failure analysis
Failure analysis is detective work on broken parts: you read the fracture surface to find out why the part failed, much as a coroner reads a body to determine the cause of death. Every broken component carries a record of how it died, if you know how to read it.
You examine the fracture surface (this study is called fractography). Beach marks and fine parallel striations mean fatigue — a crack that grew a little with each load cycle. A cup-and-cone shape covered in tiny dimples means ductile overload. A flat, faceted surface with 'river' patterns means brittle cleavage. Chevron marks form arrows that point back to the crack's origin. You then combine what you see with material tests, the actual loads, and the service environment to reconstruct the story.
This matters because it stops disasters from repeating — the brittle cracking of the WWII Liberty ships and the fatigue failures of the early Comet jet airliners were both solved this way. The honest caveat: the crack usually starts at a tiny flaw or a stress concentration, so the true 'cause' is normally design, material, AND service conditions together, not any one of them alone.
A snapped bolt is examined under a microscope. The fracture surface shows a smooth region covered in fine parallel lines curving out from one corner, then a rough final zone. The lines are fatigue striations and beach marks: the crack grew slowly, cycle by cycle, from a stress concentration at the thread root, until the shrinking sound section finally overloaded. Diagnosis: fatigue, not a single overload — so the fix is a smoother thread and lower cyclic stress, not just a stronger bolt.
The fracture surface itself records how and why the part died.
The visible break is usually the end of the story, not the start: the real cause is often an earlier flaw, stress concentration, or service condition, so blaming 'bad material' alone is usually wrong.