Most fatigue disputes turn on time as much as on mechanism. When did the crack become large enough to find, was it present at the last inspection, does it predate a repair, a sale or a change of operator. The fracture surface carries part of that record, in marks at two very different scales that are routinely conflated with one another. Reading them properly is useful. Overreading them is one of the more common ways a fatigue opinion is successfully attacked.
Beach marks and striations are not the same feature
Beach marks are macroscopic, visible without magnification, and typically curved concentrically about the origin. Striations are microscopic, generally resolvable only under scanning electron microscopy, and correspond to individual load cycles. The two are separated by orders of magnitude in spacing, and the terminology matters enough that ASTM E1823 defines it. A report that uses the terms interchangeably invites the inference that the microscopy was never done.
Beach marks record interruptions, not elapsed time
A beach mark forms where something about the loading or the environment changed: a shutdown, a load excursion, a change in temperature or humidity, a period of rest during which the crack tip oxidised. A machine run continuously at constant amplitude can produce a fatigue fracture with no beach marks at all, while a part cycled through daily start-stops may produce a great many. Their number is a count of events, not of hours.
That is also what makes them useful. Where the operating log records shutdowns, seasonal changes or a known upset, the sequence of marks can sometimes be matched to that record, giving an independent anchor for when the crack reached a given size. The match has to be argued rather than assumed, and it is strongest where the intervals are irregular enough to be distinctive.
Striations correspond to cycles, with exceptions
Where striations are present and measurable, the spacing at a point is the crack advance during one cycle at that point — the quantity that ASTM E647 measures under controlled conditions on laboratory specimens. But not every cycle leaves a resolvable striation. At low growth rates the advance per cycle falls below what the microscope can distinguish, and some materials, including many cast irons and some hardened steels, striate poorly or not at all. Corrosion, rubbing and oxidation remove striations that were once there.
Counting outward from the origin does not give a total
The region nearest the origin contains most of the cycles and is the least readable, because that is where advance per cycle is smallest and where the surface has been exposed longest. Counting striations across the readable portion and extrapolating backwards understates the early life substantially. The defensible approach runs the other way: measure spacing at known crack lengths, compare against growth-rate data for the material, and integrate to a bounded estimate that states its assumptions about stress and frequency.
The final fracture zone as a proxy for stress
When the remaining ligament can no longer carry the peak load it separates in a single event, and the size of that final zone relative to the section is informative. A small final zone means the part was carrying a low nominal stress and tolerated a large crack before separating, which usually implies a long growth period. A large final zone means the section was highly stressed and gave way while the crack was still comparatively small. The comparison is qualitative, but it is a useful cross-check on any timeline derived from mark spacing.
Frequency is the assumption that converts cycles into dates
Every translation from cycles to elapsed time depends on knowing how often the part was loaded, and that number is frequently estimated rather than measured. Rotating equipment has a defensible cycle rate; a structure loaded by traffic, wind, process transients or operator action does not. Where the frequency is assumed, the resulting date carries the uncertainty of that assumption, and the opinion should say so rather than present a single date.
The record degrades after the failure
Crack faces rub against one another under compressive portions of the load cycle while the crack is still growing, so the oldest region is often already damaged before the part separates. After separation, corrosion, handling, transport and any attempt to fit the pieces back together continue the process. A fracture surface examined promptly, dry and untouched retains features that the same surface will not retain after weeks in a box in a maintenance shop.
Anchoring the surface record to outside evidence
A fracture-surface timeline is strongest when something independent agrees with it. Prior non-destructive inspection reports, vibration or process data, photographs taken during earlier maintenance, and records of repairs or component changes all constrain when a crack of a given size could have existed. Where two independent lines converge on the same window, the conclusion holds up; where the surface is the only source, the honest output is a range.
How these opinions are challenged
Predictably: that the marks described as striations were never imaged at a magnification capable of resolving them, that the loading was assumed to be constant amplitude when the service record shows otherwise, that the cycle frequency was taken from nameplate data rather than operation, and that the count was extrapolated across a region nobody could read. An analysis that identifies each assumption and shows how far the answer moves when it changes survives that. A single figure for elapsed time does not.
This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.