When a landlord uses a 'butcher' not a welder to modify a PEMB column 2

A landlord owns several large warehouses in my area, with long-term tenants. The landlord keeps the buildings in good shape, and is proactive in upgrading and maintaining the roofs, wall envelope etc.

He recently upgraded the roof drainage system, providing new 6" Ø PVC pipes to remove roof gutter water, and routed the vertical pipes through the warehouse and then laterally through the two perimeter internal PEMB tapered steel columns.

For some reason (money I assume) he hired the local 'butcher' and not the local experienced fabricator, who proceeded to use a 5" grinder to cut a 8.5" ø hole in the web of the column, so as to install a pipe-segment reinforcement which was to be welded to the web after installation.

The hack-job resulted in gaps of 1"+ around the scribed 'cut', as follows:

A welder he is NOT!

 

[dik]

I think the only repair is to remove the pipe section and grind the one side flat and to weld a new plate over the existing hole. The plate to the flange with a fillet weld and the hole area to the plate also with a fillet weld. the hole can then be cut using a circular hole saw drill and then weld a new pipe in place. The fabrication should be designed so that the strength of the PEMB member is not compromised.

Dik

 

[Ingenuity]

Follow-up: I got to visit the project up-close last week, and was able to determine that the web thickness is 5/32" (4mm), flanges are 1/4" thick x 6" wide. Interesting that the web/flange weld is continuous to one side only - no weld whatsoever to opposite face - is this typical for web-tapered/fabricated members for PEMB's, assuming small magnitude of loads/actions?

 

[dik]

Quite common weld since they don't have to 'flip' the assembly to weld the other side and it can be set up automatically.

Dik

 

[KootK]

Really? That seem like a flange stability nightmare. Any tendency for the flange to rotate would put the welds in pure bending w/o any other alternatives.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Lomarandil]

If the web is only 5/32, the web itself just past the weld isn't much better (even if you were to weld both sides).

Granted, you theoretically have a "crack" starting right at the root of the weld, but maybe practically you end up with enough incidental weld penetration to mitigate that.

----
The name is a long story -- just call me Lo.

 

[KootK]

Poor man's partial pen?

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Ingenuity]

dik: Yes, I think you are correct.

KootK: Yes, bending about the long't axis of the weld was my concern.

Lo: I guess when you only have 5/32" of thickness to work with.

Just looks 'off-putting' - but not as 'off' as the web hole

 

[TLHS]

Koot

The single sided welding issue is addressed in Newman's "Metal Building Systems Design and Specification" in the section "Some Contention Issues of Design and Fabrication."

The general thrust seems to be that some people don't like it, but testing seems to show that it's generally okay with an exception of a specific seismic case.

"According to a test program performed by Prof. Thomas M. Murray, single sided welds do not reduce the ultimate structural capacity of the primary frames, except in the end-plate connections where seismic loading is involved. The simulation of cyclic seismic forces in the test program produced repeated local buckling, which resulted in fracture of the single sided welds in the frame rafters near the end plates. Some feel that single sided welding may be acceptable for static loads but not for frames subjected to lateral forces, concentrated loading or fatigue, where double sided welds should be used. Naturally, most rigid frames must resist both gravity and lateral loads."

Since the description of the item that was seeing failure is confusing, I chased the references. See this letter from Dr. Murray, which gives more context (first letter on the page):

https://www.aisc.org/globalassets/modern-steel/archives/1997/11/1997v11_news.pdf

His letter is in response to this article,

https://www.aisc.org/globalassets/m...97/05/1997v05_fabricators_metal_buildings.pdf

 

[Ingenuity]

TLHS: Nice research work. I shall read those references. Thanks.

 

[JoshPlumSE]

TLHS: I'm going to disagree slightly here about the single sided welds. There isn't a lot of code literature (or research) that penalizes them right now. But, there have been some efforts (related to single sided welds for HSS connections) which show that these welds do not exhibit regular increase in strength when they're put into tension. REference the AISC design guide on HSS connections.

They're beginning to realize that this may not just be an issue for HSS shapes, it's just that not many people have really tested them outside of HSS connections. Not saying that I know what's going to come of all this. I just want to add a caveat about this type of weld and why "some people don't like them" as you say. The theory (as KootK suggests) is that bending / torque about the long axis of the weld may be a problem when the connection is in tension.

Now, this connection one-sided connection weld may not be bad because it may be pure shear. So, I'm not wholly disagreeing with your comments. Just offering some caution about this type of weld.

 

[canwesteng]

I don't know that I'd trust testing to show that the single side weld is okay - your fusion face is well past the joint depending on the quality put into the weldment, so you are getting some degree of PJP here. I can see that working ok in testing, but I wouldn't want to rely on it in practice since it's hard to quantify and technically a better welder will do the weld faster and leave less fusion into the joint.

 

[Ingenuity]

The single sided welding issue is addressed in Newman's "Metal Building Systems Design and Specification" in the section "Some Contention Issues of Design and Fabrication."

I have Newman's 1997 edition and could not find this section - what edition is this in, TLHS?

 

[ajh1]

Most PEMB web/flange welds are done as a submerged arc process (semi-automatic weld equipment). This results in a fillet weld that has relatively deep penetration into the web effecting as noted above a somewhat PJP weld. Years ago AISC actually allowed a reduction of 1/16" of actual weld throat for such a weld to recognize the penetration. That allowance has now been removed, although specific testing in regard to sectioning samples produced by the weld procedure could allow someone to take the penetration into account. Dr. Joe Yura at U. Texas has tested both small PEMB beams as well as some bridge box girders that utilized single sided welds and found no issues with their use. The one area where we mandate a double sided weld would be for a 3-plate fabricated crane beam assembly due to the fatigue issues. The fabricated crane beam is a rarity unless requiring such for a long span beam (40-50') in a special situation.

 

[TLHS]

Josh,

The discussion was, as I understood, about flange to web welds. I would agree that welds that see significant tension should be on both sides. So, barring research to the contrary, if there's a point tension load on the lower flange, I would expect that the eccentricity of the weld would result in a strength loss from the resulting moment. For example, I'd never weld a shear plate on the column side on just one side.

Personally, I also wouldn't weld a web to a flange on just one side, but I have trouble calling it bad practice.

Ingenuity,

My copy is the second edition (2003), so perhaps yours doesn't include it? In mine, it's at the end of Chapter 16 "Avoiding Construction Problems"

 

[canwesteng]

I think it's bad practice to count on the fusion zone to extend pass the heel of the fillet, and I think without it the single weld wouldn't do so well in testing.

 

[dik]

A few comments:

Welding from one side for thin material is not unsafe. The weld is substantially subjected to shear and the distances are usually far enough apart to resist any flexural moment issue. For thicker web members, the issue becomes a little more murky.

There is economy in only welding from one side, albeit small, but if you manufacture a kazillion of them, the savings adds up.

Unless I have tension issues, I detail base plates so that the fillet weld can be made without 'flipping' the column; the fillet weld is on one side only. Again, a small savings, but, a savings nonetheless.

In the second paper, there was a discussion about how a fabricator can be more competitive with PEMB's... I don't think it's possible since a PEMB manufacturer designs everything to the limit for cost effectiveness. If a fabricator were to take the same approach he would become a PEMB manufacturer and no longer a fabricator.

I think of PEMB's as 'throwaway' buildings, and after 25 or so years are good for scrap (I know there are a lot older). They have limited use except for an initial enclosure. I work with a lot of industrial buildings, and it is not easy to modify them or analyse them. I generally advise the client to stay away from them.

Dik

 

[KootK]

@TLHS: thanks indeed for the information and for taking the time to educate me/us on this topic. It's reminiscent of a recent thread dealing with insufficient welding in the built up beams of a special moment frame. There, seismic steel codes specify -- and research supports -- that full penetration welding is required adjacent to the moment connections but is not required elsewhere along the span. Kinda seems like the amped up, inelastic rendition of this same concept. I do like it when stuff ties together.

While Murray's research findings are very convincing, one of Nawrocki's salient points was that of the flanges getting knocked away altogether under forklift impact. That still strikes me as a valid criticism. That said, perhaps once you're down to to a 5/16" web and a 1/4" flange, you've got that same problem regardless of how you prosecute the welding.

While I've been somewhat skeptical of this particular PEMB practice here as a result of my own ignorance, I should clarify that I hold the PEMB industry, and it's more skilled practitioners, in high regard. As with any field, the worst of them are terrible. But the best of them are pretty much steel design maestros in my experience. And their product generally does a great job of meeting the requirements of the clients, irrespective of the ire of the renovators that sometimes come later.

@Ingenuity: for your viewing pleasure... It is from the second edition should you be interested in acquiring a copy.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Ingenuity]

My copy is the second edition (2003), so perhaps yours doesn't include it? In mine, it's at the end of Chapter 16 C

Thanks. My 1997 Edition has the chapter "Avoiding Construction Problems" but no reference to single-sided welding.

@Ingenuity: for your viewing pleasure... It is from the second edition should you be interested in acquiring a copy.

Thanks. I notice there is now a 3rd Edition. Added to my wish-list.

perhaps once you're down to to a 5/16" web

I wish it was 5/16"...this one is only 5/32"!

A Cat 980 loader is a 'beast' in the hands of an operator.

 

[KootK]

I suppose it would be odd to have a flange thinner than the web.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[dik]

KootK... high shear? <G>

Dik