Pulsating tension stress high cycle fatigue tests were carried out on mild steel plates with various Unotches (form factor α=1-6) cut at base metal and at welded bond, and the following conclusions were obtained.
(1) Fatigue limit of crack initiation (σI) is slightly higher in weld than in base metal, and the α-dependency of σI is strong. But with an increase in α, a decrease in σI shows a tendency to saturate gradually.
(2) In welded specimens, fatigue limit of crack propagation (σp) is considerably greater than σI because of crack propagation being inhibited; and σp is seemingly independent of α and nearly constant at 14kg/mm
2. It seems that the non-propagating crack in weld is dependent on structural change, hardness change and residual stress due to welding and loading conditions etc. In this experiment, σp branched off from σI at around α=3 (branching point).
(3) Generally, in weld the crack propagates ladder like at a far lower velocity than in base metal. Therefore, it can be considered that the resistance to crack propagation is greater in weld than in base metal.
(4) The notch sensitivity factor for crack initiation (qI) increases rapidly from α=1 to about 3, in the range of 3<α<6 being nearly constant (0.61-0.65) regardless of α. While the notch sensitivity factor for crack propagation (σp) almost linearly drops with an increasing α (3<α<6).
(5) From the facts described above, we may conclude that the geometrical form of notch has decisively effect on crack initiation, while the various changes due to welding have greater effect on crack propagation than on crack initiation.
(6) Fatigue limit (σI) at least at α<1.3 is supposed not to be very different between notched and plain specimens.
(7) ε
0 tends to be smaller, as α is larger, and as the stress gradient is steeper or the notch radius ρ is smaller when α is same.
(8) In the first stage of crack propagation, the crack propagating rate is slower as the stress gradient at the notch is steeper.
View full abstract