ABOUT THE INFLUENCE OF RESIDUAL-STRESSES, WELD GEOMETRY AND MULTIAXIAL STRESS STATES ON THE OPERATIONAL FATIGUE-STRENGTH OF WELDED CONSTRUCTIONS FROM STRUCTURAL-STEELS

Authors
Citation
Cm. Sonsino, ABOUT THE INFLUENCE OF RESIDUAL-STRESSES, WELD GEOMETRY AND MULTIAXIAL STRESS STATES ON THE OPERATIONAL FATIGUE-STRENGTH OF WELDED CONSTRUCTIONS FROM STRUCTURAL-STEELS, Materialwissenschaft und Werkstofftechnik, 25(3), 1994, pp. 97-109
Citations number
32
Categorie Soggetti
Material Science
ISSN journal
09335137
Volume
25
Issue
3
Year of publication
1994
Pages
97 - 109
Database
ISI
SICI code
0933-5137(1994)25:3<97:ATIORW>2.0.ZU;2-4
Abstract
For a fatigue design of welded structures among other influences also the influence of residual stresses, weld geometries and multiaxial str esses must be taken into account. Knowledge about the influence of res idual stresses in the high-cycle fatigue region cannot be transformed unconditionally to the behaviour in the finite-fatigue life region or to variable amplitude loading with exceedances of the endurance limit, because the fatigue behaviour depends also on the stress concentratio n in the weld toe and the related stress relief. Principally, the fati gue strength is improved by a better weld geometry, e.g. by TIG-dressi ng, by introduction of radii which are big enough. In order to transfo rm data obtained on specimens to components of bigger size criteria li ke structural or local equivalent stress, first technically detectable crack and size effects must be considered. For a fatigue life calcula tion for structures under variable amplitude loading a damage sum of D = 0.5 is recommended. While for the evaluation of multiaxial stress s tates with constant principle stress directions the von Mises criterio n can be applied satisfactorily using structural or local stresses in the weld toe. However, conventional hypotheses fail by an overestimati on of fatigue life, when the principle stress directions change, e.g. due to a phase difference between normal and shear stresses. Presently , in such cases only an experimental proof of the fatigue behaviour ca n be performed.