El Haddad, MH
Research engineerMember of ASTM, Mechanical Research Department, Ontario Hydro, Toronto, Ontario
ProfessorMember of ASTM, University of Waterloo, Waterloo, Ontario
Research assistant, University of Cambridge, Cambridge,
An elastic-plastic fracture mechanics solution for fatigue cracks initiating from weld toes is introduced that admits plasticity by replacing the conventional stress term with a strain term. It accounts for the propagation of very short cracks by the introduction of an effective crack length equal to the actual length increased by an amount l0, where l0 is a constant characteristic of the material and material condition. Consideration is also given to the effect of mean stress and crack front shape on an intensity factor derived from this solution. Crack growth results for cracks in both elastic and plastic strain fields of welded specimens, when interpreted in terms of the intensity factor, show excellent agreement with elastic long crack data. This intensity factor, when combined with a propagation model that includes all stages of crack growth, also successfully predicts the entire life of butt- and fillet-welded specimens for two steels. The threshold stress corresponding to the failure of butt-welded specimens is equal to the smooth specimen fatigue limit stress divided by the elastic stress concentration factor. However, for fillet welds, the stress level corresponding to failure is higher than the fatigue limit stress divided by the elastic stress concentration factor. At stresses between these two stress levels, cracks start at fillet-weld toes but do not propagate to failure.
Paper ID: JTE10629J