Metal fatigue is a microscopic process of atomic rearrangement, that always occurs at stress levels below the elastic limit. No macroscopic changes are discernable. Normally it's well below the elastic limit. The elastic limit is always higher than the stress at failure when metal fatigue occurs.
Metal that is stressed beyond it's elastic limit at RT is simply cold worked. In the case of this bridge, cracks were noted. This is mild steel, so those cracks were not likely due to corrosion. Those cracks indicate the fatigue limit of that steel had been reached, and that result could be applied to the rest of the same steel in the rest of the structure. That is the new ultimate limit to be used for that steel. Values for new steel no longer apply and an effective elastic limit doesn't exist, because with fatigue, it's probable that the elastic limit is the same as the ultimate limit. ie, the steel will not bend, it fractures.
you state: “ie, the steel will not bend, it fractures.” But as Rosie has informed us, it doesn’t melt!