Fatigue testing of welded joints by means of local criteria: a comparison between the fictitious notch radius-based criterion and the criterion involving strain energy density in a finite volume

摘要

In order to evaluate (estimate) the high cycle numberrnfatigue strength of welded joints with different geometries,rnRadaj [1] suggested the use of effective notch stresses,rncalculated after having introduced a fictitious notch radiusrnrf =1.0 mm at the weld toe and root. The radius value,rnvalid for common low–strength construction steels, hasrnbeen determined by Radaj using the Neuber expressionrnrf=r + s r*. Conservative estimates based on a real notchrnradius r=0 and a microstructural length r*=0.4 mm (forrncast iron), combined with a multiaxiality factor s = 2.5,rnhave been shown to be realistic for welded structuralrnsteel joints, with nucleation and propagation of fatiguerncracking starting from the weld bead toe or root [2].rnUsing the criteria based on generalised Notch StressrnIntensification Factors, or NSIF, as formalised byrnLazzarin and Tovo3, in the evaluation of welded jointrnfatigue strength, the weld toe is modelled with a sharprnV–shaped notch (r=0) and the local stress distributionsrnin the flat transverse sections are given as a function ofrnthe Mode I and Mode II generalised stress intensificationrnfactors, indicated below by the symbols K1 and K2.rnThese factors express the asymptotic stress distributionrnintensity in accordance with the theoretical solutionrnobtained by Williams, valid in the hypothesis of flatrnstress or deformation4. In cases where it is possible tornassume an angle of 135 degrees at the weld toe, whichrnis certainly the most common value in angle bead joints,rnonly the contribution of Mode I is singular while thatrnof Mode II is not (indeed, it should be remembered thatrnthe contribution of Mode II is only singular for openingrnangles less than 102.6°).