6061-T6 Aluminum Weld Crack Repair

[designrider]

I recently designed an aluminum structural frame for a 5’x160’ long pedestrian bridge using various sizes of extruded 6061-T6 aluminum sections. Most connections were aluminum welded. I specified 5356 filler wire and mainly 1/8” fillet welds (which fabrication upsized to 3/16-1/4” nearly everywhere). The fabrication shop performed a very sub-par job and failed inspection because of cracks in dozens of welds. The inspector called us to specify a repair plan; he thought most of the cracks were the result of inadequate base metal preparation and pre-heating. Does anyone have suggestions on the best repair option(s).

AWS D1.2 (Struct Weld Code-Aluminum) section 4.20 shows that the welds can either be repaired or removed and replaced. Removing the entire weld would be a cumbersome project for the amount of welds that are cracked. However, removal of the “unacceptable portion” is likely what needs to take place. Would specifying that the cracked weld portion be removed plus a distance 1/8" past end of crack, then replacing weld be adequate? I have also heard of stop-drilled holes to arrest the crack, but the size of hole recommended (3/4” or greater) seems excessive and a threat to the structural sections themselves, and likely more applicable to butt welds rather than fillet welds.

Where redundancy exists in my design, can I permit that some welds remain cracked, and essentially assume the cracked weld has become nothing more than ‘debris’ on the frame? Or does a cracked weld introduce problems regardless of its structural “need” (ie. a cracked weld is worse than no weld at all).

Finally, where a weld group exists, can a longitudinal crack in one weld propagate into an overlapping weld that runs perpendicular?

Thanks!

 

[metengr]

designrider;
Here is my advice considering the application, REMOVE all defective welds. Local defect removal is recommended, followed by Liquid PT to assure defect removal, re-weld and perform final Liquid PT.

 

[metengr]

designrider;
I would hope that the shop that performed this work at least had a qualfied welding procedure long with qualified welders? If not, find another repair vendor that can weld aluminum.

 

[designrider]

metengr;
Thank you for your feedback.

Yes, their certifications were checked prior to the start of fabrication. However, their professionalism and quality of construction left something to be desired. The weld sizes for the original fabrication did not require continuous inspection, but we are requiring such (from a third party) for all the repair work (at a cost to the fabricator). Ultimately, I am most interested in life safety, especially because this is a public sidewalk along a highway bridge. If they cannot produce acceptable results the second time around then the project will seek a new fabricator entirely, the fabricator only gets paid on completion of an accepted product.

 

[Ron]

I gather from your weld sizes that you are using relatively thin aluminum sections. Your fabricator also apparently used a GMAW welding process (you mentioned filler wire). GMAW in aluminum puts a lot of heat into the weld as compared to the more controllable GTAW process for aluminum.

Using thin sections, pre-heat and post-heat were probably not necessary, unless the shop environment was subject to significant temperature swings, such as can occur in an open shop during the winter. Tight jigging during fabrication can exacerbate the restraint of the members and increase cracking in welds.

I would grind the welds (be careful to use grinding wheels with no metallic components) and replace them completely. Further, I would require a qualified weld procedure specification and use the GTAW process with appropriate filler. You might also check hardness of the aluminum in the heat affected zone to see if you are getting unnecessary hardening or changes in the temper of the aluminum.

 

[designrider]

The base metal sections that show the most cracking are where 3/8" web stiffener plates are welded inside a C12x10.4 (which has a web thickness of 0.51") The welds are 3/16" at these locations and the fabrication did use a GMAW process. Base metal preparation/preheat was in suspect because some welds show a complete crack and detachment from the base metal of the thicker part joined along the side of the weld leg. Most cracks are hairline fractures 1/8-1/4" long at the weld ends and are propagating longitudinally down the center (throat) of the weld. However, I am open for comment on alternative origins of this problem.

The idea to "completely remove the welds" sounds excessive considering the amount of welds under consideration. For a scope...likely 100+ welds show a crack as described above. Welds are between 3-10" in length for a total of over 500" in welding. To remove these would be quite a project. Most cracked welds will never see their design stress capacity. They are mostly attachement welds to hold stiffener plates in place, etc. That is why I asked if some of the welds could be left in cracked condition if I (as the engineer) evaluate that they are not structurally needed. If localized repair is acceptable it would be prefered...but I do not retract my previous comments about life saftey being of primary importance.

 

[unclesyd]

Is the cracking on just one end or both ends of the weld?

I see a myriad of possibilities for your cracking problem as described. Usually when a weld cracks longitudinally in the throat this says that there isn't enough weld metal present or the filler metal strength is low. Other causes are filler wire is to small for the welding technique being used, there is not enough restraint holding the two components together while welding, the welder isn't depositing enough weld metal, in hurry. There a couple of others which I don't think apply to your case.
I wouldn't leave the cracks in place even though they might be insignificant when it comes to strength as this can open up large can of worms.
Repair would normally be to removed the cracked portion and reweld. You might want to look at just removing the cracked portion and leaving as is if the remaining weld is strong enough.
I would get the best Al welder and tell him what you are trying to do and discuss the repair procedure with him and get his input as there are several ways to approach the repair job.

 

[designrider]

CRACKED WELD REPAIR & INSPECTION:

1. WHERE ANY WELD CRACKS EXIST, THE DEFECTIVE WELD SHALL BE REMOVED AND EXCAVATED TO A DISTANCE NO LESS THAN 1/2" BEYOND THE DETECTABLE CRACK TIP TO ENSURE DAMAGED MATERIAL IS REMOVED. USE GRINDING WHEELS WITH NO METALLIC COMPONENTS. USE PT, RT OR OTHER NDT TO VERIFY CRACK REMOVAL. IF PT IS USED, ALL TRACES OF PENETRANT AND DEVELOPER SHALL BE REMOVED BEFORE ANY REWELDING IS BEGUN.

2A. IF THE REMAINING WELD IS 70% OF THE ORIGINAL WELD LENGTH OR GREATER AND BASE METAL LOSS DUE TO GRINDING IS LESS THAN 1/16”: PREPARE REMAINING WELD TO BE INSPECTED AS FINAL WELD PRODUCT. NO REWELDING REQUIRED.

2B. IF THE REMAINING WELD IS LESS THAN 70% OF THE ORIGINAL WELD LENGTH OR WHERE BASE METAL LOSS DUE TO GRINDING EXCEEDS 1/16”: WELD LENGTH REMOVED SHALL BE REPLACED BY A WELD SIZE MATCHING THE DESIGN PLANS OR SHALL MATCH THE DIMENSION OF THE REMAINING WELD (WHICHEVER IS GREATER). ALL REPLACEMENT WELDS SHALL BE IN CONFORMANCE WITH THE SPECIFIED WPS. USE FILLER TO MATCH EXISTING WELDS.

3. PERFORM A FINAL TEST USING PT, RT OR OTHER NDT TO VERIFY QUALITY OF REPAIRED WELD.

 

[designrider]

Unclesyd and Others;
Thanks for your comments. I appreciate your firmness regarding leaving any cracked weld in place. I had the same sentiments, but wanted to verify that I wasn’t excessive in my repair plan. I considered one scenario to be that in the exterior exposure, water leaks in at the weld cracks, freezes, expands, and puts unexpected strain on remaining uncracked welds. …regardless of the statistics involved with this actually occurring, it is a change I will not take.

You do present an interesting concept in just removing the defective area (+ a conservative amount past), then evaluating the remaining weld available. For this I propose the repair plan shown above. I am open for comment.

 

[metengr]

Looks acceptable with one addition - I would consider portable hardness testing to assure bulk properties of the 6061 T6 are, as specified, after all re-work is completed.

 

[designrider]

Yes, this would be an ideal way to verify the integrity of the effected structural sections. Does ASTM have an acceptable range of hardness for 6061-T6 welded aluminum? I know the mechanical properties of 6061-T6 show a target Brinell Hardness=95 and Rockwell B=60, but I assume they allow a variable range. Also, does welding alter hardness? If so, does the hardness have to remain in the same range of unwelded sections?

 

[Ron]

Welding changes the hardness of aluminum...sometimes up, sometimes down. Hardness does not have to remain same across weld and won't. The design should have accommodated a reduced allowable stress within 1 inch of the weld.