Cross Frame Fatigue Check in MDX

[OSUCivlEng]

I am designing a skewed plate girder bridge with continuous cross frames (not staggered) in MDX using a the Plate & Eccentric Beam System Analysis. I am working on the design of the intermediate K shape cross frames and the sizes of the angle members are getting pretty big, L5x5x3/4, due to the check for fatigue stress. I turned off the fatigue check and changed the member sizes to L4x4x1/2 and they are good for strength.

I think the problem is is the fatigue category for the gusset plate to angle member weld. I selected Category E because the thickness of the connected members is less than 1 inch. I know there are a lot of bridges out there with cross frames composed of members much smaller than 5x5 angles that don't have a bunch of cracked welds. Has anyone else ran into a similar problem? I am contemplating ignoring fatigue for the intermediate cross frames.

 

[EQguys]

OSUCivilEng: I would suggest to check the ADTT value in MDX program if you havent already. I think 4000 trucks is the default setting which is a very high number. Depending on the type of highway, you may be able to reduce it, and thus reduce the number of fatigue cycles. I also will suggest to specify a FATRNG data item. This overrides the default category values with allowable base metal fatigue range. I hope this helps.

 

[bridgebuster]

OSU - I just finished listening to an AISC webinar on bracing. The speaker was Todd Helwig from University of Texas. He said commercial software tend to over estimated bracing stiffness, unless you use shell elements, which isn't cost effective, by 50 to 80%. He cited a bridge designed by TxDOT that had a "fatigue problem" in the frames. However, since the stiffness was overestimated it's unlikely there was a real problem. They ended up adding a girder anyway.

 

[TehMightyEngineer]

Just watched the same webinar and also thought to check back here. A very interesting webinar, some really surprising things that aren't explicitly checked by the bracing equations in appendix 6 in AISC such as system stability for double girders systems and similar.

Regarding OPs problem, it does sound almost exactly like that TxDOT bridge they mentioned. It's worth noting that the reason they found over-estimates of stiffness was due to the eccentricity in single-sided connections giving them the 50-80% over-estimate when comparing analytical models to line elements in canned software. Shell element FEA analysis included the eccentricity which allowed for almost perfect correlation between the FEA model and the analytical models. However, when you looked at connections without eccentricity (knife plate connections to tubes, double angle braces, etc.) they found that there was no over-estimation.

It's also worth noting that while over estimating the stiffness is overly conservative for fatigue, it's under conservative for actual bracing.

Try running your braces as double angles and see what you get for member size and fatigue. See if you can watch a copy of that webinar or find the research papers cited in it. I found the webinar very interesting and informative and will definitely watch out for more presentations from Mr. Helwig.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/