Sounds like a possible engineering problem and the amount of torque used when installing the bolts.
The S.A.E. has ratings on bolts, and it seems the engineering for the bridge should have allowed for worst case conditions -- not best case conditions.
In the 1970's, I worked for a computer company (Microdata) that built a computer based on best case conditions for circuits. It did not account for slow parts.
Of course, semiconductor devices were just started to be used for memories as well as other devices.
The original 1K MOS memories were terrible devices that came out of Silicon Valley then...
They came loose because they were elongating beyond elastic limits or even breaking under tension.
>> Sounds like a possible engineering problem and the amount of torque used when installing the bolts.
Or unexpected forces on the assembly.
How does one accurately torque an 18’ bolt?