Desperately Need Published Examples of Fatigue Checks For Built Up Members

[jmggks]

I need help finding published examples of fatigue checks for built up members.

We have an HSS16x8x3/8" serving as a crane runway. Using HSS in this application was a terrible idea ... no need to tell me that. The cranes run on polyurethane coated wheels which are not as wide as the flat surface of the tube. Local bending of the top surface of the tube across the 8" nominal width is causing longitudinal fatigue cracks in the top wall of the tube. The cranes are part of a fully automated industrial line. They cycle very quickly. Runways see 75,000 full load cycles per year and as many as 450,000 cycles per year at 75% of full load.

My company engaged a local AEC firm to consult on what to do. They are recommending replacement of the tube with an identical tube that has a 9"x1/2" cap plate stitch welded onto the top using flare bevel welds. The runways are continuous, so the top of the runway sees tension at supports. The tensile bending stress range adjacent to the proposed weld will be 7.6 ksi.

The outside firm is arguing that the tube and stitch welded cover plate will achieve unlimited fatigue life because they are sizing the weld such that there will be <2.6 ksi shear stress in the weld, and 2.6 ksi is the threshold stress for "coverplates wider than the flange" in AISC table A-3.1 item 3.6

My understanding, and please tell me if I am wrong, is that the bending stress in the member adjacent to the weld needs to be checked against the AISC fatigue parameters i.e. you don't just check the shear stress range in the weld, you check the stress range in the base metal (due to bending) adjacent to the weld. There is nothing in table A-3.1 that matches the proposed solution, but the threshold for intermittent fillet welds is 4.5 ksi (item 3.4), and the threshold for continuous CJP welds is 75% of the threshold for continuous fillet welds, so I would speculate that the threshold stress for intermittent CJP welds is on the order of 3.4 ksi. With bending stress in the tube of 7.6 ksi, we are way above the threshold value.

To support my argument, I found a 2nd quarter 2002 AISC Engineering Journal article where the author works an example and checks the bending stress in the runway per the AISC fatigue limits. I also found a paragraph in AISC Design Guide 7 "Cap Channels and Cap Plates" that is pretty negative about the whole idea of cap plates, and notes that intermittent welds mean that the base metal must be checked per stress category E (table A-3.1 item 3.4) whereas continuous welds are checked per stress category B (table A-3.1 item 3.1). It points to table A-3.1 item 8.2 for checking the shear stress in the weld.

My management does not believe me and is inclined to do what the outside firm is saying.

Can anyone point me to other published examples that would support my case that the bending stress range in the base metal must be compared to the AISC fatigue parameters?

Thanks for your time and consideration. I really appreciate it.

 

[Retrograde]

Might be of some use:

https://www.sciencedirect.com/science/article/pii/S2352012420301077?via=ihub

 

[r13]

AS the runways are continuous, they subject to stress reversal, so both the built up parts and the weld are to be checked for fatigue.
I personally don't like the idea of using stitch weld for built-up runways, continuous weld should be the minimum, especially for the lacking of smoothness of flare bevel weld, which may cause stress concentration and cracks. Is there any reason that a wide flange shape can't be used?

 

[SteelPE]

Good luck with management. It seems like you know what you are talking about. AISC DG7 is pretty much my go to document when designing industrial buildings with cranes. I have made the mistake before of proposing a stitch welded channel cap to a WF beam. I was lucky (ok.... I mistakenly over figured the design of the wheel loads for the runway with ended with the runway being over designed) in the fact the client decided to pass the design of the runway to the crane manufacturer. They came back with a design using a continuous fillet weld of a channel cap to a WF beam.

 

[jmggks]

Thanks to everyone for your replies. @retired13, replacement with a wide flange would be the right thing to do. In my original statement I simplified the situation a little bit. What they are actually proposing is weld repair of the cracked tubes before adding the cap plate ... which is another questionable aspect of this. I did not mention the weld repair because I wanted to limit the discussion to stitch welded cap plates.

 

[r13]

You may want to get a copy of the linked article. Link Another very useful document is CMAA Specification #74.

Another thought is, if the damage on the top surface is the only problem, why not just weld a thicker metal strip. Preferably the strip shall be narrower than the HSS, and can be welded to the flat surface.

 

[Agent666]

From my all but limited exposure to fatigue checks, I'll make the following comments on what has been outlined...

1. Agree you are checking the stress (or more particularly the stress range) in the base material and the weld. Both require individual assessment as the detail category for each will be different. As you've noted base material isn't exempt.

2. A cover plate is invariably one of the worst fatigue details as far as the detailing category goes (in my local code anyway)

3. Intermittent weld are invariably a worse detail category for fatigue than a continuous weld.

4. Weld across the ends of the cover plate are worse than the longitudinal welds.

5. That seems like a significant number of annual cycles. How long did the previous detail last? Does this back calculate out in terms of the current detail and fatigue life? Does the cover plate significantly alter this, I'm curious to know as it would seem like you'd need to reduce the stresses significantly to get an improvement if you have no control over the load cycles and you noted they are putting back the exact same hollow section?