U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
| REPORT |
| This report is an archived publication and may contain dated technical, contact, and link information |
|
 |
| Publication Number: FHWA-HRT-17-110 Date: January 2018 |
Publication Number: FHWA-HRT-17-110 Date: January 2018 |
This section contains a brief discussion of the results for each individual beam test as well as a discussion and analysis of the results.
Table 3 gives the fatigue testing matrix for the beam specimens. The stress ranges given in the table correspond to the stress at the web–flange weld, or the stress at the inner surface of the tension flange, as calculated using elementary beam theory. Note that, prior to beginning testing, the fatigue resistance of the beams was unknown. It was assumed at the time that the fatigue resistance would fall somewhere between an AASHTO category A detail and a category B fatigue detail, inclusive.( 1 ) The choice of test stress ranges for the later tests was thus partially guided by the results from earlier tests. The desire was to achieve a fatigue failure for each beam test at less than 2 million cycles due to time considerations. A high stress range was thus used in the first test to avoid a runout situation that might have occurred because of loading below the constant amplitude fatigue limit for a category A detail.
The displacement limit for beam 1 tripped at around 702,000 load cycles. Once restarted, a large crack was noted approximately 3 inches inside the load-bearing clamping plate edge. Cycling was discontinued at 705,098 cycles when the actuator pushed sideways inside the load frame due to the now asymmetrical cross section. Figure 8 shows a picture of the entire fracture surface. As can be seen from the overall perspective, the right side of the flange had completely fractured, though the left side of the flange and web were only partially fractured. Also shown on the overall perspective is a dashed arc that indicates nearly equal crack growth across the left half of the flange and up the web, indicating that this crack likely initiated at the right flange tip. The fracture surface at the right flange tip had impact damage from the fatigue cycling, and a zoomed-in view notes the likely crack origin was a small flaw from the shear-cut edge.
The displacement limit for beam 2 tripped at 885,073 load cycles. Once restarted, a crack was observed in the tension flange near the inside edge of the load-bearing clamping plate. Instability of the actuator occurred shortly thereafter at 885,955 cycles, at which point the cycling was stopped. The observed crack plane was directly parallel to the edge of the clamping plate. As mentioned previously, the clamp-type load bearing for beam 2 oftentimes had to be tightened to halt a squeaking noise that developed as loading commenced. Failure was determined to be caused by fretting of the clamping plate. Figure 9 shows a photo of the web-to-tension flange junction where the crack was observed. The photo indicates that the ratchet marks on the bottom of the flange were likely five to six small fretting cracks that coalesced together, and the fatigue crack grew upward toward the web-to-tension flange junction.
The displacement limit for beam 3 tripped at 2,157,969 load cycles. Once restarted, cracks were observed at multiple locations. The largest of the cracks occurred in the web–flange junction approximately 10 inches inside of the inside edge of the left load-bearing plate. Failure progressed most rapidly at this location. An additional smaller web–flange crack was observed almost directly under the left support. A third crack was observed on the edge of the flange approximately 2 inches from the right load-bearing plate within the constant moment region of the beam. Actuator instability occurred at 2,167,158 load cycles, and the test was terminated. Figure 10 shows a zoomed-in view of the web-to-tension flange junction of the largest crack. Semicircular fatigue crack growth rings obviously emanate from the weld flash on the right-hand side of the weld, indicating the flash was the initiation point.
The displacement limit for beam 4 tripped at 1,483,802 load cycles. Once restarted, complete fracture of one-half of the flange was observed approximately 18 inches inside of the centerline of the left load bearing. The test was stopped due to pending instability of the actuator at 1,484,999 load cycles. Figure 11 shows a zoomed-in view of the web-to-tension flange junction. The overall crack was a coalescence of three smaller cracks at slightly different locations along the beam length. Two of the small cracks initiated at each tension flange tip and grew inward to the weld, and the third crack initiated at the web-to-tension flange weld. The view in the picture focuses on the weld, and the other cracked planes appear out of focus because they are not coincident. Fatigue growth rings indicate the crack in the weld initiated from the weld flash.
The actuator turned itself off at 856,622 load cycles when testing beam 5. At this time, a full flange fracture was observed that extended approximately one-third of the way up the web. A view of the full fracture surface is shown in figure 12. Large fatigue growth striations can be seen near the left flange tip and near the crack tip in the web. A dashed arc is drawn between these last striations, indicating a growth origin near the weld. However, complete fracture of the right side of the flange indicates the crack originated from the right flange tip, grew toward the weld, and then continued to grow into the left side of the flange and up the web. A zoomed-in view of the flange tip is provided, indicating that the initiation of the crack was a discontinuity from shear cutting at the lower left corner of the flange tip. Finally, the left flange tip was necked through the plate thickness, indicating this was a ductile overload fracture, and was the last event that had tripped the limit.
The displacement limit for beam 6 was tripped at 1,057,914 load cycles. A crack was detected, this time at the web–flange junction approximately 15 inches on the inside edge of the load-bearing plate. Restarting the test to further propagate the crack resulted in actuator instability at 1,061,520 load cycles. The test was stopped at this point. Figure 13 shows the web-to-tension flange junction where the crack was first observed. Fatigue crack growth indicates initiation started in the weld flash on the right side.
The fatigue data are plotted in stress versus cycles to failure format consistent with the AASHTO fatigue design philosophy in figure 14. The three weld failures are plotted as blue triangles falling between category A and B performance. However, to be consistent with the AASHTO methodology, a lower-bound resistance is shown as the heavy, blue dashed line in the figure representing the lower 95-percent confidence limit. The lower-bound resistance is less than category B. For completeness, the one fretting failure and two flange tip failures are also plotted on the graph.
