-ESTIMATION OF FATIGUE STRESS CONCENTRATION FACTOR IN THE REGION OF ELASTIC-PLASTIC STRAIN

Damir Jelaska, professor Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, R. Bokovia b.b., 21000 Split, Croatia Tel. +385 21 305874; Fax. +385 21 563877 E-mail addres: damir.jelaska @ fesb.hr

1.Introduction In the absence of the experimental results, designers are forced to estimate the value of stress concentration factor in the region of elastic- plastic strain. The most popular method is based on empirical investigations of Hardrath and Ohman [1], concentration factor Neuber [2] and Wetzel [3] which propose the relation between fatigue stress

* k

and the strain concentration factor

be* . Intersection of corresponding curve with cyclic

stress-strain curve determines the magnitude of local stress and strain and consequently the real values of stress and strain concentration factor. It was perceived that values of

* k

obtained after [1] and [2] are greater than real ones, and those

obtained after [3] less then real ones. That is why this method is unreliable. Whereas it is also time- consuming, it became necessary to make effort to obtain simple and correct explicit formula for fast estimation of the stress concentration factor in the region of elastic- plastic strain. 2. Derive of stress concentration factor formula in the zone of elastic- plastic strain On the basis of entire worldwide experimental investigations, including author's, the S-N curves for the unnotched specimens and parts (same as notched specimens) might be desplayed as straight lines (Fig. 1). This is very good approximation for most of steels and might be taken as valid for most of metals. The limit N gr between finite and infinite life is taken to correspond with the limit between elastic and elastic- plastic strain region, and it is also taken to be same for the notched and unnotched specimen. For the materials which do not have the endurance limit, it might be taken N gr = 10 . Same holds good for the steels at elevated temperatures.
7

The limit N q between region of quasistatic fracture and region of fatigue fracture is also taken to be same for both notched and unnotched specimen.

Fig.1- Stress concentration factor computing model In the region of finite fatigue life ( N q N N gr ) , the equations of both S-N curves are:
m m N q Rq = N m a = N gr R1 ' m' m' N q R 'm q = N an = N gr ( R1 / k ) .

(1) (2)

According to definition

* k =
from (3) and (4) follows:

a Ran = , an Ra
1 1 m' m

(3)

= k ( N / N gr )
* k

(4) (5)

or
1 m ' /m * k = k ( R1 / k an )

From (3) and (4), according to Fig.1, it is easy to derive the ratio:

log( N gr / N q )1/ m m' = m log( N gr / N q )1/ m + log( k / q )

(6)

Whereas the ratio N gr / N q for the most materials is close to 10 , the exponent ratio might be approximated as

m' 3 = m 3 + m log( k / q )

(6a)

Stress concentration factor q at quasistatic fracture bondary remains the only unknown in expressions (3) to (6a). On the basis of regression analyses of great number of experimental investigations, f.e. [4,5,6,7], the stress concentration factor at quasistatic fracture limit for steels might be estimated as

q = 1 + (0,00038 R M 0,1)( k 1) .
So determined fatigue stress concentration factor

(7)

* k , involved in cyclic stress-strain curve

1

a n a = a + E k pl

(8)

enables estimation of the strain concentration factor in the region of elastic- plastic strain:
1/ n 1/ n 1 * = k + Ek pl ( k an )

(9)

For nine steels having cyclic strenghtening characteristics k pl and n after [8] , the values of for various magnitudes of hyperbola low

are calculated

an .

It was observed that

* * k

product is not in accordance with Neuber's

* 2 * * 2 * k = t , neither with modificated Neuber's hyperbola low k = k .

For that reason, by help of "Excel" computer program, new relation between stress and strain concentration factor in the region of elastic- plastic strain is obtained for the steels:
2 + 0 ,57 ( * * k = k
an / R1 1/ k

(10)

This equation might be generalized for all materials:

2 + a ( * * k = k
an / R1 1/ k

(10a)

where a is material constant, which could be determined by experiment. From this equation it is clear that product of stress and strain concentration factors in the region of elasticplastic strain increases with stress amplitude

an

and it is equal to

2 k

(after [3]) only at N gr limit. This new

relation represents a hyperbola located between original Neuber's hyperbola and modified one, and it determines stress and strain concentration factor if the cyclic stress- strain curve is known (Fig.2).

Fig.2- Determining of local stress and strain

For the supposed exponential relation between stress and number of cycles of the notched and unnotched specimens in the region of quasistatic fracture, it was easy to derive the expression for the approximate estimation of stress concentration factor:
log 4 N log 4 N q

= s ( q / s )
* q

(11)

For materials that fail in brittle manner,

factor depends on

and notch sensitivity at static fracture

related to the level of the plastic deformation at static fracture, and there is inside limits 1 s q . For fully sensitive materials (such as titanium, beryllium and most of its alloys) is For ductile materials, including mild steels with

s = q = t .
the plastic strain is so high at

M < 700 N / mm2 ,

static fracture (N=1/4) that stress concentration effect is relieved by yielding, i.e., it might be taken that and consequently
log 4 N . * q = q
q

s =1,

log 4 N

(11a)

N q is an exsclusive characteristic of a material and vary from 103 to 105 for the metals, just like
particularly for the steels.

3. Concluding remarks Stress concentration factor calculated in the way described above, is as close to its true value, as suggested computing model is close to the real one. It also depends on basic data reliability, particularly on stress concentration factor boundary values

q,k

and

s .

The special attention has to be paid to

q ,

the

suggested calculation of which is good approximation for only the steels. For other materials the additional investigations have to be done to obtain the applicable formula for determination of

q .

For that reason the

author proposes to apply the known procedure after Fig.2 using obtained formula (10), because this method becomes now much more precise. In such a way the usage of using it.

is unnecessary, and results are not worse then

NOMENCLATURE a- material constant E- modulus of elasticity, MPa k pl - coefficient of cyclic strengthening, MPa m - Whler-curve exponent of the unnotched specimen m'- Whler-curve exponent of the notched specimen N - number of endurable cycles N gr - boundary number of cycles corresponding to endurance limit R1

N q - boundary number of cycles corresponding to quasistatical fracture limit

n - cyclic strengthening exponent

R M - ultimate strength of a material, Mpa Rq - unnotched specimen strength at quasistatic fracture boundary, MPa
' Rq - notched specimen strength at quasistatic fracture boundary, MPa

R1 - endurance limit of material at -1 stress cycle asymmetry number, MPa

t - theoretical stress concentration factor k - fatigue stress concentration factor in the zone of elastic strains * k - fatigue stress concentration factor in the zone of elastic-plastic strains q - stress concentration factor at quasistatic fracture limit
* q - stress concentration factor at quasistatic fracture zone

 s - stress concentration factor at static fracture a - local stress amplitude, MPa

an - nominal stress amplitude , Mpa

REFERENCES [1] Hardrath F.H.,Ohman L.:"A Study of Elastic and Plastic Stress Concentration Factors Due to Notches and Fillets in Flat Plates" , NACA Rept. 1117 , 1951. [2] Neuber H.:"Theory of Stress Concentration for Shear-Strained Prismatical Bodies with Arbitrary Nonlinear Stress-Strain Low" ASME Trans.,N.York , 28(1961) , p.544-550. [3] Wetzel R. M. :"Smooth specimen simulation of fatigue behavior of notches", J. Materials (1968), p.646- 657. [4] "Fatigue Investigations of Higher Strength Structural Steels in Notched and in Welded Conditions", Komission der Europischen Gemeinschaften, Luxemburg, EUR- Bericht Nr. 5337 (1975). [5] Heywood, R. B. "Designing Against Fatigue", Chapman and Hall Ltd. , London 1962. [6] Olivier R., Ritter W.,:"Whlerlinien katalog fr Schweiverbindungen aus Bausthlen", Fraunhofer Institut fr Betriebfestigkeit, Berichr Nr. FB-151 (1979-1985) [7] Osterman H., Grubii, V., Einflus des Werkstoffes auf die ertragbare Schwingbeanspruchung, Chapter in "Verhalten von Stahl bei Schwingender Beanspruchung", Hrsq. Verein D. E., Dsseldorf 1979. [8] Troshchenko V.T. et al. :"Cyclic Strains and Fatigue of Metals" (in Russian) Vol 1., p.77 , Naukova Dumka , Kiev , 1985.