Weld Inspection of Pipe to Pipe - T, Y, & K Connections 2

[HereTwoLearn]

Question 1: What is the best approach to weld inspection for T Y & K connections (pipe to pipe fishmouth welds)? (For example: Maybe the prepared connection gets visually inspected prior to welding, and then the root pass gets magnetic particled, and then the fill and cap passes get ultrasonic tested)

Question 2: If the welds were all completed, and you wanted them inspected, what methods would you suggest?

(The welds are either CJP or PJP per AWS D1.1. The application is a typical, static, structure.)

Thanks for your opinion and input.

 

[metengr]

Visual for fit-up, wet MT or visual after the root is deposited and for the completed weld - wet MT or liquid PT.

 

[metengr]

I should have prefaced my response with... the code of construction NDT requirements must be followed. If none are specified, I would use what I mentioned above.

 

[Duwe6]

For inspection of existing welds, use UT in the manner developed by Am. Welding Soc. - AWS - for T-K-Y connections. See AWS D1.1

Other option is to hydrotest your welds to double or triple the Design Pressure. This will reveal [rather dramatically] design errors, in addition to gross weld defects.

Personally, I would perform the UT and follow it up [after any weld repairs engendered by the UT testing] with a heavy hydro.

 

[Ron]

I would visually inspect the fit-up and root pass. UT the final weld. Could use MT or PT for final weld, but risk inclusions that wouldn't be detected.

 

[HereTwoLearn]

Thank you for the responses.

 

[gtaw]

There is no mention of the diameter or the wall thickness of the material being fabricated. They are both worthy of consideration when determining what NDT method is to be used. The nature of the loads should also be a factor when determining if NDT other than VT is necessary.

Volumetric examination is required by D1.1 when certain conditions exist. If the welded connection is a CJP weld that is transverse to the primary stress, subject to fatigue and categorized as Category B or C, and the stress range includes excursions in tension, volumetric NDT is required.

MT is not considered to be a volumetric test method. It has limited "penetration" and will essentially detect linear discontinuities that are surface breaking or slightly subsurface. Small linear discontinuities that are slightly subsurface have a low probability of detection. I would not use wet magnetic particle testing if the weld is in-process. I would be concerned that the weld would have to cool to permit the wet magnetic particle test which means preheat would be lost. Another concern would be the potential for cracking when the root bead and the second layer are deposited and susceptible high unit stresses during the welding cycle and the cooling cycle before being subjected to wet MT. Dry MT would not require cooling between weld passes to the same extent wet MT would.

UT is not permitted by D1.1 unless the material is greater than 5/16 inch thick. While it is possible to test large diameter tubulars, smaller diameter tubulars are very difficult to test with UT. Appropriate procedures have to be developed and approved by the Owner's Engineer. UT on partial joint penetration welds is questionable at best.

Continuous VT can be employed and it will detect the majority of surface breaking discontinuities as each weld bead is examined. There are certain pipe service categories that permit the substitution of continuous VT in lieu of RT when working to ASME B31.X piping codes.


Best regards - Al

 

[connectegr]

I second gtaw's comment.

http://www.FerrellEngineering.com

 

[Duwe6]

Just because D1.1 thinks that UT suddenly goes blind at 5/16" thickness is a poor reason to refuse to use it to confirm the quality of your T-K-Y welds. UT is used on the fab welds of all offshore tubular drilling platforms. The industry started doing this when using D1.1 "exactly" [older ed., did not specifically address T-K-Y] allowed a couple of disasters to occur.

Yes, if you use the 5/8 x 3/4" 2.25 MHz "AWS" transducer, you cannot get decent results even on 3/8" flat plate. If a 3/8 or 1/4" round 5.0 MHz 'ducer is used, with an ASME-style DAC curve, I can now 'see' what needs to be seen. Come look over my shoulder anytime. I can prove my assertions.

As gtaw alludes to, waiting to perform MT/PT on a cold carbon steel weld root is a leading cause of root cracking. Don't do it - period. If T-K-Y joints are handled like regular open-root pipe welds, your problems are reduced to "damn few".
Bevel & fit-up is visually inspected AND SIGNED FOR by an experienced welding inspector [NOT the production foreman]
Preheat verification with a pyrometer is mandatory, overseen by a roving weld inspector.
All 'new' welders are 100% volumetrically inspected [UT in this case - geometry won't allow RT] for their first month
After 1-2 'cutouts' or 2-5 minor weld repairs per year, the offending welder is dismissed/demoted.
The welding inspector cannot be overridden by Production or Engineering [unless the engr. is also an experienced welding inspector, qualified to evaluate NDT reports, and WILLING TO SIGN THE ACCEPTANCE.

It is so simple that it is hard.

Very good-to-excellent welders make very good-to-excellent welds, especially with welding inspectors to 'keep them honest'. Cheap, inexperienced welders and a welding inspector that sleeps in his chair or will 'knuckle under' to Production will guarantee a significant quantity of serious weld defects. These cold facts have been documented for 60+ years in ASME boiler and vessel shops. It's simple.