Content uploaded by Stephen Dance
Author content
All content in this area was uploaded by Stephen Dance on Aug 04, 2015
Content may be subject to copyright.
ICSV20, Bangkok, Thailand, 7-11 July 2013 1
NOISE REDUCTION IN COMMERCIAL
REFRIGERATORS – A PRACTICAL APPROACH
A. C. Marques, L. Gomez-Agustina, S. Dance, E. Hammond and I. Wood
Acoustics Group, Department of Urban Engineering, London South Bank University, 103
Borough Road, London, SE1 0AA, United Kingdom
Adande Refrigeration, 45 Pinbush Road, South Lowestoft Industrial Estate, Lowestoft, Suffolk,
NR33 7NL, United Kingdom
e-mail: marquesc@lsbu.ac.uk
An Adande refrigeration unit originally designed for use in the commercial catering industry
was redesigned for use in households. This sector is more sensitive to refrigeration noise, fol-
lowing the introduction of the EU noise labelling directive. A practical noise control ap-
proach was taken consisting of benchmarking the existing commercial unit, diagnosing the
primary noise sources, redesigning the system components without affecting the refrigeration
performance and assessing improvements. The aim was to reduce noise emissions and im-
prove sound quality to those of frost free household refrigerators. Value engineering was
used to optimise the performance gains such that the new unit suitable for the domestic mar-
ket would be also used in the commercial sector. The sound power reduction achieved was
greater than 4 dB. The sound quality of both the existing standard refrigerator and the opti-
mised prototype unit were evaluated by a jury in a real living environment. The subjective
exercise showed that the optimised prototype was perceived as being quieter and of improved
sound quality compared to the standard refrigerator.
1. Introduction
Noise pollution is considered an important issue in the modern world; it impacts on our daily
lives affecting our wellbeing, health and social behaviour.
In the household, appliances are the main noise sources and refrigerators in particular can be a
cause of annoyance due to their near constant operation. The EU has introduced a new energy label
in 2011 [1] which includes the mandatory declaration of the sound power level (Lw) of household
appliances. The test code and procedures for determining and verifying declared noise emissions
values of household appliances are covered in the EN 60704:2006 [2]. The EU noise labelling di-
rective is driving manufacturers to address appliance noise, since this is another factor of product
differentiation in the extremely competitive household market.
This paper presents a practical investigation on the redesign of a commercial refrigerator for
use in a domestic environment. The aim was to reduce noise emissions and improve the sound qual-
ity to those of frost free domestic refrigerators in order to supply the high-end household market.
20th International Congress on Sound and Vibration (ICSV20), Bangkok, Thailand, 7-11 July 2013
ICSV20, Bangkok, Thailand, 7-11 July 2013 2
According to the database of the European Committee of Domestic Equipment Manufacturer’s
(CECED), the Lw of refrigerator and freezers typically vary between 32 and 48 dBA [3]. Frost free
refrigerators have more noise sources than static appliances, e.g. evaporator and condenser fans,
making their Lw values higher than static refrigerators.
The Adande commercial unit is a patented frost free refrigerator with an insulated drawer sys-
tem that retains cold air when opened. The unique modular design can be stacked up to three draw-
ers high and has a variable temperature capacity from fridge to frozen storage conditions. A prelim-
inary trial in a domestic kitchen has demonstrated that the unit offers a comparable storage capabil-
ity and easier access to foodstuffs than conventional door cabinets. However, the unit was perceived
as being noisy; therefore noise reduction was considered a major requirement to reach the house-
hold market.
The noise reduction analysis started by benchmarking the existing commercial refrigerator
and diagnosing the primary noise sources. Novel components with lower noise ratings were then
selected and their position and mounting in the appliance were redesigned to achieve lower noise
and vibration levels, whist maintaining refrigeration performance.
The refrigerator noise emissions were investigated with the appliance working on “full mode”
(all components operating) and for each individual component. Finally, the sound perception of
both the existing standard refrigerator and the optimised prototype unit were evaluated by a panel of
jurors in a real living environment following an identical procedure to that reported by Baas et al.
[4] and Altinsoy et al. [5].
2. Benchmark of standard units
The benchmark consisted of measurements of Lw to EN ISO 3745:2012 [6] (Fig. 1) and
measurements of sound pressure levels (SPL) taken at the London South Bank University 200 m3
anechoic chamber. SPL measurements were taken for diagnosis and comparative purposes at a ref-
erence receiver position located at 1.2 m from the chamber grid floor and 1.5 m away from the out-
let of the engine compartment (Fig. 2). This reference position was considered a representative lis-
tener position. The refrigerator has two operating stages: “early mode” where the only sources of
noise are two small fans placed in the unit lid next to the cooling evaporator and a “full mode” stage
where the compressor and condenser fan are also switched on (these two components are located in
the engine compartment). Both Lw and SPL were measured for the refrigerator working on “full
mode” and “early mode”. Additionally, the measurements were repeated for a double unit, consist-
ing of two single refrigerators units stacked (Fig. 2).
Figure 1: Single unit in the anechoic chamber Figure 2: Double unit in the anechoic chamber
20th International Congress on Sound and Vibration (ICSV20), Bangkok, Thailand, 7-11 July 2013
ICSV20, Bangkok, Thailand, 7-11 July 2013 3
Fig. 3 shows the benchmark Lw for the single and double unit at both operation modes. Diag-
nostic testing was carried out to determine the main contributors to the overall unit noise (Fig. 4)
and obtain representative SPL which can be correlated to sound perception scores.
Figure 3: Sound power level spectrum of the Figure 4: Standard unit SPL spectrum of each
Standard unit individual component and full mode
The measured sound power level for a single unit was 50.4 dBA, working on full mode and
41.6 dBA on the early mode. For the double unit the sound power increased to 54.1 dBA (full
mode) and 43.1 dBA (early mode).
As can be seen in Fig. 4 the condenser fan is the main contributor to the overall unit noise.
The evaporator fans show two frequency peaks at 80 – 100 Hz and at 160 - 400 Hz, whilst the com-
pressor produces only significant energy in a band between 200 – 315 Hz.
3. Optimisation techniques
3.1 Noise reduction
Table 1 present the specification of the standard unit components and the proposed alterna-
tives for a noise reduction prototype.
Table 1: Components specification of standard unit and proposed prototype unit
Component Standard unit Prototype unit
Compressor HFC refrigerant model (back) HC refrigerant model (forward)
Condenser fan Speed: 2500 rpm
Size: 120×120×40 mm (W×H×D)
Speed: 900 rpm
Size: 260×242×60 mm (W×H×D)
Evaporator fans Speed: 8200 rpm
Size: 40×40×20 mm (W×H×D)
Speed: 8200 rpm
Size: 40×40×20 mm (W×H×D)
The standard unit uses a HydroFluoroCarbon (HFC) refrigerant; the prototype unit was
charged a hydrocarbon (HC) refrigerant, which is the norm in household refrigerators. The conden-
ser fan was replaced in the prototype unit by a slower model with four times the surface area of the
standard unit fan. In order to fit the larger condenser fan the position of the components in the com-
partment was rearranged, Fig. 5 and 6 show the standard and prototype engine compartments re-
spectively.
20th International Congress on Sound and Vibration (ICSV20), Bangkok, Thailand, 7-11 July 2013
ICSV20, Bangkok, Thailand, 7-11 July 2013 4
Figure 5: Standard unit engine compartment Figure 6: Prototype unit engine compartment
Fig. 7 and 8 compare the frequency spectrum of the standard and prototype condenser fan and
compressor respectively.
As can be seen in Fig. 7 replacing the condenser fan by an optimised model resulted in a sig-
nificant SPL reduction at all frequencies above 80 Hz which dropped the overall dBA value by 10.9
dB. Fig. 8 shows that the compressor in the prototype unit peaks significantly at 80 - 100 Hz, caus-
ing an increase in the overall SPL (A) of 1.2 dB. It is currently being investigated if the new com-
pressor location excites resonant modes of the base and or side panels.
In an attempt to minimise sound radiation from the diffuser plate where the evaporator fans
are mounted an analysis of alternative fans was carried out in terms of rotation speed and airflow
capacity. Due to the lack of alternative fans with similar airflow and pressure drop, it was decided
to maintain the current 40 mm evaporator fans and apply optimised vibration isolation mounts. The
effectiveness of the optimised mounts can be seen in Fig. 9 for the prototype unit. Fig. 10 compares
the SPL frequency spectra for both the standard and prototype unit working on full mode.
Condenser
Compressor
Condenser fan
Condenser fan
Compressor
Figure 7: Condenser fan frequency spectrum Fi
g
ure 8: Com
p
ressor fre
q
uenc
y
s
p
ectrum
Figure 9: Evaporator fans frequency spectrum Figure 10: Unit on full mode – frequency spectrum
20th International Congress on Sound and Vibration (ICSV20), Bangkok, Thailand, 7-11 July 2013
ICSV20, Bangkok, Thailand, 7-11 July 2013 5
Fig. 9 shows that the optimized vibration isolation caused a shift on the frequency peaks from
100 Hz to 160 Hz and from 400 Hz to 500 Hz; the reduction in SPL (A) due to the optimized
mounts was 3 dB. Comparing the standard and prototype units on full mode shows that there was an
overall SPL (A) reduction of 5.5 dB in the prototype. The SPL reduction at frequencies above 800
Hz was due to the use of 50 mm slabs of acoustic foam lining the engine compartment walls.
3.2 Vibration reduction
Changing the position of the compressor in the engine compartment has resulted in significant
vibration of the refrigerator stainless steel panels. Several measures were undertaken to reduce the
vibration levels of the appliance. These included increasing the copper pipe length between the
compressor and condenser, fitting vibration isolation mounts in the L-shape bracket that holds the
condenser to the refrigerator chassis, stiffening the side panel with L-shaped aluminum strips and
soft mounting the compressor to the unit base plate. These measures combined significantly reduced
the appliance vibration to barely perceptible levels.
4. Benchmark of the prototype units
The prototype unit benchmark sound power spectra are presented in Fig. 11. The overall Lw
measured for a single prototype unit working on full mode was 46.3 dBA, which corresponds to a
4.1 dB reduction compared to the standard unit. For the double unit the overall Lw was reduced
from 54.1 dBA (standard) to 50.8 dBA (prototype).
There was a slight increase on the early mode sound power in the prototype caused by the
evaporator fans peak at 160 Hz. Fig. 12 compares SPL of the standard and prototype units measured
in a real living environment at 1.5 m from the unit.
Figure 11: Sound power levels of prototype unit Figure 12: SPL of standard vs. prototype in a real
living environment
The SPL measurements shown in Fig. 12 demonstrate that there was a significant improve-
ment in the prototype unit for frequencies above 125 Hz compared to the standard unit. The back-
ground noise in the kitchen was 25.4 dBA (LAeq,5 min), the standard unit SPL was 53.1 dBA whilst
the prototype was 43.9 dBA.
By further optimizing the evaporator fan vibration isolation mounts an additional 10 dB re-
duction in the 160 Hz band can be achieved, reducing the humming tone. This would make the peak
inaudible. It is expected that the compressor peaks at 50 and 80 - 100 Hz will be reduced by damp-
ening modal resonances of the base plate and improving vibration isolation between the compressor
20th International Congress on Sound and Vibration (ICSV20), Bangkok, Thailand, 7-11 July 2013
ICSV20, Bangkok, Thailand, 7-11 July 2013 6
and the unit base plate. These two measures combined would yield a further drop of 1 dB in the Lw,
reducing the overall Lw by 5 dB compared to the standard unit.
Fig. 13 and 14 illustrate the prototype unit sound directivity on early mode (just evaporator
fans) and full mode, respectively.
Both the prototype early mode and full mode sound directivity highlight that the evaporator
fans are now the main noise contributor as indicated by the higher SPL magnitude of the at 125 -
160 Hz peak (Fig. 12 and Fig. 13). The evaporator fans are mounted on a stiff and light plastic dif-
fuser plate, which distributes the air cooled by the evaporator into the drawer compartment. Future
work will include modelling the diffuser panel to find and attenuate suspected structural modes re-
sponsible for the high sound radiation efficiency of this panel.
Figures 13 and 14 show that the sound directivity of the prototype unit in the early mode is
more directional toward the front than in the full mode. The full mode presents a rather uniform
radiation in all directions at almost all frequency bands with a slight tilt towards the 90°, which is
where the engine compartment (compressor and condenser fan) is located.
5. Sound perception assessment
Sound power level is a convenient measure to compare appliances since it is independent of
the measurement distance and acoustic properties of the room, but to achieve product differentiation
perceived sound quality is very important. The psychoacoustic factors of refrigerator noise in the
real living environment need to be taken into account to develop a product with a high consumer
acceptance.
The subjective sound perception of both the standard and prototype units was evaluated by a
jury in a real living environment. The jury was composed by fifteen people with normal hearing
abilities. The average age was 46 years, which is the target age for consumers buying premium
range appliances. The assessment was carried out in a furnished kitchen of 41 m3 volume with most
of its surfaces being acoustically hard. The single standard and prototype units were presented in
turns (there was no obvious distinction between their exterior appearance) and assessed individually
at distance of 1.5 m away from the engine compartment outlet (as measured during diagnosis).
The questionnaire presented to each jury member was divided into three questions. The first
question (Fig. 15) concerned the noise produced by the unit, and can be related to loudness and
therefore to measured SPL(A) (reported in Fig. 12). The second question (Fig. 16) addressed an
aspect of sound quality and the third question (Fig. 17) attempted to describe the noise character
produced by each unit (this can be related to measured SPL spectrum, Fig. 12).
Figure 13: Prototype sound directivity – Early mode Figure 14: Prototype sound directivity – Full mode
20th International Congress on Sound and Vibration (ICSV20), Bangkok, Thailand, 7-11 July 2013
ICSV20, Bangkok, Thailand, 7-11 July 2013 7
As can be seen in Fig. 15, 47% of the jury considered the standard unit noisy; whist 40%
rated the prototype unit quiet. This result correlates well with the measured SPL (A) reduction of
9.3 dB obtained for the prototype unit compared to the standard unit. Fig. 16 presents the jury sound
quality perception of both units.
Fig. 16 shows that the prototype unit sound was perceived as being more acceptable and also
less annoying when compared to the standard unit. These results can be clearly attributed to the
overall smoother spectrum (shown in Fig. 12) and lower SPL (A) measured. The third question de-
scribed the perceived character of noise produced from the units and the results are presented in Fig.
17.
Figure 15: Comparison of subjective noise rating from standard and prototype units
Figure 16: Comparison of sound quality for the standard and prototype units
Figure 17: Comparison of sound description for the standard and prototype units
20th International Congress on Sound and Vibration (ICSV20), Bangkok, Thailand, 7-11 July 2013
ICSV20, Bangkok, Thailand, 7-11 July 2013 8
The results were identical for both appliances with an average even sound; however the pro-
totype was slightly more humming, which could be related to the evaporator fans peak at 160 Hz
and higher level compressor peaks at 50 Hz and 80 Hz.
Conclusions
Noise and vibration control techniques were applied to the Adande commercial refrigerator in
order to reduce its noise emissions and improve sound quality to those of household frost free re-
frigerators. A reduction of 4.1 dB of Lw (A) and 9.3 dB of SPL (A) was achieved by applying noise
and vibration isolation, rearranging the components position and replacing the initial main noise
source (condenser fan) by an improved design scheme. The optimised Adande single unit sound
power level is 46.3 dBA and it has been shown that it be further reduced to 45.3 dBA by decreasing
the 160 Hz peak to inaudibility and therefore enhancing the overall sound quality of the appliance.
The standard Adande single unit had a sound power level of 50 dBA and that is now the overall
sound power of two optimised units.
A sound perception assessment performed on both the standard and optimized prototype units
has shown that most respondents considered the standard unit noisy and the prototype unit quiet and
significantly less annoying. These results show a good correlation with the significant overall SPL
(A) reduction of 9.3 dB.
The sound quality of the refrigerator will be further enhanced in the future by reducing radia-
tion efficiency of the plastic air diffuser and reducing the aerodynamic noise by optimising the
evaporator fans airflow.
REFERENCES
1 Supplementing Directive 2010/30/EU of the European Parliament and of the Council with regard to
energy labelling of household refrigerating appliances. Official Journal of the European Union, Com-
mision Delegated Regulation (EU) No 1060/2010.
2 EN 60704:2006 Household and similar electrical appliances - Test code for the determination of air-
borne acoustical noise – Procedures for determining and verifying declared noise emission values.
3 AEA Energy & Environment (2008). Discussion Report: EU Ecolabel for Refrigeration. [Online.]
available:
http://ec.europa.eu/environment/ecolabel/ecolabelled_products/categories/pdf/discussion_refrigeration
.pdf
4 Baars, E., Lenzi, A., Nunes, R. A. S. Sound quality of hermetic compressors and refrigerators, Pro-
ceedings of the 16th International Compressor Engineering Conference. West Lafayette, USA, July,
(2002).
5 Altinsoy, E., Kanca, G., Belek, H. T. A comparative study on the sound quality of wet-and-dry type
vacuum cleaners, Proceedings of the 6th International Congress on Sound and Vibration. Copenhagen,
Denmark, 5 – 8 July, (1999)
6 EN ISO 3745:2012 Acoustics - Determination of sound power levels and sound energy levels of noise
sources using sound pressure - Precision methods for anechoic rooms and hemi-anechoic rooms.
