Vous êtes sur la page 1sur 5

Tips for Avoiding

Floor Vibrations
by Thomas M. Murray, Ph.D., P.E.

M
ore and more frequently, Tip 1: Don’t blame vibration ability problems as LRFD. The
engineers who have problems on LRFD–it’s not the designer needs to accept that 50 ksi
never encountered a floor cause of serviceability prob- steels, higher strength concrete,
vibration problem with their lems. optimized computer-based de-
designs are finding themselves Sure, LRFD results in lighter signs, longer spans and much
searching for fixes. I have been floor systems, especially if compos- lighter actual live loads result in
involved with the floor vibration ite construction is used. Sure, the lively floors (as the British say),
serviceability for over 30 years. For profession has a hang-up about and therefore require a little more
the first 28 years, I heard of perhaps LRFD. Yes, composite systems design time. Better yet, think of it
one problem floor a year; now it’s rarely satisfy floor vibration crite- as the need for a little art in your
one a month and sometimes more. ria, but that’s not the fault of LRFD. floor designs.
Why? A stretched-to-the-limit ASD de-
First, the reason is not LRFD. sign will result in the same service-
Yes, LRFD results in lighter floors,
especially if they are composite,
but nearly identical designs can be
obtained with ASD. The reasons
are longer beam spans (40 ft today
versus 25 ft a decade ago), less
actual live loading (11 psf versus
perhaps 25 psf), and less damping
because of open office areas and
less paper, heavy desks, and file
cabinets. The bottom line is that
floor vibration serviceability can
control the design. We may have to
accept the fact the may need to add
steel to control the problem just like
we do to control drift in high-rise
buildings.
Following are ten tips to help the
designer with the floor vibration
problem. I assuming the reader is
somewhat familiar with floor
vibration analysis, if not, I recom-
mend a study of the AISC/CISC
Design Guide 11 Floor Vibrations due
to Human Activity. Figure 1. Resonance response (Figure 1.3 of the AISC/CISC design guide).

March 2001 / Modern Steel Construction


Tip 2: Use the AISC/CISC design But why learn the new criteria? 24K8 joists at 24” on center and
guide criteria. All floor vibration criteria have two spanning 38’. The joists are sup-
Study the American Institute of parts: a prediction of the floor ported by W24×76 girders span-
Steel Construction and the Cana- response and a human tolerance ning 30’. Nothing about this
dian Institute of Steel Con- level. Furthermore, all criteria must system is really unusual except that
struction’s Design Guide 11: Floor be calibrated and thus are empiri- the live load deflection for the joists
Vibrations Due to Human Activity. cal in nature (the necessary funda- is less than L/480. The Modified
Unlike older publications such as mental studies of human response Reiher-Meister Criterion predicts a
Modified Reiher-Meister Scale, Mur- to low frequency/very low ampli- “slightly perceptible” floor. The
ray Criterion and the Canadian Stan- tude vertical vibration have not Murray Criterion requires 4.1%
dards Association Rule that use a been done). The Modified Reiher- damping, which is easily justified.
heel-drop impact, the new criteria Meister, Murray Criterion and Cana- The Design Guide predicts a peak
are based on resonance with walk- dian Rule were all calibrated using acceleration of 0.66%g, which is
ing. floors built at least 25 years ago. greater than the office environment
Resonance can occur when the However, construction and the tolerance acceleration of 0.50%g
exciting frequency (rate of walk- office environment have changed. and is an unacceptable floor. The
ing) or a multiple of that frequency Today, we use lighter structural framing shown is nearly identical
(harmonic) equals the natural fre- members, thinner concrete decks to a recently investigated floor
quency of the floor system. Reso- and longer spans. Actual office live where the building occupants had
nance results in very large loads are probably less than one- complained quite vigorously and
amplitudes of displacement, veloc- half of what they were 25 years where damping posts were in-
ity or acceleration, as seen in Figure ago, and permanent partitions are stalled to reduce vibration.
1. The criteria ensure that reso- more scarce resulting in less damp-
nance does not occur for the first ing. The older methods simply do Tip 3: Consider the conse-
three harmonics associated with not account for these changes. For quences of an electronic office.
walking. That is, if a person is instance, the Modified Reiher-Meister The electronic office is virtually
walking at 2 steps per second (2 Scale assumes 5 to 8% log decre- paperless; I have been in one where
Hz), the floor system is checked for ment damping, a level very the only papers were a few news-
resonance at 2, 4 and 6 Hz. unlikely for today’s floors. papers (mostly the financial sec-
The design guide criteria state Consider the floor framing tion) scattered around the
that a floor is satisfactory if the fol- shown in Figure 2. The structural computer terminals. The result is
lowing inequality is satisfied: system is 3¼” normal-weight con- much less live load and much less
crete on 0.6” C deck, supported by damping. Desks, filing cabinets
ap Po exp( −0.35 fn ) ao
= <
g βW g

where ap/g is the predicted peak


acceleration of the floor due to
walking as a function of gravity,
ao/g is the tolerance acceleration
for the environment, Po is a con-
stant force representing the excita-
tion, fn is the natural frequency of
the floor system, β is the modal
damping in the floor system and W
is the effective weight that moves
because of the excitation. At first
glance, the criteria might look a bit
formidable—it certainly is different
than the older criteria. In reality,
only fn and W require calculations.
Po is 65 lb for office floors, ao/g is
0.005g (0.5%g) for office environ-
ments and β is a number between
0.01 and 0.05 (see Tip 3).
Figure 2. Floor framing.

March 2001 / Modern Steel Construction


and bookcases are live load and damping. Critical damping is the Joists and joist-girders have
great sources of damping. In their damping required to bring a sys- another problem—they are not fab-
absence, the potential for annoying tem to rest in one-half of a cycle. ricated with work points. Panel
floor vibrations mounts. Adding to That is, if you hit something and it point eccentricities of up to 2”, as
the problem are modern floor lay- has 1.00 or 100% critical damping, shown in Figure 3, are common.
outs–open, with few fixed parti- it will come to rest without oscillat- Surprisingly, this has no effect on
tions, widely spaced demountable ing. For offices with fixed parti- strength although member stiffness
partitions or no partitions at all. tions, a good estimate is 0.05 or 5%; is reduced, especially if the span-
Atrium-type areas are more com- for conventional or paper offices, to-depth ratio is less than about 18.
mon and curtain walls are less stiff. i.e. good old structural engineering The design guide offers the follow-
What’s the solution? Use the offices, with demountable parti- ing expressions that are used to
AISC/CISC design guide methods, tions, use 3%; and for the paperless predict the effective moment of
assume actual live loads in the 6 to or electronic office, I recommend 2 inertia of joist and joists girders:
9 psf range, and modal damping of to 2.5%. Note again that these num-
• for angle web members with
2 to 2.5% of critical. bers are much less than those rec-
6 < L/D < 24:
Recently, because of an annoy- ommended for heel-drop based
ing floor, the office contents in one criteria. Cr = 0.90 (-e-0.28(L/D))2.8
building were actually weighed— • for round rod web members
the result was an equivalent weight Tip 5: Do not design floors with with 10 < L/D < 24:
of 8 psf! Throw in the humans, and a natural frequency below 3 Hz
you may get 9 psf! The floor design Walking speed in an office can Cr = 0.721 + 0.00725 (L/D)
live load was 125 psf. Do we need be 1.25 to 1.5 steps per second (or where L is the member span and D
to change our code live loads? Hz). Resonance at the second har- is the nominal depth; and
Probably, but that’s a question for monic, 2.5 to 3 Hz, is then a real
the ASCE-7 Committee. possibility if the floor’s natural fre- Imod = Cr Ichords
What about damping? Read on. quency is below 3 Hz. I have This moment of inertia is then
caused a floor to vibrate at its natu- used to calculate the effective trans-
Tip 4: Don’t mix-up Log Decre- ral frequency by running a shaker formed moment of inertia of the
ment and Modal Damping. (an electrically-powered oscillating composite section. The above
Now for some jargon: log decre- mass) at one-half of the floor fre- expressions were developed using
ment damping was used to quency. The result is quite unset- static analysis and tests and apply
develop the older heel-drop-based tling; if this happened in an office equally well to static live load
floor vibration tolerance criteria. building, complaints would be deflections.
Unfortunately, log decrement loud and clear. For many years, I maintained
damping overestimates the damp- However, a 3 Hz or less floor can that joist seats provided enough
ing as it measures not only energy be made to work if it is made very stiffness so that the supporting
dissipation (true damping) but also heavy, say 100+ psf. girder or joist-girder could be con-
the transmission of vibrational sidered fully composite for floor
energy to other structural compo- Tip 6: Remember that joists and vibration analysis. I was very
nents. The design guide criteria use joist-girders require special wrong. Using floors constructed in
modal damping or “true” damping consideration. the Virginia Tech Structures and
(it’s interesting that we call modal The stiffness of trusses is Materials Laboratory, we found
damping “true damping” when we affected by shear deformations in that joist seats are not, in fact, very
cannot measure it very accurately, the webs. An age-old rule-of- good shear connectors. The design
at least in floors). What’s the differ- thumb is that the effective moment guide recommends that the com-
ence? Only about 50% to 100%, so of inertia of a parallel chord truss is posite moment of inertia of a girder
be careful! Modal damping is one- 0.85 times the moment of inertia of or joist girder be approximated
half to two-thirds of log decrement the chords. This rule is used to using:
damping, so if you are accustomed compute the L/360 deflection limit
to estimating damping for heel- live load in the Steel Joist Institute Ig = Inc + (Ic - Inc) / 4
drop based criteria, you will need load tables. This rule works well if where Inc and Ic are the non-com-
to adjust your design office prac- the span-to-depth ratio of the truss posite and fully composite
tices. is greater than about 18; if the ratio moments of inertia, respectively.
What are good modal damping is less, the deflection will be greater Recent field tests have shown this
estimates? Damping is usually than predicted. expression is a bit conservative if
expressed as a ratio of critical

March 2001 / Modern Steel Construction


the joists are closely spaced, say not usually more than the frequency- made measurements on composite
more than 30”, and unconservative caused increase, resulting in a bet- joist supported floors with spans
if there are only two or three joists ter floor. Second, you can first between 40’ and 118’ (that’s not a
being supported by the girder or stiffen the member (beam or typo, there truly is an office build-
joist-girder. Testing is currently girder) with the lower frequency ing with a 118’ span). The design
being conducted to develop better until both frequencies are approxi- guide criteria predicted the floors
approximations. mately the same. If the system is would not be annoying and they
still not satisfactory, member types were not.
Tip 7: Improve a design that can be stiffened until a satisfactory The bottom line is that floors of
does not satisfy the criterion. design is achieved. My experience any span can be designed such that
The criteria for heel-drop based has shown that the latter method is occupants will not feel annoying
methods indicates that increasing more cost effective for most vibrations. Just be sure the design
the stiffness has very little effect on designs. satisfies the design guide criteria
the floor performance. With these and the frequency is above 3 Hz.
methods, the only way to effec- Tip 8: Don’t believe the myth
tively improve a proposed floor that certain beam spans should Tip 9: Be careful when designing
design is to increase the mass. A be avoided. crossovers (elevated walks).
different result is found when the In the late 60s or early 70s a Atrium crossovers can be a
design guide methods for office paper was written describing a design challenge. Crossovers typi-
floors are used. With this method, number of joist-supported problem cally have long spans; therefore,
the tolerance criterion can be satis- floors where the joist spans were in the frequency is quite low. Further,
fied by either increasing the mass the 24’ to 28’ range. Somehow this there is very little damping, gener-
or increasing the stiffness. A stiffer was interpreted to mean that bays ally about 1% modal damping. The
floor is always a better floor so the with beam or joist spans in this result is that deep, stiff supporting
latter result is logical–no one has range should be not be designed, members are required.
ever had a vibration problem with and this belief has become part of Also, the location of the slab
a 10’ span. the folklore (if I may use that term) needs to be considered. I know of
If the design guide method is of the structural engineering com- two problem crossovers where the
being used and a proposed framing munity. Even some joist manufac- structural engineers relied on pre-
scheme does not satisfy the toler- turer engineers will tell you to vious experience with floors of
ance criterion, e.g. 0.5% of gravity, avoid these spans. The problem similar framing and did not check
there are two approaches to floors described in the original the crossover design. In both cases,
improving the design. First, you paper were typical of the time, complaints were received even
can increase the mass by adding meaning that the spans and the before the buildings were opened.
concrete or changing from light- problems were connected. But, in The major cause of the problems
weight to normal weight concrete. fact, there is no correlation between was that the crossover slab was
This approach will result in a span and occupant complaints. located between the supporting
slightly lower fundamental fre- Span alone is not the reason a par- beams at about mid-depth as
quency but a larger effective ticular floor is annoying to occu- shown in Figure 4. The result was
weight, W in the criteria. The lower pants. that the moment of inertia of the
frequency will increase the pre- Likewise, long span floors, say crossover was twice the moment of
dicted acceleration and the larger spans greater than 40’, are not inertia of the supporting beams,
effective weight will decrease it, inherently problem floors. I have which, of course, is much less than

2”

Typical Joist Configuration Finite Element Model


Figure 3. Joist panel point eccentricity.

March 2001 / Modern Steel Construction


the composite moment of inertia weight-lifting and aerobics near cept of a floating floor is similar to
would had been if the slab was on offices are 5 to 10%, 2% and 0.5%, that used for isolating machinery
top of the beams. The result was a respectively. It turns out that vibration. A floating floor is simply
much lower frequency than weight-lifters are sensitive folks, a separate floor supported by very
expected and an expensive fix in thus the lower limit. Also, some of soft springs attached to the struc-
both cases. them are big, so you have to be tural floor. The natural frequency
extra careful! The required floor of the floating floor should be quite
Tip 10: Be even more careful frequencies for the three conditions low, less than 2 to 3 Hz, which gen-
when designing health clubs in and a 100 psf floor are 8.8 Hz, 9.2 erally requires a heavy slab, 50 to
office buildings. Hz and 16 Hz. For a 50 psf floor, the 100 psf. Also, the space between
Aerobics classes are part of any corresponding frequencies are 9.2 the two floors must be vented or
health club’s activities, and an aer- Hz, 10.6 Hz and 22.1 Hz. the change in air pressure due to
obics class is probably the most If the spans are less than, say 30’, the movement of the floating floor
severe building floor loading for use of lightweight concrete and will cause the structural floor to
vibration concerns. The energy closely spaced, deep joists will move.
from aerobics can travel much far- result in a floor frequency in the
ther than you might expect. I know range of 10 to 12 Hz without too A Final Thought
of an instance where aerobics on much expense. The floor system A number of structural engi-
the 60th floor of a building were would be satisfactory for aerobics neers have told me that they now
felt on the 40th floor but not on the alone or in conjunction with design for serviceability and then
floors in between or below the 40th weight-lifting but not near offices. check strength. As Hardy Cross
floor! Aerobics in one corner on the Generally, it is cost prohibitive to once wrote: Strength is essential
second level of a two-story strip design a floor system that supports but otherwise not important.
mall has been felt several hundred both aerobics and offices.
feet away. Solutions are costly: a If the aerobics activity cannot be Thomas M. Murray is the Mon-
400% increase in steel weight over moved to a slab on grade, then I tague-Betts Professor of Structural
the strength design would have suggest either a separate framing Steel Design at Virginia Tech, Blacks-
been required in a strip mall to system for the aerobics floor or the burg, VA and President of Structural
solve the problem (the owner use of a floating floor. Separate Engineers, Inc., Radford, VA. He can
decided to move the health club to framing is an easy solution for two be reached via email at:
the lower level instead). story buildings. tmmurray@floorvibe.com or via tele-
The design guide has criteria for When using this approach, the phone at 540/231-6074.
designing floors supporting rhyth- aerobics floor slab must be com-
mic activities. Basically, the floor pletely separated from the sur-
frequency must be above a limiting rounding slabs, and the ceiling
value that depends on an accelera- below cannot be supported from
tion limit, which is determined the aerobics floor framing. Separate
considering the activity and what cold-formed framing connected
is called the “affected occupancy” only to the columns has been used
and the weight of the floor. The to support the ceiling below.
acceleration limits for aerobics Floating floors may be the only
alone, aerobics in conjunction with solution in a tall building. The con-

Figure 4. Crossover cross-section. Note: transverse members allow the deck to run parallel to the girders as shown.

March 2001 / Modern Steel Construction

Vous aimerez peut-être aussi