4 over 1 Podium Overturning Analysis

[driftLimiter]

Hello fellow engineers.

Today I want to see what others are doing to address the overturning of a light frame wood building ontop of a concrete podium structure.

There are two approaches I have seen in the past.

Smeared - Take the global overturning of the upper structure, scale by R factors according to dual-stage analysis technique, then apply a uniform moment to the podium level.

Point Loads - Model actual hold down overturning forces in the location the are planned.

The elements I am going to be looking into are the podium shear walls, and foundations. The slab design will include point loads from the shear wall hold downs.

The goal of this is both for podium wall design, and foundation design forces. For podium walls I want to capture any loads opposite of compression because that will be worst case for the wall interaction.

Any other ideas? or Pros/Cons you guys can think of doing this either way I have proposed.

One major Con of Point Loads is having to go in and locate all the locations and apply appropriate loading. I would prefer the smeared technique but I'm worried that effects of overturning are more localized, (i.e. the slab isn't 'rigid' in the out of plane direction.

 

[BAretired]

I don't like the smeared technique unless it can be justified by the particular geometry of the wood structure. If the wood structure is designed to rest on soil using footings, individual attention is given to loads and footings; the same individual attention is needed when the structure bears on a podium.

 

[phamENG]

Smeared is good for preliminary analysis to determine your starting point. You'll have a good idea of what sort of overall structure and foundation you'll need. Then use well defined point and line loads to capture it in detail.

 

[driftLimiter]

Thanks for the input from you both. I understand the need for accuracy but I am starting to question whether the effect on the walls and foundations will be any different between the two methods.

The biggest difference I can see is that using the point loads will require a semi-rigid slab to be modeled so that the slab can distribute the overturning moments based on it's actual (modified) stiffness.

At the end of the day, the sum of all of the couple moments created by the hold down couples will be equal to the global overturning of the upper structure.

Obviously the slab flexural design needs to take into account the point loading from the hold downs, that is not the issue at hand. Its the wall reactions below and their reactions on the footings.

 

[milkshakelake]

I agree with BAretired and phamENG. Though honestly, there is a good argument for the smeared approach because either approach will have similar rebar for concrete podiums (unlike steel podiums, where it's a huge difference). I tend to use top and bottom mesh of rebar for podiums so the localized point loads in the middle of a span or off by some margin isn't a big deal. However, applying the actual loading would be more defensible in something like court.

In terms of practicality, a SAFE or ETABS model should make short work of it by adding the line and point loads.

 

[phamENG]

Most of it depends on geometry. If it's a really simply box with a corridor and party walls and a straightforward parking/commercial arrangement below, you're probably safe with smeared. But a lot of those things get pretty crazy. I did one a few years ago that was a 5/2.5 - z shaped with a courtyard, parking core in the base with commercial around the perimeter. Neat project, but what a mess. I'm just glad I didn't have to do the podium on that one. I do know the guy who did used precise point loads as I provided all of them to him in a really detailed spreadsheet for entry into whatever software he was using.

So as long as your lateral and overturning loads are sufficiently "smeared" already, then you're likely fine. Sort of like calling a series of joists at 24" on center a uniform load.

 

[KootK]

I would prefer the smeared technique but I'm worried that effects of overturning are more localized, (i.e. the slab isn't 'rigid' in the out of plane direction.

I think that's the crux of it right there. Very often, the shear wall overturning will be resisted by nearby columns and other vertically stiff elements operating over only whatever lever arm separates them (usually much less than the entire building width). This is robustly true even for regular corridor buildings which will tend to utilize the columns adjacent to the parking isles for overturning resistance. Seismic loads are small on most of my projects such that it's nearly unheard of for this to create net uplift on interior foundations. Not so so when I do it on Vancouver Island (seismic analog to WA state).

I can sometimes, sort of justify smearing for the gravity design of podium slab depending on the layout. I struggle to imagine a practical situation where I would support that for OT due to lateral loads.

 

[Eng16080]

I wouldn't use the smeared approach except for preliminary design. Point loads due to seismic overturning effects will possibly require the use of the overstrength factor (in addition to the R_upper/R_lower factor), which would be applied in designing the podium framing at these point load locations. I don't know how you'd account for that with the smeared approach. I've found that these amplified forces can sometimes control the design of podium members.

 

[jerseyshore]

We discussed this a bit in another thread last week I think, but around here, I don't know anyone who is analyzing point loads from holdowns on podiums, especially on concrete slabs. We're in a low seismic area so those holdowns just feel like a pinch to the 12-18" concrete podium slabs. We use ETABS, model the podium, then add the upper wood floors as dummy floors. Wind is applied by the program and we add in the mass loads at each wood floor for seismic analysis. Run the whole thing at whatever R value the lower level needs and go from there.

 

[driftLimiter]

I like the idea of using the dummy walls. TBH the hold downs don't really do much (if anything) to the podium flexural design with a 4 over 1. Obviously anchorage design and yes we need to account for overstrength, but Like I said in that other post, for the continuous threaded rod we are using 1.2 x Nsa rather than omega.

The point loads I will be inputting will all be 1.2 Nsa.

I gotta figure a way to do with with RAM and I am not good at RAM so its gonna be a bit of a pain.

 

[Eng16080]

That's fair, I suppose. I don't know where this particular project is located and what the seismic demands might happen to be. I had a project with light/moderate seismic loads (SDC C) where these effects weren't negligible.

Maybe I'm the only one using the overstrength factor in this situation. I don't know. It is required per ASCE 7 as best I can decipher, and I don't have a good justification not to follow the code there. Sure, the earthquake will never likely happen. Same goes for most loading conditions we design for.

 

[driftLimiter]

Its high seismic sds ~ 1.25. The thing on overstrength is, using 3 or even 2.5 is overkill. The through rods have a well known yielding behavior. The last podium I did I used overstrength and the hold down loads were still just a pinch like jerseyshore said. I know what ASCE7-16 12.3.3.3 says but I don't really see the point in designing the slab for a higher load than the anchorage of the tiedown.

 

[driftLimiter]

Eng16080 here's a link from the light frame master (Terry Malone).

He says plainly that he agrees with you the podium needs omega on the discontinuous shear walls.

I will take a look but I suspect it wont change the slab reinf much if any.

 

[jerseyshore]

I tried it in RAM also but it's not nearly as user friendly with concrete. Steel buildings RAM all the way. For the concrete ones we always use ETabs. But ram is pretty easy to do dummy floors as well. I do that all the time with various structures.

 

[Eng16080]

driftLimiter, thanks for the link. Mu understanding is that the intent is to protect the podium from the maximum load that could be transmitted to it from the LFRS above. That way, gravity framing members of the podium don't fail before the LFRS does. That's roughly what the code commentary states. Even though a hold-down connector might only be designed for 1,000 lbs, I assume the overstrength factor is estimating that it could really transmit 3,000 lbs to the podium before failing. If the connection is ductile, like you mention above, and would deform to absorb the seismic energy rather than transmitting it to the podium, it seems like that would also be a means to accomplish the same thing.

 

[driftLimiter]

Yep that's my understanding as well. I did a few spot checks and the 1.2 Nsa comes out very close to omega E so it's sensible.

 

[milkshakelake]

It's been a while since I used RAM, but I remember it being difficult to add point and line loads in specific locations and orientations. ETABS is much better for weird shapes, like shear walls on a podium. I do remember RAM has a dwg or dxf import option that made it much easier, though you do have to follow their line types exactly. This would be useful for making the dummy walls, or hand calculating a few critical cases and applying the same loads everywhere. The dummy wall idea is better than hand calculating the loads because you don't have to recalculate things by hand if the architect moves walls and supports. Hopefully this will reduce the burden and not make it a calculation nightmare with millions of spreadsheets.

For the dummy walls themselves, there's an approximate approach by modeling them as 1" concrete walls or something, releasing out-of-plane moments, and providing one-way slabs between them. Jambs/posts/columns can be modeled as pinned gravity columns, like an HSS. But the way I do it is by modeling them similar to CFS shear walls, with cross bracing. It doesn't have to be nonlinear; for these purposes, the cross bracing can take compression to speed up the analysis. Between them, I model studs or columns at 2'-0" on center. It's a lot easier to do than it sounds (though maybe not in RAM). It makes the holddown/compression point loading at ends more accurate than using a concrete shear wall, and still applies the gravity loading as uniform loads. If you don't want to go through all that, I still think modeling dummy walls as concrete is similar to your smeared approach but more accurate.

 

[driftLimiter]

Thanks for the modeling advice milkshakelake. Looks like I'm going to be RAMming my head into the wall for a few days here.

 

[driftLimiter]

I gotta figure a way to do with with RAM and I am not good at RAM so its gonna be a bit of a pain.

I take it back. Its a major pain.

Between all of the analysis and loading spreadsheets, and all of the different seperate applications in the RAM SS, this has been a nightmare.

My current workflow is modeling the podium in RAM with generic 'other' walls for the masonry.

Making dummy moment frames one story over the podium and adding the base shear and chord forces to each end from my hand analysis of the lightframe lfrs.

Then go to RAM frame, get the podium seismic loading, make load combinations, and run a lateral analysis.

Then manually extracting all of the wall forces and using RAM Elements to design each wall at a time.

After this I think I'm just gonna buy RISA 3d for the next building (even if my boss doesn't want to pay for it), I'm 100% certain I could have easily modeled up this podium with the applied loading much much faster.

The amount of clicking around in RAM SS is obscene, no spreadsheet entry should be a thing of the past. Then the error codes in RAM Modeler, then more in RAM frame, then more in RAM Concept.

All in all I am not a fan of this platform and I would not recommend it for podium analysis.

 

[jerseyshore]

Like I said above, I tried it in Ram but it was a disaster so we gave up. Ram is excellent for regular steel buildings. Once you start introducing different elements it gets wayyyyy too complicated. There's no reason Ram needs to be so difficult inputting loads and things like that. Lots of clicking for sure. Hell, it's even a pain to move around and zoom in & out.