Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
284985 | Journal of Constructional Steel Research | 2013 | 13 Pages |
•A micromechanics damage method is presented for fracture prediction of welded joints.•Micromechanics damage model parameters are calibrated with small scale test results.•Fracture process of welded joints under tensile loading and cycle loading is simulated.•Micromechanics damage model is validated against test results of the welded joints.
Moment-resisting steel frames often fail by fracture failure of beam-to-column welded joints during a strong earthquake. This paper provides a numerical methodology based on micromechanics damage model instead of traditional fracture analytical methods to investigate the crack initiation and propagation of welded beam-to-column joints subjected to monotonic loading and ultra low cycle loading conditions. Firstly, parameters used in the micromechanics damage model for steel base metal, heat affected zone and weld metal were calibrated, respectively, against uniaxial tension test results and cycle test results of notched specimens. The evolution of void growth in the notched specimens under different loading conditions was compared. Secondly, fracture process of the welded joints subjected to tensile loading was simulated based on the micromechanics damage model. The predicted load displacement response agrees well with the other researcher's test results. Finally, the micromechanics damage model was applied numerically to investigate the ultra low cycle fatigue fracture behavior of the welded joints under constant amplitude as well as variable amplitude inelastic cyclic loading. According to the distribution and evolution of void in the welded joints obtained from finite element analysis, crack initiation and propagation were presented and the number of cycles to fracture was predicted. It is shown that the fatigue life predicted from finite element analysis based on micromechanics damage model agrees well with the other's test results.