Live Load Reduction for Connections 3

[TidePoolJunkie]

Hi all,

What are engineers out there doing for long span situations where a girder is eligible for a live load reduction (floor live load, trib area>400sf--ASCE 7), but the trib area for the connection at either end of the girder is <400sf (ie 700sf for the girder, 350 sf for the connection).

Thanks,

Kyle

 

[haynewp]

There may be a case where less of a trib loaded area gives a worse case in the member/connection (think of skip loaded beams or full floor live on half a beam instead of reduced live on the entire beam). But that would affect both the beam and connections alike. I never thought that the code meant to treat/design connections for different loads than the member unless it is for something special like high seismic or blast.

 

[RobertHale]

I have to agree with everyone that is pro-influence area. I was fortunate to take a special topics course in graduate school with Bruce Ellingwood, where we got into the nitty-gritty of the statistics of live load reduction. Once he introduced the concept of influence area as opposed to tributary area, the "scales fell from my eyes" so to speak. The code even references the use of influence area to calculate the Kll factor in the commentary (along with a handy diagram and chart that derives the Kll factors in the code chapters). As it stands, if you look at my calcs, I identify and use influence area when calculating the allowable live load reduction factor. I would be more than willing to defend the use of influence area to any code official.
With regards to the use of reduced live load in connection design, I agree with the points made by KootK and PhamENG about considering the areas of the code where connections are required to be stronger or more ductile. I would also add that the live load reduction factor only applies to reducible live loads. Generally speaking, the magnitude of the reducible live loads (as specified by the code) is way out in the tail of the bell curve (see table C4-2). For more critical loadings, the connections can't be reduced because the loadings are ineligible for reduction. Just my two cents. My drawings require minimum connection strengths that I generate from the member end forces, and those forces absolutely consider the load reduction.

Robert Hale, PE, SE

 

[haynewp]

I think there may be a loaded area case that controls a connection while there is another case of a higher loaded area with a reduced live that controls the overall member. It's also interesting that ASCE 7 only uses trib areas for roof members and not influence. A 6ft long roof beam with 100 ft of trib per foot (extreme example case) gives a higher end reaction half loaded with 18 psf than fully loaded with the reduced 12 psf live. For floor beams, the connection and member have the same influence area for a load condition, but there may be a lower influence area that has a higher live load that controls the connection and that member end. That is what I was trying to say above. Outside of the mandatory skipped live loads, I doubt anyone is checking anything similar to a partial live load like that.

 

[driftLimiter]

The influence area for the connection remains the same in both of your examples. Not clear why the load magnitude would be different in each case.

 

[JoshPlumSE]

Hayenwp's post is more about tributary area than influence area. But, there is another aspect to it. There are different type of live loads. There's Roof Live load, Live Load, and "garage" or "heavy" Live Load (that only gets reduced when a member supports multiple floors.

There are cases where portions of the floor are subject to different types of live loads. Or, maybe subjected to an unreducible type of live load. So, the tributary area is less and the loading is less.... at least for each individual type of live load.

 

[phamENG]

haynewp - I see what you're getting at. I don't entirely agree, but you bring up an interesting point. Also, good observation about trib vs influence for roof live load reduction. That hadn't clicked before. The commentary is silent on why they left out a comparable K_LL there.

If I understand you correctly, you're saying that, in essence, the lower bound connection capacity for a beam using live load reduction should be the non reduced live load multiplied by the threshold tributary area for beginning live load reduction applied as a uniform load to the end of the beam commensurate with the trib width - right? I can see the logic there. I may have a 6 foot beam there, but I could have a 3 foot beam. And that 3 foot beam would actually have a higher roof live load reaction than the 6 foot beam.

But here's where I disagree: you've created a double standard. The beam shear and the reaction are the same. So if you're using reduced live load for your member size you're designing the beam for a shear capacity of 3.6kips. But then you're forcing your connection to be 4kips. While it doesn't make sense at this scale, if you factor it up by an order of magnitude you could easily get yourself into a situation where you have a beam that works (maybe it has a shear capacity of 38 kips) but then your connection calculations won't work because the you can't get your 40kips of shear resistance at your connection.

The other issue is that I think you're giving the statistical variations too much credit. The listed live loads tend to be localized maximums - it's why we get to reduce them (or most of them, anyway) for large influence areas. Those loads are defined to create a statistical reliability. While I've started delving into this in more detail (and discovering it's not as widely discussed/understood as it should be), I'm far from an expert. But my hunch is that these statistical reductions leave us with a viable lower bound. That is, the beam shears are accurate or better and the beam moments remain further to the conservative side. That hunch is reinforced by the fact that most reducible loads are based on a 50 year reference period, and live load reduction has been around for about 50 years now (according to the ASCE 7-10 commentary and my own arithmetic). Have you heard of any beam shear failures resulting from live load reduction underestimating the loading concentrated to one end?

 

[haynewp]

The live load being imposed on a beam doesn't know that the beam and connections have a certain influence area. I don't see why a live load can't be partially placed over an area as shown which would cause the maximum end reaction. Similar to skip loaded continuous beams. The "tributary" loaded area is less for a roof beam in the second case shown.

I see there being two different loaded area possibilities for the two examples, with the member and connection having the same influence area for each case (but the first case is twice the load TRIBUTARY area of the second). If we use only the first example, then we wouldn't be getting the worst case of a partial live placement. I used a roof beam in the examples for simplicity. To me it is similar checking a truss web member locally for full live load and then the full truss for reduced live.

 

[haynewp]

If I understand you correctly, you're saying that, in essence, the lower bound connection capacity for a beam using live load reduction should be the non reduced live load multiplied by the threshold tributary area for beginning live load reduction applied as a uniform load to the end of the beam commensurate with the trib width - right?

Yes.

Have you heard of any beam shear failures resulting from live load reduction underestimating the loading concentrated to one end?

No, I haven't. But a lot of the connections I have seen were more conservative than they needed to be, at least for steel. Not that it's a bad thing.

I don't think anyone is actually checking beams for what I am saying. I just see it as being a possible loading condition.

 

[phamENG]

I don't think anyone is actually checking beams for what I am saying. I just see it as being a possible loading condition.

I agree it's possible, but it's neither practical nor rational to design for all possible loading conditions. Unless somebody has some compelling evidence that live and roof live load reduction has been missing some high probability load configurations, I'm not going to lose sleep over not checking an infinitely variable load to loaded area combos for each beam and connection.

 

[haynewp]

I agree, but in reality I think floors are seeing partial patterns that we don't account for, but yes buildings still work. The code is somewhat silent on it, except for skip loaded continuous beams. It could be taken to more of an extreme like the below.

https://ascelibrary.org/doi/10.1061/(ASCE)SC.1943-5576.0000387

 

[TidePoolJunkie]

Amazing feedback! Thank you. I did not catch the influence vs tributary area differentiation in ASCE but now I understand.

 

[haynewp]

Found this in ASCE 7-16, which appears to put it back on us to account for it.

 

[phamENG]

which appears to put it back on us to account for it.

It is intended that the full intensity of the appropriately reduced live load over portions of the structure or member be considered, as well as a live load of the same intensity over the full length of the structure or member. (Emphasis mine)

While "full intensity of the appropriately reduced live load" first comes off as a little ambiguous, possibly meaning you need to look at a full 40psf live load for part of the span when LLR lets you take it down 25psf. But the commentary clears that up with my emphasized line above. If the interpretation I just mentioned is correct, this section of the code is completely negating live load reduction, because you have to consider the same load intensity over the full member.

This code section seems to apply more to continuous beam systems, geometrically unique systems (such as the mentioned arches) that are very sensitive to how they are loaded as they can sometimes be simultaneously stabilized and destabilized by different parts of the same uniform load. Not so much to simple beam systems and their connections.

 

[haynewp]

Yes it is saying the same intensity as if not reduced, but it also mentions partial length loads on a simple span beam having higher shear on that portion. I don't know how you get that without using a higher live load on the partial span.

It does seem to be saying otherwise to not use the tributary reduction for a partial case. That's where I still say that the live load doesn't "know" what the influence area or tributary area of the beam it is being applied to below. Back to my example above.

 

[phamENG]

but it also mentions partial length loads on a simple span beam having higher shear on that portion.

But it doesn't say which part. Look at mid-span. Full length load, shear is zero at midspan. For a partial length load the zero shear point shifts to the left or right and increases the shear on part of the beam. This doesn't really matter in most beams, but if you're designing around a web hole in a wide flange, repairing a notch in a wood girder, or sizing the webs of a truss, that difference becomes non-trivial. But it has no influence on the connection strength.

 

[271828]

Yes it is saying the same intensity as if not reduced, ...

I don't think that's what it's saying.

"The full intensity of the appropriately reduced live load applied only to a portion of a structure or member shall be accounted for if it produces a more unfavorable load effect than the same intensity applied over the full structure or member."

If they were saying what you're saying, then wouldn't they word it something like the following?

"The unreduced live load applied to a portion of the structure shall be accounted for if it produces a more unfavorable force effect than the reduced live load over the full structure or member."

 

[phamENG]

the live load doesn't "know" what the influence area or tributary area of the beam it is being applied to

I agree. But I would assert that it doesn't matter, because we don't design for "real" loads in most cases. We design for theoretical and statistically derived loading conditions that are loosely based on real world load surveys (where/when available) and that provide a sufficiently reliable structure when exposed to real and largely unpredictable loads.

Some exceptions: assembly areas where we know that exceeding 90psf is exceedingly uncomfortable to the vast majority of people, but cramming people and stuff into some rooms is a common occurrence. Crane loads. Design of specific pieces of equipment, etc. But none of these fall into the topic of live load reduction as they are some of the rare exceptions that prove the rule - we know what these are, but the others we're just guessing at anyway.

 

[driftLimiter]

Are you guys interpreting S4.3.3 to say that you need to use the unreduced load? It reads to me that the full intensity in this context is the reduced live load. I can't see how that interpretation jives with live load reduction in General like PhamENG said.

For me it helps to separate how I think about tributary area and influence area. Tributary area is a mathematical construct used with simple spans to make the statics really easy. Influence area is a physical region.

I guess what I am getting at is just because you decide to load a partial region of the beam, that doesn't change the influence area or the tributary area of that beam. Tributary loading is basically just for uniform distributed loads on simple spans. The partial loadings' statics don't work out the same.

I don't think I agree with increasing the live load based on some difference in the loaded area of the floor. The code says its based on influence area of the whole beam.

 

[271828]

Are you guys interpreting S4.3.3 to say that you need to use the unreduced load? It reads to me that the full intensity in this context is the reduced live load. I can't see how that interpretation jives with live load reduction in General like PhamENG said.

No, I think the text is saying "the full intensity of the reduced live load."

LOL - it's an example of using too many words. They should've just said "The full intensity of the appropriately reduced live load applied only to a portion of a structure or member shall be accounted for if it produces a more unfavorable load effect than the same intensity applied over the full structure or member." It's funny that someone thought it was a good idea to put the word "full" there. LOL

Like phamENG mentioned, there would be no such thing as live load reduction if this section was talking about the full live load.

 

[haynewp]

271828, yes I said it backwards. I agree it meant REDUCED live load applied to either the partial or full beam span which comes out to the same value (load intensity) for each case as clarified by the commentary. But I still say it's not logical to me to apply such a reduced live load over a small area of a large trib beam when we used the entire beam area for the live load reduction to be applied in either the partial or fully loaded case.