shrinkage of large girders with stiffeners

Hi, I'm looking for a small formulae or some rules of thumb to predict the longitudinal shrinkage of large girders (mostly for bridges). We use only fillet weld. I know that shrinkage can be produce by the web-flange welds, the flange-stiffener welds and the web-stiffeners welds. All the formulae given in Blodgett or in Lincoln are not representative of our experiences. It always shrinks a lot less than what those formulae tells us.

For the web-stiffener welds, I have to be able to draw, on the flanges, all the places were the stiffener goes before anything is welded.

thanks

 

[ENC]

Can you describe in detail the experiences of shrinkage that you have? Under what conditions does this occur?

 

Actually, the hard part is mostly the web-stiffenenr welds. For example, if there's 15 stiffeners on the girder, welding the first one will make the girder shrink just a little bit, so the second draw will be a bit before it is suppose to be. This shrinkage times 15 makes the 15th draw to be at a completly wrong place, and the girder is smaller.

If a could calculate the shrikage made by one fillet weld, I'd multiply this by 4 to have the shrinkage of one stiffeners. With that, I'll be able to draw the lines were the stiffeners goes, a little bit in front of were it would be without the shrinkage. (for example,if one weld cause .35mm of shrinkage, one stiffeners cause 1.4mm. So the first one will be in front by 0.7mm, the second one will be in front by 2.1mm and the 15th one will be in front bye 20.3mm. The all girder will be smaller by 21mm)

What can I say more... we mostly use SAW on 350AT steal, usually the preheating T° is 100°F, all the other parameters always changes, so all those parameters should be in the formulae.

I hope this help. I know I'm not to good in english but if there's some other questions, don't be shy, I'll try my best to answer it since I'll do anythng to solve my problem.

thanks

 

[GRoberts]

It may or may not apply directly to your situation, but you might want to check out WRC bulletin 455 "Recent Progress in Analysis of Residual Welding Stresses – Report 1: Modeling of Weld Residual Stresses and Distortion: Computational Procedures and Applications"

http://www.forengineers.org/pvrc/wrcpvrcbulletin.htm

 

Thanks, but I'm not sure this question worth 65$... Anyway, I'm just a trainee and I'm not able to spend some money on my project. Any free advice would be great...

 

[welding1]

The best way to control distortion is to reduce the amount of welding; use stitch weld instead of continues ones, if possible and use smaller sized welds where possible.

The problem with predicting distortion is that it is dependant on so many variables and it will occur in all three dimensions, causing twisting as well as shrinkage. Try to position joints near the neutral axis of the fabrication and try and balance the welding; wherever possible form items rather than weld plates together. Also try putting assemblies back to back so that the distortion in one pulls against the other.

I think the easiest solution is to embark upon your own research, do some trials on scrap and see what sort of distortion levels you get.

Regards

John

http://www.gowelding.com