Steel has fatigue resistance. When one looks at a strength vs. cycles chart, the steel's strength decreases with the number of cycles, until at around 57% of the original tensile strength, the curve goes flat. So as long as the maximum load does not exceed that 57% value, the part does not fail. For aluminum the strength vs. cycles curve continues down until it hits zero.

That's the simplified version. There are other factors which affect the steel, such as surface condition and size effect, the first two that popped into my mind

And yes, I know that stress is cumulative, at least stress cycles are. But the question of what constitutes a cycle hasn't been defined very well, at least in my research. Saying that we replace aluminum parts after so long is like saying ice floats on water. The fact that it does gives no explanation of why it does.

R.