This will complicate your structure design.

But fatigue failure is important.


Can you break your structure without reaching the failure stress?


Wait, what? Did I fudge you on with this one?

I said the structures are safe below the failure stress, right?


Well, I did. And it’s true.


But only for loads that happen once or limited amount of times.

We call these static loads.

However, it's different with loads that happen repeatedly over long periods of time.

We call these cyclic loads.


Every material has its limit of how many times you can stress it, even at very low levels. This is shown in the S-N curve.


S-N Curve

Here's an example, from Wikipedia:

By AndrewDressel at English Wikipedia, CC BY-SA 3.0,
By AndrewDressel at English Wikipedia, CC BY-SA 3.0,

The vertical axis shows the stress levels in your part/structure.

The horizontal values tell the number of cycles per each load. The logarithmic scale is your first clue for millions or at least hundreds of thousands of cycles at reasonable stress levels.



You have an Aluminum part, you know your loads and you calculated the stress level at 20ksi (~138MPa).

How many repeated loadings can you do before the part fails?

From the red curve we can do about 10 million (10 to the power of 7) before the part fails.


If your part is made out of Steel, for the same stress levels you could go for ever.

This is due to Endurance Limit property (flat curve at larger cycle numbers) which Aluminum doesn't have.


What happens at larger stress levels?

The number of allowable cycles falls (going left on the curve) and this is expected when getting closer to the failure stress.

This is the reason for relatively low stress levels at cyclic loads.


Live and simple example

I would probably never break it this way...

Low stress levels, can go for a long time
Low stress levels, can go for a long time

But this way it broke after around 40 cycles:

40 times and it's gone
40 times and it's gone

What about real life structures? Maybe from esk8 world?

Let's do it next time.


Meanwhile, this is an excellent introduction to fatigue and S-N curves.

And, if you google it, make sure to use "mechanical fatigue" or "material fatigue" instead of "fatigue" only.




Like what you read? Please subscribe or forward this to a friend.

You can also follow me on  Instagram.