Anda di halaman 1dari 10

VIBRATIONAL BEHAVIOUR OF TIMBER FLOORS

Ivan Glisovic1, Bosko Stevanovic2

ABSTRACT: A complete design methodology for timber floors must address the problem of annoying vibration
caused by the occupants themselves. Proper controlling relies on good understanding nature of floor vibrations induced
by human movements. This paper reviews the major findings on characteristics of the footstep forces and discusses
mechanism of footstep-induced vibrations in timber floors. A numerical modelling procedure, which is based on the
finite-element method, for obtaining the natural frequencies, mode shapes and vibration response of timber floor
systems is presented. Using this numerical model, the effect of change in various floor parameters on vibrational
performance of timber floors is investigated.
KEYWORDS: Timber floor, Footstep, Vibration response, Finite element modelling, Construction details

1 INTRODUCTION 12
Annoying floor vibrations are common in many types of
building structures. Problems of this nature have been
reported in private dwelling houses, office buildings,
schools, restaurants, etc. Although floor vibrations can
result from many sources (e.g. use of machinery,
explosions, external traffic) the most common and
problematic are caused by the occupants themselves by
their everyday activity. Such forces are particularly
problematic because they cannot be easily isolated from
structure and they occur frequently. Therefore, excessive
floor vibrations due human-induced loading have been
characterized as probably the most persistent floor
serviceability problem encountered by designers [1].
For heavy floors which utilize concrete as a deck,
vibrations produced by normal human movements are
generally less noticeable. In comparison the amplitudes
of vibration response found in timber floors are
relatively high. This is because amplitudes of response
are inversely proportional to the self-weight of the
structure being vibrated. As human bodies are generally
sensitive to vibrations, this high level response can cause
discomfort and undue disturbance of occupants.
Consequently, the requirement for designing against
disturbing vibrational performance is particularly
important for light-weight floors built from materials
such as wood.
In past, many design guides and codes of practice have
used point load deflection and fundamental frequency as
a measure of acceptable floor performance [2]. However,
1

Ivan Glisovic, Faculty of Civil Engineering, University of


Belgrade, Bulevar Kralja Aleksandra 73, 11000 Belgrade,
Serbia. Email: ivang@grf.bg.ac.rs
2
Bosko Stevanovic, Faculty of Civil Engineering, University
of Belgrade, Bulevar Kralja Aleksandra 73, 11000 Belgrade,
Serbia. Email: bole@imk.grf.bg.ac.rs

modern standards have now moved away from these


traditional measures of vibration performance and
require that designer make realistic estimation of level of
floor response that will be encountered in practice by
considering the walking excitation directly and
comparing this with human annoyance criteria. This is
primarily a result of a considerably reduced cost of
computing, but also reflects an increasing awareness that
floors are dynamically very difficult structures to model
heaving complex geometry, boundary and continuity
conditions, stiffness distribution and interaction of close
modes of vibration requiring multiple degree of freedom
modelling.
The general goal of this research is to gain a better
understanding of timber floor vibration phenomena in
order to apply it to a better prediction then what exists
for real systems. A finite-element approach is utilized to
model timber floor systems under dynamic loads
resulting from normal human activities. The numerical
model takes into account the various complexities in
timber floor construction: orthotropic sheathing and
semirigid sheathing-to-joist connections. Using this
numerical model, the influence of construction variables
on floor performances is evaluated. Evaluation is based
on comparison between the reference and modified
floors.

2 DYNAMIC LOAD
Acceleration and deceleration of the (mass of) human
body during various human activities is causing dynamic
forces. Forces depend upon many factors including the
characteristics of the person or persons, the activity
being undertaken (e.g. walking, running, jumping), the
number of people, whether activities of different people
are coordinated and the characteristics of the floor
surface.

There are very many variations of rhythmic body


movements leading to large variety of dynamic loads.
Loads differ in their nominal frequency, frequency
content and time function. Activities generating
synchronized rhythmic movements such as those due to
several or more people dancing or exercising are
especially problematic. Several people acting
synchronously for 20 seconds or more can lead to
approximately periodic loads that produce almost steady
state structural vibration [3].
Annoying vibration of timber floors is commonly
associated with walking and running forces. Vertical
forces due to an individual foot impact have been
measured and characterized (Table 1 and Figure 1).
Table 1: Data on walking and running (Bachmann and
Ammann, 1987) [4]

Activity
Slow walk
Normal walk
Fast walk
Slow running
(jogging)
Fast running
(sprinting)

Pacing
frequency
f (Hz)
1.7
2.0
2.3

Forward
speed
V (m/s)
1.1
1.5
2.2

Stride
length
L (m)
0.6
0.75
1.0

2.5

3.3

1.3

>3.2

5.5

1.75

As can be seen, there are two peaks in a force-time


history with the first corresponding to heel strike
contact and the second to toe-lift off contact. For
normal walking the peak force is about 1.2-1.5 times the
static force, and has duration of between 0.5 s and 0.8 s.
[3]. As the individuals gait increases from a walk to a
run, the peaks merge and peak force is much higher, but
the duration shorter, for running than for walking. Also,
in running, periods when both feet were off the ground
were clearly observable between the forcing peaks. On
the other hand, in periods of normal walking forces
induced by left and right leg overlapped, which means

that both legs were periodically on the ground. In


addition to the type of movement (walking or running),
the weight of a walker plays a significant role in
determining the magnitude, but has less effect on the
actual shape of the force fluctuation. The type of
footwear and traversing surface were found to have little
effect on the forces applied by human movements [1].
Dynamic forces from groups of people walking about the
floor at random would seldom cause serviceability
problems [5]. The static deflection of the floor naturally
would increase. However, the excitation would lack
coherence unless the group was walking in step, and thus
the dynamic component of motion would be small.
Therefore, a certain general conclusion can be made that
individual excitation represents an adequate model for
development of vibration serviceability criterion.
Frequency content of walking excitation was studied by
Ohlsson [6]. The force is composed of two different
component types:
Low-frequency components (0-8 Hz) which origin
from the step frequency and its harmonics;
High-frequency components (8-40 Hz) which origin
from impacts when the heel contacts the floor
surface.
Therefore, if the floor is low-frequency having a first
mode of vibration which is less than approximately 8 Hz,
the low-frequency loading components may create a
strong resonant vibration response. On the other hand,
higher frequency floors are more responsive to the
impulsive component of walking.

3 DEFINITION OF THE DYNAMIC


SYSTEM
A typical arrangement of components in a joisted floor is
shown in Figure 2. Most timber floors employ
equidistant joists as primary structural members. Joists
are sheathed (overlain) by a wood or non-wood layer of
structural material, the top of which is the floor surface.

Figure 1: Detailed forcing patterns for different types of human activities (Wheeler, 1982) [1]

In the structural sense whole floors with such


construction behave as a thin plate reinforced by series
of ribs. The plate layers can be either isotropic or
orthotropic depending upon the material used. Floors are
often stiffened in the across-rib (across-joist) direction
by addition of row of bridging or blocking.
Whenever possible, engineers conceptualize systems as
simple, usually rectangular, plan arrangement with
idealized supports. However, definition of the dynamic
system for any floor is often not simple. Plan geometry,
including openings for stairs and services, needs to be
known, as does the nature of the structural system for the
building as a whole. Characteristics of both supporting
and supported components will influence the response of
the floor [3]. Physical and mechanical properties of
structural and non-structural components take effect on
modal characteristics of a floor system (stiffness, mass
and damping).

4 FINITE ELEMENT MODELLING


Numerical analysis was carried out using the finite
element method (FEM). Use of finite element models for
considering floor vibration offers opportunities to the
designer, by allowing a more realistic consideration of
floor structures then can be achieved with simple hand
methods of analysis. Also, numerical modelling allows
easier and more accurate treatment of human-induced
walking excitation. In this study, commercial software
packages for static and dynamic structural analysis SAP
2000 was used. The presented numerical model is
intended primarily for dynamic analysis of timber floors.
The model can be used for both frequency and time
history response analyses of floor systems.
Consider a reference floor with the dimensions and

Figure 2: Typical joisted floor construction: a) sheathing


on lumber joists; b) sheathing on wood I joists [3]

properties shown in Table 2. This floor, which is typical


residential timber floors, was designed to satisfy stress
and deformation conditions (l/250) in accordance with
current SRPS regulations.
The floor cover (sheathing) was assumed to be
semirigidly fastened to equidistant joists to produce an
assembly capable of composite action, which together
vibrates under dynamic loads as a stiffened plate. The
stiffness of connectors controls the level of composite
action between the sheathing and joists. In general, the
sheathing has different elastic properties in the directions
parallel and perpendicular to the joists and was modelled
as an orthotropic plate. The material properties for the
joist were considered isotropic.
Timber joists were modelled using Solid elements
(spatial three-dimensional elements with 8 nodes), while
Shell elements (plane elements with 4 nodes) were used
for sheathing. The given elements have 6 degrees of
freedom in each node, three displacement and three
rotations - ux, uy, uz, x, y, z.
Each connector was modelled by vertical Linear Link

Table 2: Dimensions and properties of reference timber floor

Floor property
Span length
Joist spacing
Number of joist
Joist depth
Joist width
Modulus of elasticity for all joists
Mass density of joist
Thickness of sheathing
Modulus of elasticity for sheathing* EL / ER / ET
Shear modulus for sheathing* GLR / GLT / GRT
Poissons ratio for sheathing* LR / LT / RT
Mass density of sheathing
Connectors
Nail spacing
Nail vertical stiffness**
Nail horizontal slip stiffness**
Nail rotational stiffness**
Uniformly distributed load
Bridging
Support conditions

Value

3.6 m
60 cm
9
16 cm
12 cm
1000 kN/cm2
500 kg/m3
4.8 cm
1000 / 86 / 51.8 kN/cm2
73.5 / 69.1 / 7 kN/cm2
0.37 / 0.42 / 0.47
500 kg/m3
Nail (two rows)
100 mm
1.2 x 109 N/mm
1200 N/mm
180 N-mm
0 (unloaded)
No
All joist simply supported
on both ends
Note: *Suggested by Bodig and Jayne [7] **Suggested by Folz and Foschi [8]

elements (linear springs) having horizontal slip stiffness


(slip modulus) in two directions, one vertical stiffness,
and rotational stiffness. The adopted assumption that the
connectors are characterized by a loading-slip linear
relationship does not seriously limit the application of
this numerical model, since the vibrations of timber
floors is a serviceability problem where the structural
integrity of the floor is not jeopardized. In a strict sense,
the vertical stiffness value should be different in tension
(i.e. nail pullout) versus compression. However, because
deflections were assumed to be quite small, the model
was simplified by using a single value that was very
large in comparison with the horizontal and rotational
stiffness values [9]. Values of nail horizontal slip
stiffness and rotational stiffness as suggested by Folz and
Foschi are given in Table 2. The links had a length equal
to one-half of the sheathing thickness.
In the floors used for this study, all joists were assumed
to be simply supported (i.e. pin and roller) at their ends.
The sides of the floor (parallel to the joists), primarily,
were not supported. The presence of bridging was
simulated by restricting the torsional deformation of the
joists.
The finite-element modelling strategy of timber floor is
shown in Figure 3.

modes of strongly orthotropic plates exhibit a common


shape (usually a single half sine-wave) along the joists,
but have different shapes across the joists (see Figure 5).
Table 3: Natural frequencies of reference floor

Mode
number
1
2
3
4
5
6
7
8
9
10

Frequency
(Hz)
16.98
19.17
23.47
32.32
47.33
62.39
65.89
68.55
69.74
76.04

Human pedestrian excitation was considered using timehistory analysis based upon Newmark time integration
approach. A typical non-dimensional force-time relation
for a single footfall is shown in Figure 4a. For time-step
dynamic analysis, damping was included by use of
Rayleigh damping, in which the material damping matrix
is assumed to be directly proportional to both the
stiffness matrix and the mass matrix.
The analysis comprised two excitation cases: walking in
place in the centre of the floor and walking across the
floor. The excitation of a floor system by one person
walking is shown spatially and temporally in Figure 4b.
The force pulses due to successive footsteps will
normally overlap by roughly 0.1 sec. The step length
was taken as 0.75 m.

Figure 3: Finite element modelling strategy

5 DYNAMIC ANALYSIS
Dynamic analysis comprised determination of natural
frequencies, corresponding vibration modes and floor
response to excitation by human footsteps. The first 10
natural frequencies for a given reference floor are shown
in Table 3. It can be noticed that the values of natural
frequencies are quite closely spaced. This result is
typical for strongly orthotropic structures. Despite the
propertis of the floor plate and presence of bridging,
joisted timber floor are inherently orthotropic. This
promotes a tendency towards clustering of the first few
modal frequencies, with the second modal frequency
often being only 15-20% higher than the first [3]. Modal
clustering has the effect of increasing the amplitude of
vibration response, and thus velocity and acceleration
levels experienced by a person or object located on a
floor. Because modes are clustered, it is not generally
reliable to base vibration assessments of timber floors
only on the fundamental mode. The extent to which
modes are clustered depends upon parameters such as
floor shape, span and width, and flexural stiffness along
and across joists. For recangular plan floors the first few

Figure 4: Force due to walking on floor: a) force from


single footstep, b) forces on span from one person
walking [5]

Based on above understanding of the nature of the


footstep force, it is concluded that the two components in
walking excitation can initiate two types of vibrations in
floors, i.e. transient vibration and resonance, depending
on the dynamic properties of the floor system, rather the
floor system stiffness, mass and its capacity to dissipate
vibration energy (damping) [10]. These properties are

Mode 1
f = 16.98 Hz

Mode 1
f = 17.34 Hz

Mode 2
f = 19.17 Hz

Mode 2
f = 23.76 Hz

Mode 3
f = 23.47 Hz

Mode 3
f = 36.59 Hz

Mode 4
f = 32.32 Hz

Mode 4
f = 56.35 Hz

Figure 5: Typical mode shapes for a rectangular joisted floor: simply supported on two edges (left); simply supported on
all edges (right)

determined by floor material, design and construction.


The fundamental frequency of a floor built with a
material having a high ratio of strength to mass such as
wood is most likely above 8 Hz, which is above the
footstep frequency and its harmonics. Therefore, the
vibration induced by footstep forces is most likely
dominated by a transient response caused by the

individual heel impact force from each footstep. Time


records of response in the centre of the reference floor
are shown in Figure 6. The figure shows that the floor
response was a train of transient vibrations. It can be
observed that each transient vibration signal contains a
high initial peak and quickly decays. The peak values of
a transient vibration are governed by system stiffness

0.1

0
0

0.5

1.5

2.5

-0.1

0
0

Displacement (mm)

Displacement (mm)

-0.2
-0.3
-0.4
-0.5
-0.6

0.5

1.5

2.5

3.5

4.5

3.5

4.5

3.5

4.5

-0.1

-0.2

-0.3

-0.4

-0.7

-0.5

-0.8

-0.6

-0.9

Time (s)

Time (s)
0.012

0.015

0.01
0.01

0.008
0.006

0
0

0.5

1.5

2.5

-0.005

Velocity (m/s)

Velocity (m/s)

0.005

-0.01

0.004
0.002
0
-0.002

0.5

1.5

2.5

-0.004
-0.006

-0.015

-0.008
-0.02

-0.01

Time (s)

1.5

1.5

Acceleration (m/s )

Acceleration (m/s )

Time (s)

0.5

0
0

0.5

1.5

2.5

-0.5

0.5

0
0

0.5

1.5

2.5

-0.5

-1

-1

-1.5

-1.5

Time (s)

Time (s)

Figure 6: Time history floor response in centre of the floor: due walking in place (left); due walking perpendicular to joists
(right) (damping ratio of 3%)

and mass, but not by damping. Damping is a measure of


how quickly the response of vibrating system decay.
When floors are light, response includes time variation
of static deflection due to a moving repeated load, as
well as decaying natural variations due to footstep
impulses. A point load stiffness criterion is appropriate
for the static deflection component and a criterion based
on footstep impulse vibration is appropriate for the
impulsive component of loading.

individually, while holding all the other parameters


constant at their reference-floor values.
The results of parameter analysis given in Tables 4 show
the values of the fundamental frequencies of floors and
values of first four modal separation factors. MSFn, n-th
modal separation factor, is defined as [8]:

6 PARAMETRIC STUDY

where fn+1 and fn are the (n+1) and n-th floor natural
frequencies, respectively. Clearly, improvements in floor
performance, from the perspective of human perception,
can be achieved by raising the fundamental frequency
and increasing the spacing between adjacent frequencies
(i.e. increasing the MSFn).
The levels of vibration can be represented by
displacement, velocity or acceleration. In this study, the

In order to investigate the influence of different floor


parameters on natural frequencies and on vibration
response of a referent floor, a parametric study was
carried out using the previously described procedure of
numerical modelling. The procedure for parametric
study involved varying each significant floor parameter

MSFn =

f n +1
fn

(1)

value of root mean square (rms) of acceleration was


adopted as a referent value in accordance with ISO
standard. This parameter is able to account for human
sensitivity to amplitude, rate of decay and frequency
components of the vibration. Since people can tolerate
higher vibration magnitude at higher frequencies versus
lower frequencies, the calculated rms acceleration is
frequency weighted by an appropriate factor. According
to international standard ISO 2631-2: 1989 [11], for
frequencies between 8 and 80 Hz, the factor is equal to
8/f0, where f0 is the floor fundamental natural frequency
in Hz. The Arms value is calculated as:
1 T

Arms = a 2 ( t ) dt
0
T

1
2

(2)

where a(t) is the frequency-weighted acceleration in time


t and T is total duration of the vibration. The duration of
vibration was taken as 1 sec.
A heel-drop test was developed to achieve impacts
similar to those resulting from human footfall impacts.
The person produced impacts standing on his toes then
dropping his heels rapidly through a distance of about 65
mm. A typical shape of the force-time function for heeldrop impact is represented graphically in Figure 7
together with the simplified rectangular representation
used in following analysis. This approximation leads to a
considerable simplification in computation but only a
small conservative error in the final solution. The size
of the rectangular impulse varies according to such
factors as the weight and build of the person producing
it. Experiments by Chui and Smith [12] have shown that
t1 is in the range 0.05-0.07 s and P0 can be taken as 70%
of the weight of the person applying heel-impact. In the

given analysis, t1 = 0.05 s and P0 = 500 N, which is 70%


of the weight of a man whose mass is 70 kg, was
adopted.

Figure 7: Shape of forcing function of a heel-drop impact

The root-mean-square values of frequency-weighted


acceleration (Arms) of a vibration caused by a footfall
impact are shown in Table 4 (damping ratio of 3%).
Construction details have a strong influence on the
dynamic behavior of lightweight timber floors and the
acceptability of particular floors to users of buildings.
Intensive laboratory measurements as well as in situ
measurements were performed on real floors in order to
evaluate the influence of various structural parameters
on static stiffness and transient response of the floor
[3,8,12-18]. Results obtained by parametric study
completely correspond to the results of this research.
Reducing the joints spacing mainly increases the
stiffness of a floor in the joist span direction. This results
in an increase in the fundamental natural frequency,
which implies an improvement in floor performance, and
reduction in the spacing between natural frequencies. As
already mentioned, reduction in the spacing between
frequencies may have a negative effect on vibration
response as closely spaced vibration modes may interact
to produce motion with relatively high amplitudes. Thus

Table 4: Results of parametric study

Floor parameter

Fundamental
frequency f (Hz)

Joist spacing (cm)


40
18.66
60*
16.98
80
15.72
Joist depth (cm)
12
12.08
16*
16.98
20
22.08
Sheathing thickness (cm)
2.4
19.94
4.8*
16.98
6.0
15.85
Nail spacing (mm)
50
17.40
100*
16.98
200
16.61
Support condition
Supported two edges*
16.98
Supported four edges
17.34
Bridging
No bridging*
16.98
Bridging
17.03
Note: *Reference floor values

Modal separation factor


MFO1 MFO2 MFO3 MFO4

Rms acceleration
Arms (m/s2)

1.11
1.13
1.15

1.20
1.22
1.25

1.33
1.38
1.41

1.42
1.46
1.49

0.291
0.380

1.19
1.13
1.10

1.37
1.22
1.15

1.55
1.38
1.26

1.47
1.46
1.37

0.380
0.284

1.07
1.13
1.11

1.17
1.22
1.32

1.18
1.38
1.51

1.21
1.46
1.12

0.380
0.323

1.15
1.13
1.11

1.27
1.22
1.19

1.39
1.38
1.37

1.45
1.46
1.49

0.380
0.408

1.13
1.37

1.22
1.54

1.38
1.54

1.46
1.05

0.380
0.266

1.13
1.13

1.22
1.23

1.38
1.44

1.46
1.49

0.380
0.292

smaller joist spacing does not, as commonly believed


always ensure satisfactory vibrational performance.
It is also interesting to observe that increasing the depth
of the joist increases the fundamental frequency, but also
causes a clustering of all the frequencies. This potential
problem cannot be detected from a static analysis.
Increasing the thickness of the sheathing can greatly
reduce static deflection under concentrated load. This
does not automatically mean that dynamic behaviour is
improved, because the mass increases as well as the
stiffness and natural frequencies can actually reduce. The
degree of reduction depends primarily on the
relationship between modulus of elasticity of the
sheathing in two orthogonal directions and it is larger for
larger anisotropy of the sheathing material. The given
conclusion is contradictory to the opinion in practice that
the vibration problem can be simply solved by increasing
the thickness of the sub-floor.
Reducing the mechanical connectors spacing (which is
equivalent to increasing both horizontal and rotational
stiffness of connectors) increases the fundamental
frequency and the modal separation factors. An increase
in the mechanical connector horizontal stiffness
increases the fundamental frequency, but little effect on
the modal separation factors. On the other hand, an
increase in the mechanical connector rotational stiffness
has little effect on the fundamental frequency, but
increases the modal separation factors. Due to
exceptional orthotropic of the floor observed in the
parameter study the given effect is not expressed.
The benefits of having all four edges supported instead
of two are demonstrated. Although this practice has little
effect on the fundamental mode of vibration it stiffened
the floor in the direction perpendicular to the span and
thereby raises the higher natural frequencies. The mean
level of response is also observed to be lower in the floor
with all edges supported.
Between joist bridging (blocking or cross-bracing) in the
interior of the span often has a very beneficial influence
on the stiffness in the across joist direction. Although
bridging has little influence on the first three natural
frequencies, it improves the modal separation of the
higher natural frequencies. When the joists are tall and
slender and have low torsional rigidity, bridging is
necessary to prevent torsional movement in joints.
Considering the influence of various floor parameters, a
general conclusion can be made that the fundamental
frequency of a floor is the predominant function of
stiffness in the span direction, while the interval between
adjacent floor frequencies is governed by the ratio of
stiffnesses in two orthogonal directions and that the
intensity of the dynamic response depends on the mass
of the entire system. Based on the foregoing,
improvements in floor performance may be attained by
raising the natural frequencies, increasing the spacing
between adjacent frequencies, or reducing the
amplitudes of the response.

7 CONCLUSIONS
Dynamic behaviour of timber floors in buildings is
complex and interdisciplinary in nature, with many

factors contributing to their excitation, response and


acceptability of resultant vibrations. Human-induced
footstep loading has proved to be major source of floor
vibration disturbance, as it happens frequently and, in
practice, cannot be isolated. Vibrations induced by
footsteps in floors do not cause damage or structural
failure, but can human annoyance and malfunction of
sensitive equipment. An awareness of the nature of
vibration can help avoid or alleviate vibration-related
problems. It is advantageous and less costly to consider
the effect of vibration on the building and its contents
during the design process. It has been noted that
annoying vibrations of timber floors can be effectively
controlled through a design approach that accounts for
stiffness and mass. The choice of appropriate structural
arrangements and detailing is also important in order to
achieve good results.

REFERENCES
[1] Pavic A., Reynolds P: Vibration serviceability of
long-span concrete buildings floors. Part 1: Review
of background information. The Shock and
Vibration Digest, 34(3):191-211, 2002.
[2] Hu L. J., Chui Y. H., Onysko D. M.: Vibration
serviceability of timber floors in residential
construction. Progress in Structural Engineering
and Materials, 3:228-237.
[3] Smith I.: Vibration of timber floors: serviceability
aspects. In: Thelandersson S. and Larsen H. J. (eds)
Timber Engineering, Johan Wiley & Sons Ltd,
Chichester, England, 242-266, 2003.
[4] Newland D. E.: Pedestrian excitation of bridges recent results. In: 10th International Congress on
Sound and Vibration, Stockholm, Sweden, 2003.
[5] Ellingwood B. M., Tallin A.: Structural
serviceability: floor vibrations. Journal of Structural
Engineering, 110(2):401-418, 1984.
[6] Ohlsson S. V.: Serviceability limit states - vibration
of wooden floors. In: STEP 1, Centrum Hout, The
Netherland, A18/1-A18/8, 1995.
[7] Bodig J., Jayne B. A.: Mechanics of wood and wood
composites. Van Nostrand Reinhold Company, New
York, 1982.
[8] Filiatrault A., Folz B., Foschi R. O.: Finite-strip
free-vibration analysis of wood floors. Journal of
Structural Engineering, 116(8):2127-2141, 1990.
[9] Al-Faqahaa A. A., Cofer W. F., Fridley K. J.:
Vibraion design criterion for wood floors exposure
to normal human activities. Journal of Structural
Engineering, 125(12):1401-1406, 1999.
[10] Hu L. J., Desjardins R., Chui Y. H.: Nature of
vibrations induced by footsteps in lightweight and
heavyweight floors. In: 9th World Conference of
Timber Engineering, Portland, USA, 2006.
[11] ISO 2631-2:1989. Evaluation of human exposure to
whole-body vibration: Part 2: Continuous and
shock-induced vibration in buildings (1 to 80 Hz).
International Organisation for Standardization,
Geneva, Switzerland, 1989.

[12] Smith I., Chui Y. H.: Design of lightweight wooden


floors to avoid human discomfort. Canadian
Journal of Civil Engineering, 15:254-262, 1988.
[13] Smith I., Chui Y. H.: Construction methods for
minimizing vibration levels in floors with lumber
joists. Canadian Journal of Civil Engineering,
19:833-841, 1992.
[14] Foschi R. O., Gupta A.: Reliability of floors under
impact vibration. Canadian Journal of Civil
Engineering, 14:683-689, 1987.
[15] Chui Y. H., Smith I.: Dynamic response of lightweight floors with wood I-joists. Proceeding of the
1991 International Timber Engineering Conference.
London, UK, Vol. 4, 4.378-4.385, 1991.

[16] Chui Y. H.: Evaluation of vibration performance of


light-weight wooden floors. Proceeding of the 1988
International Conference of Timber Engineering.
Washington State University, Seattle, USA, Vol. 1,
707-715, 1988.
[17] Ohlsson S. V.: A design approach for footstepinduced floor vibration. Proceeding of the 1988
International Conference of Timber Engineering.
Washington State University, Seattle, USA, Vol. 1,
722-729, 1988.
[18] Weckendorf J. et al.: Assessment of vibrational
performance on timber floors. In: 9th World
Conference of Timber Engineering, Portland, USA,
2006.

Anda mungkin juga menyukai