Chronology of Crack Initiation

For many years I ran a Fracture Mechanics testing and analysis laboratory. Circulating among the cognoscente was a tongue-in-cheek chronology of events associated with crack initiation*. This sequence of observations contains an unintended grain of truth that illustrates the random behavior of even well-behaved physical phenomena.

Stages in Fatigue Crack Detection

  1. Possibly see where a crack might possibly start.
  2. Possibly see where a crack will definitely start.
  3. Definitely see where a crack might possibly start.
  4. Definitely see where a crack will definitely start.
  5. Possibly see a possible crack.
  6. Possibly see a definite crack.
  7. Definitely see a possible crack.
  8. Definitely see a definite crack.
  9. Possibly see the crack growing.
  10. Definitely see the crack growing.
  11. Specimen fails along a crack other than the one being observed.

* The mathematics of Fracture Mechanics (F/M) requires the existence of a macrocrack (a crack that is large with respect to the surrounding microstructure), so it is convenient to divide the essentially continuous process of fatigue crack formation into “propagation” (described by F/M) from what precedes it, “initiation.” Unfortunately this has led to considerable semantic hair-splitting and the insistence by some that “initiation” doesn’t exist since there is no measurable signal of it. We will avoid such nonproductive debates here.