One of the participants of the two-day India seminar that I taught for the Chennai Chapter of ASM International and the Madras Metallurgical Society in February 2017 sent an interesting question. “Valar” was one of my three Indian “sisters” who attended, along with 97 of my Indian metallurgist “brothers.”

Valar had come across some literature that stated that if smooth portions of cylindrical components (same diameter everywhere nearby) get a fatigue crack, the curvature of the crack front does not change, even if the stress level is quite high. On the other hand, the article stated, if there is a change of diameter (stress concentration), then the curvature of the crack arrest marks (“beach marks”) reverse toward the final separation event. Let’s take a closer look at some nominally uniaxial tension loaded, cylindrical areas of (steel) components.

Many fatigue cracks start out with a visible “thumbnail” shape. See the dark crescent at the bottom of the crack surface in Figure 1. This is because the crack grows in all directions at once early in the crack process. There are two directions sideways and one inward. So the crack initiation thumbnail feature tends to be about twice as wide as it is deep.

I have to say that I looked and looked in my old files to find a “classical,” and easy-to-see, fatigue-initiation thumbnail. It appears that this part was fairly highly loaded because the fatigue crack became critical when the fatigue portion was around 10% or less of the total load-bearing cross section. You can’t see it, to protect the identity of the component, but this was actually loaded in uniaxial tension (and not reversed). There could have been a bit of bending in this application. In fact, it is often very difficult to tell pure tension (rare) from tension combined with bending (much more common) from a fracture surface alone, rather than thinking about the application. Furthermore, once the crack is sizeable, even if the applied force is in one direction (along the cylinder axis) a bending will naturally occur as the crack opens and closes.

Figure 2 shows a somewhat different fracture surface. The initiation is not smooth and rounded like Figure 1. This may be due to the fact that there are several individual initiations along what is shown as the left side of the fracture surface (as seen by the small steps along the edge). The bend in the part shape (lower left) shows that this fracture probably had a combination of tension and bending.

We will take a look at more examples next time in part 2.