Theory 21/Manufacturing & Metrology /Manufacturing science/3 min
The Surface Outvotes the Alloy
Design teams argue about the alloy. The crack starts in the last ten micrometres.
Fatigue life of a machined component is set more by surface roughness than by bulk material properties, so two parts from the same alloy can differ in life by an order of magnitude, and surface specification should dominate alloy selection.
01The evidence
- Surface roughness and a non-martensitic surface layer were shown to affect fatigue life, with the surface condition controlling initiation rather than the bulk.JSEM
- A gradient-based multiaxial criterion for fatigue crack initiation has been developed for components with surface roughness, which means the roughness effect can be predicted rather than only observed.International Journal of Fatigue
02The tension
Material selection is driven by bulk properties: strength, toughness, corrosion resistance. Surface finish is specified as a manufacturing detail, often with a single roughness number and little attention to its direction or the subsurface layer it leaves. Failures then occur at stresses the bulk data said were safe.
03The theory
Fatigue cracks initiate at stress concentrations. A rough surface is a field of them, with each groove acting as a small notch. The local stress at the root of a groove can exceed the nominal stress by a factor that depends on the groove geometry, not on the alloy.
If initiation dominates life, then the surface is the controlling variable. That predicts that a rough part in a strong alloy can fail before a smooth part in a weaker one, and that the ranking of alloys by fatigue life should collapse when roughness is controlled.
The prediction is measurable with existing test machines: hold the alloy fixed and vary roughness, and the life should span an order of magnitude. Then hold roughness fixed and vary the alloy, and the span should be much smaller.
The manufacturing consequence is that surface specification should be treated as a design variable with the same status as alloy choice, and that finishing processes should be selected for the residual stress they leave as well as the roughness they produce.
04What this is not
Bulk design uses S-N curves from polished specimens and applies a surface factor. This makes the surface the primary variable and the alloy secondary, which predicts that alloy ranking is unstable across finishes.
05Predictions
- P1Fatigue life should vary more with roughness than with alloy at fixed geometry.
- P2Alloy ranking by fatigue life should collapse when roughness is controlled.
- P3Surface residual stress should predict life better than bulk strength.
06What would refute it
Life tracking bulk strength at controlled roughness would restore the material-first account.
07What it would change
It would move surface specification into design, and change how finishing processes are selected.
08Evidence base
- Effects of surface roughness and surface layer on fatigue life — JSEMhttps://www.jstage.jst.go.jp/article/jjsem/11/Special_Issue/11_Special_Issue_s251/_article
- Gradient-based multiaxial criterion for fatigue initiation with surface roughness — IJFhttps://www.sciencedirect.com/science/article/abs/pii/S0142112318303943