LOW CYCLE FATIGUE CRACK INITIATION LIFE ASSESSMENT OF HY-100 UNDERMATCHED WELD

Abstract

This effort was focused on the methodologies of the predictions on low cycle fatigue crack initiation for H Y 100 welds of an undermatched weldment. Several approaches are evaluated for their effectiveness in predicting the initiation of low cycle fatigue cracks at welded structural details in application. Finite element analyses and experiments using various types of low cycle fatigue specimens were conducted, and the results were compared with those using theoretical algorithm, such as Neuber's rule. A two-surface cyclic plasticity algorithm was formulated and implanted in the finite element code ABAQUS user subroutine to simulate cyclic stress-strain behavior of the material under cyclic loading conditions. 2-D and 3-D finite element models for the notched specimens were made corresponding to the experiments such that the local stress-strain responses can be obtained to predict the crack initiation life. The finite element results were also used to evaluate the analytical notch root strains predictors. Fatigue tests ranging from small, standard smooth specimens to notched cylindrical specimens with notch constraint using specific notch root radius were used to evaluate the fatigue properties of H Y 100 steel with under-matched weldments. In addition, full scale (2 thick) butt beam specimen with undermatched weld, machined from gas metal-arc (GMAW) welded plates, were also tested under cyclic zero to tension loading conditions. The fatigue crack initiation life defined as a sudden drop of specimen compliance was estimated for all small cylindrical specimens using strain-controlled fatigue testing methods. Tests of these small specimen with notches were also conducted to study the fatigue notch effect produced by sharp notch radius. Comparisons for the prediction of fatigue crack initiation life ware made between finite element analyses, analytical predictor and experimental results. notched specimen or component (the local strain are measured at the root of the notch) and an unnotched specimen (where the local strains are applied on smooth specimen to determine the corresponding stresses). It is assumed that the notch root material and the unnotched specimen behave in a similar manner and show the same crack initiation life behavior. The material stress-strain response and the damage accumulation are simulated on an actual smooth specimen, which should yield quite realistic results. The determination for the notch root strain usually is difficult to be achieved due to the fact that the effects of notch root constraint by the surrounding elastic material is quite significant. Current analytical notch root strain prediction models either derived from plane stress or plane strain situations. With 3 dimensional effect, the local strains predicted using the current analytical method yield unsatisfactory results. Two broad approaches can be adopted to determine notch root strains for comparison with the predictions of analytical models fall on the finite element method and experimental techniques. Experimental results are used to evaluate strain data obtained from 2-D and 3-D finite element elasto-plastic analyses. It has been shown that, for fine grained materials and moderately complex geometries, the FEA and experimental results show good agreement. This is particularly true for the strain data along loading direction.f2] Two typical analytical stress concentration strain prediction models widely used are Neuber f31 and Glinka relation.r4] Their strain and stress relations are shown below: