My only observation is that thats a rather thick deposited weld thickness for one pass. The welding sequence if continious from one end of the joint to the other can also contribute to distortion. If no physical restriction is placed on the members of a tee joint they will distort.

A recommended practice is to balance the weld around the neutral axis of the member. In general practice on a weld the amount of heat buildup in a joint can be easier distributed with a backstep sequence and balanced welding. If an object is welded with no physical restraint the stress in the weld is low. If its restrained by balanced welding and physical restrained the stresses are locked into the weld joint (Not always a bad thing). The assembly procedure and the welding sequence could probably be modified to reduce the distortion but without knowing all the details it's hard to say.

I don't have much experience with box girders but distortion is what caused your problem and the ways to control distortion are by physically restraining the object and controlling the welding sequence/heat input.

There are books available from various sources including AWS that address distortion control specifically. I would recommend getting a few of those and thinking about some of the replies you get here. I think that there are some people on the forum with experience in this or maybe some welding engineers that may help.

I hope you get your problem resolved.


Gerald Austin

Weldinginspectionsvcs.com

Let's first look at possibly fixing the distortion experienced:

All distortion is caused by the area surrounding the weld becomming "shorter" than the material that is further away from the weld. There are many "rules of thumb" for removing distortion, but the bottom line is to try and shorten the weldment in the areas where it is now "too long". You therefore need to sit down and figure out where you need to "shorten" the girder to bring it in line with the welds. Typically this will be in the centre of the flanges. This can be done by "flame shinkage". (Heating this area in "spots" to around 600°C with an Oxy-Acetylene torch and cooling down again.)

If you have a "pure" twist, without any bowing, then the twist is most likely due to a "buckling" distortion. (The fact that the material is so long compared to the thickness, would suggest that this is the case.) This type of distortion is very difficult to control, because even if you constrain the girder during fabrication, it will just "spring" when you take off the constraint, resulting in exactly the same amount of twist. (This twist can be in any direction.) The best way to prevent this sort of twisting, is to either put as little welding in as possible:

1) Small welds, performed at high speeds.
2) Space welds, rather than continuous welds.

Or to design the stiffness into the girder in the first place:

1) Use thicker flanges.
2) Tack weld boxed sections into the girder before final welding.
3) Place stiffners or braces on the girder in strategic places.

Most welding handbooks will give you some guidance on preventing distortion, but if you make a study of the different texts, you will soon find out that the "quantitative" advice varies so widely that it becomes rather useless. None-the-less, you may want to look at "The Procedure Handbook Of Arc Welding" (Published by Lincoln Electric) for some guidance on preventing twisting distortion.

Hope this helps.

Niekie

Dear Jorge,
I would be most pleased of being of service to you. Let me know your postal address and I'll mail you a few pages of the book of the Course of Specialization for Welding Engineers of Petrobras (the Brazilian State Oil Company), which speak exactly on warpage of box girders. It's written in Portuguese, but this is not a problem for you (is very similar to Spanish).
Cordially
Giovanni S. Crisi
Sao Paulo - Brazil