Page 67 - Biaxial Multiaxial Fatigue and Fracture

52 E LABESSE-JIED ET AL.

or the incremental hole method. These residual stresses are then combined with the local
stresses in order to determine the total operational stress states from which the fatigue life
procedure is performed:

The plasticity correction is performed when it is necessary with the cyclic plasticity
parameters of the critical area. The multiaxial fatigue criterion function and the Miner's
damage amount are calculated on each material facet. By this way the fatigue life on the facet
maximising the damage is assessed.

EVALUATION OF THE METHOD ON WELDED COMPONENTS

The proposed fatigue analysis method has been performed on welded structures. Figure 10
shows experimental data on butt-welded joints. Two series of tests corresponding to loading
ratios equal to -1 and 0.5 respectively are plotted in the figure. These results have been
obtained for the 6 mm thick 16MnNiCrMo5 high strength steel plates welded in 2 layers with
the MAG process.
The real geometry and angular distortion of the joint has been measured and then modelled
by the Finite Element method. Figure 7a shows the stress states involved when the joint is
clamped in the jaws of the fatigue testing machine and subjected to nominal axial loading. The
weakest point of the joint is focussed on the weld toe. The local stress states are calculated by
the Finite Element method using the actual geometry of the weld. For an axial cyclic loading
they are obtained by combining the cyclic nominal stresses modified by the stress
concentration matrix and the initial stress states due to the distorted welded specimen clamped
in the jaws of the testing machine.
The local mechanical properties of the overheated thermally affected area corresponding to
the weld toe are presented in Table 2. These properties come from cyclic loading tests
performed on specimens on which a heat treatment simulating the overheat of the thermally
affected area has been realised and validated using the inverse method [ 121.

Table 2. Mechanical properties of the thermally affected material

=Y K' n' O'f b
(ma) @@a) (MPa)
565 1150 0.111 2080 -0.13

As the welded joints are subjected to high-temperature stress relieving, the residual stresses
are disregarded. The fatigue strength of the joint under tensile cyclic loading corresponding to
the ratios R = -1 and R = 0.5 respectively are calculated. They are plotted with straight lines in
Figure 10. The comparison against experimental fatigue tests results shows a rather good
agreement of predicted fatigue behaviour despite the fact that the scatter of these results is
rather wide.