Sound transmission in buildings. Flanking transmission. 337
section). We shall, as above, not reproduce Mechel’s theory, just present one or two
examples.
In addition, it might be interesting to compare these results with predictions based
on a simpler model based on the diffuse field approach combined with a treatment of the
plenum as a lined duct. In reality, this concerns two different models developed by
Mechel (1980), of which we shall just present the one-dimensional variant.
The transmission factor for the transmission path by way of the ceiling and the
plenum may be defined as
(^) cl R,d
S,d
,
W
W
τ = (9.11)where WS,d and WR,d are the diffuse sound power incident on the ceiling in the sending
(source) room and the radiated sound power from the ceiling in the receiving room,
respectively. An alternative definition could also be used by referring to the power
incident on the partition. This may be a more suitable definition if the task is to add
contributions from several transmission paths. In this case; referring all transmission
factor to a common surface area, e.g. the surface area of the partition, we may directly
add the different transmission factors. The relationship between these alternative
definitions of the transmission factor is
(^) cl,p cl S
S
,
L
H
ττ=⋅ (9.12)where the extra index p indicates reference to the partition. The quantities LS and HS are
the length and height of the sending room, respectively (see Figure 9.10).
Correspondingly, the relationship between the reduction indexes will be
(^) cl,p cl S
S
10 lg.
H
RR
L
=+ (9.13)
9.2.3.1 Undamped plenum (cavity)
We shall start by showing a comparison between measured and predicted results with the
aid of Mechel’s (1995) modal theory, where the suspended ceiling is made of 9.5 mm
plasterboard and no absorbers in the cavity. Mechel does point to the fact that it is often
difficult to find measured results containing sufficient specifications for materials and
dimensions. He is therefore using data collected from several sources but even so, some
data have to be estimated as well. Another problem concerning the sound reduction
indexes is the reference surface chosen for the incident power. As discussed above,
different definitions may be used. We shall, in each case, point to the definition used.
Another important assumption made in the prediction models is that there is a
structural break both in the ceiling (and in a prospective absorber) above the partition.
This implies that there is no direct coupling between the ceilings in the two rooms; there
is no flanking transmission according to our strict definition. Whether this assumption is
valid in practical cases is open to discussion but suspended ceilings are often an
assembly of smaller units that may result in a less stiff structural coupling.
Measured and predicted results for the sound reduction index Rcl are shown in
Figure 9.11, where the length of the ceiling is the same in both rooms, i.e. LS and LR are
equal to 4.75 metre (see Figure 9.10). As stated above, the ceiling is made of 9.5 mm