Greetings to all from beautiful Thailand.
Just starting on the $50 billion Gorgon LNG project and I am busily reviewing specifications.
Chevron are the client and project management is a joint venture - KBR, JGC, Hatch and Clough.
Hopefully someone can assist with my query.
We are fabricating CS, KCS, 304L,316L and Duplex piping for utilities.
PMI is called for on s/s piping and I can only assume it is to ensure we don't mix the 304,316 and duplex pipe / fittings etc.
The specification calls for 5% and 10% testing.
My question is what is the point of doing PMI if it is not 100% ?
I understand the premise behind random radiography - 5% may comply but 95% may have defects. That is why if you want defect free piping systems you stipulate 100% testing.
But hypothetically 19 pieces out of 20 (95%) of 304 in a duplex piping system - surely then you have a serious problem ?
Any thoughts,
Regards,
Shane
By Superflux 
Date 01-29-2011 08:52
Edited 01-29-2011 08:58
Is this for receiving of prefabbed spools, or joints of new pipe? On prefabbed parts it would be to make sure that the fabricator was conscientious enough not to mix and match. If the random sampling demonstrates complete adherence then it is "assumed" that all is good. Most likely the fabricators were also required to perform some PMI. Any deviations could give cause for further investigation if say for instance you PMI a piece that they did and come up with different numbers.
In the case of the latter, it is to ensure or develope a "faith" in the mill that they did their chemistry homework. Even if all the tests indicate compliance, if there are gross variations in the %'s , there might be cause to further the investigation and up the % to ensure the products integrity. If there are any rejects, then be assured that the owner or project manager is going to step in, and then you'll be on major overtime or hiring some more field QCs. As always, we hope for the best, because one non conformance means piles of extra paperwork.
As in all things, somewhere a bean counter has allotted a certain amount of funding for every phase of a project. Some projects allot the funding for 100% testing.
I have tentative plans to make it over to Thailand some time this year for some R&R.
Shane,
is that a Bechtel project in Map Ta Phut, Thailand? I have a friend working there for Bechtel.
If it is a Bechtel spec. then the reason is that they expect the field welding engineers to match CMTR's to heat numbers on material and then match chemistry against ASME or ASTM to confirm complience, and the random PMI is just a precautionary measure. I'm sure if you find non complient material then the requirement goes from 10% to 100%.
hope that helps
Jim
I am working on contract for Exxon right now and one of the things I do is PMI procedure audits for vendors. The Exxon specs call out either 100% PMI or random PMI depending on their view of how critical the component is.
My take is that any aspect of welding engineering is subject to something less than 100% overchecks. I mean, if you tell a welder the settings to use on a WPS what is to say he does use those settings? Compliance can only be assured through audits and audits by nature end up being less than a 100% overcheck. Same with PMI and material labeling.
Thanks for the replys gentlemen,
Not working for Bechtel Jim, working in Bangpakong - fabricating modules for shipment to site in Australia.
Regards,
Shane
Your understanding of 5% RT is all wrong....
Obviously NO owner is happy with only 1 out of 20 is in compliance with codes and/or spec.
A (any) random check is to verify procedures are in place, and actually working as intended.
One can only ASSUME that the remaining 95% are just as good as the 5% which has been checked, providing same welder has done the welds following the same procedure.
The above dont hold true when dealing with ASME - the Americans are just too stupid to realize that.
With the best wishes of a great new year from Shanghai
3.2
3.2,
Is this excerpt from B31.3 not what I described above ?
"Random or spot examination will not ensure a fabrication product
of a prescribed quality level throughout. Items not examined in a
lot of piping represented by such examination may contain defects
which further examination could disclose. Specifically, if all radiographically
disclosable weld defects must be eliminated from a lot
of piping, 100% radiographic examination must be specified."
If I follow all the correct procedures - review and accept MTR, check heat number on pipe, review and approve procedure for transferring heat numbers when cutting, witness random transferring of heat numbers etc there is still a possibility (however small) that 19 pieces of 304 pipe (out of 20) - have been installed in a duplex line.
Unless you witness 100% of every step of fabrication you cannot "guarantee" to the client that every s/s item in that piping is what it is supposed to be.
I have used 19 out of 20 purely as an example.
Surely one piece of 304 or 316 in a duplex line is a matter for concern.
How can you guarantee that one mistake has not been made unless you are performing 100% PMI testing of all pipe (fittings should not need it if they have appropriate identification.
Regards,
Shane
Shane,
PMI machines lie to, then what do you do? I think the best thing to do is develope a site specific procedure and get it approved by your client and follow it. For you it might be 100% PMI. For us its CMTR and heat number verification and random PMI. so far it's like novacain worked everytime we use it.
Jim
