Minimum number of cycles to consider fatigue

[jgrady]

AISC states that for steel structures, fatigue does not need to be considered for less than 20,000 cycles. This never made sense to me, if you look at a S-N diagram, the fatigue strength begins to decrease before this. Is the thinking that if you have your safety factors against static, the fatigue strength won't be below this until at least 20,000 cycles? But it seems like you could have some part with a high stress concentration factor that could cause something to fail in fatigue before 20,000 cycles.

Does anybody know where the 20,000 number comes from?

 

[BrianE22]

Just guessing, but maybe it has something to do with high cycle fatigue versus low cycle fatigue.

 

[MGZmechanical]

The eurocode design for steel the fatigue starts at 10 000 cycles (if I'm not wrong). The european pressure vessel code considers fatigue from 500 cycles.

 

[XL83NL]

The European pressure vessel code, EN 13445, looks at equivalent pressure cycles. What that means in short is that it compares the actual cycle, expressed in pressure, with the MAWP of a vessel. actual cycles thus may be higher than 500, without consideration of fatigue.

 

[IRstuff]

The objective of any stress limitation is to increase the number of fatigue cycles, so the number of fatigue cycles is where you start, designing the system to achieve the a maximum stress everywhere in the system to prevent early fatigue. Given that a steel structure experiences a minimum of one cycle of thermal stress per day, 20,000 cycles works out to 55 years, not counting wind loading, etc.

So, yes, you could have a stress concentration that would cause failure before 20,000 cycles, but that means that you didn't do the design correctly to achieve the design requirement of 20,000 cycles.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[drawoh]

jgrady,

Why are you asking this?

If I have a large safety factor and not a lot of loading cycles, I don't worry about fatigue. If I have a low safety factor, I consider and analyse fatigue.

--
JHG

 

[jgrady]

@drawoh

I am thinking along the lines of some threaded part, where the stress concentration can be almost 4X. It guess it just doesn't make sense to me to say if you are designing something for 10,000 cycles you don't need to consider fatigue but for 20,000 cycles you do. When the fatigue curve shows that it is possible to fail in fatigue @ 10,000 cycles. I am wondering if others have experience in designing for this range of cycles.

 

[RolMec]

from a practical p.o.v.
Looking into some applicable standards:
e.g. FEM 1.001 sec 4.1.3.1 the limit cycle no. to require a fatigue investigation is 8000. That's less than 10k as per EC.
e.g. FEM 2.131 sec 3.4 sets this limit to 2.5 10exp5.

The accompanying wording is that the fatigue loads should be LC I loads and fluctuate significantly.

So, if I see the fatigue loads
- to be predominant by magnitude over the stress spectrum (near or even to LC II and LC III)
- to be fluctuating with a range of > 10%
- have a cyclic loading cycle time taking the larger part of the overall lifecycle (as e.g. in aerospace)
--> I'd do a fatigue investigation regardless of any such number.



Roland Heilmann

 

[HS_PA_EIT]

The AISC 360 fatigue provisions deal with high cycle fatigue with elastic stresses. Check out the AISC 360 commentary for an in depth description.