Page 200 - Biaxial Multiaxial Fatigue and Fracture

I84 T LAGODA ETAL.

distribution (cruciform specimens and thin-walled cylinders), the criterion of the maximum
normal strain energy in the critical plane, including the positive energy parameter under tension
and the negative energy parameter under torsion was a suitable quantity for describing the test
results of lOHNAP and SUS304 steels. All the analyses were based on a parameter of strain
energy density which has been presented in details in [IO, 11 1.
In [12 - 141 the mesoscopic approach to fatigue life estimation was proposed. The papers
concern uniaxial tension-compression with torsion in cylindrical solid specimens made of
35NCD16 steel, and the proposed approach is based on the concepts of Orowan [15], Papado-
pulos [16] and Dang-Vang [17]. In this paper the authors will reanalyse the former test results
of 35NCD16 steel and assess if the proposed energy parameter can correlate these experimental
data.
As it results from the previous papers [4 - 91, this parameter seems to be very efficient for
correlation of the experimental data obtained for steels IOHNAP, 18G2A, SUS304 and 12010.3
as well as cast irons GGG40 and GGG60.
FATIGUE TESTS

The tested 35NCDl6 steel has the following chemical composition: Si - 0.37%, Mn - 0.39%,
P-O.O1%,S-<o.oO3%,Ni-3.81%,Cr-1.7%,Mo-O.28%,C-O.364%, therestFe. Young’s
modulus is E = 205 GPa and Poisson’s ratio is v = 0.3, the yield point is R0.2 = 930 MPa, ten-
sile strength R, = 1070 MPa, contraction Z = 20.7 %. Cylindrical specimens were tested under
constant and variable-amplitude loading. In the case of constant amplitude loading, tests were
performed under tension-compression and pure torsion. In the case of variable-amplitude
loading, tests were performed under tension-compression, combined proportional tension with
torsion for the ratio of shear and normal stress T(t)/b(t) = 0.5 and pure torsion [I2 - 141. All the
fatigue tests were realized under controlled force and twisting moment. Basing on the tests un-
der constant amplitude uniaxial tension-compression it was possible to determine the regres-
sion models of the fatigue characteristic according to [18]

logN, =A-m.logo, (1)

Parameters of the determined Wohler curve are as follows: A = 21.07, m = 5.86, the fatigue
limit oaf= 450 MPa, and the corresponding number of cycles No = 3.328 . lo5. The results of
fatigue tests under torsion were approximated by a similar regression model:

logN, =A, -m,-log.c, (2)
where A, = 44.51, mt = 15.08, the fatigue limit z,f = 330 MPa, and the corresponding cycle
number Not = 3.395 lo6.
It is important to note that the numbers of cycles No and Not for the analysed material differ
by one order. The results for variable-amplitude loading are shown in Table 1. In the tests the
standard course CARLOS - lateral - Car Loading Sequence with 95180 cycles in a block was
used. The cycles with low amplitudes were eliminated so as to obtain the block of fl type (with
a high reduction of low amplitudes) and f2 type (with a low reduction of low amplitudes). The
loading of CARLOS type is a typical loading used in tests of car elements realized in France
and Germany (specially LBF Darmstadt). Its course is characterised by a high irregularity coef-
ficient.