Fatigue: why things break under repeated load
Most parts don't fail from one big overload — they fail from millions of small loads that each seem harmless. What fatigue is, why it's invisible until it isn't, and how the S–N curve turns cycles into a design limit.
Breaking below the breaking point
Bend a paperclip back and forth and it snaps after a handful of cycles — at a force far below the pull it would take to break it in one go. That is fatigue: repeated loads, each well under the static strength, slowly growing a crack until the part fails suddenly.
It is not a curiosity. Fatigue is how most real structures actually die — axles, gears, aircraft, bridges, wind-turbine blades. A part that would never fail under its worst single load fails anyway, because it sees that load, or a fraction of it, again and again.
How a fatigue crack grows
A fatigue crack almost always starts at a stress concentration — a hole, a sharp inside corner, a machining mark, a scratch — where the local stress is amplified well above the average. Each load cycle pushes a microscopic crack forward a tiny amount.
For most of the part's life nothing looks wrong: the crack is small and the part behaves normally. Then, once the remaining cross-section is too small to carry the load, it breaks in a single final overload — often with no visible warning. That silence is what makes fatigue dangerous, and why inspection and design margins matter so much.
The S–N curve
To design against fatigue, engineers test many identical specimens, each at a different stress amplitude, and record how many cycles each survives before it breaks. Plotting stress (S) against cycles-to-failure (N), usually with N on a log scale, gives the S–N curve: the higher the stress, the fewer cycles it lasts.
For some materials — notably steels — the curve flattens into an endurance limit: a stress below which the part essentially never fails, no matter how many cycles. For aluminium, and for most polymers and composites, there is no such floor, so life is always finite and you design for a specific number of cycles.
Mean stress, and why it matters
Real loads rarely swing symmetrically about zero. A rotating shaft also carries steady weight; a pressure vessel cycles between a high and a low pressure, not between plus and minus. That steady component is the mean stress, and a tensile mean stress makes fatigue worse — it holds the crack open.
The Goodman relation corrects the allowable stress amplitude for the mean stress present. The fatigue simulator here combines Basquin's law (the power-law S–N line) with a Goodman mean-stress correction, so you can see how a steady preload eats into the cyclic life.
Why blades and aircraft care most
Some structures are dominated by fatigue rather than by peak strength. A wind-turbine blade flexes with every rotation — of order a hundred million to a billion cycles across a twenty-year life. A transport aircraft sees tens of thousands of pressurisation and flight cycles. For these, surviving the single worst gust is easy; surviving the cycles is the hard part, and the whole design is shaped by it.
Composites are well suited to fatigue when the layup is right: the fibres carry the load along their length, and matrix micro-cracks are contained rather than running straight through as they would in a brittle metal. It is a large part of why composites dominate wind blades — a case worked through in the blade article and the energy branch of Make.
1. What is fatigue failure?
2. On an S–N curve, a higher stress amplitude means…
3. Why is a wind-turbine blade called a 'fatigue-driven' design?