Flanking transmission in light weight timber houses with elastic
flanking isolators
Anders Ågrena
Fredrik Ljunggrenb
Division of Engineering Acoustics, Luleå University of technology, SE-97187 Luleå, Sweden,
Åsa Bolmsvikc
School of Engineering, Dep. of Timber Engineering, Linneaus University, SE-35195 Växjö,
Sweden
Kirsi Jarneröd
SP Wood Technology, SP Technical Research Institute of Sweden, SE-35196 Växjö, Sweden
There is a strong trend to industrially produce multi-storey light weight timber based
houses. This concept allows the buildings to be manufactured to a more or less
prefabricated extent. Most common types are volume/room modules or flat wall and floor
modules. When assembling the modules at the building site, elastomer isolators are used in
several constructions to reduce flanking transmission. The sound insulation demands in the
Nordic countries are relatively high and therefore the flanking transmission must be well
controlled, elastomer isolators is one choice. Decoupled radiation isolated walls is another.
There are though no working studies or mathematical models of the performance of these
isolators. They are only treated as simple mass-springs systems that operate vertically. In
this paper there is a first approach of an analysis of the structure borne sound isolating
performance of elastomer isolators that are separating floors and walls. The performance
dependence of structure type is also presented. An empirically based regression model of
the insertion loss is derived. The model will be based on measurements of more than ten
field installations. In this paper the first three are presented. A goal is that the model can
be used for input in future SEN prediction models for modeling of sound insulation.
a)
email: anders.agren@ltu.se
email: fredrik.ljunggren@ltu.se
c)
email: asa.bolmsvik@lnu.se
d)
email: kirsi.jarnero@sp.se
b)
1
INTRODUCTION
Flanking transmission or flanking isolation between floors in light timber houses is
traditionally improved by adding additional wall plates with distances on the sending room
and/or on the walls of the receiving room. Other ways are to add a resilient layer on the floor
board or hang the ceiling in resilient channels. A further method is to add damping on the floor
board in order to reduce vibrations from impacts to reach the walls /Ljunggren/. A recent method
is to put the floor on elastomer isolators or to put the whole room on resilient isolators. This
method introduces both possibilities and new challenges to solve. This paper presents the first
results where the flanking transmission with this method is analyzed.
2
ELASTOMER FLANKING ISOLATORS – THE PRINCIPLE
The isolators are used as spring and damper between the loading structure element, which
in the presented examples can be either a floor package or the whole room volume including
floor, walls and ceiling, figure 1. The isolators that are being used are in case a) put as a strip
around the whole flank. In construction b) the elastomers are put as pads of 0.1x0,1 m under
each wall beam.
a)
Elastomer
strip
b)
c)
Figure 1 -. Elastomer flanking couplings for decoupling of floor from the supporting structure.
Construction A: a) Separate CLT floor and ceiling. Elastomer strips loaded by the floor
cassette.. b) Coupling for apartment separating walls. c) Outer wall with a CLT load bearing
wall plate and a radiation protection isolating layer.
Elastomer pad
Figure 2.- Elastomer flanking couplings for decoupling of floor from the supporting structure.
Construction B:. Elastomer isolator pads loaded by whole room volume units.
The isolators are treated and normally dimensioned as vertical point loaded mass-springisolators in one degree of freedom; although they in practice also act as shearing isolators in the
horizontal plane and that they are not point loaded. They are optimized by taking the mass
load/area of the construction unit and optimizing it with the stiffness of the damper. Varying
loads around the floor is taken into account where that is necessary, e.g. heavy bathrooms.
Typically the stiffness is chosen to give a resonance frequency of 13-17 Hz. The mechanical loss
factor is varying depending on which elastomer material is chosen, typical values presented by
manufacturers are 0,08 to 0,4 in the frequency range from 10 to 1000Hz..
3
ELASTOMER FLANKING ISOLATORS – AT LOW FREQUENCIES
The flanking coupling influences low frequencies as well as higher. In a study by
/Jarnerö/ it is shown experimentally in lab how the dynamics of a floor package is influenced by
the way it is supported. The test setup was done with a floor element from construction A of 1,5
x 5,1 m, which was supported by a line elastomer, figure 1, on glulam beams on the long sides
and on steel beams on the short sides. In figure 3 the driving point accelerance with various
elastic supports can be seen and compared to the floor freely lying on the supports and bolted to
the supports.
Figure 3.- Driving point accelerance of the floor with different isolators around the supporting
frame./Jarnerö/
Some results of this setup are that a floor supported on a line elastomer compared to a freely
lying or bolted support:
-
-
The first bending mode is increased with increased elastomer stiffness, i.e. in this case
around 11, 13, 14 and 18 Hz. While the first bending mode of the freely lying and the
bolted floors were 20 and 21 Hz.
The frequencies of the modes 4-7 are somewhat reduced.
The damping of the first bending mode was increased. Approximately from 1,5% to 12%.
The damping of modes 4-7 are approximately increased by 2-3 times.
A note though, is that damping in a real structure will be clearly higher than in the lab
setup.
4
FLANKING TRANSMISSION BELOW 100 HZ – FE MODEL AND
MEASUREMENTS IN LAB
Measurements and FE calculations of the vibration transfer over the junction between a
floor and a supporting beam has been carried out by /Bolmsvik/on the same floor module from
construction A as mentioned in the previous chapter. The floor was positioned on a frame
structure with the floor as a load over a line elastomer on the supporting frame. In this study the
focus was on the vibration insertion loss from the floor to the supporting frame at the side that is
perpendicular to the beams. In this operation mode the elastomer will not work in the vertical
direction which it normally is dimensioned for, but will instead work in the horizontal direction
in a shear motion.
117
119
112
110
Figure 4.- Insertion loss from the floor to the wall beams in the vertical direction, relative to a
freely lying floor./Bolmsvik/
IL point 110 and 112x xdirection relative free conditions rel 5e8 ms2
Insertion Loss [dB10 rel 5e-8 m/s2]
50
IL
IL
IL
IL
IL
40
30
pink 110o112x
orange 110o112x
green 110o112x
purple 110o112x
blue 110o112x
20
10
0
-10
-20
-30
0
10
20
30
40
50
60
Frequency [Hz]
70
80
90
100
Figure 5. - Insertion loss from the floor to the wall beams from z- to x-direction, relative to a
freely lying floor./Bolmsvik/
IL point 110 and 112x xdirection relative fixed conditions Exp and FE
Insertion Loss [dB10 rel 5e-8 m/s2]
50
IL pink 110 and 112x Exp-measurements
IL pink 110 and 112x FE-analysis
40
30
20
10
0
-10
-20
-30
0
10
20
30
40
50
60
Frequency [Hz]
70
80
90
100
Figure 6. - Measured and FE-modeled vibration insertion loss in point 110 and 112 from z- to
x-direction relative to bolted conditions /Bolmsvik/..
Some results of this setup, figure 4, 5, 6, are that this type of floor supported on a line elastomer
compared to a floor bolted to the support:
5
The vibration insertion loss from vertical floor vibrations to horizontal vibrations in the
wall beam will be zero or even negative in the frequency range 35-65 Hz
The vibration insertion loss will reach around 25 dB at 15-25 Hz.
The vibration insertion loss will reach 10-20 dB at 75-100Hz
MEASURED VIBRATION INSERTION LOSS IN SITU
A simplified field method for approximating the flanking transmission is used in
AkuLite. The simplified method is chosen as it is a field method with time limitations and as a
large number of objects will be measured. Six accelerometers are mounted; on the floor along a
line 0,2 m from the wall, on the ceiling along a line 0,2 m from the wall and on the receiving
room wall 0,2 m from the ceiling, figure 7. The vibration level difference between floor and wall
is then calculated to approximate the flanking transmission, and the difference between floor and
ceiling to show the direct transmission. This is a simplified method and is not expected to give
the same precision of flanking transmission measurement as when the vibration level over the
whole receiving wall is measured /Villot et.al./ or when surfaces are covered as in /King et.al./.
Response
Source
Beam direction
0,2m
Figure 7 . - Accelerometer positions for the simplified flanking transmission field measurements.
5.1 House type A – Elastomer isolation in a cross laminated timber plate type house
In house type A the flanking transmission over to the two walls represents two very
different cases. The receiving room wall parallel with the beams is an inner wall and the wall
perpendicular to the beams is an outer wall. The outer wall takes a larger load from the building
above compared to the inner wall and thus has a stiffer elastomer, see figure 2 and 4.
A very clear difference is seen between the two directions. The constructions are quite
different and must therefore be treated as different cases. The elastomer properties, the static load
as well as the construction of the inner wall are different. The inner wall /receiving wall has a
clearly lower insertion loss than the load bearing outer wall, figure 8. At frequencies below 80
Hz it is practically zero.
70
60
50
dB (acc)
40
TL floor to wall,
perpendicular
to beams
30
20
TL floor to wall,
parallel w
beams
10
0
-10
Frequency
Figure 8. - Comparison of vibration level difference over the flanks from floor to wall in two
directions, perpendicular to and parallel with floor beams. i.e. outer wall and inner wall.
Construction type A, CLT with elastomer.
5.2 House type B – Elastomer isolation and traditional beam plate construction in volume
modules
In this case the two receiving walls are identical. The isolation is designed as of 0.1 x
0,1m elastomer pads, placed on top of a horizontal plank above the wall studs. The load
distribution becomes different in the load bearing direction perpendicular to the beams compared
to parallel with the beams. A higher load is distributed along the beams to the walls in that
direction. In figure 9 it can be seen that the flanking isolation is in the frequency range 80 to
1000 Hz is weaker in the direction parallel with the beams. The reason for this unsure and has to
be further examined.
50
40
TL floor to wall,
perpendicular to
beams
30
20
TL floor to wall,
parallel w beams
10
3150
2000
Frequency (Hz)
1250
800
500
315
200
125
80
50
31.5
20
0
12.5
Vibration level difference (dB)
60
Figure 9. - Comparison of vibration level difference over the flanks from floor to wall in two
directions, perpendicular to and parallel with floor beams. Construction type B.
5.3 House type C – Flanking transmission isolation though radiation reduction from
additional wall layers
House construction C is a concept without elastomer isolators. Instead a radiation
isolation additional wall without any direct contact with the load bearing walls is used. The layer
comprises 50 mm mineral wool and two layers of plaster board and is connected through metal
channels to the ceiling and the floor. In figure 10 it can be seen that, in contradiction to the
elastomer constructions, the flanking transmission is quite independent of direction.
70
60
50
TL floor to wall,
perpendicular to
beams
TL floor to wall
parallel w beamsl
40
30
20
10
3150
2000
1250
800
500
315
200
125
80
50
31.5
20
0
12.5
Vibration level difference
(dB)
80
Frequency (Hz)
Figure 10. - Vibration level difference over the flanks from floor to wall. Comparison of
perpendicular to the beams wall and parallel with the beams wall. Construction type C.
As a comparison of direct path transmission towards flanking transmission a comparison
of vibration level difference from floor to walls compared to floor to ceiling is shown in figure
11. With the reservation that only the vibration levels close to the walls are measured, the direct
transmission is dominating by about 10 dB above 100 Hz while the direct vibration transmission
Vibration level difference
(dB)
is very strong in the 20 -40 Hz range, where e.g. the resonance frequency of the suspended
ceiling is.
3150
2000
1250
800
500
315
200
125
80
50
31.5
20
TL Floor to
wall
TL Floor to
ceiling
12.5
80
70
60
50
40
30
20
10
0
-10
-20
Frequency (Hz)
Figure 11. - Comparison of vibration level difference over the flanks from floor to wall and from
floor to ceiling. Non-Load bearing side, parallel w floor beams. Construction type C.
5.4 Conclusions of vibration difference measurements over the flanks
6
The difference in flanking transmission from floor to wall in the receiving room is larger
in the elastomer constructions.
The flanking transmission is clearly dependent on the receiving wall construction.
It can as I s expected clearly confirmed that there is a strong direct transmission over the
floor to the ceiling at frequencies below 40Hz.
It is also seen that elastomer isolation can have zero isolation due to shear resonances at
low frequencies.
Elastomer support of floor will increase the damping of the floor modes.
EMPIRICALLY BASED LINEAR REGRESSION
FLANKING TRANSMISSION ISOLATION
APPROXIMATIONS
FOR
There is a need for data of flanking transmission for prediction methods like SEN 12354.
One of the objectives with the measurements in the AkuLite project is to deliver empirically
based statistics with which the flanking transmission can med approximated. Based on the results
above expression of vibration insertion loss through the vibration level difference have been
derived, one for the transmission along the loadbearing floor beams to the walls perpendicular to
the beams and one for the side that is parallel with the beams.
Figure 12a shows two regression expressions for the CLT-based construction C. The
different walls give rise to also very different expressions.
In figure 12b the regression lines for the volume based construction C two different floor
to wall transmissions can be seen. The more pronounced differences in load bearing directions
can be one of the reasons behind the differences.
Vibration level difference dB
(acc)
70
60
50
TL floor to wall
perpendicular
to beams
40
30
TL floor to wall
parallel w
beams
20
10
0
-10
Frequency
Linear regression:
Direction perpendicular to beams 1) y = 1,4x + 9,5, R2 = 0,73
Direction parallel w beams
2) y = 1,6 x -3,4, R2 = 0,83
Vibration level difference
(dB)
Figure12a.- Construction type A, CLT floor plate, glulam beams and elastomer isolation.
Linear regression lines for flanking vibration level difference in two directions over the flanks to
an outer CLT wall (red) and inner wall (blue).
60
50
40
TL floor to wall,
perpendicular to
beams
TL floor to wall,
parallel w beams
30
20
10
Linear regression:
3150
Frequency (Hz)
2000
1250
800
500
315
200
125
80
50
31.5
20
12.5
0
Direction perpendicular to the beams 1) y = 1,7 x + 11, R² = 0,92
Direction parallel with the beams
2) y = 1,4x + 10, R² = 0,7
Figure 12 b. - Construction type B, Plaster boards, glulam beams, volume construction and
elastomer isolation. Linear regression lines for flanking vibration level difference in two
directions over the flanks.
TL floor to wall,
perpendicular to
beams
TL floor to wall
parallel w beamsl
3150
2000
800
500
1250
Frequency (Hz)
315
200
125
80
50
31.5
20
Avg TL floor to
walls
12.5
Vibration level difference
(dB)
80
70
60
50
40
30
20
10
0
Linear regression on average of two couplings: y = 2,2x - 0,56
R² = 0,89
Figure 12 c. - Construction type C, board, glulam beams, uncoupled inner walls. Linear
regression for the average flanking vibration level difference of two directions over the flanks.
In figure 12c the regression lines for radiation isolated construction C in two different
floor to wall transmissions can be seen. Construction B and C are rather similar in the way that
they use about the same construction elements and dimensions. It is interesting therefore to
compare them as two related alternatives. The radiation isolation layers give almost identical
flanking isolation behavior in both directions. A disadvantage can be seen at low frequencies
where the flanking isolation is 5 dB worse than construction B.
7
CONCLUSIONS
This work is a first step with the objective to simplified characterize and give prediction
expressions for determination of impact sound flanking transmission in light weight timber
constructions. These three measurements will be followed by approximately six more, which will
give a better statistical trust in the estimations. Further studies will also be made on how to
correct for assumed biases caused by parquet top layers and by the distance to the walls.
The results show that in real constructions where walls will be loaded with different loads
and where wall constructions vary, that there will also be large variations in flanking
transmission in the different directions.
Elastomer isolators show preliminary in this study to give larger variations between
directions than a solution with radiation isolation layers.
Supporting a floor by elastomer line strips changes the fundamental vibration modes of
the floor. The frequencies are reduced and damping is increased. The relative importance of this
for the vibration and impact sound pressure in situ is so far not clearly known.
8
AKNOWLEDGEMENTS
This work has been sponsored by Vinnova and Formas in the AkuLite research program and by
TCN (Wood Centre North) together with timber construction industry.
9
REFERENCES
Jarnerö K., Bolmsvik Å., Brandt A., Olsson A., Effect of flexible supports on vibration
performance of floors. EuroNoise 2012, Prague
Bolmsvik Å., Linderholt A., Jarnerö K. FE modeling of a light weight structure with different
junctions. EuroNoise 2012, Prague
Ljunggren F., Ågren A., Potential solutions to improved sound performance of volume based
lightweight multi-storey timber buildings. Applied acoustics 72 (2011) 231-240
Homb A., Austnes J.A. Experiences with sound insulation for cross-laminated timber floors.
BNAM (2010), Bergen , Norway.
King F., Schoenwald S., Sabourin I. Characterizing flanking transmission paths in the NRC-IRC
flanking facility. NRC report. NRCC – 51389. (2009)
King F., Schoenwald S., Govern B.N. Effect of some floor-ceiling construction changes on
flanking transmission. NRC report. NRCC – 53618. (2010)
Villot M., Guigou-Carter C. Prediction method adapted to lightweight constructions and
related laboratory characterizations. Forum Acusticum (2005). Budapest.
Table 1 . Summarized linear regression insertion loss relative to frequency for the three
constructions.
Construction
Frequency range
10 Hz > f < 31,5 Hz
>31,5 Hz
Insertion loss
Approximation
10 dB
4,3 dB/octave
10 Hz > f < 31,5 Hz
>31,5 Hz
10Hz > f < 25 Hz
>25 Hz
0 dB
5 dB/octave
10-12 dB
4,3 dB/octave
wall 2)
10Hz > f < 25 Hz
10-15 dB
C, Plaster boards, glulam beams,
radiation isolation wall layers
>25 Hz
10 Hz > f < 40 Hz
>40 Hz
A, CLT plate glued to glulam T
beams. Freely suspended ceiling,
wall 1)
D:o
wall 2)
B, Plaster boards, glulam beams,
volume elements,
wall 1)
D:o
5,2 dB/octave
8 dB
6 dB/octave