What maks us want to do fatigue analysis? 1

[BeBtrEngeer]

Hi everyone,

I have just joined (although I've been using the knowledge here for a few years already).

I've got a very short question necessitated by wildely differing approaches that have appeared at my job (by two analysts). One has compiled a report focusing on fatigue calculations only, the other proceeds with static nonlinear (due to contacts) only for certain structure.

The question - what set of circumstances makes it - (a) required (b)potentialy wanted - for fatigue calculations to be had for a structure? It's a general question.

I've searched the internet but couldn't come up with the guidelines as to when do the calculations (blurred lines to be expected)and when to forgo them.

Thanks!

 

[FEA way]

Well, if the structure is going to be subjected to relatively long-lasting cyclic (repetitive) loads then it would be nice to check it for fatigue. Many industries require this for various components. Often physical testing is necessary but at least fatigue analyses on FEA models should be carried out if possible (and if loading is too complex to rely on hand calculations).

 

[rb1957]

If you think the part will be subject to significant oscillating loads ... then you should do a fatigue analysis.

If fatigue analysis is the tool in your toolbox ... then you will do a fatigue analysis.

Different people have different interests, and exploit those interests in their work. If the boss allows this situation, then learnt what you can !

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[SWComposites]

It depends highly on the application. There simply is no universal answer. If its a one time use thingy, then probably no need to bother with fatigue analysis. If its designed for 40 years, and going to see lots of cyclic loads, then its probably a good idea to do a fatigue analysis.

The other thing is what material is being used, and what is the design stress level? Some materials have a fatigue endurance limit (most steels) and if you stay below that stress level, then there (typically) is no fatigue issue. Other materials (aluminum) do not have an endurance limit and thus are more sensitive to fatigue loads.

So, a) determine the loads on your structure, b) determine the material(s), c) determine the stress levels at design ultimate and at fatigue load levels.

Oh, and get the book, Fracture and Fatigue Control in Structures, Barsom and Rolfe and study it.

And lastly, this is NOT a FEM issue, it is a structural analysis issue.

 

[rb1957]

guess you guys don't like "ncode" ?

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[BeBtrEngeer]

Thanks for the answers!

I sort of expected that it's hard to quantify this. I know the basic thory behind it but it's hard to justify or forgo the calculations.

In this case it's a heavy, high strength steel (650 MPa yield point) structure loaded in one dirction a couple of houndred times a day with a normal force and a big bending moment, designed for infinite life expectancy. Consists of thick plates welded together, mostly loaded in their plane via contact with nuts and washers.
Contact stresses are decently high (up to350-ish MPa) while stresses far from the contact area go down to insignificant levels quickly. Max displacement slightly above 0.5mm.

I guess a similar load case would be an excavator constantly lifting it's near max capacity load (a granite block for example?).

What I mean is - I know it's being loaded and unloaded in certain fashion, it can't fail, it's welded, it's definately oversized for static load case but if I know the max load for these circumstances how can I determine if the static FEA results warrant fatigue asessment or not?
Or perhaps the reverse approach would be the better one? To calculate the stress levels for which the structure achieves infinite life span right off the bat knowing it's being loaded this way and then just compare it to structural static results?

 

[GregLocock]

" designed for infinite life expectancy". Not happening in steel structures. Nothing can prevent grain boundaries moving in response to changing loads. The experiments indicating an endurance limit in steel were not sufficiently well analysed

https://www.geartechnology.com/ext/resources/issues/1105x/gigacycle.pdf?1633054982

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[SWComposites]

BeBtr - you need fatigue S-N curves for your material. Then given an R-ratio (min/max fatigue load) you can compare your static stresses to appropriate S-N curve to see where you fall re potential fatigue failure.

 

[IRstuff]

Interestingly, to GregLocck's point is a paper from 1999 "There is no infinite fatigue life in metallic materials" by C. Batthias, but infinite is in the eye of the beholder. Nevertheless, the attached illustrates SWComposites point:

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

 

[rb1957]

"infinite" life in steel is well known for decades (since before i was in uni)

maybe in '99 Batthias meant in Aluminium (ie in aerospace metallics, tho' yes aerospace uses steels) ?

the parallel concept of "crack growth threshold" (ie no crack growth below some value of stress intensity) is also pretty well known (these days).

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[geesaman.d]

In my line of work, we deal with continuously rotating components that operate for years/decades. So fatigue calculations (backed up by fatigue data) won't even clear our warranty periods.

There was a merger/buyout involving two companies who make essentially the same product, and it was interesting to see that one company used fatigue calculations and the other used static stress calcs and endurance limits. The fatigue life calculation method didn't generate better reliability in this comparison. Looking back at the history suggests the fatigue approach started because failures displayed fatigue behavior, not because the fatigue life was a required output. In fact, the problem would have been solved more simply by adding a suitable stress concentration factor to the existing static stress calculation and not the entire fatigue life method.

IMHO, seeing a fatigue fracture does not imply fatigue life calculations when you need 10^8 or greater cycle life. A static stress analysis with a reasonable endurance limit is effective.
Only perform a fatigue life calculation when 1) you have s-n data of the appropriate cycle count and 2) there is value to knowing a calculated fatigue life below the 'infinite' life criteria. Calculating fatigue life where infinite life is the effective requirement is both a distraction and an obfuscation. It's harder to optimize and quickly understand the design using an overly complicated stress theory.

Another factor is the noisiness of the load data. For a horizontal shaft with gravity loading, the alternating stresses are clear as a bell and the fatigue life calculation is straightforward. In other cases where load amplitude, direction, frequency, etc are varying, and a first-pass application of Miner's rule doesn't fully capture the richness of the data, the fatigue life method is probably out of reach and isn't adding anything.

That said, fatigue calculations make extensive use of stress concentration factors. Those are equally important for static and fatigue stress calculations.

As for your case, where in the FEA is the maximum alternating tensile stress? Feed your FEA numbers into a fatigue life analysis and see what that means to you. Your cycles are just into 10^7 territory so s-n data may exist.

 

[rb1957]

sorry, but this is going down a very familiar rabbit hole.

we start with a very innocent, general, question comparing the analytical approaches of two colleagues ... "I've got a very short question necessitated by wildly differing approaches ...".

Now its "In this case it's a heavy, high strength steel (650 MPa yield point) structure loaded in one direction ..." and "it can't fail, it's welded, it's definitely oversized". Seriously, did you really mean to write that !? "it can't fail" ??

But I know what you mean, it has a high MS against static loads. Ok, fine. But, as you're thinking, fatigue is different ... different structural concept, different analysis, different loads, different allowables, different everything from static.

And so someone's done a fatigue analysis. ok, great, why the question ? is the question "why did one guy do a fatigue analysis and another not ?" what are their job titles ? did they split the job up between themselves ?? Even if the two are working in isolation, the guy analyzing contact isn't also saying "hey, no problems with fatigue bro".

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[BeBtrEngeer]

Thanks to all the new contributors!

I actually remembered incorrectly and the peak stress was in the (350-530) MPa region, albeit quite local and far from welds, sorry about that.

@GregLocock I was referencing, as I assume, the ubiquitous understanding of the "infinite life" as exceeding the 10^6 - 10^7 number of cycles.

@geesaman.d Thank you for the insights.

@rb1957 I guess this could've been a controversial statement, although innocent in my view - "it can't fail" was supposed to mean that in case it fails it's potentially cathastrophic (as in loss of life).

The question comes from someone (me) who's new in the field - learning - and who's faced with potentially having to calculate similar things soon-ish. I've encountered two approaches by two colleagues (same job title, different times, different software, different factory) calculating very similar, albeit not exactly the same, structures. One responded with "there are no fatigue critical junctions", the other is gone. Hence the questions. I figure there's no harm in asking if someone's willing to take a minute to read and share some thoughts. Had I found a straight answer in literature (or colleagues) the thread wouldn't have been created.

And the structure description is just for the context in which it's easier to present the views. As I've said before - the question is general, not specific.

 

[Roman.S]

You're asking a quite general question and probably you won’t get a single "solid" answer, but I'll try to summarize the key points, where you would definitely need to do the fatigue analysis.

Fatigue calculations are required in cases where:

1) Cyclic Loading: the structure is subjected to repeated or cyclic loads, which can lead to progressive material deterioration over time. For example, structures such as aircraft, bridges, cranes and of course excavators(!) which you mentioned (or similar), are subjected to continuous loading and unloading cycles, which you do have in your case - "...structure loaded in one dirction a couple of houndred times a day with a normal force and a big bending moment...".

2) Standards and Codes: certain industry standards or codes may explicitly require fatigue analysis (e.g., Eurocode, ASME, API, etc.), just like in your example with excavator where you would need to use appropriate lifting appliance standard, e.g. EN12999/EN13000/13001 etc. These codes set limits on fatigue life and provide guidelines for structures that will undergo cyclic stresses.

3) Stress concentrations: if there are significant stress concentrations due to notches, welds, holes or connections, fatigue life may be significantly reduced (having repetitive loading). Fatigue calculations are necessary to account for these stress raisers.

4) Safety-critical structures: for structures where failure could lead to catastrophic consequences, such as in aerospace, energy, and transportation industries, fatigue calculations are mandated to ensure safety and structural integrity.

Btw, here are some nice articles about fatigue theory and usecase examples, which might be useful:

What is Fatigue? Definitions, Types, Causes

https://sdcverifier.com/structural-engineering-101/what-is-fatigue-definitions-types-causes/

Fatigue Strength and Limit: Understanding Materials-Specific Data

https://sdcverifier.com/structural-...-limit-understanding-materials-specific-data/

Fatigue Life: Key Influencing Factors and Advanced Prediction Methods

https://sdcverifier.com/structural-engineering-101/fatigue-life-structural-engineering/

Fatigue Stress and Its Role in Structural Failure

https://sdcverifier.com/structural-engineering-101/fatigue-stress-structural-failure/

Hope it helps a bit.