It seems as per the article that they are (or were not) entirely forbidden. However see what Bruneau has to say...

"CONCLUSIONS
Partial Penetration Splices
A half penetration butt weld connecting two heavy jumbo
steel sections, and tested in pure bending, has been able to
develop and exceed its nominal design capacity. However, it
failed in a very brittle manner partly as a consequence of a
severe stress concentration created by the unwelded part of
the flange, as well as due to a series of concurrent
metallurgical and welding factors inherent to these type of
details when used for heavy steel section splices. Continuous
visual inspection during welding and ultrasonic testing of the
base metal and welds was not sufficient to avoid this
undesirable failure mode.
Where such welds are to be used, care must be taken to
ensure that (1) splices would not be expected to yield under
realistic loading conditions or (2) a brittle failure of the
splice would not impair the stability of the structure."

http://www.eng.buffalo.edu/~bruneau/AISC%20EJ%2019...

I personally would try to avoid the partial penetration flange splices. Ensure that all the combinations and fatigue, stress concentration, size effects etc are well represented in your factored forces -or ASD equivalents- prior to accept the detail.

A normative answer to your question reduces to examination of the ruling code respect welds, normally the weld chapter in AISC 360 and seismic spec ... I have not looked at them at the moment.

Why is he splicing the beam? If this is at a planned splice, then your details should show how the connection is to be made. If he is splicing short pieces together to make one beam, the question should be "Will I allow splicing when I have specified a one piece member"?

The reference that partial penetration welds of jumbo sections or thick material results in bending of the weld, is related to the size of the partial penetration weld specified. It is obvious if an E = 1/4 weld is the strength required for a 3" plate, a tensile force will result in bending at the weld root. A weld of sufficient thickness related to the material thicknes does not have the same effect. Even if a conservative partial penetration weld is select due to the material thickness, other advantages remain. Reduced NDT inspection requirements and backgouge/re-weld of the root pass is not required.

If designed correctly, a PJP weld can be used for a WF splice.

www.FerrellEngineering.com

connectegr - thanks, the reduced inspections and backgouge reqmts are exactly why he wants a pjp.

The beam is a 14x120 that only requires about M=175kip-ft (ASD). If I am not missing anything I just check the base metal and the weld based on table J2.5 (13th edition), with tension normal to weld being the controlling condition correct? I am using a 45 degree single bevel with an effective throat of 1/2" (5/8" - 1/8"). Using E70xx this gives me a weld strength of about 160kips for a M = 200 kip-ft, i.e. it works. Anything else I am missing or should be checking? Thanks

bookowski
The weld detail is (AWS) BTC-P4. Is this a field weld or shop splice? If field welded the bottom flange weld can be difficult in an overhead position. Welded from the top, consideration needs to be made for the area of the beam web and fillets.

www.FerrellEngineering.com

It is a field weld, however I think they can do both welds from above by flipping the beam at site. This is a beam that is being added on top of an existing roof top to support some new equipment. The splice is to allow them to get the beam through doors, up the stairs etc (no crane). Once on the roof however they will be able to weld, then roll it over and weld.

So it sounds like I'm not missing anything obvious in the calcs?

Another thought, since you mention mechanical equipment. Is the loading cyclical? If so, some additional consideration of the welds may be required. (fatigue)

Basically the calc is correct. Is there any shear force? This can be taken in the beam web. A bolted connection in the web may also help with field alignment.

Note that the root opening for the PJP is zero. A tolerance of 1/8 is allowed "as fit-up". This is another reason CJP welds are often used, they allow for increased field tolerances. Just want you to be aware.

www.FerrellEngineering.com

I misspoke slightly, it's actually a new swimming pool - so not cyclical. It's a pool being installed on an existing roof deck, the base structure is a flat plate concrete building. So loading is static, no fatigue issues. I am using a bolted web plate.

Thanks for you help connectegr.

If it is a site connection, and you are using a bolted connection of the webs, why aren't you using a fully bolted connection? Doesn't make sense to me.

There is a pool going on the top of the beams, we can't have any projection for a connection in this area.

Bookowski
Are you painting or galvanizing the steel? I recently consulted on a residential problem with a pool. The $4 million home had a pool built on a cliff side. The concrete patio and pool where supported by steel trusses, which received only primer. After 7 years the trusses deflected 8" as the top chord rusted away. Although the concrete deck was supported by galvanized decking, the salt water pool had no problem reaching the trusses and beams. Thick salt formations showed where the water leaked in and over the steel. Just another consideration.

www.FerrellEngineering.com

Thanks connectegr - It is not being galvanized. I am only responsible for designing the connections for the fabricator, I'm not the eor for the beams - so I dont' have much say. I had similar thoughts.

I wonder if the conractor has a guy qualified for full pen welds or not? I assume not and that is why they are calling for a partial. He could use backing bars and just weld it in the flat(rolling the beam to keep each weld in the flat position).

Just thinking out loud. Thats how the fab shop does it here. Less qualifications required.