factor remain constant. The variations
of range K and the mean value K m of stress intensity factor over the load history could be accounted
by applications of damage accumulation theories (Sorensen, 1969) and (Wheeler, 1972).

6
 Unification proposals for fatigue crack propagation laws

Furthermore, it is assumed that the cracks with the initial length  exist in the material. The cycle
count starts with the beginning of crack propagation from its initial length:
(2) n a   0.

As experimentally recognized for some alloys there is a theoretical value for stress amplitude below
which the material will not fail for any number of cycles. Namely, there exists a threshold value of
amplitudes of stress intensity factor below which fatigue cracks will not propagate (Ritchie et al, 1999).
At the other extreme, approaching the fracture toughness, the material fails almost immediately. The
unified functions that reveal the mentioned behaviour are known from the literature. A unified relation
that covers all three regions was proposed by (Freudental, 1973):

  1/ p
 
ln K 2 / K  K1  K 2
 exp   
da / dn , K0 
ln K 2 / K1 
,
  da / dn K  K   2
  0  
p 1 is the fatigue exponent,

K2 is the short-term threshold limit,

K1 is the endurance threshold limit ( 0  K1  K 2 ).

Alternative form of unified relation was proposed by (Schwalbe, 1980):

da  K K 
 c arctanh  arctanh 1  .
dn  K2 K2 

Unfortunately, both mentioned unified relations do not lead to closed forms of solutions for spring
propagation. The use of the above mentioned laws requires the numerical solutions of nonlinear dif-
ferential equations.
The paper Miller and Gallagher (1981) presents eight different methods to prediction the fatigue life.
Each presented method is applicable to describe the three regions of the crack growth rate curve. The
second fatigue crack growth rate (FCGR) description reads:
m1
K 
  1
 c  1  m2 ,
da K
dn  K 
1  
 K2 

m2  1 is the exponent at the short-term limit,

m1  1 is the endurance limit exponent.

7
 Unification proposals for fatigue crack propagation laws

3 The unification of Paris law type I

In this paper the method of representation of crack propagation functions through appropriate ele-
mentary functions is employed. The choice of the elementary functions is motivated by the phenom-
enological data and covers broad region possible parameters. With the introduced crack propagation
functions differential equations describing the crack propagation are solved rigorously. The resulting
closed form solutions allow the evaluation of crack propagation histories on one side, and the effects
of stress ratio on crack propagation, on the other side.
Two new unified functions (type I and type II) are proposed in this paper. The functions are suggested
in the form that incorporates the three commonly accepted stages of fatigue. The advantage of the
newly proposed functions is the closed form solution of crack propagation. The solution immediately
delivers dependency of crack length over load history and to the number of stress cycles that a speci-
men sustains before failure occurs. The proposed functions allow fitting of acquired experimental data.
Most known generalized laws follow from the newly introduced functions as the special cases.

The new type I unified propagation function U I for the damage growths per cycle accounts the Paris
propagation law together with transition regions at high and low amplitudes of stress intensity factors:
m2
 K 
1   
 K 2 
(3) U I K   K p
m1
.
 K1 
1  
K 
The main advantage of the proposed function is the possibility of closed form analytical solution for
the unified propagation law. On one side, the closed form analytical solution facilitates the universal
fitting of the constants of the fatigue law over all stages of fatigue. On the other side, the closed-form
solution eases the application of the fatigue law, because the solution of nonlinear differential equa-
tion turns to be dispensable. Moreover, the mean stress dependence is the immediate consequence
from the law (3).
The proposed function delivers the closed form solution of the ordinary differential equation (1) with
the initial condition (2):

n  na,  ,   .
Precisely, the function na,  ,  delivers the number of cycles for growth of the crack length from
the initial value  to the given value a , assuming that the stress range  and mean stress

 max   min
m 
2
remain constant.
From the equation (1) two ultimate crack sizes could be determined immediately. Firstly, the value:

a2  K 22  2
is the critical crack length at which instantaneous fracture will occur. Secondly, the value:

8
 Unification proposals for fatigue crack propagation laws

a1  K12  2
is the initial crack length at which fatigue crack growth starts for the given stress range. The number
of cycles to failure at a1   is infinite. For a finite number of cycles before fracture the initial crack
length  must satisfy the condition:

a1    a2 .
The relations between the crack growth rate

dn U I K 
 , with K  a 
da c
and the range of stress intensity factor K for simulated materials are shown on the Fig.1. The endur-
ance and short time threshold exponents were experimentally acquired (Schwalbe, 1980), alloys AlZ-
nMgCu0,5, Ti6Al4V, X35NiCrMo31) and (Albrecht, 1976), aluminum alloys X-7075 (180°), X-7075
(100°), 7075 (100°). For these materials the adopted parameters of fatigue law are summarized in the
Table 1.
The solution of the ordinary differential equation (1) with the initial condition (2) delivers closed form
analytical expression for the remaining number of cycles to fracture:

(4) na, ,   nI a,   nI  ,  ,

where the auxiliary functions are:

2aK12 p  K m1 
(5) n I a,   N I  m ,
c m1 K 2  K1
1

m2
K 
(6) N I (q)   1   2 q   1 q  ,
 K2 
 2  m1  p   2  p  m1  m2 
(7) 1 q   Β q  ,0 ,  2 q   Β q  ,0 .
 m1   m1 
In the expressions (10) appear the incomplete beta-function Β q x, y  (Pearson, 1968), which relates
to the hypergeometric function 2 F1 (Abramowitz, Stegun, 1972) as:
q
qx
Β q x, y    t x 1 1  t  Β q x, y  2 F1 x,1  y , x  1, q 
y 1
(8) dt , .
0
x

The functions 1 and  2 are related to each other as:

(9) 1 a,   lim  2 a,  .
m2 0

9
 Unification proposals for fatigue crack propagation laws

The following example illustrates the behavior of the metallic material with the unified propagation
function (3). The diagrams that express the dependence of cycles to failure, S-N curves (4) are pre-
sented on the pictures on left side of Fig.2 and Fig.3. The same initial length of the crack was assumed
for definiteness in all cases   10 4 m . The stress intervals for calculation depend on threshold stress
intensity factors:

1     2 ,  1  K1  ,  2  K 2  .
The dependencies of crack lengths on cycles, a-N curves are drawn on the right pictures. The number
of cycles was calculated as the function of length of the crack in the interval

107 m  a  102 m .
The S-N curves and a-N curves are shown on the Fig.2 for aluminum alloys X-7075 (180°), X-7075 (100°),
7075 (100°) from (Albrecht, 1976).
The S-N curves and a-N curves are displayed on the Fig.3 for aluminum alloy AlZnMgCu0,5, titanium
alloy Ti6Al4V, and steel X35NiCrMo31 (Schwalbe, 1980).

4 Limit cases for cycles to fracture for the type I

There are three corresponding limit cases for the function (3):

  K  m2 
(10) U I ,1 K   lim U I  K p
1     ,
K1 0
  K 2  
1
  K1  m1 
(11) U I , 2 K   lim U I  K p
1     ,
  K  
K2 

(12) U I ,3 K   lim lim U I  K  p .

K1 0 K2 

The crack growth rate U I ,1 (Eq.10) express the fatigue with zero the short-term threshold limit:

K1  0 .
The crack growth rate U I , 2 (Eq.11) describes the materials with the infinite endurance threshold:

K2   .
The crack growth rate U I ,3 (Eq.12) corresponds to Paris-Erdogan fatigue law with both K1  0 and
K2   .

The Eq. (4) could be simplified for the limit cases of vanishing short-term threshold K1  0 and of
infinite endurance threshold K 2   .

The corresponding expressions for the remaining number of cycles to fracture for three limit cases are:

10
 Unification proposals for fatigue crack propagation laws

2aK  p    K  m2 
n I ,1 a,    lim n I a,    m   p  2    1,
 p  2 p  2  m2 c  2
(13)
K1 0   K2  
 

for the materials with the short-term threshold limit K1  0 (Eq. 10);

2aK12 p  K m1 
(14) n I , 2 a,    lim n I a    1  ,
K 2  c m1 K 2  K1m1 

for the materials with the endurance threshold limit K 2   (Eq. 11);

nI ,3 a,    lim lim nI a,   

2a
2  p cK p
(15)
K10 K2 

for regime with both K1  0 and K 2   (Paris-Erdogan law, Eq. 12) .

The function n I ,1 represents the fatigue curve in the region of Low Cycle Fatigue (LCF), but possesses
no endurance limit.

The function n I , 2 possesses the endurance limit, but does not represent the fatigue curve in the region
of Low Cycle Fatigue (LCF).

The function n I ,3 possesses neither endurance limit nor the short-term threshold and represents the
Paris law. In the logarithmical coordinates the function

2  1 2p 1 
p
n I ,3 a,  ,   
2a 2
  p/2 a   
2  p c a   2  p c   
2
p p
 2  p c p  
 
is straight line:

ln nI ,3 a,  ,    p ln   ln nI ,3 a,  ,1,

with

 1 2p 1 
p
n I ,3 a,  ,1 
2
 a   2 
 p / 2 2  p c  .

The outline of the expressions for the unified propagation law I displays the Table 2.

5 The unification of the fatigue law type II

An alternative generalization of the fatigue law is achieved by the introduction of the unified propaga-
tion function type II:

11
 Unification proposals for fatigue crack propagation laws

m2

  K k 
1    
  K 2  
(16) U II K   K  p m1
,
  K1  k 
1    
  K  

k is the dimensionless material constant.

The additional the dimensionless material constant permits more precise fitting of the experimental
data.
The solution of the ordinary differential equation (1) with the initial condition (2) and function (16)
delivers closed form analytical expression for the remaining number of cycles to fracture:

(17) na, ,   nII a,   nII  ,  .

The auxiliary function nII a,   in the equation (17) is:

km1
K
n II a,   
2a
  N II ( K ), with K   a ,
cK p  K1 
(18)
e i  m1   2  km  p 2  k  km1  p   K   K  
k k

N II ( K )  ,m2 , m1 , ;  K  ,  K  .
1
F1
2  km1  p   k k  2  1 


The transcendental function in the Equation (18) is known as the Appell hypergeometric function of
two variables (Erdélyi, 1950) and (Kampe de Feriet, 1957):

C 
1
F1 A, B1 , B2 , C ; x, y   t A1 1  t  1  tx  B1 1  ty  Bs dt .
 AC  A 0
C  A 1

The second fatigue crack growth rate (FCGR) description due to Miller and Gallagher (1981) is the spe-
cial case of the equation (16) for the following values of parameters k  1, p  m1 :
m2

 K
1  K 
U II* K   m1  2
1
(19) m1
.
K1  K 
 K  1
 1 
The solution of the differential equation (1) provides in this special case the expression for the remain-
ing number of cycles to fracture:

(20) na, ,   n *II a,   n *II  ,  .

The auxiliary function n *II a,   reads in this case as:

12
 Unification proposals for fatigue crack propagation laws

 K K  i  m1
n *ÎI a,    F1  2,m2 , m1 ,3;
a
(21) , e .
c K1m1
 K 2 K1 

6 Limit cases for cycles to fracture for the type II

The expressions could be simplified for the limit cases of vanishing short-term threshold K1  0 and
an infinite endurance threshold K 2   as well. There are three corresponding limit cases for the
function (18):
m2
  K 
k

(22) U II ,1 K   lim U II  K p
1     ,
K1 0
  K 2  

 m1
  K1  k 
(23) U II , 2 K   lim U II  K p
1     ,
  K  
K 2 

(24) U II , 3 K   lim lim U II  K  p .

K1 0 K2 

The matching expressions for the remaining number of cycles to fracture for three limit cases of Eq.
(20) respectively are:
2 p k
2 p  K 
n II ,1 a,    lim n II a,   
2aK 2
(25) Β q2  ,1  m2 , q2    ,
 k 
K1 0 2
ck K  K2 
for the materials with the short-term threshold limit K1  0 ;

2 p k
2 p  K
n II ,2 a,    lim n II a,   
2aK1
(26) Β q1   m1 ,1  m1 e i  m1 , q1    ,
 k 
K 2  2
ck K  K1 
for the materials with the endurance threshold limit K 2   ;

nII ,3 a,    lim lim nII a,    nI ,3 a,   

2a
,
2  p cK p
(27)
K10 K2 

for regime with K1  0 and K 2   . The later expression turns again into the Paris-Erdogan law,
which is given by Eq. (15).
The three limit terms for the function (19) are:
m2

 K 
(28) U *
II ,1 K   Klim
*
II
1
U  m1 1   ,
1 0 K1  K2 

13
 Unification proposals for fatigue crack propagation laws

 m1
1 K 
(29) U *
II , 2 K   Klim U  m1   1
*
II ,
2  K1  K 1 

(30) U * II , 3 K   lim lim U II*  K  m1 .

K1 0 K2 

The matching expressions for the second fatigue crack growth rate (FCGR) description due to Miller
and Gallagher (1981) respectively are:
2m1
(31) n *
II ,1 a,    Klim n a,   
*
II
2aK 2
Β q2 2  m1 ,1  m2 , q2 
K
,
1 0 cK 2 K2

for the materials with the short-term threshold limit K1  0 ;

2  m1
n *II , 2 a,    lim n *II a,    Β q1 2,1  m1  exp i  m1 ,
2aK1 K
q1  ,
K 2  cK 2 K1
(32)
Β q1 2,1  m1  
m1  q1  m1q1 q1  1 ,
1  q1 m 2  m1 1  m1 
1

for the materials with the endurance threshold limit K 2   .

The second fatigue crack growth rate (FCGR) description can also represent the regime with K1  0
and K 2   :

n*II ,3 a,    lim lim n*II a,   

2a
.
2  m1 cK m1
(33)
K10 K2 

The summary of the solutions for the unified propagation law II and its limit cases represents the Table
3.

7 Influence of stress ratio

The load ratio is defined as the ratio of the algebraically minimum over the maximum load, Eq. (1). The
experiments demonstrate that the load ratio affects the fatigue crack growth and threshold behavior.
Namely, the fatigue crack propagation rate and threshold value vary with the applied load ratio
(Walker, 1970). If the load ratio is positive, the experiments reveal that the necessary stress intensity
factor range for growth decreases with increasing positive values. In the region of the negative load
ratio the required stress intensity factor range for growth (threshold stress intensity range) decreases
as load ratio decreases. Reaching a definite negative value, known as saturation point, the required
stress intensity factor range for fatigue crack growth stabilizes.
Consider now the variation of load ratio for the proposed propagation laws. The closed form solution
for the cycles to failure for a given initial crack length upon the stress amplitude could be found for
given load ratio for both type I and type II of unified laws.
Firstly consider the phenomenological approach to this problem. In accordance with the phenomeno-
logical approach (Klesnil, Lukas, 1972), the damage growths per cycle depends upon the stress ratio:

14
 Unification proposals for fatigue crack propagation laws

   
def
c0  cR  0.
1
(34) c R*  1  R* c0 ,

The exponent  accounts the influence of the stress ratio on the crack growth rate, 0  R*  1 is an
arbitrary stress ratio.
According to (Forman, Kearney, Engle, 1967), the crack closure decrease the fatigue crack growth rate
by reducing the short-term threshold:

(35)   
K 2 R*  1  R*  2
K 20

is the function of the load ratio R * , where

def
K20  K2 R  0

is the threshold value at R  0 and  2 is a material dependent constant. Alternative relationship

K 2 R*   1  R* K 20
2
(36)

has been proposed by (Branco, Radon, Culver, 1976).

The endurance threshold limit might be also considered as a function of stress ratio:

K1 R*   1  R*  K10 .
1
(37)

Here
def
K10  K1 R  0

is the endurance threshold value at R  0 and  1 is a material dependent constant.

The mentioned above equations (34)-(37) are the empirically determined phenomenological depend-
ences.
Secondly, instead of an empirical approach, in this paper we use the equations (1) to derive the influ-
ence of the stress ratio on the crack growth rate explicitly. For this purpose consider two harmonically
varying loads with equal amplitude, but with the different mean values. The mean values and ampli-
tudes are assumed to be positive, such that:

K m  0, K m*  0, K 0.

Both loads lead to the harmonically varying stress intensity factors. The harmonically varying stress
intensity factors correspondingly are:

(38) K t   K m  12 K sin t , K * t   K m*  12 K sin t .

For both considered cyclic loads (38) the ratios are equal correspondingly to:

15
 Unification proposals for fatigue crack propagation laws

K m  12 K K m*  12 K
(39) R , R*  .
K m  12 K K m*  12 K
Consider at first the positive load cycles

K t   0, K * t   0 .
For the positive load cycles the stress ratios are in the ranges:

(40) 1  R  1,  1  R*  1 .
The damage caused by the varying load depends linearly upon the stress intensity factor in the power
p along the cycle. Consequently, the relation between fatigue coefficients, which correspond to the
different stress ratios cR  and c R* , reduces to:  
c R 

 R, R *   c R * 
,
2
c R    K  12 K sin t  dt ,
c0 p
(41)
2
m
0
2
c R   0  K  12 K sin t  dt.
*c * p

2
m
0

Notable, that the integrals in (41) could be expressed in term of the hypergeometric function:

2  1  p p   K  2 
0 K m  K sin t  dt  2 F1   2 , 2 1,  2K m   K m .
1 1 p p

2
2
 
From (36) follow the following formulas for the parameters of Eq. (31):

1 R K 1  R* K
(42)  ,  .
1  R 2K m 1  R 2 K m*
*

The use of the expressions (42) delivers the relation between fatigue coefficients of the Eq. (1) for two different
load ratios is:

(43) 
 R, R *   RR. *

The function in equation (43) accounts stress ratio for  1  R  1 :

 1  p p  1  R 2  1  R p
(44)  R2 F1   , 1, 
2  .
 2 2  1  R   1  R 
p

Consequently the dependence of the fatigue coefficient upon the load ratio the for positive load cycles
reads:

16
 Unification proposals for fatigue crack propagation laws

(45)   
c R *  cR /  R, R * 
The plot of the function ln(R) for different values of p is shown on the Fig.4. With this function
the mean value sensitivity expresses directly from the first principles.
Thirdly extend the validity region for the fatigue coefficient upon the load ratio from positive to the
negative values. For this purpose the effective stress intensity concept is implemented, following the
results (Elber 1970, 1971). In these papers the fatigue crack in a center-cracked-tension panel sub-
jected to zero-to-tension loading was studied. It was demonstrated that a fatigue crack is fully open
for only a part of the loading cycle. This effect was attributed to contact of residual plastic deformation
around the crack tip. The cause for plasticity induced crack closure is the plastic wake which is devel-
oped during crack propagation. The faces of crack could be subjected to considerable compressive
stresses. In the article (Boyce, Ritchie, 2001) was proposed that the crack tip stress intensity factor
range, have to be modified by an experimental factor. The experimental factor accounts the exposed
crack closure effect.
The ranges (40) could be extended using the introduction of effective stress intensity for the proposed
unified propagation functions. The simplest formulation for the experimental factor leads to the effec-
tive stress intensity factor range K eff , effective mean value of stress intensity factor K m.eff and effec-
tive stress ratio Reff correspondingly:

K eff  min K max , K max  K min ,

K K  K min 
(46) K m.eff  max  max , max ,
 2 2 
K  K
1
Reff  m.eff 12 eff .
K eff m  2 K eff

The equations (43) to (45) could be used also in the case K min  0 with the introduced effective values
according to Eq. (46). Namely, the stress intensity factor range K , mean value of stress intensity factor
K mf and stress ratio R should be replaced respectively by effective stress intensity factor range
K eff , effective mean value of stress intensity factor K m.eff and effective stress ratio Reff .

8 Conclusion
The paper introduces the closed form analytical expression for crack length over number of cycles.
Two new functions that express the damage growth per cycle are introduced. These functions allow
the unification of different fatigue laws in a single expression. The unified fatigue law provides closed
form analytical solutions for crack length upon the mean value and range of cyclic variation of stress
intensity factor. The solution expresses the number of cycles to failure as the function of the initial size
of the crack and eliminates the solution of nonlinear ordinary differential equation of the first order.
The explicit formulas for stress against the number of cycles to failure are delivered for both proposed
unified fatigue laws.
The different common expressions, which account the influence of the stress ratio, are immediately
applicable. For the proposed unified propagation functions, the ranges for the stress load factor are

17
 Unification proposals for fatigue crack propagation laws

extended using the introduction of effective stress intensity. The solution leads to the effective stress
intensity factor range, effective mean value of stress intensity factor and effective stress ratio.

18
 Unification proposals for fatigue crack propagation laws

References
A. Carpinteri, Ed. (1994) Handbook of Fatigue Crack Propagation in Metallic Structures, Elsevier Science
B.V. Philadelphia, PA
Suresh S. (1998) Fatigue of materials, Cambridge University Press, Cambridge
Fatemi A., Yang L. (1998) Cumulative fatigue damage and life prediction theories: A survey of the stat
of the art for homogeneous materials. International Journal of Fatigue, 20(1):9–34
Richard H.A., Sander M. (2012) Ermüdungsrisse, ISBN 9 78-3-8348-1594-1, DOI 10.1007/978-3-8348-
8663-7, Springer-Vieweg, Berlin
Krupp U. (2007) Fatigue Crack Propagation in Metals and Alloys: Microstructural Aspects and Model-
ling Concepts, ISBN: 978-3-527-31537-6, Wiley-VCH, Berlin
Totten G. (2008) Fatigue crack propagation, Advanced Materials & Processes, May 2008, p. 39-41
Leese G.E., Socie D., Eds. (1989) Multiaxial Fatigue, Analysis & Experiments, SAE Pub. AE-14, Warren-
dale, PA
Paris P., Erdogan F. (1963) A critical analysis of crack propagation laws. J. Basic Eng. Trans. ASME, 528–
534
Beden S. M., Abdullah S., Ariffin A. K. (2009) Review of Fatigue Crack Propagation Models for Metallic
Components, European Journal of Scientific Research, ISSN 1450-216X Vol.28 No.3, pp.364-397
Donahue, R. J. Clark, H. M., Atanmo, P., Kumble, R., McEvily, A. J. (1972) Crack opening displacement
and the rate of fatigue crack growth. Int. J. Fract. Mech. 8, 209–219, DOI 10.1007/BF0070388
Kanninen M.F., Popelar C.H. (1985) Advanced Fracture Mechanics. Oxford University Press, New York
McEvily A.J., Groeger J. (1977) On the threshold for fatigue-crack growth. Fourth International Confer-
ence on Fracture, vol. 2. University of Waterloo Press, Waterloo, Canada, pp. 1293–1298
Xue L. (2008) A unified expression for low cycle fatigue and extremely low cycle fatigue and its impli-
cation for monotonic loading, International Journal of Fatigue, 30, 1691–1698
Pugno N., Ciavarella M., Cornetti P., Carpinteri A. (2006) A generalized Paris’ law for fatigue crack
growth, Journal of the Mechanics and Physics of Solids. 54, 1333–1349
Mishnaevsky L. Jr., Brøndsted P. (2007) Modeling of fatigue damage evolution on the basis of the ki-
netic concept of strength, Int J Fract, 144:149–158, DOI 10.1007/s10704-007-9086-1
Pugno N., Cornetti P., Carpinteri A. (2007) New unified laws in fatigue: From the Wöhler’s to the Paris’
regime, Engineering Fracture Mechanics, 74, 595–601
Castillo E., Fernández-Canteli A., Siegele D. (2014) Obtaining S–N curves from crack growth curves: an
alternative to self-similarity, Int. J. Fract., 187:159–172 DOI 10.1007/s10704-014-9928-6
Gumbel, E.J. (1954) Statistical theory of extreme values and some practical applications. Applied Math-
ematics Series. 33 (1st ed.), U.S. Department of Commerce, National Bureau of Standards
Buckingham, E. (1915) Model experiments and the forms of empirical equations. Transactions of the
American Society of Mechanical Engineers, 37: 263–296.
19
 Unification proposals for fatigue crack propagation laws

Blasón S., Correia J.A.F.O., Apetre N., Arcari A., De Jesus A.M.P., Moreira P., Fernández-Canteli A. (2016)
Proposal of a fatigue crack propagation model taking into account crack closure effects using a modi-
fied CCS crack growth model, Procedia Structural Integrity 1, 110–117, XV Portuguese Conference on
Fracture, 10-12 February 2016, Paço de Arcos, Portugal
Correia J.A.F.O., Blasón S., Arcari A., Calvente M., Apetre N., Moreira P.M.G.P., De Jesus A.M.P., Canteli
A.F. (2016) Modified CCS fatigue crack growth model for the AA2019-T851 based on plasticity-induced
crack-closure, in: XV Portuguese Conference on Fracture and Fatigue, Theoretical and Applied Fracture
Mechanics, Volume 85, Part A, October 2016, Pages 26–36
de Castro J.T.P., Meggiolaro M.A., Miranda A. C. (2009) On the estimation of fatigue crack propagation
lives under variable amplitude loads using strain-life data, in: Proceedings of COBEM 2009 20th Inter-
national Congress of Mechanical Engineering, ABCM, November 15-20, 2009, Gramado, RS, Brazil
Chen F., Wang F., Cui W. (2011) Fatigue life prediction of engineering structures subjected to variable
amplitude loading using the improved crack growth rate model, Fatigue & Fracture of Engineering Ma-
terials & Structures, 35, 278–290, doi: 10.1111/j.1460-2695.2011.01618.x
Wheeler, O. E. (1972) Spectrum loading and crack growth, J. Basic Engng. 94, 181–186
Noroozi A.H., Glinka G., Lambert S. (2005) A two parameter driving force for fatigue crack growth anal-
ysis, International Journal of Fatigue, Volume 27, Issues 10–12, October–December 2005, Pages 1277–
1296
Smith K.N., Watson P, Topper T.H. (1970) A stress–strain function for the fatigue of metals. J. Mater.
5(4): 767–78
Hutchinson J. (1968) Singular behaviour at the end of a tensile crack in a hardening material, Journal
of the Mechanics and Physics of Solids, 16, p. 13-31.
Rice J., Rosengren G. (1968) Plane strain deformation near a crack tip in a power-law hardening mate-
rial, Journal of the Mechanics and Physics of Solids, 16, p.1-12
Kujawski D., Ellyin F. (1984) A fatigue crack propagation model, Engineering Fracture Mechanics, 20,
695-704
Ellyin F. (1986) Stochastic modelling of crack growth based on damage accumulation, Theoretical and
Applied Fracture Mechanics, 6, 95-101
Shi K., Cai L., Bao C. (2014) Crack growth rate model under constant cyclic loading and effect of differ-
ent singularity fields, Procedia Materials Science 3, 1566 – 1572
Noroozi A.H., , Glinka G., Lambert S. (2008) Prediction of fatigue crack growth under constant ampli-
tude loading and a single overload based on elasto-plastic crack tip stresses and strains. Engineering
Fracture Mechanics, Volume 75, Issue 2, January 2008, Pages 188–206
Correia J.A.F.O., Blasón S., De Jesus A.M.P., Canteli A.F., Moreira P.M.G.P., Tavares P. J. (2016) Fatigue
life prediction based on an equivalent initial flaw size approach and a new normalized fatigue crack
growth model, Engineering Failure Analysis, Volume 69, November, Pages 15–28
Sorensen A., A General Theory of Fatigue Damage Accumulation, J. Basic Eng 91(1), 1-14 (Mar 01, 1969)
(14 pages) doi:10.1115/1.3571021

20
 Unification proposals for fatigue crack propagation laws

Ritchie R.O., Boyce B.L., Campbell J.P., Roder O., Thompson A.W., W.W. Milligan(1999) Thresholds for
high-cycle fatigue in a turbine engine Ti–6Al–4V Alloy, International Journal of Fatigue, 21, 653–662,
DOI: 10.1016/S0142-1123(99)00024-9
Freudenthal A.M. (1973) Fatigue and Fracture Mechanics, Engineering Fracture Mechanics 5(2):403–
414 · June, DOI: 10.1016/0013-7944(73)90030-1
Schwalbe K.-H. (1980) Bruchmechanik metallischer Werkstoffe, Carl Hanser Verlag, München, Wien
Miller M.S., Gallagher J.P. (1981).An Analysis of Several Fatigue Crack Growth Rate (FCGR) Descriptions.
Fatigue Crack Growth Measurement and Data Analysis. S.J. Hudak, Jr. and R.J. Bucci (Eds.).ASTM STP
738, American Society for Testing and Materials, pp. 205-251.
Albrecht J., Martin J.W.R., Lüthering G., Martin J.W. (1976) Influence of micromechanisms on fatigue
crack propagation rate of Al-alloys, 4th Int. Conf. on the Strength of Metals and Alloys, Conference
Proceedings Nancy, 30 August-3 September 1976, E.N.S.M.I.M, Nancy, 1408 P.
Pearson K. (1968) Tables of Incomplete Beta Functions, 2nd ed. Cambridge, England: Cambridge Uni-
versi-ty Press
Abramowitz M., Stegun I.A. (1972) Handbook of Mathematical Functions with Formulas, Graphs, and
Mathematical Tables. Chapter 6, New York: Dover Publications
Erdélyi A. (1950) Hypergeometric functions of two variables, Acta Mathematica, January 1950, Volume
83, Issue 1, pp 131–164
Kampe de Feriet J., Fonctions de la Physique Mathématique, Formulaire de Mathématiques à l'usage
des Physiciens et des Ingénieurs (1957) Centre National de la Recherche Scientifique, Paris
Walker K. (1970) The effect of stress ratio during crack propagation and fatigue for 2924-T3 and 7075-
T6 Aluminum, in: Effects of environment and complex load history on fatigue life, ASTM STP 462
Forman R.G., Kearney V.E., Engle R.M. (1967) Numerical analysis of crack propagation in cyclic loaded
structures, J. of Basic Engineering, Trans. ASME, D89, 459-464
Klesnil M., Lukas P. (1972) Effect of stress cycle asymmetry on fatigue crack growth, Mat. Sci. and
Engng. 9, 231-240
Branco C. M., Radon J. C., Culver L. E. (1976) Growth of fatigue cracks in steels. Metal sci., 10, 149-155
Elber W. (1971) The Significance of Fatigue Crack Closure, STP486, DOI: 10.1520/STP26680S, Annual
Meeting ASTM, Toronto, ASTM International
Elber W. (1970) Fatigue Crack Closure under Cyclic Tension, Engineering Fracture Mechanics 2(1):37-
45 · July 1970, DOI: 10.1016/0013-7944(70)90028-7
Boyce B.L., Ritchie R.O. (2001) Effect of load ratio and maximum stress intensity on the fatigue thresh-
old in Ti6Al4V, Engineering Fracture Mechanics, 68, 129±147, DOI: 10.1016/S0013-7944(00)00099-0

21
List of symbols

Kmax maximum stress intensity factor per cycle

K min minimum stress intensity factor per cycle
K range of stress intensity factor
   max   min stress range
 max maximum stress per cycle
 min minimum stress per cycle
Y dimensionless geometry parameter
c  c R  material constant for a given stress ratio R
R  K min K max  1 stress ratio of cyclic load
K m  K max  K min  / 2 mean value of stress intensity factor
U I K  unified propagation function of type I
U II K  unified propagation function of type II
p 1 fatigue exponent
m2  1 exponent at short-term limit
m1  1 endurance limit exponent
K2 short-term threshold limit ( K2 K1)
K1 endurance threshold limit( K1 0)
n  na,  ,  number of cycles for growth of crack length
U I ,i K , i  1,2,3 limit cases for propagation function of type I
U II ,i K , i  1,2,3 limit cases for propagation function of type II
n I ,i a,  , i  1,2,3 limit cases for number of cycles to fault (type I)
nII ,i a,  , i  1,2,3 limit cases for number of cycles to fault (type II)
influence parameters for the stress ratio
 , 1 ,  2
 
 R, R *   R  /  R *  
relation between fatigue coefficients
Implemented special functions
incomplete beta function
q
qx
Β q  x , y    t x 1 1  t  dt  2 F1 x ,1  y , x  1, q 
y 1

0
x
hypergeometric function
2 F1 x, y z , q 
Appell hypergeometric function of two variables
 C 
1
F1  A, B1 , B2 , C ; x , y    t A 1 1  t  1  tx  B1 1  ty  B s dt
C  A 1

  A  C  A  0
 Unification proposals for fatigue crack propagation laws

Table 1. Mechanical properties of simulated materials (From Schwalbe, 1980)

Alloys Aluminum Alloys Ti-Alloy Steel Alloy
Al Al Al Al Units
Alloys type Ti6Al4V X35NiCrMo31
7075 X-7075 X-7075 (180) AlZnMgCu0,5
s
460 430 390 397 912 1348 MPa
Yield strength
B
590 730 560 481 1209 1588 MPa
Ultimate strength
E-Modulus 68700 113800 202086 MPa
 , Poisson ratio 0.33 0.34 0.28
P
3.6 3.6 3.6 3.6 3.6 3.6
fatigue exponent
cR K 2p
Material constant
0.512  106 0.104  105 0.402  106 0.1976  106 0.219  105 0.1488  104 MPa  m  mp

R ,Stress ratio 0.2 0.2 0.2 0.2 0.1 0.1

K1
10.0 12.0 15.0 11.0 12.6 10.6 MPa  m
endurance threshold limit
K2
40.0 50.0 60.0 26.0 91.4 96 MPa  m
short-term threshold limit
 10 µm
1 56 67 84 62 71 59 MPa
2 225 282 338 146 515 541 MPa
Table 2. Propagation function U I K  and the corresponding number of cycles to failure as function of stress intensity range

Propagation function Eq. Number of cycles to failure as function of stress intensity range K   a Eq.

m2
2 aK 12  p  K m1  K 
1 n I a ,    N I
 K m1 , N I (q)   1   2 q   1 q  ,
  K 
m2
   K  m1  c m1 K 2  1   K2 
U I K   K p
1     1   1   3 5
  K 2     K    2  m1  p   2  p  m1  m 2 
 1 q   Β q  ,0 ,  2 q   Β q  ,0 .
 m1   m1 

  K  m2  2aK  p    K  m2 
K1  0 U I ,1 K   K p
1     10 nI ,1 a,    m   p  2    1 13
  K 2    p  2 p  2  m2 c  2   K2 



1
  K 1  m1  2aK12 p  K m1 
K2   U I , 2 K   K p
1     11 nI , 2 a,     1   14
  K   c m1 K 2  K1m1 

K1  0 2a
U I , 3 K   K  p 12 n I , 3 a ,    15
K2   2  p cK p
Table 3. Propagation function U II K  and the corresponding number of cycles to failure as function of stress intensity range

Propagation function Eq. Number of cycles to failure as function of stress intensity range K   a Eq.

km1
2a K 
n II a,      N II ( K ),
cK p
m2

  K  k
  K   k  m1
 K1 
U II K   K  p 1     1   1   16 18
  K 2     K   e i  m1   2  km  p 2  k  km1  p  K 
k
K 
k

N II ( K )  F1   1
, m2 , m1 , ;  K  ,   
2  km1  p   k k  2  K1  
 
m2 k
  K  
k
2aK 2
2 p
2 p   K 
K1  0 U II ,1 K   K p
1     22 n II ,1 a,    2
Β q2  ,1  m2 , q2    25
  K 2   ck K  k   K2 
 m1 k
  K1  k  2aK1
2 p
2 p  K
K2   U II , 2 K   K p
1     23 n II ,2 a ,    2
Β q1   m1 ,1  m1 e i  m1 , q1    26
  K   ck K  k   K1 

K1  0 2a
U II , 3 K   K  p 24 n II , 3 a ,    27
K2   2  p cK p

m2
 m1
1  K  K  a  K K  i  m1
U K   m1
*
1  K   K  1 19 n *ÎI a ,    F1  2,m2 , m1 ,3; , e 21
II
K1  2  1  c K1m1
 K 2 K1 

m2
2  m1
1  K 2aK 2 K
K1  0 U II* ,1 K   m1 1   28 n *II ,1 a ,    Β q2 2  m1 ,1  m2 , q2  31
K1  K 2  cK 2 K2

 m1 2 m1
K  2aK1 K
K2   U *
II , 2 K   1m1   1 29 n *II , 2 a,    Β q1 2,1  m1  expi  m1 , q1  32
K1  K1  cK 2 K1

K1  0 2a
U * II , 3 K   K  m1 30 n *II , 3 a ,    33
K2   2  m1 cK m1

1
 Unification proposals for fatigue crack propagation laws

U I K   1 
U I K   1 
c  m 
c  m 

AlZnMgCu0,5 X35NiCrMo31
7075
(24h, 100°C)

X-7075
(24h, 100°C)

X-7075
Ti6Al4V
(48h, 180°C)


K MNm 3 / 2  
K MNm 3 / 2 
Fig. 1 The plot of the relations between the crack growth rate dn / da  U I K  / c and the range of stress intensity factor K for different materials (Schwalbe, 1980,
Albrecht, 1976)
 Unification proposals for fatigue crack propagation laws

X-7075

 , Pa a, m (24h, 100°C)
X-7075
X-7075 
(24h, 100°C)
(48h, 180°C)
X-7075


/2
X-7075
X-7075

7075 
7075 (24h, 100°C)
(24h, 100°C)


X-7075 7075

(48h, 180°C) X-7075 

 7075



N
N

Fig. 2 The dependencies between the cycles to failure N for the given initial crack lengths upon the stress amplitude  (s-N-curve, links) and cycles to failure N for the given
stresses upon the length of crack (a-N-curve, rechts)
 Unification proposals for fatigue crack propagation laws

X35NiCrMo31

Ti6Al4V

(24h, 100°C)
Ti6Al4V
X35NiCrMo31 


Ti6Al4V



Ti6Al4V
X35NiCrMo31

AlZnMgCu05 
AlZnMgCu0

AlZnMgCu0


AlZnMgCu0
X35NiCrMo31



N
N

Fig. 3 The dependencies between the cycles to failure N for the given initial crack lengths upon the stress amplitude  (s-N-curve, links) and cycles to failure N for the given
stresses upon the length of crack (a-N-curve, rechts)
View publication stats

Unification proposals for fatigue crack propagation laws

ln( R )
ln( R )

Fig. 4 The plot of the function ln(R) for different values of p