Familiar theme - you have to "play" with the phenomenon to usefully know it enough to act upon it effectively.
As just mentioned - same link - 2011 research on fatigue resistance of welds.
From my "Career" section...
The effect of "playing" with fatigue-testing: leading for Hyundai's construction division on fatigue-resistant design for the 3rd Bosphorus Bridge project in Turkey, 2015 (1.4km span - a massive bridge).
This is an extensive writing on various topics. If you are interested in fatigue you might jump to the subsection "The story of fatigue-resistant weld design".
* Actively in development *
No guarantees this will be confirmed - but if so, would be amazing.
For a start, this would double the cyclic stresses a bridge could take.
Invitation to take interest in high-performance steel structures for cyclic-loading ("fatiguing") applications
"High performance welds - non-welding, non-fatigue explanation"
The extensive Proposal : research path to high-strength steel fatigue-resistant welded structures (in Structures directory)
"Fatigue performance evolution in time - simple explanation"
If you know
"Weld fatigue performance - bimodal - EXC2 and EXC3"
Which - if the observation is confirmed - explains how the current Standards get it so wrong.
Contact me link