2022-28-0439 Published 09 Nov 2022

An Experimental Study for Reduction of

Refrigerant Flow Induced Noise in a Passenger
Vehicle
Rachit Sharma, Mukesh Kumar, Shubham Kumar, and Subhashree Patra Maruti Suzuki India Limited

Citation: Sharma, R., Kumar, M., Kumar, S., and Patra, S., “An Experimental Study for Reduction of Refrigerant Flow Induced Noise in a
Passenger Vehicle,” SAE Technical Paper 2022-28-0439, 2022, doi:10.4271/2022-28-0439.

Received: 03 Aug 2022 Revised: 12 Sep 2022 Accepted: 26 Sep 2022

Abstracts
This paper elaborates about the study conducted on a

A
ir conditioning these days has become an indispens- HVAC system mounted on a calorimetric bench in a semi-
able part of human comfort due to rising global anechoic chamber to understand the noise signatures and
temperatures. In order to achieve thermal comfort in acoustic characteristics of refrigerant induced noise. This
confined environments like residences, car passenger cabins, research investigates potential causes and solutions for noise
offices, etc., air conditioners are used. As the air conditioning generated by refrigerant flow. Additionally, using a typical
units employ dynamic processes to maintain thermal comfort NVH study setup, it assesses several mitigation strategies that
it creates many unwanted noises which lower the acoustic have been investigated to eliminate or reduce noise caused by
comfort. One of these unwanted noises is the refrigerant flow refrigerant flow. This study also tries to categorize and evaluate
induced noise inside passenger cabin of an automobile when the decline in sound quality caused by the noise that is contin-
the air conditioning is switched on in a thermally uously generated by refrigerant-f low inside a car's
soaked vehicle. passenger compartment.

1. Introduction
FIGURE 1 Various noise sources in an automotive car.

A
s living standards rise, consumers are becoming more [Image source: Link - https://www.bksv.com/en/instruments/
conscious of the level of comfort provided by the production-testing-systems/discom/discom-mobile]
products/services they pay for. Because of the rising
temperatures around the world, achieving thermal comfort
has now become crucial. Summertime temperatures in India
can reach up to 50°C, making air conditioning essential for
achieving thermal comfort.
The HVAC system is one of the major sources of acoustic
discomfort in passenger cabin. Apart from HVAC System, the
major sources for acoustic discomfort for passenger vehicles
are shown in figure 1:
Thermal comfort is achieved through a variety of dynamic
mechanisms that are fundamental to car air conditioning
technologies. The engagement of these dynamic processes has
an impact on acoustic comfort in enclosed environments. The
various noise generated by the air conditioning system are
frequently heard within a car's passenger compartment. These
noises include compressor noise, airflow noise, HVAC motor
noise, clutch noise, noise caused by refrigerant flow, etc.
Because of its proximity to the occupant and high acoustic
sharpness, refrigerant flow induced noise can be very notice-
able. As Refrigerant flow noise is an abnormal noise, it is
preferred to eliminate it through various means and measures.
If, However, it is not possible to suppress the noise completely,
the noise levels should be kept as low as possible. This can
 2 AN EXPERIMENTAL STUDY FOR REDUCTION OF REFRIGERANT FLOW INDUCED NOISE

be judged by subjective observation and objective data capture The analysis for refrigerant flow induced noise was done
using standard NVH measurement modules. The nature of in a experimental setup where testing is done at a calorimetric
refrigerant noise is either continuous or transient in character. bench and other sources of noise are restricted to minimum
Most of the time, transient noise occurs over relatively brief inside a semi-anechoic chamber. We could analyze the noise
periods of time. As transient noise only occurs for extremely on standalone basis using standard colormaps and front-
brief periods, it is the constant noise that is alarming. The back graphs.
amount of subcooling in the condenser, pressure inside the
refrigerant circuit, ambient temperatures, vehicle thermal
soaking time, evaporator inlet pipe design, thermostatic
expansion valve design setting, etc. are some of the combina- 3. R
 efrigerant Flow
tions that contribute to refrigerant noise generation. In this
paper, the mechanism for generation of refrigerant flow
Induced Noise
induced noise is explored briefly through two phase flow
theory and bubble dynamics theory. Efforts are put to establish
Mechanism
a correlation of these theories with the experimental findings. The two-phase flow theory presents the most plausible expla-
The studies and techniques used to mitigate refrigerant nation for the occurrence of the refrigerant flow induced noise.
flow induced noise at the bench level are the main topics of A two-phase flow pattern is characterized by the simultaneous
this study. Additionally, it covers potential measures that movement of two separate media phases inside a closed region.
could be employed in this field of future research to reduce Figure 2 shows a two-phase flow inside a tube (liquid and
refrigerant noise. The noise quality signatures of the base and gaseous states):
upgraded samples are also discussed and compared in the
paper. Sharpness and loudness are the basic noise quality
metrics that are utilized for comparison. Standard NVH FIGURE 2 Two phase flow for separated flow model.
analysis software was used for signal processing and analysis.

2. S
 ignal Processing and
Analysis Methodology
Employed in Software In the refrigeration cycle, the high pressure and high
temperature liquid refrigerant undergoes throttling in the
expansion valve. When throttling is finished, the refrigerant
The noise amplitude is evaluated through measurement of
possesses dual phase (gas + liquid). Refrigerant flow induced
sound pressure level. The value of the sound pressure level is
noise is created when different flow patterns created by the
determined by the sound pressure and reference value as
system interact with the inner surfaces of the evaporator input
below equation [1].
pipe and the evaporator core. There are mainly four types of
P  flow patterns that can happen in closed tube as shown in figure
L  20log 10   dB (1) 3 [2] & table 1.
 P0  In 1977, Taitel and Dukler established that the refrigerant
Where, P0 =20 μ Pa which is the minimum threshold of flow induced noise is encountered in case the flow inside the
sound pressure audible to human ears. & P = measured rms tube is plug or churn or bubbly. From here it is only logical to
pressure fluctuation due to noise sources. say that the flow transition happens in below pattern as the
Such calculations are acceptable for noise analysis of refrigerant passes in the refrigerant circuit: [3]
single source of noise, However, in case there are multiple
noise sources we must isolate the noise sources relevant to our Liquid → Bubbly → Plug → Churn → Annular
problem area for effective analysis and comparison. The refrig- The noise is not present in case the flow is annular in
erant induced noise occurs in presence of various other noise nature. Hence the chances for refrigerant flow induced noise
sources such as engine idle running noise, compressor occurrence decreases with the refrigerant approaching the
running noise, HVAC airflow noise, compressor running suction pipe.
noise etc. Thus, if we consider a simple use case of vehicle idle
for analysis the generic formula used in such cases is mentioned
in below equation [1]
4. E
 xperimental
 L1 
Verification and
L2 L3 Lm
Ls  10 log10  10 10  10 10  10 10 10 10  dB (2)
 
 
Where, Ls = Overall sound pressure level generated due
Iterations
to all sources. Li = Sound pressure level generated due to ith The explanation makes it quite evident that the refrigerant in
source. And where i = 1,2,3,…n. the pipe is in a two-phase state, which is what creates the
 AN EXPERIMENTAL STUDY FOR REDUCTION OF REFRIGERANT FLOW INDUCED NOISE 3

FIGURE 3
closed tube.
Displays 4 types of flow patterns in a
4.1. T
 est Equipment and
Operating Conditions
The HVAC unit (with integrated TXV & Blower Motor),
Compressor, and Condenser make up the test unit setup as
shown in the figure 4. Here, the compressor (and the motor
operating the compressor) and condenser unit are positioned
outside the semi-anechoic chamber while the HVAC unit is
placed within. As the HVAC fan is running at the lowest
airflow setting by an external power supply, the noise genera-
tion by it is assumed to be constant. In this case, when the test
is performed on a typical AC cycle, the noise level is only
dependent upon the flow of refrigerant in the evaporator inlet
pipe & core assembly.
For measurement of Sound Pressure Level (SPL), two
microphones were installed at a standard measurement
distance from center of evaporator core child parts (side plate
and tank header) in vertical alignment as shown in figure 5.

FIGURE 4 A schematic setup for measuring noise caused

by refrigerant flow. (Image courtesy of gomechanic-blog-car
AC working)

TABLE 1 Various flow pattern in closed tube

Flow Type Definition

Bubbly The flow in which the gas bubbles are
approximately uniform in size.
Plug (Slug) flow The gas flows in the form of large, bullet-
shaped bubbles.
Churn Flow Highly unstable flow of an oscillatory nature.
The liquid near the tube wall continually
pulses up and down.
Annular Flow Like the horizontal flow.

refrigerant flow-induced noise in a normal automotive HVAC

system. Considering the thermodynamics cycle of the AC
FIGURE 5 Position of Microphones and Accelerometer on
system in the car, the region of the two-phase flow is predomi-
bench set-up. Microphone – MIC 1: At tank header; MIC 2: At
nantly in the evaporator-inlet, evaporator core and condenser.
side plate. Accelerometer – 3: At tank header, 4: At evaporator
As the condenser is placed outside the vehicle, the refrigerant
inlet pipe
noise impact on passenger cabin is miniscule. Therefore, the
main area of interest is the evaporator inlet hose and evapo-
rator core because the bubbles that develop collapse with the
internal child parts of pipes and evaporator core. They also,
merge and oscillate creating the distinct refrigerant induced
flow noise.
For verification of this hypothesis, AC system calori-
metric bench was setup inside a semi-anechoic environment
to emulate the exact conditions of vehicle passenger cabin.
The set-up can operate continuously at the typical air condi-
tioning cycle conditions where the refrigerant-induced noise
occurs. In the preceding sections, the estimated representation
of refrigerant noise is done basically in terms of its prominent
octave frequency bands and its source level acoustic ampli-
tudes. Then the experimental verification of theoretical
concepts is done to arrive to a solid countermeasure for reduc-
tion in refrigerant induced noise.
 4 AN EXPERIMENTAL STUDY FOR REDUCTION OF REFRIGERANT FLOW INDUCED NOISE

Additionally, two uniaxial accelerometers were also TABLE 2 Child parts with higher thickness proposal for
installed at inlet pipe and tank header to ascertain child part reducing noise induced by refrigerant flow
vibration pattern as shown in figure 5. SL. NO. Proposal Background
The test was done keeping various parameters such as the
1 Increased thickness of Increase in the
discharge pressure, inner room temperature, outer room Evaporator inlet/outlet thickness of evaporator
temperature, soaking time, airflow as constant to keep the AC pipes inlet pipe will increase
cycle as constant for evaluation of all countermeasures at the mass and stiffness
standard running condition. This would ensure there is no of pipe.
change in air-conditioning cycle when a particular iteration 2 Increased thickness of Thickness of the
is done and for homogeneity. evaporator side plate evaporator side plate
was increased to
increase its mass &
4.2. Experiments to Reduce stiffness.

Refrigerant Induced Flow

Noise To evaluate the effect of geometry of evaporator pipes on
refrigerant flow induced noise amplitude, below trials were
While maintaining standard test conditions, basic iteration planned as per table 3:
data was captured and compared to background noise to
obtain the dominant refrigerant flow induced noise frequency TABLE 3 Geometry change proposal for reduction of
octave bands. refrigerant flow noise
Figure 6 shows the different frequency band representing
the refrigerant flow induced noise compared with the back- SL. NO. Proposal Background
ground noise for microphone position 1. 3 Use of straight To evaluate the effect of
As can be seen from the graph, refrigerant flow-induced evaporator inlet / usage of straight evaporator
outlet pipes pipes on refrigerant flow
noise dominates over background noise mainly in the noise.
f requenc y ra nge of 1250 –1250 0 Hz for bot h
4 Placement of TXV To evaluate the effect of
microphone positions. closer to evaporator lower length of evaporator
The SPL dB(A) difference in the different octaves here pipes on refrigerant flow
represents only the refrigerant induced noise since the test is noise.
carried out in an acoustically sealed setting (no other source
of noise except the blower fan which is run at the lowest
constant speed possible in background and base iterations). It is evident that more are the bends in the evaporator
As already explained, the flow of refrigerant causes noise pipes, the louder the refrigerant flow-induced noise will be as
because of the impact on the internal walls of the child parts refrigerant flow collides with the pipe bends. Afore mentioned
of the evaporator. To minimize the noise generated by these studies were carried out to support the same theory.
impacts, an experimental evaluation was planned in order to
reduce vibrations in child parts by increasing their stiffness. 4.2.1. Effect of Increasing Evaporator Inlet/
Proposals listed in table 2 were evaluated. Outlet Pipe Thickness As a countermeasure to mitigate
To evaluate the impact of these trials on refrigerant noise, refrigerant flow noise, the evaporator inlet / outlet pipe thick-
separate verification was conducted. ness was increased as shown in figure 7.
To minimize the vibration in the inlet/outlet pipes,
we replaced the thin gauge pipes with thick gauge pipes in the
evaporator core assembly to mitigate noise at source level.
The figure 8 depicts that there is a miniscule decrease in
vibration level with increase of evaporator inlet pipe thickness.
FIGURE 6 Represents the background noise relative to the To check further, we measured the sound pressure level
base noise of refrigerant flow through human sensitive at microphone position 1 & 2 (as per figure 9).
frequencies at Mic 1. As shown in the figure 9,
Improvement of 2 dB(A) is noted in the frequency range
of 6300–8000 Hz at microphone position 1.
And an improvement of 3.1 dB(A) in the octave of sound
pressure level can be seen (4000-5000 Hz) for microphone
position 2.
The sound pressure level figure 9 depicts the prominent
octave ranges in which refrigerant noise is generated.
Improvement is found with application of higher
thickness evaporator.

4.2.2. Effect of Increasing Evaporator Side Plate

Thickness Based on the same principle, the evaporator’s
 AN EXPERIMENTAL STUDY FOR REDUCTION OF REFRIGERANT FLOW INDUCED NOISE 5

FIGURE 7 The accompanying graphic shows the intake and FIGURE 10 Evaporator side plate with increased thickness.
output pipes for the evaporator. (Image courtesy of gari. (Image courtesy of boodmo.com website)
pk website)

side plate thickness was increased from the existing width to

increase its mass and stiffness to check whether it will lead to
any reduction in refrigerant flow induced noise.
To increase the stiffness of the evaporator side plate addi-
FIGURE 8 Vibration level at accelerometer position 3 (as tional aluminum strips were brazed to the base side plate as
per figure 5). shown in figure 10.
After analyzing the data, it was discovered that increasing
the thickness of the side plates had very little effect on the
vibration level.
Similarly, as the thickness of the side plates increases, the
noise amplitude decreases minutely. The conclusion we could
reach was that the contribution of side plates to refrigerant
flow noise in the evaporator core was negligible.
Figure 11 shows that increasing the thickness of the side
plate has no significant effect on noise performance measured
at both microphone positions.

4.2.3. Use of Straight Evaporator Inlet/Outlet

Pipes The evaporator pipes in a typical HVAC design typi-
cally have many bends to accommodate layout needs. Study
FIGURE 9 Sound pressure level at microphone (MIC) was done to determine the effects of straight-pipe evaporator
position 1 & 2. layouts on noise caused by refrigerant flow. That is why straight
pipes were used in place of bend pipes in the evaporator as
shown in figure 12.
Figure 13 illustrates the findings for noise amplitudes in
prominent refrigerant flow noise octaves.
A significant difference [2.9 dB(A)] is observed in the
frequency range from 5000–8000 Hz.
For microphone position 2, an improvement of 3.5 dB(A)
is seen in the frequency range of 1600-3150 Hz.

4.2.4. Effect of TXV Placement Closer to

Evaporator Placing the TXV close to the evaporator
reduces the interaction of the evaporator pipes with the
 6 AN EXPERIMENTAL STUDY FOR REDUCTION OF REFRIGERANT FLOW INDUCED NOISE

FIGURE 11 Sound pressure level at microphone (MIC) FIGURE 14 The evaporator pipes are brazed to the
position 1 & 2. evaporator core with minimum possible length of the pipes.
(Image courtesy of ClearMechanic.com)

FIGURE 12 Evaporator I / O bend pipe layout changed to

straight I / O pipes.

FIGURE 15 Sound pressure level at microphone (MIC)

position 1 & 2.

FIGURE 13 Sound pressure level at microphone (MIC)

position 1 & 2.

4.2.5. Use of Noise Damping Material on

Evaporator Side Plate As already discussed, for cases
where the design of the HVAC is already fixed there is a little
scope of part geometry modification. The application of noise
damping material is explored to mitigate the refrigerant flow
induced noise.
In figure 16 we can see that over the distinct octave (1250
to 10000 Hz) of microphone position 1, the noise improvement
refrigerant flow by leaving refrigerant in the pipes for a little in refrigerant flow induced noise is significant [3.5 dB(A)].
duration. For the same reason, shorter evaporator pipes are
expected to produce less noise due to refrigerant flow. The
sample evaporator core manufactured for the same purpose
FIGURE 16 Sound pressure level at microphone position 1.
is illustrated in the figure 14.
As shown in the figure 15,
The reduction in refrigerant flow induced noise is signifi-
cant in all frequency ranges and maximum improvement [2.0
dB(A)] is observed in the frequency range (1250-10000 Hz)
for microphone position 1.
For the microphone position 2, maximum noise reduc-
tion of 1.1 dB(A) is observed in the octaves (800-10000 Hz).
Although, all the above solutions can be evaluated for
implementation in the design stage, for models that are already
in production stage a ready solution is required. For this
purpose, the use of noise damping material was evaluated by
applying it to different subcomponents of evaporator.
 AN EXPERIMENTAL STUDY FOR REDUCTION OF REFRIGERANT FLOW INDUCED NOISE 7

TABLE 4 Improvement in sound quality 1. The increase in stiffness of evaporator pipes by

increasing thickness reduces the amount of
refrigerant flow induced noise. However, increasing
the rigidity of the evaporator side plates does not
affect the noise levels much. Further research is
needed to study the effect of increased stiffness in
other child parts of the evaporator core.
2. When the evaporator pipes are made straight and
shorter, there is a noticeable decrease in noise
amplitude. This is due to less bends in the pipes, so
Similarly, for microphone position 2, significant improve- when the pipe length is short and straight, there are
ment [3.6 dB(A)] occurs across the dominant octaves fewer refrigerant bubbles colliding with the inner walls
(800-10000 Hz). of the evaporator pipes due to the smooth flow of the
We have evaluated these proposals based on their decibel refrigerant. Noise amplitude is reduced in important
value [dB(A)] relative to all the different countermeasures that octave ranges when the tube length is reduced because
have been described. However, sound quality characteristics there is generally less interaction between the
such as sharpness, loudness and articulation index were calcu- refrigerant flow and the pipes’ inner walls.
lated to make it more relevant to customer NVH perceptibility. 3. The use of damping material might be considered for
It has been seen that for all the countermeasures evalu- production models where it is not feasible to change
ated, 4.2.5 showed the highest improvement in sound quality the design of the evaporator and child parts in order
(using noise damping material on evaporator side plate). A to reduce the noise caused by refrigerant flow. The use
comparison analysis of sound quality parameters was carried of noise damper offers a quick and easy solution.
out to better observe the effect of the refrigerant flow-induced However, the use of noise deadening material has
noise reduction on sound quality within passenger cabin. The weight and cost implications and should therefore
evaluated parameter of loudness and open AI showed a signifi- be avoided on models where noise reduction is
cant improvement as shown in the table 4. achievable through the aforementioned
design changes.

Therefore, as a recommendation, keep the TXV as close

5. Summary/Conclusions to the evaporator as possible to ensure minimum interaction
of refrigerant f low with pipes. Evaporator pipes should
In general, the high frequency octave bands containing the be designed with minimum bends. If the noise problem
refrigerant flow-induced noise content range from the original persists, noise damping materials can be tried.
core frequency of 1600 Hz to 12500 Hz. The sound quality
inside the passenger cabin is negatively impacted by refrig-
erant flow-induced noise because the noise content is in the
high frequency range, which is where human hearing is sensi-
References
tive (human hearing is most sensitive between 2000 Hz and
5000 Hz). 1. Zwicker, E. and Fastl, H., Psychoacoustics: Facts and Models
For data comparison and analysis purpose, microphone (Heidelberg: Springer, 1999)
placed at position MIC 2 was considered as a primary refer- 2. Whalley, P.B., Two Phase Flow and Heat Transfer (New York,
ence data due to its higher noise amplitudes. USA: Oxford University Press, 1999)
The table 5 below summarizes the improvement in dB(A) 3. Taitel, Y. and Dukler, A.E., “Flow Regime Transitions for
levels in the high frequency range, where human hearing is Vertical Upward Gas-Liquid Flow: A Preliminary Approach
most sensitive. Through Physical Modelling,” in AIChE 70th Annual
Meeting, Session on Fundamental Research in Fluid
TABLE 5 Summary of sound pressure level Mechanics, New York, USA, 1977.
dB(A) improvement
Improvement In Contact Information
SPL [dB(A)]
S.No. Proposal MIC 1 MIC 2 AUTHOR:
1 Effect of increasing Evaporator 2.0 3.1
inlet / outlet pipe thickness Rachit Sharma: - Manager
2 Effect of increasing evaporator 0.4 0.9
Contact No. +91 8447670289
side plate thickness Rachit.Sharma@maruti.co.in
3 Use of straight evaporator inlet 2.9 3.5 CO- AUTHORS:
/ outlet pipes
4 Effect of TXV placement closer 2.0 1.1 Mukesh Kumar: - Assistant General Manager
to evaporator Contact No. +91 9999175750
5 Use of noise damping material 3.5 3.6 K.Mukesh@maruti.co.in
on evaporator side plate
 8 AN EXPERIMENTAL STUDY FOR REDUCTION OF REFRIGERANT FLOW INDUCED NOISE

Shubham Kumar: - Deputy Manager

Contact No. +91 7062171565
Definitions/Abbreviations
Shubham.Kumar4@maruti.co.in HVAC - Heating, ventilation, and air conditioning
Subhashree Patra: - Junior Engineer NVH - Noise, Vibration and Harshness
Contact No. +91 8895555650 Ug - Velocity of gas
Subhashree.Patra@maruti.co.in
Ul - Velocity of liquid
TXV - Thermostatic expansion valve
Acknowledgments SPL - Sound pressure level
We would like to express our sincere gratitude to Mr. MIC - Microphone
Shriganesh Umbarkar, the head of the HVAC (Heating,
Ventilation and Air Conditioning) engineering department
at Maruti Suzuki India Limited for his guidance and
continuous motivation.

© 2022 SAE International and SAE India. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form
or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE International.

Positions and opinions advanced in this work are those of the author(s) and not necessarily those of SAE International. Responsibility for the content of the work lies
solely with the author(s).

ISSN 0148-7191