Issue 14

B. L

obato da Silva

et alii, Frattura

ed Integrità Stru

tturale, 14 (201

0) 17-26; DOI:

10.3221/IGF-ESI

S.14.02

Figure 2 : S

pecimen 2 for

sample B.

Specimen / 1 / A 1 / B 2 / B

Sample a

15 15 15 (mm) b (mm 1.42 63.7 1.13 61.5 2.40 58.8 1

) c (mm)

d (mm) e 10.00 12.00 12.50

(mm) f (m 6.00 48. 7.00 28. 7.00 56.

m) g(mm) 00 50.00 00 50.00 00

le 3 : Specimen 24.00 7 28.00 7 34.66 he MTS 810, of necessary oduction of 5 chosen str ast 58% of re ack growth u efined as lim o loads, R, -1 sists in the u ng life, N . A mean stress according squin’s equa model was ad shows a resu

Tab

data

Fa Th 90 the cur obs Th stre stre Str Th me cha equ app end and alte

tigue tests e fatigue test [9] and ASTM critical value ve, 2 specim erved the tes e S-N curves ss levels. Th ss, S-N curv ategy for evalu e strategy us an stress, S m racterize the ation when lication of t urance limit, S ar Basquin sh rnating stres

s under axial 739-91 [10 s of design i ens were tes ts were repro were obtaine e stress relat es were desig ation of mean ed to evaluat , alternating mean and S m = 0 , the e he value of r called S ar Basq ould be iden s, S ar Model .

loads were p ] , the minim s 12 specime ted for each duced, guara d considerin ed to the in ned for the f stress models e models’ ad stress, S a , an alternating s quivalent fa esulting life, uin Then, if th tical statistic

erformed in t um number ns with repr one of the nteeing at le g the total cr finite life is d ollowing rati ’ adherence herence con d the resulti tress in one tigue strength N , in the Ba e prediction ally. Tab. 4

universal tes specimens to 50 to 75%. T ess levels. In production. nder dynami it of fatigue , 0, 1/3 and 2 se of three p ccording to model allow to specific m tion allows t herent to th me of the e

ting machine obtain a cur hen, for a pr the three le c loads, repe . In order to /3. arameters th Fig. 3, the a s to evalua odel, called o estimate a e experiment quations use

. According ve S-N in or eliminary an vels where la ating the pro evaluate th at characteri pplication o te, through S ar Model . In a new value fo al results, the d to estimat

to ASTM E 4 der to determ alysis of the rger scatter cess for diffe e effect of m ze a fatigue f the data wh extrapolation similar way, r the equiva values of S ar e the equiva

68 ine S-N was rent ean test: ich of the lent Model lent

Extrapol ar

ate data

S



 ,

a m  

S



 ,



 

Model ar

Experimental d

ata

ar

a m

Fatigu experime data

e ntal



S

m

m

S

tatistical analysis

a 

S

ar

N

S

Basquin ar

N

n stress model Cycle numbe

s’ adherence r

Figure 3 : St

rategy for eval

uation of mea

20