Without a doubt, the most common failure mode I see in my lab is fatigue. Unlike overload failures, which occur more or less immediately when a load is applied, fatigue failures are progressive in nature. In other words, they develop over time as a result of repeated loads. If you've ever bent a paperclip back and forth until it snapped, you've seen fatigue at work.
The word fatigue implies that the material is somehow getting tired, or that its properties are becoming exhausted by repeated application of loads. Of course, this is not actually the case. If you want to convince yourself of this, you can cut a small sample from of a part that failed in fatigue, and perform mechanical tests on the sample. In almost all cases, you will find that the strength is the same as it was before the part was put into service. This demonstrates that fatigue is not a matter of the material losing its strength or degrading over time. So why does fatigue occur?
Fatigue failures start with the formation of a small crack, which grows with repeated loads. Once the crack gets big enough, the part breaks.
Crack growth can be understood as an energy balance. When you apply stress to a part, you increase its strain energy. You can think of this as the energy stored in a stretched spring. Now let's think about a part with a crack in it. The surface of the crack has a surface energy associated with it. Making the crack larger increases its surface energy. So a part that has a crack in it can lower its strain energy by increasing the size of the crack by converting strain energy into surface energy. This is why a crack can grow under repeated loads. (Incidentally, it also explains why you can cut glass by scribing a line in it with a glass cutter, then bending it.)
But what causes the initial small crack to form, especially if the overall stress is low? Most parts contain stress risers, where the local stress may be significantly higher than the overall stress. Cracks initiate at these stress concentrations.
Some stress risers are geometric in nature, and can be minimized by good design. These include sharp corners, notches, and holes, among others. Other stress risers may come from manufacturing processes, and can be minimized by good manufacturing practices. These include casting porosity and other kinds of defects. Still, other stress risers may be internal to the material itself. For example, most steels contain nonmetallic inclusions, which act as internal stress risers. Bearing quality and aircraft quality steels are produced using special steelmaking techniques to minimize their inclusion content.
Really great explanation on the basic principles that cause fatigue failures--one which even non-engineers like myself can understand. Beyond the key equations you provided at the end of your piece, are there other sources of information about materials properties or specific design tools (perhaps CAE software) that can help engineers with this design challenge?
@Beth: There are a number of CAE tools for fatigue analysis. I don't have direct experience with any of them, but some I've heard of include fe-safe, MSC Fatigue, and nCode. I'm sure there are others. Some companies have their own in-house fatigue codes.
As far as sources for fatigue data, the American Iron and Steel Institute has an excellent, free, on-line database of fatigue data for a wide variety of steels. (Registration is required). Of course, it's important to understand what you're looking at. Not only do you need to find the data for the correct steel; it also needs to be the correct condition. I've seen engineers use data for a given steel in the annealed (soft) condition when the actual part is hardened, and vice versa.
Often, though, it's worthwhile to do your own testing rather than relying on published data. The main drawback of this is that a solid fatigue testing program can take months to complete. For example, at 30 Hz -- which is pretty fast for a hydraulic load frame -- a single 20 million cycle test will take over a week, and you need to perform several such tests in order to have statistically valuable data. The advantage, though, is that you can test the alloys you actually use, in the conditions you actually use them.
That extra info will be extremely helpful to the community. Thanks so much for providing it, Dave. Your point about having to do your own testing rather than rely on third-party material stuck me, however. Given that so many companies across industries are dealing with time-to-market pressures and there are so many aspects of a product that need vertification and testing, how do you prioritize what's required for fatigue testing, vs. say what's necessary for structural testing or something else? Can development teams really afford to take the time to do this, but I guess the bigger question is can they afford to not?!!!
@Beth: We try to do the best we can using published data, while at the same time accumulating our own body of test data about the materials we use most often. This way we don't hold up the development cycle waiting for test results; we just constantly fill in the gaps in our knowledge. I suspect that this is what a lot of other companies do, too.
We also try to avoid component-level fatigue testing where possible. For example, if you understand how die-cast 380 aluminum behaves in fatigue (and the effect of casting porosity on the properties), you can apply that knowledge to any die-cast 380 component; you don't have to repeat the test for each part. But for some components, like crankshafts, component-level testing is a must.
Thanks, Dave, for such a complete intro to fatigue failures. I also find it especially interesting to read about all the CAE tools for fatigue analysis. Beth's second question and your response is also intriguing. It sounds like there's a need for more centralized fatigue databases of materials and/or parts made with them. Each company doing all this on its own and building up its own database seems awfully wasteful of time and energy. Is this info just too hard to centralize and keep updated?
Dave, loved the article, and this response. I think you should qualify the use of software further though. You pointed out you need to use the right material, in the right condition. Anyone who reaches for fatigue software is likely NOT to find the correct material for their analysis at one point or another. Using a similar material as "close enough" is also likely to lead to erroneous and dangerous results.
@Ann: I think TJ's response sums up why it's often advantageous for companies to build up their own internal databases of fatigue data: "close enough" often isn't.
McDonough and Braungart explain this point of things getting overdesigned in their cradle to cradle design concept. The lack of data leading to overdesigning is really a point worth noting
On April 21, NASA launched a novel project, putting into orbit three satellites that employ an off-the-shelf commercial smartphone as the control system.
The legacy endpoint devices that control our critical infrastructure (utility systems, water treatment plants, military networks, industrial control systems, etc.) are some of the most vulnerable devices on the Internet.
In a switched-capacitor filter, capacitors and switches take the place of resistors and accurately reproduce the characteristics of continuous-time Bessel, Butterworth, and elliptical filters.
From Dell / Intel® New Paradigms in Design Work Scott Hamilton, vertical market strategist for Dell Precision workstations, 5/2/2013 5
Early in my career, I worked as a draftsman and remember the days of drawing on vellum with numbered pencils and Mylar with plastic lead. This was a fun experience in the sense that I ...
I've been using workstations for more than 10 years and love finding ways to get more performance from my system. With demanding professional applications that require more power each ...
A lasting memory from my first job as an engineer in an auto assembly plant is standing on hard concrete at six in the morning, vending-machine coffee clutched in hand, listening to ...
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 radio show will show what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.
To save this item to your list of favorite Design News content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Tweet This
11 
