The fracture discovered on May 11th by Michael Baker International, occurred in a welded splice between two plates in the tie girder of the bridge. Upon closer examination of the specimen it was discovered that the initial fracture formed in an area of the weld where two weld repairs had been performed during fabrication. The weld repairs were more susceptible to cracking because of the type of steel and the welding method used in the fabrication of this bridge in the 1970s [probably sensitized and no post-weld heat treatment (PWHT). Inspection (NDT) methods/equipment were not as robust in the 1970s as they are now]. In all likelihood the cracking in the weld occurred within hours of its completion but was not detected by any post-weld repair fabrication testing and remained unchanged for a number of years.
In the 1980s, the potential for cracking in welds was identified at a national level because of defects found in other similar bridges, and in 1982 an ultrasonic testing inspection of the Hernando de Soto tie girder welds was performed. The defects at this weld location went undetected.
The fracture report described how the crack propagated in three phases from the cracking of the weld repairs to the eventual fracture that was discovered in May of 2021. The initial fracture occurred on the interior face of the box where it was not visible by conventional inspection. The second phase of crack propagation fractured through the remaining thickness and was later identified in the 2019 drone video. The third fracture event propagated up the remaining web, across the top flange, and arrested in the flange to web weld of the tie girder. It is likely the existing weld cracks became unstable as a result of a unique combination of low temperatures, increasing tie girder stress, and the effects of live loads (aperiodic medium to high stress with low frequency) to which the bridge had not been previously subjected.
Given the 48-year service life of this structure, the fact that the subject fracture occurred in three separate phases over several years, the recent inspection efforts to identify welding defects, and no evidence of observed fatigue crack growth during this study, it is highly unlikely that a similar fracture will occur. However, it is prudent to continue arms-length fracture critical inspections with a focus on identifying any new visible cracks at these welds. Given that all other anomalies investigated originated on the inside surface of the tie girder, ultrasonic testing should be performed on a periodic basis.
