Question about UT testing of CJP weld

[OSUCivlEng]

I recieved an RFI for a steel frame I designed that is constructed of HSS. This frame holds salt spreaders that go in the back of dump trucks. The fabricator (who is also the erector)wants to shop build the frame in two pieces, and them ship them to the site. I think this is a good idea and have no problem with it. He will have to get permits and all that, it's his problem.

The fabricator is proposing a CJP weld at each of the 18 columns. This is fine with me as well. He has submitted several examples that demonstrate his experience in constructing similar HSS frames with field welded joints like this. He has also submitted his AWS prequalified welding procedure for the joint and the qualifications of his QC/QA personnel.

My only concern is that he states the CJP weld will not be able to be UT'd because of the minimum material thickness of 5/16". The wall thickness of the HSS is 1/4". The fabricator's QA/Safety and Quality Management Systems Director has offered to be onsite and visually inspect each pass and produce an inspection report with photos of each splice, weld layer and finish. This person is Class A-D inspector as well as trainer. Additionally the 3rd party welding inspector for the job would be on site as well for the duration.

I am not sure if this is good enough or not because my understanding is that all CJP welds have to receive more than just a visual inspection. Is there another kind of NDT that can be done for this weld?

 

[holdpt]

Assuming you are fabricating to D1.1, UT testing can be done per Annex S which gives guidelines for less than 5/16" and more than 8". It will be a separate procedure that can be requested from the NDT contractor. I would be curious how he is making a CJP weld. Is he using backing bars at the joints? Most likely not able to perform a backweld due to inaccessibility. D1.1 prequalified joints do not recognize one-sided CJP welds without some type of backing.
Anything beyond VT should be listed on the drawings or other contract documents.

Just my two cents.

 

[OSUCivlEng]

Yes he is using backing bars at the joint. Actually I think the way he's doing it is fairly innovative. The backing bars actually taper down to act as an erection guide to line up the column. I have a attached a picture of how this fabricator has done it before.

The structural steel specifications state that CJP welds have to be inspected with UT, radiographic testing or magnetic particle inspection. I'm not sure how common the latter two methods are.

Are many welding inspectors familiar with Annex S?

 

[holdpt]

You don't normally see UT, RT, or MT. Commonly it is MT, and then UT or RT. When given the choice, we UT in order to not lose production time by shutting down an area.
If an inspector has ever requested NDT procedures from whomever is performing the work, then I would say yes, he would be familiar.

 

[gtaw]

D1.1 only requires volumetric examination when the connection is loaded in tension and subject to fatigue. Even then, the fatigue category must be of a specific type. In short, the project specification is stringent to say the least.

There is no reason not to UT the connections, but as mentioned, the procedure must be acceptable to the Engineer.

If I were involved, I would suggest the UT be conducted in accordance with JIS standards. I've used the JIS for examining rail car frames. It is different than AWS, but AWS isn't the answer in every case. JIS is based on a DAC curve that is constructed using a standard calibration block consisting of a plate with drilled hole. Easy, quick, but not exactly like AWS.

Best regards - Al

 

[OSUCivlEng]

I confess I wrote the specifications. However the only planned CJP weld was on a moment connection on a canopy that is subject to fatigue due to reverse wind loading.

Perhaps the ultrasonic testing is overkill for this splice. I will have to see how much tension would actually occur at the splice location.

I don't think introducing Japanese Industrial Standards would be a wise idea at this point.

Thanks for your help.

 

[DekDee]

OSU,
Firstly, a piece of advice - never write UT, RT or MT in a specification as you are only going to get one thing from the contractor - MT.
UT & RT are volumetric whilst MT is a surface method examination - big difference in costs and more importantly the chances of finding internal defects in the weld.
It should be UT or RT and MT or PT (if stainless steel)
I have been involved with structural fabrication (in the Southern Hemisphere) for over 30 years and every time a contractor wants to splice something not shown on the drawings it must be examined by UT or RT and MT or PT.
There is no reason why these connections couldn't be RTed by the contractor (other than the cost due to access) but that is your call as the designer based on the criticallity of the joint.
Basically, if it falls down then it comes back to you,
Cheers,
DD

 

[Duwe6]

Your NDT contractor is probably not familiar with JIS-spec ultrasound. You might try giving the option of JIS or ASME for the procedure, and an acceptance criteria from D1.1 - Static table. Further, the UT tech has to be very familiar with testing of joints with backing bars. A large number of structural welds get rejected on UT because the tech saw something 'funny and linear' at the root of the weld, and then fell back on the NDT mantra of "When in doubt, throw it out".

RT of backing bar welds has the same problem; excellent welds rejected due to too much doubt.