Shrinkage in aluminum extrusion parts after welding.

[DrGonzo]

Hello,
This is my first post and I have checked for any related topics but don't find any that suit my need. Therefore, please forgive any redundancy.

The problem we are having is that after we weld our 6061-T6 extrusion framework we are getting unacceptable overall length dimensions due to weld shrinkage. It is box extrusion with a few features and a flange. The corners of the frame are 45 degree miters and are 360 welded and the outer faces are ground. We are using GMAW. The reason for 360 degree welding is due to design requirement for E3 (Electromagnetic shielding). The square framework (approx. 88"x108&quot has various square aluminum tubing welded to it creating, a grid work so to speak, for mounting and eventually becomes a bonded panel with 0.025" aluminum skin on both sides.

I am researching various solutions to retain our dimensional stabililty without inducing much cost. I am aware that other welding methods exist but we are limited to TIG or MIG.

We have thought about internal bracing with a TEE extrusion miter cut and welded so the base is actually the interior vertical wall on a 90 degree angle and the flange is outboard. Plug welding these in place and then welding the joint. I know this will increase my residual stresses, but I am not sure how this will affect the weld integrity.
Any help would be appreciated.
Thanks,
Erik

 

[Jeffro]

Hi Drgonzo,
put quite simply aluminium will always shrink no matter what you do, lowering heat input will minimise shrinkage but it can never be eliminated. The only thing you can do is balance weld to maintain uniformity of the shrinking, keep gaps to a minimum & allow for shrinkage when fabricating i.e. fabriate slightly bigger. As it sounds like you have already made at least 1 unit you should have some idea of how much shrinkage you are getting and therefore how much extra to allow.

I hope this will help
jeffro

 

[smcadman]

What you will have to do is find out how much shrinkage you are getting first with some test pieces. On some close-tolerance frames that we did, we found out that each weld would shrink a piece 1/2 milimeter. So if a piece is welded on both ends, we would have to allow for 1 whole millimeter. We did this by adding our "shrink" allowance to the length when cutting.
On a "T" shape, the horizontal piece would shrink less than the vertical piece. Our vertical is one that would shrink .5 millimeter. These welds were welded solid all the way around the piece. Find out what your shrinkage is, for butt joint and miter joints.
Also, some of our tapped holes on our frame would stay within tolerance, while others accross welds would lose tolerance because of shrinkage. After a couple of test runs we were able to determine which holes could be drilled and tapped on our CNC before welding, and those that couldn't would be drilled and tapped afterward. We made a couple of templates out of aluminum plate, and the holes in the template were were punched on our turret punch. After the welds were cooled, we would straigten our frame out within tolerance, and then clamp our template and well you get my drift.

Flores

 

[CWIC]

Heat input (60EI/V)would have to be consistent to determine how much extra length would be required to obtain your target dimension. The welder would have to weld the part using the same amps, volts and arc travel speed (WPS) to get any continuity with the results.

It is common practice to allow for shrinkage by adding to a dimensional property. Aerospace companies do everyday. There are also formulas to determine how much your part will deviate from the original-unwelded dimension.

ANSI/AWS D1.2-97, section 3.4: Control of Distortion and Shrinkage denotes several options for various aluminum alloys.