All the margins detailed in our ECSS-E-ST-32-19C calculation guide cover static loads. But a joint subjected to cyclic loading (vibration, pressure/depressurization cycles, repeated loading) can fail by fatigue well before reaching its static limit. This article covers the method and the most common pitfalls.
Why fatigue is a distinct failure mode
A positive static margin guarantees a bolt withstands an applied load once. It says nothing about what happens after tens of thousands, or even millions, of load-unload cycles. The fatigue mechanism is different: micro-cracks progressively initiate at a stress concentration point (typically the first engaged thread), then propagate to fracture — often well below the material's static yield strength.
This is why a fatigue check is essential whenever a joint is subjected to repeated variable loads over time, regardless of the static margin results.
Preload's protective role
An often counter-intuitive point: a properly sized preload reduces the cyclic stress amplitude seen by the bolt — it doesn't increase it. As long as the joint remains in full contact under external load (no separation), most of the cyclic load is absorbed by a decrease in compression in the clamped parts, rather than a proportional increase in bolt tension — thanks to the stiffness ratio between the bolt and the parts (see our article on δb and δc compliances).
This is precisely why insufficient preload is one of the most common causes of fatigue failure: as soon as the joint partially separates under load, the bolt directly absorbs the entire external load variation, with a much higher stress amplitude.
Fatigue margin calculation principle
The calculation compares the alternating stress amplitude actually seen by the bolt (derived from the cyclic external load and the joint's stiffness ratio) to the material's endurance limit for the geometry considered — often represented by a Haigh or Goodman diagram, combining mean stress (linked to preload) and alternating stress (linked to cyclic load).
Unlike static margins, which use a single operating point, fatigue verification requires knowing the actual service load spectrum (amplitude and number of cycles), generally derived from the global structural analysis rather than a simplified assumption.
The effect of thread manufacturing process
The thread manufacturing method has a major, often underestimated, impact on fatigue life: a rolled thread (formed by plastic deformation) introduces surface compressive residual stresses, which significantly delay fatigue crack initiation. A cut thread (formed by material removal) doesn't have this beneficial effect, and generally shows a noticeably lower endurance limit for an otherwise identical geometry.
This difference can represent a factor of 2 or more on fatigue life — a parameter to explicitly verify on the datasheet of the bolt used, not to assume by default.
Stress concentration and location
Fatigue failure of a bolt almost always initiates at the same location: the first engaged thread on the loaded side, where stress concentration is maximal due to the thread's geometric discontinuity combined with concentrated load transfer on the first threads. The under-head fillet is a second frequent concentration point, particularly for bolts with a tight fillet radius.
Common pitfalls
- Only checking static margins on a joint subject to known vibration or cyclic loads — the most frequent and costly omission.
- Underestimating the required preload thinking it reduces stress — the opposite is often true in fatigue, an insufficient preload worsens the stress amplitude seen by the bolt.
- Ignoring the thread manufacturing process (rolled vs cut) when selecting the endurance limit used.
- Using an unrepresentative load spectrum, particularly an underestimated amplitude from a simplified structural analysis.
Automating this with BoltCore
Fatigue verification requires correctly combining mean stress, alternating stress, and the joint's stiffness ratio — a calculation prone to error when done manually.
BoltCore automatically calculates all 11 static safety margins from your joint parameters, consistent with the ECSS-E-ST-32-19C methodology.
Try BoltCore for free →Frequently asked questions
Is higher preload always better for fatigue?
Only up to a point. Sufficient preload keeps the joint closed and limits the cyclic stress amplitude seen by the bolt, but excessive preload brings it closer to the static yield limit — there's an optimum, not a "higher is always better" rule.
How do I know if a thread is rolled or cut?
This is generally stated on the manufacturer's datasheet or inferable from the standard manufacturing process for the bolt class concerned — if in doubt, confirm with the supplier, as this information isn't always visible to the naked eye.
Does every bolted joint need a fatigue check?
No — only if the joint is subject to repeated variable loads over time (vibration, mechanically-coupled thermal cycles, dynamic loading). A purely static joint doesn't need this additional verification.