Fatigue Performance of High-Frequency Welded Steel I-Beams

SPECIMENS

Six HF-welded, doubly symmetric I-beams consisting of ASTM A769 grade 50 steel were attained to be fatigue tested.⁽²⁾ The specimens were identified sequentially as beam 1 through beam 6. The nominal geometric properties of the steel I-beams are given in table 1. A picture of a representative partial beam cross section (one flange and most of the web) is shown in figure 1. Most noticeable in this picture is the horizontal offset of the web; while it is joined at the midwidth of the flange, the forging force during welding caused a shift in the web to the left. This distortion of the cross section was observed in all the beams and did cause stability issues that will be described in more depth later. The weld itself, shown in figure 2, appeared very uniform, though the manufacturer of this I-beam did not remove the flash that is ejected from the forging event, and this is noted in the figure and highlighted, as it did affect results described later.

Tension Properties

Tension testing was performed in accordance with ASTM Standard E8 to confirm the steel mechanical properties.^{( 3 )} Four standard-sized, sheet-type tension specimens were taken from the tension flange of beam 1 after fatigue testing. The samples were taken from a flange near the support that experienced low stresses throughout the fatigue testing period. Strain was monitored with a clip-on extensometer over the 2-inch gauge length of the specimen. Due to software issues while running one of the tests, one specimen was discarded.

Figure 3 shows plots of the tension test results from the three valid specimens tested, and succinct results are presented in table 2. Based on testing averages, the beam specimen steel exhibited a 0.2-percent offset yield strength of 51.6 ksi and an ultimate strength of 71.8 ksi. The fracture location for specimen 2 was very close to one of the extensometer contact points; thus, the elongation percentage at failure for this specimen was not captured accurately by the extensometer readings. This is illustrated by the behavior shown for specimen 2 in the plot of figure 3. The average elongation percentage for the two specimens with valid fracture locations was 42.6 percent, and the average reduction in area was 71.8 percent.