SM50A鋼突合せ溶接継手の疲れ強さのばらつきと余盛形状について

抄録

It is well known that the fatigue strength of reinforced butt welded joint has a always large scatter. It this case, the fatigue crack almost originates at the toe of a weld. Therefore it is considered that the fatigue strength of the welded joint is primarily related to the shape of reinforcement, and the scatter of the stress concentration due to that shape causes the scatter of the fatigue strength of welded joint.
In this report, the above correlation was investigated and the result was examined by a statistical analysis. The material used in this study was 6mm thick plate of a steel for welded structure SM50A. V-groove was used for edge preparation and after arc welding with covered electrode, only the parts of sound weld were used as specimens from the result of X-ray examination. Specimens for welded joint and base .metal were machined to 25mm width. After measurement of the height of reinforcement h and the toe radius of weld r were chosen as the shape parameters. It was found that the statistical distribution of h was normal and that of r was log-normal. It was also shown that the effect of sampling the specimens from different welding lots could be neglected.
Fatigue tests were carried out in axial loading with R=-1 using a Vibrophore type fatigue testing machine. After each tests, the point of crack initiation was confirmed and the two parameters h and r were measured at that point. The stress concentration factor Kt was obtained from h and r using a formula due to Heywood. From the experimental relation between Kt and Kf (the fatigue notch factor) for the base metal, Kf for the welded joint was obtained. Then the scatter of fatigue strengths of the welded joint revised by Kf was examined statistically.
It was found that the scatter of fatigue strengths revised by Kf was of the same order of magnitude as that of the base metal, and was much smaller than that of non-revised fatigue strengths of the welded joint. Therefore, it was confirmed that the scatter of the stress concentration due to the shape of reinforcement caused the scatter of fatigue strengths of the weldment. Moreover, from an experimental relation between the height of reinfrocement h and the toe radius of weld r, the fatigue strength of the welded joint could be estimated using only h from that of the base metal.
Also in this report, the fatigue strength at 2 × 10⁸ cycles for the welded joint was investigated statistically using the staircase method. It turned out that the estimated mean value was 12.5 kgf/mm² and the estimated standard deviation was 2.5 kgf/mm².