Yes, we work with ultrasonically driven dental tool manufacturing. You can quickly reach a fatigue failure at 30KHz if you have a surface finish that can initiate crack propagation. Excellent article.
I used to get frustrated by friends in my car hobbies discussing bolt failures of rperformance cars. Classic examples were the failures of the Corvette suspension bolts at the ends of the leaf springs. In many cases these were fatigue failures and the friends reactions were to the effect that "these are grade 8 bolts with very high tensile strength, you cannot get better than that!" Often these high strength bolts failed from brittle crack propagation from fatigue not from gross overstress. Salt pitting corrosion was often the culprit. The strength was less importent than corrosion resistance and anodic zinc plating. Sometimes aircraft hardware with controlled geometry under the head, good thread root shape, and good hole preparation was also helpful.
Thanks for highlighting the fatigue failure mechanism in an new way. Interestingly enough the same design approaches that reduce stress risers for fatigue failure generally also improve the performance of a part in heavily loaded, non-cyclical stress situations. Good design practice often yields multiple performance benefits.
Dave, thanks for that input. The attitude is one that is, unfortunately, prevalent in several industries. Glad to hear that there are movements to open up things more in materials engineering. Industry associations can play a big part in such efforts.
If the Made for Monkeys column is any indication, there are a ton of folks out there that are well willing to juryrig products or trouble shoot issues just to get a longer product life span. Interesting, often those older products have failure points that don't necessary have to do with fatigue of parts, but rather quirky design choices that lead to issues.
@kenish: You are absolutely right; any interruption to the surface is a potential site for initiation of a fatigue crack. This might mean nicks or dings which occur either during manufacturing or during use.
Fatigue testing is usually done on highly polished samples. Broadly speaking, the rougher the surface, the lower the effective fatigue strength. Surface finish is a very important parameter, which unfortunately is very difficult to incorporate into a FEA model.
Ncode/Somat has a good fatigue calculator, designed by Daryll Socie from the University of Illinois, Champaign. He also offers seminars/short courses for engineers who need to know more about designing for fatigue.
Using the NCode/Somat software to perform the calculations works fine, but you still have to know all of the properties of the materials in each situation to plug into the software in order to get accurate results.
If "overdesign" is the term for a design that includes only just barely enough to last until the warranty expires, possibly, then it is a very good attribute. The reality is that many products are sometimes used beyond the "typical" levels, and so they do need to be stronger than only enough to handle "typical". Designing only to the lower boundry of typical is why such a large portion of consumer goods are trash at the very instant that they are made.
Fatigue failure is indeed a whole lot more subtle than the other kinds, such as yield and wear failure, but it is avery important consideration in a lot of places. The article was both useful and needed.
Dave, great article. I'm a EE and this helped me visualize fatigue in a different way. Even I understood it! You mention design, manufacturing, and materials as sources of fatigue cracks. Aren't they also important to prevent minor damage in-service from turning into a fatigue site (along with periodic inspection of critical areas...I'm thinking about aircraft skins)?
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.