AWS D1.2 crack/crater repair, how much needs ground out?

[CountryBoy19]

I'm being told by a Welding Engineer that in order to repair welds on a 6061 (T6) structure the entire HAZ must be removed. The structure is primarily 1/4" wall structural extrusions with 1/4 fillet welds and some bevel welds. The structure was supposed to be welded IAW AWS D1.2, cyclically loaded structures but it was not and quality control missed it. It was sent out and cracks developed in some craters that were left unfilled. To what extent does the weld need to be removed before repair welds can be made? If we remove 1/4" of parent material HAZ beyond all weld materials we won't have structure left to weld to. Is that right? I questioned the engineer on this recommendation and he said, "Yeah, you're better off starting over."

Welding engineer says at least 1/4" of parent material beyond all weld material has to be removed in order to remove entire HAZ. Note: he was not the original design engineer, just an engineer that was consulted on the repair. I personally think he is possibly being conservative or possibly misquoting standards/specs?
D1.2 doesn't specify.
Discussion on AWS forum seems to indicate 1/2" beyond all cracks or grind and verify crack is totally removed using dye penetrant or other inspection equipment and techniques.

Anybody here have thoughts? I'm likely going to run an FEA to determine the effects of the HAZ being left in place but I also want input from others.

 

[DVWE]

If it were me, I’d get an experienced metallurgist involved. T6 does have some unique properties that can be affected by welding, but I tend to agree with you that the welding engineer is being overly cautious.

Welding engineers (like me) know just enough about metallurgy to be dangerous…… or overly cautious.

The devil is in the details; she also wears prada.

 

[SWComposites]

So you are going to believe a FEM rather than a (presumably) experienced welding engineer? ok.

 

[r6155]

1) "Yes, you better start over." I agree
2) The 1/4" fillet weld is overdone.
3) Change the welder
Change the inspector
4) Make a mock-up for a repair.
5) Forget FEA.

Regards

 

[MintJulep]

The challenge with 6061 is that temperature and time at temperature and cooling rate all come into play.

You'll never be sure what the material has been through after the repair, so you'll never be sure what the material properties after repair are.

Starting over sure isn't wrong, and if you don't start over now there's a good chance that you'll need to start over later anyway.

 

[CountryBoy19]

Originally posted by SWComposites:
So you are going to believe a FEM rather than a (presumably) experienced welding engineer? ok.

No, not necessarily, and if my post led you to believe that I apologize. I don't think engineering decisions should be made blindly. The thought is that the FEA will give more data to make an educated decision, just as the opinion of the welding engineer does.

Originally posted by r6155:
1) "Yes, you better start over." I agree

Duly noted, do you mind expanding on this? Is there just too much risk that the HAZ has affected material properties so severely that the part is too weak? Is there a point where the it would be acceptable to repair? What if the original FOS was 10? 20? 50? Does weakened/damaged parent material still cause a concern if the original FOS was extremely high?

Originally posted by r6155:
2) The 1/4" fillet weld is overdone.

Are you saying the original specification of 1/4" was too large?

Originally posted by r6155:
3) Change the welder
3) Change the inspector

Unfortunately I don't have any control over that.

Originally posted by r6155:
4) Make a mock-up for a repair.

Can you explain more? Mock-up meaning what?

Originally posted by r6155:
5) Forget FEA.

So nothing valuable to be had from an FEA?

 

[r6155]

a) We do not know complete information. I assume the full penetration welding was done for cyclic loading.
b) fillet weld size is 3/16" for cyclic loading
c) FEA, for what? If you can't make a defect-free weld, there is no FEA that can help you. Wasting time
d) a mockup is a coupon with the same production conditions, including a defect simulated and tested accordingly.

I hope to be helpful

Regards

 

[CountryBoy19]

Originally posted by r6155:a) We do not know complete information. I assume the full penetration welding was done for cyclic loading.

A safer assumption would be an engineer not very experienced with weld engineering assuming oversize is better than undersized.

Originally posted by r6155:b) fillet weld size is 3/16" for cyclic loading

Just curious, but what dictates this? Is this a general rule? I don't see it in AWS D1.2.

Originally posted by r6155:c) FEA, for what? If you can't make a defect-free weld, there is no FEA that can help you. Wasting time

Point of clarification, I'm not the welder or the inspector, I'm also not the one that has the power to control who welds or inspects it (hence my previous comments). I'm just the engineer tasked with coming up with a fix to the problem. I do have access to the original FEA; running an FEA accounting for heat aging causing severe loss of material properties would give an idea of what to expect from rewelding. I guess I don't understand what the intial weld problems have to do with an FEA.

Originally posted by r6155:d) a mockup is a coupon with the same production conditions, including a defect simulated and tested accordingly.

Gotcha, I figured it was a coupon but wanted to be sure I understood fully.

Originally posted by r6155:I hope to be helpful

I do appreciate the help. I'm not a weld engineer, we don't currently have any weld engineers with capacity to work this problem so it's been assigned to me. The one weld engineer I did consult with took 2 months before he could even entertain the idea to meet up with me and look at the welds and give advice.

 

[gtaw]

Do you have a copy of D1.2 ? If not, buy a copy. Then read what it has to say about repairs. Crater cracks in 6061 isn't that unusual if the welder doesn't fill the crater completely. The repair typically entails removing the crack using a rotary carbide burr (buy several burrs intended for soft metals such as aluminum). Once the crack is removed, it is rewelded using the same WPS used for the original welds. The most common cause of crater cracks is underfilled craters. While 6061 is time at temperature sensitive, the sensitivity is associated with multiple pass welds on thick material. Single pass welds don't heat the base metal to the critical temperature range for a long enough period of time to be a concern. The exception is when wide weave beads are used. One must also recognize that the tensile strength is going to degrade by about 40% and yield about 50% under the best of conditions (per AWS D1.2 and ASME Section IX).

Best regards - Al

 

[dik]

Do you have a good detailed photograph that shows the condition you have?

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[kingnero]

What is the primary technical reason for requiring repair? I don't want to hear reference to standards.
- Crater cracks? gtaw has given correct advice. PT is your friend here.
- Fatigue cracking? The retained welding engineer has given conservative, but defendable advice. 1/4" might be a stretch, but a mock-up and a macro will give exact dimensions.
- Liquation cracking that has grown and surfaced? Wouldn't be the first time in less than optimal base material, combined with not negligible stresses. Start over with better base material.
- other? get to know the root cause first and share all circumstances and information. The original weld procedure would also be very helpful. I've seen structural welds from 0.1 to over 1 kJ/mm, with in general .2 - .4 as optimal range in most cases. And teach the welder(s) how to avoid end craters, how is this still possible in the age of digital power sources with all bells and whistles possible to avoid this. Or even without the aid of any technology.

 

[CountryBoy19]

Originally posted by gtaw:
Do you have a copy of D1.2 ? If not, buy a copy. Then read what it has to say about repairs. Crater cracks in 6061 isn't that unusual if the welder doesn't fill the crater completely.

Yes, I have a copy of the AWS standards. Thankfully my employer pays for access to a database for standards so I can get just about any standard I need access to.

95% of the cracks are crater cracks from unfilled craters. 3 of the cracks appear to be cracks that grew from crater cracks. There are other signs that the cause of the large cracks (that presumably grew from crater cracks) was an abuse/overloading condition. Those cracks opened a can of worms and they started doing DP on all the welds on all the weldments and started finding crater cracks. When it was sent to me for my thoughts I immediately noticed the welds were not AWS D1.2, cyclically loaded like the specifications on the drawing. I think the welders just got lazy. Carbide burrs/die-grinders are already on-hand. If grinding out the weld and rewelding is a viable option (even with reduced tensile and yield strengths) that's probably going to be cheaper and still meet the requirements.

Originally posted by gtaw:The most common cause of crater cracks is underfilled craters. While 6061 is time at temperature sensitive, the sensitivity is associated with multiple pass welds on thick material. Single pass welds don't heat the base metal to the critical temperature range for a long enough period of time to be a concern. The exception is when wide weave beads are used. One must also recognize that the tensile strength is going to degrade by about 40% and yield about 50% under the best of conditions (per AWS D1.2 and ASME Section IX).

Thank you for this! I understand strength will degrade, thank you for the help and insight!

 

[gtaw]

Happy to help

Best regards - Al