Inspection of GMAW fillet welds with limited history

By Algis Date 11-17-2017 15:29

We have an issue with fillet welds produced by a fabricator using 1/16" ER70S-6 wire without any recorded welding parameters (I am awaiting data on parameters they usually use). The material is 3/4" thick plate welded to 2" thick plate. Many areas have welds that are highly convex with significant overlap on one side (other areas are flatter). It appears that a short-circuit or globular mode of transfer was used in the convex areas. Needless to say, no qualification of short circuit welds were produced.

This particular shop produces heavy plate welds on a regular basis so I am surprised if short-circuit was used - is it even possible that short circuit mode would actually result in any fusion with such thick material?

A lot of welded material has been produced as part of a donation by the company to a project. I will reject any material that is not safe or non-compliant. I would also like to ensure that any safe and compliant welds are salvaged.

The heavy welded sections have marginally warped due to weld shrinkage - a clear visual indication that the welds have fused. MT shows "some indications" that there is lack of fusion according to our inspector, but this is unclear.

Is there any way to distinguish a deeply penetrating globular weld from a non-compliant and relatively unfused short circuit weld? There is not much spatter. Would a cross section need to be destructively cut and etched to reveal penetration? Can the welds be ground back instead to check on penetration?

I am an engineer, so my practical knowledge of welding is limited.

By Lawrence (*****)

Date 11-17-2017 16:59

A couple of thoughts.

"Needless to say, no qualification of short circuit welds were produced."    I will assume you are required to work within AWS D1.1 since that is the primary code that requires procedure qualification for Short Circuit GMAW.

"This particular shop produces heavy plate welds on a regular basis"    If a shop regularly welds on heavy plate and cannot produce welding procedure specifications WPS's, Their quality program and management must be called into question before looking at the first weld.

You report that the visual profiles of the welds are different........... If they are different on top, and there are no procedures to direct the welders, my first assumption would be that they might be inconsistent at the root and sidewalls as well.

Distortion does not equal acceptable/compliant fusion.

Donated work and material... That's a shame.   It's really too bad they did not send you in for a quick project review *before* production.

Fillet breaks or Cuts and Macros can verify individual workmanship, but I don't think that's what you are really after......

By Steelslinger (**)

Date 11-17-2017 17:20

To start, I'd have to know the project specifications and requirements.

Regardless, with no Welding Procedures and no Welder Performance Qualification Reports, I'd tag it as Non-Conforming Material and document everything (reports stating lack of WPSs, pictures of all the welds, etc) and send it to the Engineer of Record and let them decide what they want to do going forward.

In my opinion, the very least is get the parameters they did use (Volts, Amps, Travel Speed, Wire designation, Shielding gas, any preheating, etc), have them run the Procedure Qualifying Tests. Based off those results, select a decent representative sample of the work completed and have it destructively tested (cut and Macroetch). Then submit all that to the EOR for acceptance.

By Algis Date 11-17-2017 21:34

Steelslinger and Lawrence

Thank you very much for your replies.

Job is per D1.1.

I am awaiting their welding quality program documentation, but I am not too hopeful.

Unfortunately I am in another state and the local inspector was not consulted with the fabricator prior to production.

I can see this helping the fabricator in the long run to get their act right, but a pity that so much has been done prior this issue being highlighted.

By fschweighardt (***)

Date 11-29-2017 20:41

I never let guys go back and qualify procedures with parameters from the already finished product

"This is how we did it, now we gotta prove it to make a document", is a NOGO at my station.