Failure Analysis: How to Recognize a Fatigue Crack

ih0521-tt-fig1-900 — Fig. 1. Fatigue crack on a fretted shaft. Fatigue cracks are usually smooth when they start growing. As the crack gets deeper, the newly created surfaces get rougher. Thus, we can see that the crack started on the left and propagated toward the right, as shown in the figure on the left. Close examination of the smooth area revealed a small crescent shape radiating outward from a roughened (and darkened) position on the surface of the shaft. See the detail at right. The white arrow shows the fretting damage where the crack started, and the red arrows show a “thumbnail” shape that is a witness to the extent of the crack early in its propagation history.

ih0521-tt-fig2-900 — Fig. 2. The black-and-white arrows show ratchet marks. The series of blue and purple arrows show some more- and less-distinct beach marks. The series of red arrows point to the inner edge of a slightly darkened feature at the edge of the crack. This is not a classical “thumbnail” shape but a series of overlapping thumbnails. This crack initiated in multiple locations, but the cracks were confined to two arcs opposite each other. The interpretation of this fracture is that the lower arc cracks either initiated before the ones shown at the top or they grew faster. The lower crack is bigger than we can see. Its far reaches extend under the flap at the top of the image.

ih0521-tt-fig3-900 — Fig. 3. Faint beach marks are visible beyond the end of the ratchet mark. This crack is very smooth. The right side of the keyway has a bit of a swoop to it, while the left side is perpendicular to the length of the shaft. This indicates that the left side was the last to separate. The reddish-brown coloration on the shaft surface is consistent with fretting damage.

ih0521-tt-fig4-900 — Fig. 4. The lower black rectangle shows the crack initiation. The spreading conchoidal marks witness the enlarging crack, which started on the inside diameter and grew in depth and width until it created a leak path for the molten metal it was containing. The rainbow, orange, red and green arrows show the beach marks at later and later stages of crack growth. The green rectangle (left) shows some faint beach marks that are mostly radial in orientation. The crack, once it broke through the wall at the bottom, made its way around the circumference. We also see quite a few ratchet marks along the inside diameter of the “cup” to the left and especially to the right of the central-initiation black rectangle.

Spectacular structural collapses sometimes happen due to inadequate strength of the material used to make the structure. But machinery components, subject to stresses from rotational motion and/or vibrations, usually break due to fatigue.

What does this mean exactly? Metal and plastic parts do not “get tired” in the same way humans do. Being in a state of fatigue, for humans and other animals, implies that we are as good as new after resting. But metal components subject to fatigue do not get their strength restored. They are permanently impaired, unless they are somehow repaired.

Fatigue loading is repeated, cyclic loading. The loading may be uniform, random or some other pattern. A truck going down a smooth highway may experience fairly uniform loading on the axles once it is loaded with whatever it is carrying. It may carry a heavy load for one trip and a lighter load for another trip. A vehicle on a dirt road with many potholes will experience a non-uniform load, even if the exact same weight and distribution of cargo is hauled every time.

What is unique about fatigue and what confused early railroad engineers trying to understand why locomotive axles were breaking is that a fatigue crack initiates and grows at stress levels below the nominal yield strength. Why would a crack initiate and propagate at such low stresses?

Eventually, people figured out that although the nominal stress (i.e., average stress) was low, a sharp corner, a corrosion pit, a scratch or dent, damage from weld heat or many of a long list of other factors could result in a local area with the actual stress approaching the tensile-stress value. Thus, with repeated load cycles creating a stress above the yield strength – even within an area as small as one or two grains – a crack may initiate. If you are unfamiliar with the idea of grains, you may wish to look up ASTM grain-size numbers.

The bottom line about fatigue cracks is that they are progressive. They happen over time. They do not happen all at once due to a single load cycle. Fatigue cracks in most common structural steels will look very different from cracks that happen all at once. Figure 1 shows a shaft from a stamping press. The crack initiated from an area roughened due to contact with the inner bearing race, which was supporting the shaft. Fatigue cracks commonly have beach marks, or alternating stripes of lighter and darker colors. The different colors result from different loading levels, different atmospheric conditions or both. Therefore, cracks that grow in uniform loading and environmental conditions do not have beach marks.

How can we tell if it is a fatigue crack if the most well-known feature of a fatigue crack is not present? Figure 2 shows a tie rod from an injection-molding press. While there are a few faint and indistinct beach marks, we have another indicator: ratchet marks. Ratchet marks are (generally) small steps along the part surface. Ratchet marks tell us that there were multiple crack initiations on both sides of the ratchet. Higher stress levels and sharp stress concentrations often result in a series of ratchet marks.

Figure 3 shows another shaft. When viewed directly, the fracture surface is very smooth, almost featureless. The overview image at left already has the contrast enhanced. The detail view at right has the contrast enhanced more to show some slight variations in color. Very smooth fatigue cracks are generally associated with low stress.

If the stress is so low that the fatigue crack kept propagating when there was barely any material left “holding on,” why did the crack start in the first place? There might have been a shock load that was not big enough to create a complete separation. There might have been some damage to the surface, as previously noted. In fact, we see some severe damage to the surface. But the area where the crack appears to have started, left and right of the one obvious ratchet mark, looks pretty clean. A mystery!

Figure 4 shows the solid end of a die-casting machine “gooseneck.” This was basically a deep cup-shaped part. The entire cylindrical wall of the cup broke off. This was sent to me for a quote, but when I saw it, I did not recognize the crescent marks as the beach marks that I now believe they are. To be fair to myself, I was younger, less knowledgeable and very suspicious about the fact that they had boiled the broken gooseneck in sodium hydroxide to remove the metal that had solidified onto the fragments. I think after sitting in my lab for a couple of decades, as part of the “museum of failure,” the marks have become somewhat more noticeable. The very smooth fracture is also consistent with fatigue.

Conclusion

Fatigue cracks are sometimes easy to recognize and sometimes much more challenging. Beach marks, ratchet marks and smooth surfaces are all keys that may point to the crack having propagated over time in fatigue due to cyclic stresses rather than due to a single, large load cycle. Beach marks are not found on cracks resulting from uniformly loading in a uniform temperature and humidity environment.

All images/graphics provided by the author

Debbie Aliya is the owner and president of Aliya Analytical, Inc. in Grand Rapids, Mich., and specializes in failure analysis and prevention. She has a BS in Metallurgy and Materials Science from Carnegie Mellon University and an MS in Materials Science and Engineering from Northwestern University. She is also an IMT associate.