Floor vibration in existing building retrofit without jacking 2

[AndEI]

We’re working on an office building that is experiencing floor vibration problems due to walking (that is, people walking, not the building). The building is 20 years old, and we are told there have not been complaints in the past. The interior was redone in the last couple of months, so the occupancy went from a typical office to a more open/electronic office. The floor system is a non-composite 2.5” concrete slab on 1” metal deck (total floor depth = 3.5”). The floor is supported by 26” deep open web steel joists spaced at 28” on center (joist span = 35ft). The joists frame into a 28” deep joist girder (span = 30ft).

Our proposed retrofit of the joists is to weld a 12” castellated beam to the bottom chord of the open web steel joists, similar to Figure 8-1(b) in the new DG11. We’ll also stiffen up the connection between the girder and the slab, so we can use the full composite section of the girder.

I can’t quite wrap my head around why the floor needs to be jacked up. I know DG11 says, “Jacking is required to introduce strain into the added elements so that the retrofitted system is ‘tight’ and able to respond to the very small loads introduced by human activity.” And I know natural frequency depends on the uniformly distributed line mass, m, supported by the beam, and the new beam doesn’t support the existing loads.

But I would think that because (1) the strains between the new beam and the existing joist will be compatible (the strain would be discontinuous, but the welds won’t slip), and (2) the steel will remain linear, the built-up section will bend as a unit from any additional load, even the very small loads introduced by human activity. If this is the case, I would think the increased moment of inertia of the built-up section would still increase the joist stiffness enough to help reduce noticeable floor vibrations.

The client is hesitant to spend the money to retrofit the floor, so they’re going to try to just move a bunch of file cabinets into the areas with vibrations first. And if we do go through with the retrofit, we’ll remove the loading anyways so we can weld to the bottom chord of the joists in tension, but I would still like to feel more comfortable with the mechanics.

Thank you in advance for your time and help.

-Andy

 

[oldestguy]

While doing this installation, is there a possibility of adding a column now and then along this new reinforcement?

 

[structSU10]

Did you have any in field vibration testing done? You may find certain areas where vibration is annoying that could be avoided as noted in your original post.

Also have you investigated a tuned mass damper instead of structural retrofit? not sure it will be cheaper or not, but it could be an option.

 

[JoshPlumSE]

Some thoughts:

1) After reading through the design guide commentary on those figures, I would think the Queen's post stiffening would be a better option.
2) Keep in mind walking excitation isn't just the joists or girders it's based on the overall PANEL vibration. And, vibration isn't the same thing as adding load or deflection. It has to do with frequency and dynamic amplification. If the reinforcement has been added, but remains un-stressed, then is it really affecting the frequency at these low levels of movement?
3) I think about it this way, a cables frequency is different based on the amount of pre-tension. Not saying it's exactly the same thing, but the restoring force is different based on the stress level in the cable. I think it's related to why they want that reinforcement at the bottom flange stressed.... So that the restoring force for vibration movement is truly present (and consistent) across the whole section, or else the entire section won't participate as effectively for vibration resistance.

 

[XR250]

You might have to jack just to straighten the sag out of the joists in order to weld the new member.

 

[WARose]

I've fixed all sorts of vibration problems in floors before and it hasn't involved jacking once. My version of DG 11 (an older one: 1997) doesn't include that statement. It says:

"The addition of rods to the bottom chord of joists is
not very effective even if the floor system is jacked-up prior
to welding of the rods. Even with jacking, the expected
increase in frequency generally does not occur because only
the flexural stiffness of the joist is increased, while the effect
of deformation due to shear and eccentricity at joints (see
Sections 3.5 and 3.6) is unchanged (Band and Murray 1996)."--DG 11, p.56)

I would think (in such a situation) that adding mass would be a viable option. I've done it with everything from floors to paper machines.

 

[AndEI]

Thank you all for taking the time to offer your expertise.

oldestguy: We could likely move some things around on the ground floor to provide a few extra columns. We're trying to not to interfere with the floor plan, but we can if we have to.

structSU10: We visited the site and felt the vibration in some areas, but we haven't had any field measurement done to get the floor frequency/damping. The floor framing is regular, and it is consistent throughout most of the floor. The areas where we felt the vibration had fewer people and office supplies than where we did not feel vibration, so adding furniture will probably help, but we can't predict the change in damping reliably. As for TMDs, we liked the idea because this floor is lightly damped. But we were quoted $15000-$17000 per bay a manufacturer. To the client, that's a lot of money to spend on something that DG11 says, "has been used with varying degrees of success." The limitations of only being effective for one mode and requiring tuning if the floor occupancy were to change again steered us towards stiffening the floor.

JoshPlum: I agree with you, I think a queen post would be a better repair. What you said makes a lot of sense, but here's whats tripping me up

If the reinforcement has been added, but remains un-stressed, then is it really affecting the frequency at these low levels of movement?

I think it would. See the attached sketch. Let's say I simplify a single joist as a SDOF with the linear spring stiffness associated with the composite joist, and the mass compresses the spring, this SDOF has a mass and a stiffness. If I then add an undeformed spring next to the deformed spring, I would expect the springs to act in parallel, even though the new spring does not have any strain. Because both springs are linear and have the same Young's modulus, their stiffnesses should not depend on if they have any strain.

XR250: You make a good point. It could be tough to weld a straight beam to a sagging joist. We're dealing with pretty small displacements under the service loads, though (about L/2000).

WARose: I am glad to hear you've fixed vibration problems without jacking up the floor. That is encouraging. DG11 still contains that language, but this is specifically in reference to cover plates and rods welded to the bottom chord, not for beams or queen posts attached to the joists.

 

[WARose]

WARose: I am glad to hear you've fixed vibration problems without jacking up the floor. That is encouraging. DG11 still contains that language, but this is specifically in reference to cover plates and rods welded to the bottom chord, not for beams or queen posts attached to the joists.

It's still applicable. When it comes to joists (and really any other type of framing member) the rotational characteristics of the joints can be critical. I've seen people ignore that in analysis.....and find themselves in trouble.

 

[Agent666]

One thing that comes to mind is running another member perpendicular to the joists at midspan and keep it continuous over a few bays. Connect it to each joist rigidly. I'm thinking this would work to break up/link the deformation of a joist to those adjacent ones.

A couple of full height partitions here and there might help as well.

Your most bang for buck is to alter the frequency of the floor. Model a few typical bays and just try stuff like noted above and see how it affects the fundamental frequency. You can easily address a single bay on site with a proposed solution and see if it has the desired effect (in terms of occupant feedback or measured results) before pulling the trigger on all of the problem areas.

I suspect getting some measurements done, will arm you for addressing the actual issues, for example the best way to address any resonance at particular frequencies. Can also do it after any fix is in place, to show it has suppressed the undesirable effects. Vibrations are quite funny in that one person can feel it and another can't, so being able to have more than someones perception to evaluate can be helpful in knowing when to stop. Once people are aware of a vibration issue, they tend to be more sensitive to noticing it as they are looking out for it, vs being oblivious to it previously.

 

[AndEI]

WARose: You're absolutely right. For our proposed retrofit, the beam welded below the joist, DG11 says, "If this type of stiffening is used for open web joists or joist girders, the effects on shear and eccentricity on stiffness need to be evaluated using the finite element method." We've done this, accounting for panel point eccentricities, etc.

When you abated floor vibrations in the past, did you typically do it by adding mass? I agree that adding mass could be preferable in this case. How did you add the mass? We're starting out by just moving file cabinets back onto the floor, but this isn't my favorite solution because (1) we can't reliably predict the reduction in vibrations because we can't really estimate the increase in damping, and (2) if there is a future change in occupancy, they will run into these problems again. I've heard full-height partitions can help a lot, and I guess thickening the slab would help (but would be very intrusive for the office).

Agent666: Thank you for the ideas. You're right, they don't need to buy all the steel at the same time, and they could save a lot of money if the initial retrofit isn't satisfactory. I will plan to get floor natural frequencies/damping from heel drop tests, too, that's good advice.

being able to have more than someones perception to evaluate can be helpful in knowing when to stop. Once people are aware of a vibration issue, they tend to be more sensitive to noticing it as they are looking out for it, vs being oblivious to it previously.

I didn't think about that, but it will definitely help to have measurements showing a reduction in floor accelerations, in case people still notice vibrations.

 

[WARose]

When you abated floor vibrations in the past, did you typically do it by adding mass?

It's been a mix of that and changing the stiffness.

I agree that adding mass could be preferable in this case. How did you add the mass?

Anywhere/anyway I could get it. I remember one time they wouldn't let me add any concrete to the base of this paper machine....and we would up adding a bunch of (open) 55 gallon drums filled with water. Ever so often they add water to the drums (as it evaporates).....and the vibration disappears.

We're starting out by just moving file cabinets back onto the floor, but this isn't my favorite solution because (1) we can't reliably predict the reduction in vibrations because we can't really estimate the increase in damping, and (2) if there is a future change in occupancy, they will run into these problems again.

I'm not sure I follow your point with #1. You should have a FEA model of this and the effects of mass addition should be apparent.

As far as #2 goes.....not much you can do there.....except maybe add a sign in the area.

 

[Agent666]

One thing, by adding mass you are not changing the damping. You are changing the frequency.

 

[BridgeSmith]

If the vibrations are being carried through the joists, adding cross bracing (diaphragms) between the joists could make a substantial difference at fairly low cost.

If it's truly a vibration (small magnitude), as opposed to a vertical acceleration, then it could just be in the slab itself. If that's the case, the slab could be positively attached to the joists (concrete anchors with clips to the joist flanges?).

The other thing to look at is whether the slab is bearing on all the joists. If not, that could be the issue. Might just require some careful shimming or injecting something into the gaps. (urethane or polyurethane caulking?)

 

[271828]

The jacking recommendation is due to the extremely small displacements that occur during floor vibration. A mid-bay acceleration of 0.5%g at 5 Hz is only 0.002 in. at mid-bay and smaller everywhere else. If you don't force new elements to engage, then they won't feel anything at such small amplitudes. The tiniest slipping, wiggle, etc. will result in the new elements not taking load, and the system won't even know they are there.

 

[AndEI]

271828: I was hoping you would see this thread. I've read many of your previous comments on the topic, so I am glad to get your input. It's a little unintuitive to me that the vibration amplitude is too small to engage new unstressed elements, but I am satisfied by your response. Thank you for the explanation.

WARose and Agent666:

by adding mass you are not changing the damping. You are changing the frequency.

This is getting a little off topic, but in theory, sure, I agree adding mass will not affect damping. Generally, though, I don't agree with this. The purpose of adding mass is not to change the frequency because adding mass lowers the frequency, right? Adding mass for the sake of mass would probably help because the panel mode mass will increase and so the acceleration will decrease, but adding mass could be counterproductive if the fundamental frequency is lowered too much. My (limited) understanding is that adding mass retroactively in office buildings is effective for reducing vibration mostly because it increases the damping of the floor. In the physical world, if we put a few file cabinets on the floor, some of the vibrational energy will dissipate through friction/heat. Looking at the different damping ratios given in DG11 for different office fit-outs seems to say, "hey, if you have a bunch of stuff around, you get more tiny 'slipping, wiggling, etc' in non-structural components that will help make the floor less bouncy." I very well could be misinterpreting what is going on, and I am interested to hear your understanding.

 

[WARose]

It's a little unintuitive to me that the vibration amplitude is too small to engage new unstressed elements,....

It's something to consider....but it's one of the reasons you try to make a lot of these modifications by welding and having the area as unloaded as possible during modifications. Jacking is great.....but not always practical.

This is getting a little off topic, but in theory, sure, I agree adding mass will not affect damping. Generally, though, I don't agree with this. The purpose of adding mass is not to change the frequency because adding mass lowers the frequency, right? Adding mass for the sake of mass would probably help because the panel mode mass will increase and so the acceleration will decrease, but adding mass could be counterproductive if the fundamental frequency is lowered too much. My (limited) understanding is that adding mass retroactively in office buildings is effective for reducing vibration mostly because it increases the damping of the floor. In the physical world, if we put a few file cabinets on the floor, some of the vibrational energy will dissipate through friction/heat. Looking at the different damping ratios given in DG11 for different office fit-outs seems to say, "hey, if you have a bunch of stuff around, you get more tiny 'slipping, wiggling, etc' in non-structural components that will help make the floor less bouncy." I very well could be misinterpreting what is going on, and I am interested to hear your understanding.

I agree that adding mass will not (in general/theoretically) change the damping. For a floor, most of the modes of vibration will be based on the flexural stiffness of the floor system. Ergo the damping of the steel will control the damping for those mode(s). (The damping level for steel is typically low.) In practical terms, you should get some help from non-structural elements (including their physical interaction). But that's hard to quantify. (That's one of the reasons people frequently suggest testing....as they have done on this thread.)

 

[JoshPlumSE]

The jacking recommendation is due to the extremely small displacements that occur during floor vibration. A mid-bay acceleration of 0.5%g at 5 Hz is only 0.002 in. at mid-bay and smaller everywhere else. If you don't force new elements to engage, then they won't feel anything at such small amplitudes. The tiniest slipping, wiggle, etc. will result in the new elements not taking load, and the system won't even know they are there.

Thank you! That's kind of what I was hinting at with my post, but explained more rationally.

 

[271828]

By the way, I'll relate some practical experience with this stuff.

When I was in grad school, we constructed a full size composite bay and ran experimental modal analysis tests on it. I ran one test and recorded the FRF. One of my pals and I went underneath with a screwjack and installed it at mid-bay to the limit of our strength. It was as tight as I think is possible without jacking. I repeated the shaker test and the entire FRF curve was indistinguishable from the first. The specimen behaved as if the screw jack wasn't even there.

I have worked on quite a few vibration reduction projects. This stuff is very tricky. Many times, I'm coming into the project after the structural engineer has take a shot or three at it, and the solutions didn't help. Some even make the vibration worse, especially as people are getting aggravated and thus more sensitive to vibration. At two projects last year, the employees were so frustrated with the floor motion that they were complaining of vibration in the 0.3%g range (long stories)! People can feel vibration far below 0.5%g, especially if they're paying attention to vibration.

In one project last year, we didn't have them jack the slab because it was not possible. The retrofit worked anyway. In that case, we were mostly adding mass, though, along with some stiffness. The main problem with that floor was lack of mass.