Extent of the Weld Joint Reject of ASME B31.3 4

[wXp1223]

Hi All,

I have an inqury regarding the weld reject for the process piping during the fabricaton.

Is there any requirement in the ASME B31.3 where if one (1) joint gone through the NDE (RT) has been rejected, another two (2) additional joints shall be selected which is consider a penalty for the rejected joints.

This is the case if the process line only required random NDE.

Appreciate your feedback

 

[XL83NL]

Review para 341.3.4 on progressive sampling - it deals with what you seem to mean.

 

[weldstan]

And it can get worse - up to 100% of the welds within a "lot".

 

[wXp1223]

Thanks a lot guys.

 

[Duwe6]

On 100% RT jobs, mediocre welders are not particularly expensive. You were going to RT every weld, so the only additional work is the RT of the repair areas. On B31.3 5% RT jobs, mediocre welders can be VERY expensive. You will find that on 4" and smaller piping the RT costs more than the weld. [I assume that that cost is the reason that 5% was chosen over 100% when B31.3 was written]

 

[XL83NL]

Good point duwe6!

 

[weldstan]

The 5% RT requirement was and is a workmanship check. Progressive sampling punishes the Employer for poor workmanship of its welder employees.

 

[BigInch]

Punish a company that doesn't know Rule #1? Isn't it really that the company just receives another lesson in, "The cheapest way isn't necessarily the cheapest way." OK, I suppose for some companies it is the hardest lesson ever to learn.

 

[wXp1223]

Good points guys.

Just to share, common practices during my involvement of upstream project for the 100% RT apply to the process/critical line usually define by the engineering guys, however for the random@5% RT only for the utilities etc and most of the rejected weldment were from the random@5% because they put the least skilled welders to save a cost eventhough it is not save after repair.

 

[weldstan]

wXp1223,
Some managers never learn! Until they cease being managers.

 

[gtaw]

"Their pain is our profit. The more it hurts, the more they are willing to make the pain go away."

In this case the pain is the cost of removing the rejected welds, rewelding, and reinspecting the completed welds.

Best regards - Al

 

[moltenmetal]

Duwe6: you have a good point, but might be too subtle for some so it needs a little clarification. 100% RT is MUCH more expensive than 5% RT, particularly for small pipe, in a typical shop environment, for the reason you've mentioned and for others as well. It is also much slower. The only time 5% RT is more expensive than 100% RT is when a joint fails in service as a result of a flaw that could have been caught if RT was done- that can obviously be greatly more expensive in life and limb safety terms as well as financially to all the affected parties. 5% RT can also become more expensive than 100% RT if your fitters, welders or procedures are terrible- or the RT examination or interpretation is being done incompetently.

To a good welder, ANY random RT serves to motivate them to do better work for fear of failing RT. To a bad welder, 100% RT alone won't make them into a good one. And RT, like any examination method, isn't foolproof- it can miss things and can also generate false positives. You still need to do all the VT and other supervision, training, procedure and welder qualification etc. in an attempt to eliminate the flaws before they're produced rather than trying to find them after the fact.

 

[Duwe6]

what moltenmetal said: " 5% RT can also become more expensive than 100% RT if your fitters, welders or procedures are terrible- or the RT examination or interpretation is being done incompetently"

I've seen 'all of the above'. Had a 'mom & pop' mechanical company come to a refinery to install some pipe. They had so many rejects from every one of their welders that the job went to 100% RT, but in 'fits and jerks'. If the job had started out as 100% RT, it could've been done efficiently. Since the extra testing was being forced by the bad welds being found, the RT crew was shooting a couple and them processing the film. Then the rejects they found would force more RT the next day; again and again. Company went broke.

Seen a BUNCH of spurious RT rejects for Lack-of-Fusion, being called on good welders working on carbon steel or austenitic steel, welding 'stick' SMAW or 'tig' GTAW. When the 'defect' was being excavated, nothing could be found. Many, many times. Because LoF is such a major defect to leave in a pipe or vessel, the guy reading the film will stare at it until he thinks he sees something. Problem is, RT film is a collection of shadows; when you stare into shadows long enough, you will see whatever you are expecting to see. Ask any hunter or soldier.

So, if your project starts getting LoF rejects on easy-to-weld c/s or s/s, on welds made with stick or tig by your good welders, throw the "B.S." flag. Call in UT-shearwave [or phased array] to recheck the RT reject. RT is terrible at finding sidewall LoF, shearwave or phased array cannot miss it. Don't start cutting welds that may be good, recheck with an NDT method that works much better for LoF. And if you can't see the defect on the film, it probably is spurious. Reading film is not magic.

 

[gtaw]

No one NDE method will find everything. RT will not detect planer discontinuities that are perpendicular to the "beam of radiation." That type of discontinuity is easily detected with UT. However, if the discontinuity is parallel to the ultrasound "beam", it will often go undetected until the transduced is moved to redirect the beam nearly perpendicular to the discontinuity.

If I was limited to one test method, I would opt for UT.

Best regards - Al