Multi-Story Pattern Live w/ Lateral

[haynewp]

In multi-story buildings, we of course design concrete beams for skipped live loads for gravity combination. Now, looking at the overall building under seismic or wind, I know high-rise engineers that routinely ignore pattern live loadings in combination with lateral.

I can't find anywhere in ACI or IBC that allows you to ignore skipping the live load on the same floor level or disregarding patterning it on or alternate levels in combination with lateral. I don't think that ETABS or RAM will even combine skipped live with lateral which may be part of the reason they are ignoring it as well.

Do you think this is being overly conservative if I combine the effects of pattern live loads with lateral for a typical office building?

 

[271828]

I've never heard of anybody skipping those along with lateral loads, but I don't know of an "out" anywhere. My personal judgment is that it's uberoverkill (LOL) because of the low chance of the skipped load happening during the lateral event.

 

[haynewp]

uberoverkill?

There are some levels with storage and mech room live (100 psf) that other levels do not have. I don't know.

What if it were a multi-level parking deck where there were a much greater chance of pattern loading during seismic? Still overkill?

 

[WillisV]

I also believe it is overkill for almost any normal situation. In my opinion, skipping live load is mainly for localized continuous beam type conditions to catch possible moment reversals etc.

I think it would be quite rediculous to try to checkerboard live loadings all over the building while simultaniously performing a lateral analysis especially considering the live load being designed for is probably 4 times what will every be there and the seismic load being checked is being determined based on a largely arbitrary R factor.

 

[UcfSE]

The only mention that I know of is in ACI 318 13.7.6.2, that for the given live load amount you may omit checking the pattern. That doesn't have anything to do with pattern live loads in combination with lateral loads, as you asked.

It does seem that patterning a live load that is already high to meet a certain probability of exceedance, with a load pattern that is unlikely to actually occur, is overkill.

Without intending to take the thread off topic, why would you consider the inelastic reponse modification factor to be arbitrary, Willis?

 

[haynewp]

OK, so if the consensus is that it is overkill (which I tend to agree), do you think it would be found acceptable in court for an engineer to decide that since the code does not specifically require a certain effect, that it can be ignored if he feels it to be overly conservative (through my judgment as an engineer, though this is not mentioned as an out by the code/standard)?

 

[271828]

I started to type uberkill, but settled on uberoverkill.

One immediate problem I see with a catmatic reading of the pattern loading code language is practicality.

I'm not sure it's even possible to pattern load the structure to ensure the absolute worst load on every element, from a roof beam moment connection down to each anchor rod. That might be a thousand load combinations for all I know.

Also, how do second-order effects come into all this? If I put the live load everywhere, that would cause second-order effects to be larger than if I leave off some LL.

In 3D, do I need to leave off live load strategically to make the most severe centroid location, simultaneously considering pattern live load, FOR every little element of the structure?!?

LOL, it's a good question, but I think this subject is best left "not thought about" any deeper than folks usually do.

 

[haynewp]

"For determining column, wall, and beam moments and shears caused by gravity loads, the code permits the use of a model limited to the beams in the level considered and the columns above and below that level."

Reading through ACI R8.9, I remember my concrete professor saying you can take the gravity moments from a model that considers the columns directly below and above the floor being analyzed as fixed and then add these gravity moments to the moments found from a separate full lateral model that is full height or approximated by portal method etc..

271828:

If I used this allowed simplified gravity model, then I can automatically disregard a full height live load pattern analysis (with or without lateral included) as we have all seen in structural analysis books to get column moments. I would only be left with the decision to combine the worst skipped moments from the gravity model at any one level with the overall lateral moments. Which I agree would still be an overkill in itself.

 

[JAE]

Good comments above. I feel that I tend to agree with the idea that alternating live loads WITH wind/seismic analyses is overkill.

And I'd like to take this opportunity to vent on our revered academians out there who write the codes. This is a prime example of how individual code provisions, when combined, create a exhausting complexity that most engineers must resolve to either ignore, or fake, in their designs.

As 271828 mentions above: "That might be a thousand load combinations..." to meet the code. This is correct.

To meet the technical intent of the code we'd have some sort of patterned load combination on every bay, odd/even bays, adjacent bays, in both directions for each floor. And technically, there are patterned combinations for each story (odd/even/adjacent stories) that would have to combine with the odd/even/adjacent bays.

My gripe is with the apparent ignorance of the code writers in understanding that their silence on this issue within the code is appalling.

 

[haynewp]

JAE,

Forgetting about the lateral part and only thinking about dead+live:

For concrete buildings, I don't think you would have to check story to story pattern loads since you can use a single floor model per ACI 8.9. But for steel buildings where there are moment frames all over the place, I think you would have to check all patterns on the same floor and floor to floor patterns since there is no single floor model assumption allowed, correct?

 

[UcfSE]

There may be cases of pattern loading producing a net uplift on a column due to dead and live. I expect you would use that on every floor to find the total net tension on your column and footing. That may be up to judgment since you are permitted to omit floor-to-floor under the conditions the ACI 318 specifies. For steel I know of no such clause either.

 

[271828]

hayne, say I have a 5 story steel moment frame, 4 bays wide. I'd really like to know how many load combinations would be needed to ensure that the following load effects were individually maximized:

1. Base plate moments and axial loads. Rememeber that some skip loading patterns will have bigger moments, but smaller axial loads. Haven't started talking about shears yet...
2. Positive moment on every beam.
3. Negative moment on every beam.
4. Unbalanced moment at every beam-column connection (panel zone check).
5. Every beam shear, both ends.
6. Column moments.

I'm sure there are others.

I'd be willing to bet $2 (LOL) that it would take hundreds if not thousands of LCs to ensure that each of those was maximized individually. It would really be interesting to start drawing influence lines of such a frame and see how the LCs start piling up.

The floor-by-floor approach seems more reasonable to me.

 

[civilperson]

Maximum seismic loading occurs with maximum live loading. Any skip pattern will DECREASE the seismic load compared to full loading. Other loadings may be maximum with a pattern load, but not seismic.

 

[271828]

civilperson, what if the moments from live load are much larger than moments from seismic load?

 

[haynewp]

271828

If I had to analyze a 5-story, 4x4 bay all moment frame steel building, I honestly would not know what to do about all this pattern checking. I was kind of hoping someone on here has done this before and could tell us what simplifying assumptions they used. (I assume we are only talking about gravity only now and ignoring the skipped live combined with lateral stuff.)

 

[JAE]

civilperson - not sure where you're coming from. Seismic load is produced by dead loads (and only some specific portions of live loads like storage). So the max. live loading doesn't really affect most structure's seismic response.

haynewp - yep - I agree with the ACI "fix the far end of the columns" provision so that technically floor-to-floor pattern loading for concrete structures can probably be ignored....but again, if you're analyzing a full moment frame for a multi-story building, I get so tempted to consider it due to its extreme influence on column moments.

 

[271828]

haynewp, if you're really interested, you could sit down and start drawing influence line shapes. After doing a few of them, you might see the same patterns happening again and again. It only takes a couple of minutes to draw one of these shapes for a given load effect for a frame that size.

 

[JKW05]

I miss DL+LL+WL with 1/3 stress increase... Were there really building failures using that combination?