Page 413 - Biaxial Multiaxial Fatigue and Fracture

A Multiaxial Fatigue Lfe Criterion for Non-Symmenical and Non-Proportional Elasto-Plastic ... 391

of-phase. It is seen that the von Mises criterion produces the same value of equivalent strain as
for the in-phase case.

COUPIRISOX BENKENCRmRU

, . , ... .. . . .... ....
......
... ...... ,.. ....' .. .. .. . .......... -
F" - - New Criterion (limed)
-.- Sonsino-GrubsC
6
i
0 ' I 2 3 4 5 6 7 00 ; 2 i 0 New Crilermn (equivaM strain) 7
6
mt [rad] mt [rad]
Fig. 1 1 Comparison between criteria
The new proposal leads to an equivalent strain higher than that of the von Mises value but
lower than the Sonsino-Grubisic approach. This means that for out-of-phase loading the new
proposal predicts a shorter life than for the in-phase case. However, longer lives are predicted
by the new approach than by the Sonsin-Grubisic methodology. In view of the analysis of the
experimental data presented in the previous section, which showed that fatigue life predictions
of the Sonsino-Grubisic method were conservative, one can conclude that the new proposal
represents an improvement with respect to the Sonsinc-Grubisic criterion.

CONCLUSIONS

The most commonly used multiaxial LCF strain-based criterion (Le. von Mises) leads to the
same equivalent strain for in-phase and out-of-phase loading of the same imposed total strain
amplitudes, resulting thus in unsafe life predictions. The original proposal of Sonsinc-Grubisic
represents a certain improvement of the preceding criteria because it summarises all the
advantages of criteria based on equivalent strain, correctly predicting the fatigue life reduction
observed under out-of-phase loading conditions. However, in view of the analysis of the
experimental results examined before, it seems that the Sonsino-Grubisic approach leads to
some extent to conservative life predictions. This could be possibly attributed to the lack of
taking into account the beneficial influence of stress gradients. The stress gradient effect is not
modelled even in the approach proposed here. Nevertheless, the new methodology developed
in this work has proved to give life predictions that are in good agreement with the
experimental results. The advantage of the new criterion relies on the fact that no elasto-plastic
material model is required and, since mean strain effects are explicitly included in the proposed
methodology, its applicability may be also extended to the intermediate fatigue region.

REFERENCES

1. ASME (1988). Cases of ASME Boiler and Pressure Vessel Code. Sec. 111, Div. 1, Code
Case n"47-23, "Class I Components in Elevated Temperature Service", Appendix T, NY.