Floor Vibration Analysis - Engineering Judgement

[southard2]

So I'm designing a simple two story classroom with CMU Corridor walls and CMU exterior walls (see attached pdf) The walls are typically spaced at 36-8.67-36 spacing. So over the classrooms I could easily use a 28LH770/300 bar joist to support a DL =94 psf and a LL = 60 psf. OK so a fairly conservative design. Using a composite slab 5" thick with 1 1/2" deck. And if I go into the Floor Vibe software and use WALLS for my girder spans and say set by girder span to 25 ft. And the overall floor length of say 120 feet or any other largely large value everything works fine. I'm assuming an church type assembly with .02 dampening. And I do have the ability to put in a few full height partition walls. So with typically that adds .02 to .05 additional dampening. If I use just .05 of additional dampening this joist size works fine. So I'm going to use full height partitions at the storage rooms between the classrooms. And with just .025 dampening it works. And even without any partition wall dampening it is within the bounds of engineering judgement. So with any additional damping it is OK.

HOWEVER the corridor walls aren't exactly straight. At the end of each classroom they jog outward a bit. And this reduces the joist span to 30 feet for about 15 feet or so. Storage rooms are located between the classrooms and this job is associated with that and allows students to enter the classrooms from an alcove that is created instead of directly from the classroom. Now if I take the same exact set up as above and reduce the Floor length to 25 feet (now matching my girder length which remember are walls with infinite stiffness) the floor vibration analysis fails badly. And since there is no direct dampening in the 25 foot width classroom the only solution is to increase the joist loadings. Which is what I did on two other schools with this exact design. And they've been in service now for a few years no complaints.

But I can't help but feel this was an overly conservative interpretation of the intent and costly. The software and associated guides have language that states that the maximum floor length is the length of the building however intervening columns or non-typical bays can reduce the floor length. If I limit my floor length to 25 feet it fails. So from the example layout the corner bay would not work unless I assume some of the adjacent spaces will vibrate with it.

Also as an experiment I change the joist spacing from 5 ft on center to 6 ft on center. The deck design is good up to 6 feet so why not. It helps a little bit in all cases since it more closely matches the overall floor joist loading with the loading the stiffness requirements for floor vibrations would want. But ultimately the floor vibration design is still driven by the assumed floor length. If I'm conservative and assume a floor length between the wall undulations it is much shorter and needs heavier joists. If I assume that 6 foot inward undulation will not restrict the vibration from moving outward and beyond then it is fine.

I suppose this is always the issue with empirical design methodologies. Because the testing is limited but the real world situations can vary vastly.

I'd just like to hear other various engineer's opinions on how they have handled situations similar to this. And I get these are opinions. I remember when designing the original buildings I actually called up Dr. Murray the author of the methodology and the software. He was very kind. Sent me extra books with the software when I upgraded my license. But even he was hesitant to give guidance and I understand that. So I just want opinions. I know that is all they are.

John Southard, M.S., P.E.

https://www.pdhlibrary.com/

 

[driftLimiter]

Floor Vibe and Dr. Murray is basically the only tool/criteria we have to address this problem. If its simply a matter of joist stiffness to get adequate performance I would lean towards just making it work by floor vibes analysis. Then at very least you can say that you did the industry standard to avoid these serviceability issues and more than likely wont ever have to hear about the floor vibration problem.

Adding mass would be my second go to. I know its not intuitive for us to think that way but it helps the vibration performance.

 

[JoshPlumSE]

Floor Vibe and Dr. Murray is basically the only tool/criteria we have to address this problem.

Well, the 2nd edition of the design guide does give alternate methods of evaluating a floor system. Specifically intended for cases where Dr Murray's traditional method doesn't work very well. Like when you have an area with cantilever beams or stairs or such. This new section was based on Brad Davis' work. So, you could check out that section of the design guide or some of Dr. Davis' published articles.

Plus, there are other programs (ETABS, RISAFloor, et cetera) that also do similar calculations per Design Guide 11 that might be able to handle this situation as well.

Caveat: I work for CSi (the company that sells SAP2000 and ETABS). So, I am not exactly an unbiased observer.

 

[JLNJ]

I have an office building design on the horizon and I'm already fretting over the floor vibrations. I've seen the architect's first pass and it seems that we will be in that 30 to 35 foot danger zone for floor vibration. So I feel your pain.

I have designed a couple dozen schools, most of which were hollowcore slabs. More mass, fewer vibration concerns.

Does Dr. Murray address bridging? I was wondering if I could add x-bridging to give the floor more two-way action as it tries to vibrate. It seems like this type of bridging might have a chance to activate more mass and help decrease the period.

 

[bookowski]

Nothing technical to help here but I have a related anecdote. Several years ago I worked on a charter school that had 2 full court bball gyms over the library, it just worked out that way and there was no changing it. They were of course concerned with acoustics and there was an acoustician on the project (very large high profile one as well). Structural steel and slab on deck. We could not come near his recommendations, the steel started getting silly as we chased stiffness and the client didn't have the budget. We even added reinforcement in the slab over the columns to design for partial fixity at the ends (aisc has a design guide on this), tried modeling bridging between beams, etc. We did fem models, not just the murray approach. Eventually the client said to just leave it with reasonable steel and they'd have to deal with it as is. After it was built they had the acoustician test it and it was completely fine, not even close.

 

[southard2]

Yes, it is just very frustrating. Because they at least the parts of the literature I've read it doesn't really explain what they mean by "non typical bays" Like what exactly is the threshold for that. The SJI technical bulletin #5 just says "The distance perpendicular to the span of the joists over which the structural framing are identical or nearly identical within adjacent bays". Not worried about the girders cause walls are walls. My joist spacing is very uniform. The only difference is those length changes. The And honestly how many buildings these days are actually constructed where all the bays are uniform. I'd venture to say less than 10% perhaps lower have uniform bays. I still suspect that even if overly expensive the ease with which bar joists are set makes them worth it even if designed for insanely high loading. Perhaps in the future I'll steer the architect to using hollow core. This building would be pretty suitable for hollow core to be honest.

Thanks for the input everyone. I'll try taking a look at some of Brad Davis's work.

In the mean time I'm not going to venture far off the previous design since I know the conservative approach worked. And honestly I was surprised but I didn't get complaints on price though I would have thought that I would have. Everything is so expensive right now it is like nobody questions it.

This whole floor vibration issue is probably something AI will solve for us in the future. I figure someone punches in the data for 50,000 buildings or so identifying the problem bays. And you'll have so many case scenarios that AI will be able to pick up on a potential problem. Of course with everyone using the same guidance we might just end up with the same parameters..... lololol

John Southard, M.S., P.E.

https://www.pdhlibrary.com/

 

[human909]

Nothing technical to help here but I have a related anecdote...

the steel started getting silly as we chased stiffness and the client didn't have the budget

We did fem models, not just the murray approach...

After it was built they had the acoustician test it and it was completely fine, not even close.

I think there often can be so much variance between expected loads, stiffness's (particularly regarding nominally 'pinned' connections) and expected sources of vibration absorption that designing in advanced can be fraught with difficulties. Chasing silly stiffness to get away from your vibrations just becomes expensive and don't even know if you are achieving your goal.


My anecdote starts with this thread I created where I started chasing stiffness ideas. When I eventually got to site to investigate I discovered the vibration was so bad that it was being transmitted over 50m away through the earth! While I did have accelerometers with me, this is what caught my attention when I was taking a break in the site office:

Anyway to cut a long story short. If I was going to address the issue structurally I needed more mass and less stiffness. Instead I provided dampers to the machines causing the issue.

I suspect for human induced floor vibrations that similar solutions might also be practical such as with products such as this. (Assuming you haven't created a stupidly flexible structure.)

 

[271828]

The video above reminded me of something that's sorta funny:

Glasses of water are our enemy when it comes to floor vibrations. They're a great "sensor" to alert the occupants to vibration. Once users are alerted to vibration, they are much more sensitive. I've seen floors that had to get down from about 0.5 %g to 0.2 %g to satisfy occupants who have been alerted. We don't want them thinking about this. If someone comments about the vibe on your floor, I'd recommend talking with them privately about that and make sure they feel heard so they don't feel the need to talk to everybody and their uncle about it. Psychology is extremely important for this. You don't want to have to cut the vibe by 3x to satisfy a group of highly sensitized occupants.

I once put an accelerometer on a typical residential floor and did some walking around and watching the water in a glass. I could see ripples to accelerations way, way below 0.5%g. I forgot the level at which they disappeared. It was something like 0.15-0.2 %g. My office has a heavy elevated concrete floor. With a water glass on my desk, I can see ripples if I stomp the floor. I can't see them with regular walking.

 

[271828]

Thanks for the input everyone. I'll try taking a look at some of Brad Davis's work.

Those methods are described in detail in Design Guide 11 Chapter 7.

 

[bones206]

Also as an experiment I change the joist spacing from 5 ft on center to 6 ft on center. The deck design is good up to 6 feet so why not.

This has been my approach lately. I try maximizing deck span to add tributary mass to each joist, even if it means going with a deeper deck profile.