Oooh this is a good one.
The behavior your friend observed isn't too surprising. When we look at fatigue due to a cyclic stress, it's a function of two things:
- The peak stress. More stress means less life.
- The stress ratio. In general, the higher this number, the longer the life.
Some explanation on stress ratio: This is the minimum stress divided by the maximum stress. Tensile stress is positive, compressive stress is negative.
- A fully reversing load (say 30 ksi tension to 30 ksi compression) would have a stress ratio of -1.
- A partially reversing load (+45 ksi to -15 ksi) would have a stress ratio of -.33
- A unidirectional load (+60 ksi to 0 ksi) would have a stress ratio of 0
As a blade goes around the wheel, the outside sees additional tension, while the inside sees reduced tension or even compression. If you know the wheel diameter, we can compute the state of stress on both sides of the blade. This is the origin of the cyclic stress which causes fatigue.
As your friend added tension to the blade, he increased the peak stress, but he also improved the stress ratio which improved the fatigue life of the blade.
Here's the S-N curve for 4130 as an example:
View attachment 398893