Repair of fine weld cracks in aluminium frame 2

[wjj15]

Can anyone advise on the possibility of repairing cracks in aluminium welds without the need to grind the weld back to remove the crack first?

I have been presented with a proposal stating that cracks on my project are small/shallow enough that in the process of welding over the existing weld the crack will simply be fused together, thus no need for preparation of the area aside from ensuring cleanliness. I have a small number of cracks both in 1/8" fillet welds of aluminium square hollow sections, and adjacent to fillet welds in the SHS itself, which has 1/8" wall thickness and seems to have caught the arc of the weld away from the weld area - there is a burn indication and localised very fine crack. Locations of defects are far from each other so are unrelated.

Obviously this saves time as there is no dye penetrant testing, grinding back etc. to prove that the crack is gone, but may lead to the risk that the new weld is made and that an amount of the crack remains, buried, and uninspectable.

Thanks in advance!

 

[Ron]

What is your application? What type of frame? Is the frame subjected to repetitive loading? What alloy is being used? How long before cracks appeared? Do you know if the alloy specified was actually provided? What welding process was used and what welding process is proposed for the repair?

What are the stress levels? Was there a reduction taken in allowable stress for the area within 1 inch of the weld?

Your question raises more questions than answers can be given without further information.

Yes, you can repair some cracks and yes they can sometimes just be welded, but you need to know the "WHY?" before you get to a solution. Keep in mind your welder only sees the cracks for a few minutes while he's welding them. He likely doesn't understand the engineering issues behind them. You have to live with the repair for a much longer time than he. Repairing cracks in aluminum takes more consideration than in steel.

 

[wjj15]

I have just started on a new project (in the UK) so am not completely on top of all the specific construction history as I would like to be. The structure is aluminium, welded at all joints, and makes up a gantry which hangs from a bridge so is subject to vibration with traffic, but very little applied live load. The material is 6082-T6 as outlined in the following link:

http://www.matweb.com/search/datasheet.aspx?matguid=fad29be6e64d4e95a241690f1f6e1eb7&ckck=1

The defects do not appear to be caused by overstress of the joints; none of the joints in question are particularly highly stressed. All cracks are fine - visible to the eye but not very obvious.

We have a crack from root to face on the end surface of a fillet weld. This was the starting end of the weld and there appears to be a lack of root fusion at this point - perhaps due to inadequate preheat?

We also have cracks in the parent metal where the weld rod appears to have struck the face of the member as the weld was being made/completed, leaving a 'dish' in the surface. There is a crack across this dished zone where the wall thickness has obviously been reduced.

Proposed repair is GTAW process, from ambient temperature with max. prescribed interpass temperature 250 degrees C, approx 150 Amps, 23 Volts.

Hope this provides a little more, as to the specifics of the present weld material it would be fairly safe to assume it is the same tungsten filler as the repair. The construction team were fairly thorough so we are certain that materials specified have been implemented.

 

[Ron]

From your description, it would appear that the cracks are related to fatigue. Traffic vibration can produce high cycle, low stress conditions. With the inherent reduction in strength at and near the welds, there could be some stress concentrations going on as well as metallurgical issues from the arc strikes as noted.

I'm confused by your "tungsten filler" statement. Tungsten is the GTAW electrode, but is not consumable. The filler for this material is probably a 4043 series filler rod.

I would sample a small area of cracking and have a metallurgical evaluation done. With this the probable cracking source can usually be determined.

 

[btrueblood]

If the wrong filler rod is used (low silicon content), the odds of porosity formation and cracking in the weld puddle are greatly increased. Subsequent fatigue could increase the cracks from there. Ron is correct that the right filler is 4043 for this alloy, but who's to say if the right filler was used? Get a met analysis done on a sample of the weld, and go from there.

 

[wjj15]

Thanks all, plenty of recommendations to go on with here.

 

[racookpe1978]

You've got extreme liability issues here (hanging loads from your platform up overhead - apparently with those people !!!! and their cargo/tools over a highway with other people below.

Vibration of the load, the platform, and from the suspending structure (the bridge ?) that is itself being moved adds to your problem -> You can't control the actual vibratio levels, and need to approximate them and assume their size and frequency.

Don't skimp by assuming that the welder will "melt everything" back together somehow as he tries to melt a cracked blob of old - uncleaned !!!! - and oxidized metal. Make him grind the old cracks, dye pen test the ground out crack to verify the crack has been removed and has not propagated further by the grinding itself, reweld (root and cover probably) and repeat the NDE to make sure the new weld has not cracked, nor that the base metal has not cracked from the heat..

 

[Ron]

I agree also with racookpe1978 except that I would not guess at the vibration frequency or magnitude. It is too easy to test for these and will give you data upon which you may act.

If your issue is high frequency low amplitude vibration, the spectral effects are cumulative and can cause fatigue cracking when you wouldn't particularly expect it. I've seen this particularly in structures subject to aeolian vibration, where any single measure of frequency and amplitude caused no issue but the cumulative effects caused a problem.

 

[ColonelSanders83]

I would be very concerned about this issue. You have aluminum loaded in a fatuige condition. The primary properties of the base material is that it has no clearly defined yield point. Furthermore all of the material has suffered some amount of fatuige damage, repairing the cracks will only lead to the structure cracking at another location within very short order. You have a significant liability issue here especially since you don't have a good characterization of the vibrations present in your part.

If this part fails and someone is injured the fact that you repaired the visible fatuiged aluminum will be a major point. Furthermore, how do you know the cracks are shallow? How can you garuntee that the overweld will fully fuse the root of the crack if you arn't %100 positive how deep it goes.

Personally I would remove a cracked unit from service and perform some destructive testing, find out how bad it really is compared to a new design. if forced into a repair I would require full excavation of the crack and I would require it pass a penetrant dye test, and even then I would be very uneasy.

A question properly stated is a problem half solved.

Always remember, free advice is worth exactly what you pay for it!

http://www.ap-dynamics.ab.ca/

 

[Guest102023]

I won't assume anything about the mechanism, even though fatigue seems obvious from superficial consideation, but I can say that the proposed repair procedure is a very poor one.

(The person who suggested it should be be served with a restraining order to keep 100 yards clear of any industrial joining operations or even to comment on them from a distance.)

I'm not familiar with the specific alloy mentioned, but is the filler metal undermatched strength-wise?

 

[wjj15]

Thanks for the help everyone.

We have had a couple of weeks of investigation, consideration, repair and reinvestigation. We have a large number of joints in the structure and it is a relatively small number (statistically) showing defects. In all cases we have been able to attribute the cause of any cracks to workmanship - ie lack of fusion in one case, a few crater cracks caused by arc-strike to parent metal next to the weld, blowholes, undercut etc.

Having had an apparent explanation for why each defect was present, we checked this explanation with an independent expert who agreed in all cases. We then commenced dye-penetration testing of all defect locations, ground back to remove any cracks, retested then welded in accordance with the original procedure for the joints. The welder making the new welds was the best of the crew of welders originally on the job. Note some contract workers were employed originally too, so it may be that they came to our job for only a short time and then left, this may be their legacy!

We have also consulted with the designer for peace of mind - no fatigue issues here as many of these joints have very low utilisation, and the structure is still young (ie less than two years old).

Still, we have made note of all locations and will monitor over time.

Thanks again!