Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations pierreick on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Let's Talk Fatigue in Steel Members (with pictures) 1

Status
Not open for further replies.

human909

Structural
Mar 19, 2018
2,072
Does anybody have some good references on understanding and dealing with fatigue failure in steel? As I understand it is exacerbated by cyclical stress (obviously) particularly high stress or alternating compression-tension stress.

I ask as I will be looking at this bottom flange & web crack.
temp_hhgr39.png


I believe it is in an area of hogging moment but likely with enough vibration for it to cycle through compression and tension. It is loaded by heavy vibratory equipment.
 
Replies continue below

Recommended for you

It requires stress fluctuating, the stress range, number of repetitions, and the detail being two of the primary inputs into a fatigue design. I'd say it's given more treatment in AISC for welding and welds than it is for the overall member. What you've photographed looks like some kind of tension fracture of the bottom flange for whatever reason.

If that crack in the web is progressing, it looks like the member will eventually shear off and "hang" on the severed piece below it.
 
The things you’ll have to do are the following.
Obtain an S-N curve for the material. It gives cycles to failure at different stress levels.
You need to generate a fatigue loading spectrum, from off condition, loading during standard usage and end of usage.
From your loading spectrum, calculate corresponding kt factored stresses.
The big step is to then conduct a rain-flow analysis. It basically generates the stress cycles and their corresponding cycle count.
Take into account surface finish and treatment conditions, as some treatments will reduce fatigue life (shot-peen and cold working increases fatigue life).
Final stage is to sum up the damages for all stress cycles, hoping the total is less than 1 (or any specified damage level based on aspects such as S-N curve used and probability required). Note, for each stress cycle, damage is usually “calculated cycles” over “cycles to failure” (Miner’s Rule).
Hope that gives some ideas!

As you are dealing with an existing failure, back-track to determine the point of initiation and, for fatigue purposes, calculate a kt factor at that location.
 
This is my goto doc for fatigue. It's under 200 pg and targeted to our space. It's not PhD level and it's not kindergarten. More like 8th grade which, I find, is about the sweet spot for most real world problems that I encounter. This allows me to sound like the smartest guy in most rooms when I need to speak to fatigues. And that, of course, is just what I want from a client confidence perspective. I have to manage the "brand" after all.

c01_cb7rpr.jpg
 
If you're looking at that crack, the first thing I'd do is jack the beam !
Then I'd cut out the fracture surface and look at it (start with an eyeball, go to an SEM as needed). When you see "beach marks" then you know you have fatigue. You'll also be able to see what initiated the whole damn thing. But I think you'll have little idea about the stress driving the crack. Maybe you can backtrack through the calcs ... for this crack to develop over this time (this many cycles) would've need a stress of X ?
Then repair the beam ... splice the web, separately splice the caps, do good fatigue design (lots of guidelines) ... others may think it's overkill.
Then install a sensor ... accelerometer for vibration, strain gauge (why not, can't hurt, may even get some good data ??).

Then you'll be able to review over the next couple of years to see what the loading really is, and make some sensible, learned, predictions.

But maybe the crack is from an overload, and you don't have fatigue, and all is "good" ?

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
 
There are two areas you can look into, one is fatigue, the other is damage tolerance (DT), which is basically crack growth. The initiation of crack growth is usually from a high kt factored critical detail, an example being a bolt hole. You have an initial flaw, which grows under cycle loading through the plate thickness, and then grows to a critical crack length. This type of analysis requires stress intensity factors, or SIFs, for each stage of crack growth (example of stages being i) through thickness, ii) if hole at free edge, growth to critical length towards free edge, and single crack tip growth away from free edge to final critical crack length). Critical crack lengths can be based on K1C or, from memory, R curves (moved on from F&DT). At the end of the day, crack growth usually needs more data input than fatigue calcs. Crack growth is usually done under fatigue cycle loading (non kt factored as SIFs are used). I would suggest looking at fatigue first.
 
The text Fracture and Fatigue Control in Structures by Barsom and Rolfe is a really good reference.

And I agree with rb and Stress_Eng above, need to determine the loading spectrum, then do a crack growth analysis. And examining the fracture surface is a good method to understand what is driving the crack growth. There are test labs with experience in this if needed.
 
Thank you all for your helpful suggestions and comments. Fatigue is something I haven't delved into with any depth. Industrial steel is my primary work so I should learn more. Thanks for the suggested books and documents.

KootK said:
More like 8th grade which, I find, is about the sweet spot for most real world problems that I encounter.
8th grade reading level with lots of pictures is what I want! [rednose]

KootK said:
This allows me to sound like the smartest guy in most rooms when I need to speak to fatigues. And that, of course, is just what I want from a client confidence perspective. I have to manage the "brand" after all.
That is what I want. I've done well for this client so far, I need to keep it that way.

CANPRO said:
this is great for some light reading on a Friday night:
That looks at least 10th grade reading level! I might have to pass Kootk's 8th grade first. But I now have both.

lexpatrie said:
What you've photographed looks like some kind of tension fracture of the bottom flange for whatever reason.
But my big question is why there? It is likely one of the lowest stressed areas as it near the support where it is welded onto the web of a column. To answer my own question maybe the crack initiated in the web or the weld at the column. (the web of the column is also cracked)

Ah... [smarty] I believe I've answered my own question by a clue in another grainy photograph supplied. This beam isn't a gravity beam, it is a horizontal member as part of the lateral support system. The system is x-braced compression members so is quite indeterminant. That area could be under quite high tension depending on the way the structure is loaded. (Also this structure supports some very heavy vibrating machinery so the cyclical cause doesn't need to be chased down.)

Additionally the member frames into a thin web of a large narrow flanged column without stiffening.
temp_jrfm9k.png



Sorry for the drip feed of the photographs. At this stage I don't have much to go on. I'll likely be attending site and investigating it myself. A good read of the material beforehand will help.

(I expect my recommendations will include additional members or stiffening. And a inspection plan. There is currently already a site patch in place performed by site maintenance with little or no engineering input.)
 
is that the same picture as originally posted ? looks different.

looks "odd" for fatigue crack, being (appearing to be) in the middle of the web ...

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
 
rb1957 said:
is that the same picture as originally posted ? looks different.

looks "odd" for fatigue crack, being (appearing to be) in the middle of the web ...

Same location, different member. On the other side of this web and this crack is the previously shown wide flange beam. It frames into the web of the narrow flange column. (Welded to the web of this column.)

The web would be under significant localised stress in this area. I'm unsure though if it initiated the crack or the crack propagated here from the previous photo. These are all the photos I have.
 
This is just going by the RHS picture in the original post, but it looks as though the crack initiated in the free edge of the bottom flange and grew into the web. With the shear loads in the web, it looks like the crack grew by the max principal stress opening the crack tip at what looks to be at an angle approx 45 degrees. The crack grew to the middle and then split into two tips, going up and down the web. Having two crack tips in the web, you would think a single crack tip would have initiated in the flange. I could be completely wrong!
 
Wow. Amazing HTURKAK. That is a great resource. Thanks.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor