A231288 PDF

00 4 4
00
The Pennsylvania State University

APPLIED RESEARCH LABORATORY
P.O. Box 30
State College, PA 16804
ACOUSTIC DIAGNOSTICS OF AN
AUTOMOTIVE HVAC SYSTEM
V..JAN 291991
i . "T. G.A.
C.Brungart
Lauchle
J.Tichy
Technical Report No. TR 91-001

January 1991
Supported by: L.R. Hettche, Director

Ford Motor Company Applied Research Laboratory
Approved for public release; distibution unlimited
IIF IIL S/
ABSTRACT
The heating, ventilation, and air conditioning (HVAC) system of an
automobile can be a source of acoustic annoyance, particularly when it
is operated under maximum airflow conditions. In order to suggest
possible design changes for the purpose of noise control, a systematic
characterization of the acoustic sources and mechanisms must be
conducted initially. This report addresses such a characterization for
a typical automotive system. The approach is based on acoustic
intensity measurements of the st.nd-alone HVAC system operating under
maximum airflow conditions in the ARL Penn State flow-through anechoic
chamber. The experimental data indicate that the centrifugal blower is
the dominant low-frequency source of noise, while separation zones and
the flow over sharp edges within the HVAC system ducting are secondary
sources of noise that become increasingly dominant as the frequency
exceeds 3 kHz. Qualification of these identified sources of sound is
aided by detailed flow visualizations of the subject system.
Recommendations for acoustic improvements to the system are provided.
A-1
TABLE OF CONTENTS
Pane Number
ABSTRACT .............................. 1
ACKNOWLEDGEMENTS .......................... 2
LIST OF FIGURES..........................5
LIST OF TABLES ........................... 9
LIST OF SYMBOLS ......................... 10
CHAPTER 1. INTRODUCTION. .................... 12
CHAPTER 2. ACOUSTIC INTENSITY TECHNIQUE .. ........... 15
2.1 Background. ...................... 15

2.2 Sound Field Quantities .................. 16
2.3 Complex Acoustic Intensity ................ 17
2.4 Meeasurement Technique.................21
2.5 Measurement Precision..................22
2.6 App,.ications. ..................... 25
2.7 Sourd Power Measurements ................. 26
2.8 Sound Source Identification. ............. 28
CHAPTER 3. EXPERIMENTAL PROCEDURES FOR THE HVAC SYSTEM .31
3.1 Acoustic Intensity Mapping ................ 31

3.2 INCE Plenum Box .................... 32
CHAPTER 4. EXPERIMENTAL RESULTS .................. 33
4.1 Nearfield Acoustic Intensity Maps of

the HVAC System .................... 33
4.1.1 Measurement Plane 1 ................ 34

4.1.4 Measurement Plane 4 .. ............. 40
4.2 Sound Power Measurements of the HVAC System . .43

4.3 INCE Plenum Box Measurements ............... 44
CHAPTER 5. CONCLUSIONS ..................... 46
CHAPTER 6. RECOMMENDATIONS FOR QUIETING .. ........... 49
REFERENCES.............................50
APPENDIX 1. HVAC System Noise Reduction

Using Resonators..................109
APPENDIX 2. Register Nnise ................... 113

4
LIST OF FIGURES
Figure Number Title Page Number
1 Intensity Vectors, Wavefronts arid Contours

of Constant Pressure .. ........... .. 51
2 Block Diagram of Measurement Technique 52
3 Intensity Field and Wavefronts of Three

Sources Separated by 1/10 Wa rplength . . 53
4 Sound Pressure Field of Three Sources

Separated hv 1/10 Wavclcngth ....... 54
5 Intensity Field and Wavefronts of Three

Sources Separated by 1.35 Wavelength . . . 55
6 Active Intensity in Nearfield of Circular

Membrane Excited to (0,3) Mode ...... 56
7 Potential Energy for Configuration of

Figure 6 ..... ................. .57
8 Reactive Intensity for Configuration of

Figure 6 ..... ................. .58
9 The ARL Penn State Flow-Through Anechoic

Chamber ...... .................. .59
10 The HVAC System Situated in the Anechoic

Chamber ...... .................. .60
11 Schematic Diagram of the 3-D Acoustic

Scanner ...... .................. .61
12 Construction Notes for INCE Plenum Box . . 62
13 INCE Plenum Box Used in Production Blower

Sound Power Measurements . ......... 63
14 Measurement Plane 1 Intensity Vector Map,

250 Hz, Blower In ... ............. .. 64

250 Hz, Blower Out ... ............ .. 65

500 Hz, Blower In ... ............. .. 66

500 Hz, Blower Out ... ............ .. 67
5
I
LIST OF FIGURES (CONTINUED) I

1 kHz, Blower Out ............... .. 68
19 Measurement Plane 1 Intensity Vector Map, I

I kHz, Blower In .. ............. .69

2 kHz, Blower In . ...... ............. 70
3
2, Active Intensity Data of Figure 20 with Color
Applied to the Levels ........... .. 71 I
22 Measurement Plane I Intensity Vector Map,
2 kHz, Blower Out ............... .. 72 1
250 Hz, Blower In ............... .. 73
24 Active Intensity Data of Figure 23 with Color

Applied to the Levels ........... .. 74 1
250 Hz, Blower Out . .............75

500 Hz, Blower In ............... .. 76

500 Hz, Blower Out . .............77

I kHz, Blower In . ..............78
3
2 kHz, Blower In . ..............79
3
2 kHz, Blower Out .. ................ 80 I
250 Hz, Blower In ............... .. 81 S
250 Hz, Blower Out . .............82 I
500 Hz, Blower In ............... .. 83 3
6£
===
LIST OF FIGURES (CONTINUED)
34 Measuremert Plane 3 Intensity Vector Map,

500 Hz, Blower Out ... ............ .. 84
35 Measurement Plane 3 intensity Vector Map,

1 kHz, Blower In ... ............. .. 85

1 kHz, Blower Out ... ............. .. 86

2 kHz, Blower In . . . . . . . . . . . . . 87

2 kHz, Blower Out ... ............. .. 88

125 Hz ...... .................. .89

250 Hz . ... ................. . . 90

500 Hz ...... .................. .91

1 kHz ...... ................... .92

2 kHz ...... ................... .93

250 Hz ...... .................. .94

500 Hz ...... .................. .95

1 kHz ...... ................... .96
47 Active Intensity Data of Figure 46 with

Color Applied to the Levels ......... .. 97

2 kHz ...... ................... .98
49 Sound Power Radiated by the HVAC System

Operating in its "Standard" Condition and
With Airflow Only ... ............. .. 99
7
I ! •
LIST OF FIGURES (CONTINUED) 3

so Measurement Positions Used in INCE Plenlun

Box Sound Power Measurements ....... 100 I
51 Sound Power Measurements of the Blower at
1700 RPM .... ................. .101 I
2000 RPM .... ................. .102 I
2500 RPM .... ................. .103

2900 RPM .... ................. .104
55 Sound Power Level as a Function of Wheel

Speed at Constant Pressure Coefficient 105 1
56 Flow Visualization in the Exit Region of
the Blower .... ................ .106
57 Flow Visualization in the Transition Region I

Between the Evaporator Core and Plenum
Section .... .................. .107 3
58 Flow Visualization Upstream of Heater Core
and Distribution Ducts . .......... .108
1.1 The 1/4 Wavelength Resonator Concept . . . 11l
1.2 Experimental Data Showing the Effectiveness

of the 1/4 Wavelength Resonator ...... 112 U
2.1 Relative Intensity Levels Measured With
All Vents Open and With the Middle Vents
Closed ..... .................. .116
2.2 Schematic of the Register Noise

Measurements ... ............... .117 I
LIST OF TABLES I
Table Number Title Page Number
2.1 Octave Band SPL for HVAC System with Air

Supplied by an Auxiliary Quiet Airflow
Source .... .................. .115 3
2.2 Octave Band SPL Measurements of Register
Noise ..... ..................... 115
i3
i m m m8
LIST OF SYMBOLS
AC air conditioning
B filter bandwidth
BPF blade passage frequency
c sonic velocity
FFT fast Fourier transform
G11 , G22 auto-spectral density for microphones 1 and 2 respectively
021 cross-spectral density between microphones 2 and 1
HVAC heating, ventilation and air conditioning
I active intensity vector component in the direction of the two

microphone axis
INCE Institute of Noise Control Engineering
I, complex acoustic intensity in the frequency domain as measured

by a two microphone probe
IT true active intensity vector component in the direction of the

two microphone axis
Iactive intensity vector
estimated active intensity vector component in the direction

of the two microphone axis
I complex intensity vector
instantaneous intensity vector
active intensity vector normal to the surface enclosing the

sound source
k acoustic wavenumber (w/c)
p acoustic pressure
P acoustic pressure amplitude
Q reactive intensity vector component in the direction

reactive intensity vector
9
I
LIST OF SYMBOLS (CONTINUED) I
r spatial coordinate 3
rpm rotations per minute
T data acquisition time
t time
U particle velocity amplitude I

u particle velocity vector component in the direction
17 particle velocity vector
V, impeller tip speed
W sound power 3
AP pressure rise across the blower in the INCE box
Ar microphone center to center separation distance 3

ASJ incremental surface area enclosing sound source
5 phase mismatch I
C normalized bias error
A acoustic wavelength
p fluid density 3
4phase angle
* pressure coefficient I
w circular frequency
10 |I
ACKNOWLEDGEMENTS
This work has been supported by the Ford Motor Company under contract
747-8-G75483. The authors are grateful to Mr. Ronald E. Kretschmer,
program monitor at Ford, for his interest and support of this work. Also
acknowledged is the expert advice provided by Mr. Mark W. McBride of ARL
Penn State and the efforts of Mr. Michael L. Sullivan in performing the
acoustic intensity mapping.
1i
CHAPTER 1
INTRODUCTION
The acoustic noise within the passenger compartment of an
automobile can be attributed to several independent sources. Of
particular concern is the noise generated by the heating, ventilation,
and air conditioning (HVAC) system. In order to suggest and implement
acoustic noise control measures, an acoustic evaluation of the system is
required. Such an evaluation should identify the individual components
responsible for noise production, and equally important, it should
identify the predominant physical mechanisms within these components
that produce acoustic energy. Once these mechanisms are understood, it
is much easier to proceed with control concepts.
The HVAC system under consideration is typical of those used in
automotive applications. The acoustic annoyance of this and all other
similar systems tends to increase as the airflow is increased. Thus, it
is known at the start that the predominant mechanism for noise
generation is unsteady flow. A major objective of the current
investigation is to identify the specific regions within the system
where the flow has high unsteadiness and then to determine which of
these regions produce radiated sound, and how much radiated sound. The
part of the study directed toward flow measurements has been completed
and documented by Denger et al. (1990). The part that is concerned with
acoustic emissions and the correlation of these emissions with unsteady
flow is reported here.
This study is experimental, as was the fluid mechanics
investigation of Denger et al. (1990). The methodology selected is the
12
two-sensor acoustic intensity method which permits direct measurements i
of acoustic power flow from a source. By performing such measurements
in the nearfield of a complex system, individual sources of noise can be U

identified. This approach was implemented in such a way that near-field e
intensity vectors were obtained in selected cross-sectional planes of
the stand alone (removed from the automobile) HVAC system. The
interpretation of these intensity "maps" is straightforward and key
sources of noise are identified, and in some cases coiielated directly !
to the unsteady flow fields measured previously in the same system. 5

Because the centrifugal blower used in the HVAC system was
identified as the key source of noise, additional studies are discussed
for the blower operating by itself in a specially designed acoustically
transparent plenum box. This study provides information on the total i

power created by the blower and the dependence of this power on
operational parameters such as head rise and wheel speed. Under low
head conditions, the blower radiates sound at the blade passage i
frequency. A resonator concept was implemented to cancel this noise and
is discussed in Appendix 1. 3
In another part of the study, the blower was removed from the HVAC
system (as to eliminate this source of noise), and then air from a
remote, quiet airflow source was pumped through the system. This
enabled secondary sources of noise to be identified that are normally
masked by the blower noise. Because passenger compartment register 3

vents were identified as important contributors in this part of the
investigation, a limited set of experimental data were acquired for
register noise and is documented in Appendix 2. £
13I
The report concludes with recommendations for noise control.
Prototype implementation of these recommendations is the subject of
future investigation.
14
CHAPTER 2
ACOUSTIC INTENSITY TECHNIQUE
2.1 Background
The acoustic part of the research on identification of the sound
generation, radiation and propagation was based on the recently
developed intensity technique. This modern measurement and analytical
tool turned out to be indispensable to measure the sound power generated
by the blower, to determine the sound radiation from the ducts, and to
reveal many details of the acoustical properties of the system.
Therefore, conclusions could be made on the overall acoustical
characteristics. The purpose of this section is to provide detailed
information on the principles and relevant applications of this new and
powerful tool.
Until recently the only acoustical quantity that could be measured
with high precision was the sound pressure. Although, in principle, a
complete knowledge of the sound pressure is sufficient to calculate
other acoustical quantities of interest, the extent of scanning the
sound field to secure a sufficient amount of data makes this approach
impractical. Therefore the required acoustical quantities need to be
measured directly.
The development of the Fast Fourier Transform analyzer, the
introduction and perfection of digital electronic technology, and the
improvement of electro-acoustic transducers has permitted the
construction of a reliable acoustic intensity meter for both laboratory
and field use. This permits a direct measurement of acoustic energy.
15
I
The ability to measure acoustic power flow substantially expands
our capability to study details of sound radiation from complex
structures, to obtain sound power radiated by acoustic sources into the I

farfield from the nearfield measurements, to study the qualitative and 3
quantitative details of sound propagation in space with complex
boundaries, to evaluate acoustic diffraction and dispersion, and t 3

simplify or perfect the measurements of acoustical properties of
materials and structures. I

Signal processing, which was developed in conjunction with 5
intensity measurements, can be easily extended to obtain sound pressure,
particle velocity, potential and kinetic energy density and acoustic
impedance. Using computational techniques, sound field maps of acoustic
intensity vectors and wavefronts, which represent sound power I

propagation, can be constructed. Other sound field maps allow us to p
visualize the standing wave formation by means of the pressure magnitude
gradient. 3
2.2 Bound Field Quantities
In most situations, we deal with sound fields that are steady over I
a sufficiently long period of time. Therefore, many useful mathematical 5
formulations for intensity and other acoustical quantities can be based
on the assumption of steady state and represented in the frequency J

domain.
Calculation of power progression in acoustic fields requires I
determination of two acoustical quantities. Acoustic intensity can be
obtained from the sound pressure and particle velocity amplitudes and
the phase between these quantities. While sound pressure can be 3

16
measured directly by small and precise microphones, there are no
suitable transducers to directly measure particle velocity.
Because particle velocity is proportional to pressure gradient, the
intensity probe approximates the gradient by the sound pressure
difference obtained from two closely spaced microphones. In addition,
we can also measure the acoustic pressure.
Assuming that both p and z can be obtained from field
measurements, the relationships among the sound field quantities needed
for extended applications of the intensity technique will be summarized.
2.3 Complex Acoustic Intensity
The power of the intensity technique is in its ability to provide
comprehensive information on all acoustic quantities of interest thus
permitting a complete analysis of sound radiation and propagation.
Usually the sound field is described in terms of acoustic pressure which
can, without loss of generality, be expressed at a point (r) by its
amplitude P(r) and phase O(r)
p(r,t) = P(r)e-1e (1)
where both P(r) and O(r) are real quantities.
The particle velocity, which can be obtained from Euler's equation,

-1I
a (r,t) - - -- Vp(r, t)= - [P(r)V#(r) + jVP(r)]e' (r)e j u t (2)
is thus expressed in terms of the pressure amplitude and phase gradient.
The instantaneous acoustic intensity, which describes the energy
propagation through a point as a function of time, can be calculated
from Equations (1) and (2)
17
I
1 (rt)= Re[p(z,t).Re O(r,t)l = 2op P2--
P(r)V¢() (3)
(3))
2
+ 1--P (r)V4(r)cos2(u)t - P (r)VP(r) sin2 (t - o(r)
2Fp 2wp
A quantity of greater practical interest is acoustic intensity,
which describes the time-averaged power flux through a point, and can be
obtained by time averaging the instantaneous intensity 5

(2r)
T
Re[p(r) . (r) 2 (r) V$(r) I

As can be seen from Equation (4), the time-independent intensity
vector Y has a direction of the phase gradient VO(r). It is high in
areas where P 2 (r) is large (high potential energy density) or where the
wavefronts are compressed so that VO(r) is large. 5

It should be emphasized that the intensity vector I(r), which
depends on r, only describes the power associated with a sound wave at a 1

field point. The intensity vector is time independent and does not
propagate with the sound field.
The concept of acoustic intensity can be extended by using the 3

complex intensity defined as
(r)+j (r) = P (r)

= Y(r) p2(r)V1(r) -jrVP(rr)].
- P(r) (5)
2 2c(p
The first term in Equation (5) is intensity I as defined by
Equation (4), which can be alternatively calculated from I

1/2 Re p(r) .T(r)].
The imaginary part of the complex intensity Y(') is called I

reactive intensity Q(r) and can be further expressed as
181
1(r) = -1 Im[p(r). P(r)] = - P(r)VP(r)= - --
lVP2(r) (6)
22op 4(ap
Similarly to f(r). 1(r) is time-independent and depends on
coordinates only.
An example of the sound field representation in the vicinity of
sound radiators is shown in Figure 1. The sources generate a sound
pressure field, characterized by surfaces of constant pressure amplitude
P(r) and constant phase 4(r). The constant phase surfaces are the
wavefronts and, as shown in Equation (4), the I magnitude is
proportional to the phase gradient, and vector f is, therefore,
perpendicular to the wavefronts (Figure 1). Knowledge of the wavefronts
is very useful in determining the source behavior. The wavefronts,
which are not measured directly, can be obtained by reconstructing the
phase from the f map.
As Equation (6) reveals, the reactive intensity vector is
proportional to the pressure square gradient and, therefore, once a
field map is plotted, areas of constant P(r) can be constructed.
At a given point, both intensity vectors f and d have a different
direction and, therefore, the surfaces of constant P(r) and O(r)
mutually intersect, as shown in Figure 1. These vectors have the same
direction only in the case of one-dimensional fields of plane waves or
spherical waves. In spherical waves, areas of constant P(r) and O(r)
are concentric spheres.
The vector d as defined in Equation (6), is proportional to the
gradient of squared pressure amplitude. Because the gradient is equal
to zero at the pressure maxima or minima, the reactive intensity vector
19
I
aims toward the minima and is pointed away from the maxima. These 3
pressure extremes are sources and sinks of 1.
Both vectors fand d are particularly useful for investigation of
sound sources and sound radiation in general. As mentioned above, the !
intensity vector f describes the sound power at a point which of course
includes the surface of vibrating objects. Most surface vibrations
consist of modes representing the standing transverse waves. A 3

classical example are plates and shells which constitute the
constructional elements of many machine parts. The duct walls of I

the HVAC system vibrate like plates.
Vector f is an invaluable tool to reveal the relationship between
the vibration and sound power radiation. The sound power radiation from 5
plates depends on the ratio of the wavelengths of the bending wave on
the radiator to the wave in the air. The maps of vector I reveal that I
if the wavelengths of the mechanical waves on the radiator are shorter 3
than in the air, areas of sound power radiation alternate with areas of
sound power absorption. The total radiated power is given by the 3

algebraic sum of both powers.
The reactive intensity vector a, which is proportional to the I

pressure square gradient, reveals on the surface of radiators the areas j
of vibration which set the surrounding air into motion. The larger the
vibration amplitude, the higher is the sound pressure at the radiator
surface. However, the radiated sound power depends also on the phase
shift between the pressure and particle velocity and, therefore, areas I
of high pressure are not generally associated with large sound power 3
radiation.
20
While hard wall vibrators are marked by high values of vector O5,
openings such as duct terminations or leaks which radiate acoustic power
have high f ad low §. This means that although acoustic pressure is
low, sound power is radiated due to the large in-phase particle velocity
component with sound pressure. This brief description of radiator
properties shows the importance of both vectors fand .
2.4 Measurement Technique
As mentioned earlier, the measurement of intensity is based on
direct or indirect determination of p and 17.
A measurement probe consists of two pressure microphones which are
closely spaced to obtain the sound pressure from the average, and the
particle velocity from the difference of the two measured sound
pressures. In the frequency domain, the complex intensity can be
expressed in terms of the auto-spectra and cross-spectra between the two
microphones
=
Ic(W) = I(W) + jo(w) p1
2jcop~r [G2 2 (w) - Gj(w) - j2rm(G2 l(4))] (7)
where G 11(w) and G22 (w) are the pressure auto-spectral densities at the
microphone locations, G2 1(4) is the complex cross-spectral density, and
Ar is the distance between the microphones. Both intensity estimates
can be obtained from

-1
I(w) - Re[T(w)] - I m[G 2 1 ,(W)] (8)
and
21
I
Q(() = Im[ (W) - 2WpAr
- [G2, () - (a)]. (9) 1
The sound pressure amplitude is obtained as the average of the two I
pressures 5
p(r,t) = P(r)e-J(r)e t - re + P 2 (r)e- (YI]eJor (10)
as
p2(W) = " (W) + G2(( ) + 2Re(G 21 ( ))] ()
The particle velocity is obtained from the pressure gradient
approximated by the pressure difference 3

f(r, 0)) - -I IVp(rt) 1 -1 [pe, )] eJu (12)
and the velocity amplitude is written as shown in Equation (13) 1

U 2 (W) = 1
4W 2 p2 (Ar)
2
[G1,(w) + G22 (w) - 2Re(G.,(W))1 (13)
3
2.5 Measurement Precision 3
The precision of intensity measurements depends on many more
factors than relatively simple sound pressure measurements. Besides the 5

quality and precision of the transducers and associated electronics,
errors can be caused by unsuitable transducer configuration which I

includes their distance, size and associated diffraction effects. In 3
addition, the measurement precision depends on the field properties
which are, of course, not known prior to their measurement. However,
both the theoretical analysis of the errors and an intensive effort of
the manufactures to develop new transducers and electronics resulted in I
22
U
the construction of an intensity meter which satisfies the demands of
noise control engineers.
The principal systematic errors, when using intensity probes, are
caused by the finite physical distance and response mismatch of the two
microphones and associated electronics. An intensity probe consists of
two transducers with Ar separation of their centers. The sound pressure
is determined from the pressure average, and the particle velocity from
the pressure difference which approximates the pressure gradient. In a
plane wave, the estimated intensity f depends on the true intensity IT
and Lr, and is given by
sin (kAr) (14)

kAr '
where k is the acoustic wavenumber. The error affects the upper
frequency limit of the probe and, therefore, a sufficiently small Ar has
to be chosen. In nonuniform fields, such as those created by monopoles,
multipoles, and other complex sources, the error is lp.rger, particularly
for measurements taken close to the sources.
Both the amplitude, and particularly the phase difference of the
response of the two probe channels measured by exposing them to an
identical sound field, cause another bias error. The response magnitude
can usually be adjusted, while the phase mismatch depends on the actual
hardware used. The normalized bias error for equalized magnitude
response and phase mismatch 6 between the microphone channels is
23
f -xr
P,(15)i 8
*12 IT
where 012 is the phase difference in the measured wave which depends on U
the measured sound field. For plane propagating waves 012 - kAr, while 5
in perfect standing waves 012 - 0. For the same 6, the error can be
decreased by choosing sufficiently large Ar which, however, limits the 3

upper frequency limit as given by Equation (14). The error depends on
the mixture of propagating and standing waves which, as shown in I

Equation (15), can be expressed in terms of P2/1 T . This ratio varies 3
from point to point in the field. The error affects the low frequency
limit of the probe due to the small value of *12 - 2wAr/X, where A is 5
the acoustic wavelength.
Although microphone pairs with 6 - 0.05@ are available today, it is I

safe to use 6 - 0.3* as an overall value, for the practical choices of 3
Ar to keep the high frequency bias error low.
Depending on the type of source signal and the goals of the 3

intensity measurements, an FFT broadband analyzer can be used for the
data processing. Alternatively, great flexibility may be achieved by 3

acquiring and processing the data by a computer. In any case, the
random error, which depends on the BT product of the filter bandwidth B
and the data acquisition time T, has to be minimized by selecting 5

sufficiently large BT. In addition, when synthesizing 1/3 or 1/1 octave
data from line spectra obtained from an FFT analyzer, the frequency 3
range of the analyzers has to be selected for a minimum of 10 lines in
the bandwidth. When taking broadband measurements, a true band analyzer
could be a better choice than an FFT analyzer. 3

24 ~I I
The above formulations for I, Q and u are valid for a component of
a particular vector having the direction of the two microphone axis.
Therefore, in order to obtain the entire vector, three measurements with
orthogonal microphone axes are needed. Similarly, the kinetic energy
density is obtained (see Equation (9)) from a dot product of two vectors
with three orthogonal components.
A practical measurement arrangement consists of two microphone
pairs with perpendicular axis so that the vector quantities in a plane
are obtained. Such measurement is very efficient in sound fields which,
for reasons of symmetry, become two-dimensional.
The above Equations (8) - (13) show that all quantities of interest
can be obtained from auto- and cross-spectral densities between two
closely spaced pressure sensors. Therefore, once these are calculated
and stored in a computer, extensive information on the sound field can
be obtained with only moderate additional computational effort. Figure
2 shows the block diagram of a universal measurement technique which
provides all acoustic quantities of interest. A variety of available
FFT and other analyzers have internal postprocessing circuitry which
usually calculates the active intensity and sound pressure level.
2.6 Applications
The main incentive leading to the development of the intensity
technique was the need to develop a better tool for sound power
measurements than the techniques based on calculating sound power from
the squared pressure data. Once a reliable sound power measurement
technique was developed, the applications have escalated rapidly and the
new technique has been adopted in practically all measurement techniques
based on sound power, such as transmission loss, sound absorption
25
I
coefficient and radiation efficiency measurement. In addition to
improving the laboratory procedures, the intensity meter is
irreplaceable for field measurements because of its abiliLy to directly I

measure the sound power. The growing inteiest in near-field data linked 5
to sound radiation and source identification provides the necessary data
base to develop new prediction techniques. Finally, the ability of the 3

intensity technique to measure sound power flux in addition to the sound
pressure and particle velocity, makes it a perfect tool for complete I

studies of diffraction, reflection and other field studies linked to 3
wave propagation.
2.7 Bound Power Measurements 5

Precise knowledge of sound power radiated from a source is needed
for its characterization and labeling, and as the input parameter for I
environmental design. The extensively researched reverberation room 3
technique, which has the needed precision to measure the radiated power,
requires large and costly reverberation rooms. In addition, it is 3

usable for small sources only. The in-situ sound measurement of large
so irces is based on determining the sound power flux over an enclosed I

surface around the source. The use of intensity, instead of pressure,
eliminates up to 10 dB of error which arises from the substitution of
p2 /pc for intensity when measuring in the nearfield of sources.
The sound power measurement procedure consists of selecting a
surface enclosing the source. Sound intensity is measured at a 3

sufficient number of poinus and Lhe suuiku poweL W is calculated as a sum
of partial power contributions from small areas ASj associated with each I
intensity vector fl, normal to the surface. 3
26I
I
n
w = f, Asi (16)
The principal advantage of the intensity technique is the ability
to measure the power flow in the farfield as well as in the nearfield.
As a consequence, the power radiated by large sources located in a
highly reverberant environment can, for the first time, be measured with
sufficient precision. Measurement procedures are provided in the
drafted standards: International ISO DIS9614, "Determination of the
Sound Power Levels of Noise Sources Using Sound Intensity Measurement at
Discrete Points," and U.S. Standard ANSI S 12.12, "Engineering Method
for the Determination of Sound Power Levels of Noise Sources Using Sound
Intensity." Work is also in progress on an international IEC Standard,
"Instruments for the Measurement of Sound Intensity" and on its U. S.
ANSI counterpart.
These standards define the optimum number and location of the
measurement points, the sttps determining the desired measurement
distance from the source surface, and the procedures for the
verification of the measurement precision.
The intensity technique permits measurement of the sound power of a
source in the presence of other noise sources operating simultaneously.
If no sound energy is absorbed inside the Volume circumscribed by the
measurement surface, the sound power fliux from external sources
penetrates and eventually exits the measurement areas so that the total
net flow is zero. This strategy can be applied to measure the sound
power output of parts of larger systems which cannot be separated
mechanically and measured in parts. The presence of reflectors, or the
proximity of hard walls, as well as highly reverberant environments,
27
results in the formation of standing waves. These environments require 3
careful selection of measurement locations, sufficiently large number of
measurement points, and a carefully calibrated system. The precision of
the measurements has been the subject of intensive research and, as a 5

result, indicators were formulated to verify the precision of the
measurements. The mentioned Standard drafts provide details on the use 3

of these indicators. Ir general, the overall precision of the intensity
technique equals the precision of sound power measurements in I

reverberation rooms and other reflecting environments. 3
2.8 sound Source Identification
Detailed measurements and mapping of sound pressure and acoustic 5

intensity in the nearfield of a radiator lead to both qualitative and
quantitative evaluation of the sound radiation from a source. The maps I

of intensity vectors reveal the areas of sound power emission and their 3
propagation into the farfield. The-, are essential in determining the
radiation pattern of a source and in deciding on noise control 3

strategies.
First, examples of the nearfield of typical radiators will be 3

shown. The vibration of plates, beams, or machine surfaces can be
expressed in terms of modes. Each vibrational mode on the source
surface is characterized by areas which are separated by nodal lines and
vibrate with 1800 phase shift. If the distance between the centers of
the adjacent areas is smaller than one-half wavelength of the radiated 3

sound, a substantial amount of the radiated energy is returned back into
the source and usually only a small fraction propagates into the I
farfield. 3
28
This "shortcircuiting" of the acoustical power flux is due to the
coupling. Its effect will be shown first on a field of three sources,
separated by one-tenth of the wavelength (Figure 3). The center source
vibrates with 1800 phase shift relative to the two outer sources. As
the intensity field reveals, the outer sources radiate sound power which
sinks partially into the center source and only a fraction of the
radiated power propagates into the farfield. The solid lines in this
figure are wavefronts reconstructed by the computer as lines normal to
the intensity vectors (see Equation 4). Figure 4 shows the pressure
distribution represented by circles proportional to P2 amplitude. This
figure reveals that the pressure measurement cannot provide any
information on actual energy flow from one source into another; it is
the result of the wave interference in the nearfield. The overall
reduction of the energy propagating into the iarfield is caused by the
strong source coupling. As shown in Figure 5, when the source distance
is increased (1.35 wavelength), th,: coupling has substantially decreased
and all three sound sources radiate sound power into the farfield.
The "shortcircuiting" of acoustic energy is characteristic of
radiation from vibrational modes at frequencies lower than the critical
frequency for which the distance of the adjacent "source centers" is
smaller than one-half wavelength. Figure 6 shows the active intensity
in the nearfield of a circular membrane in a baffle which is excited to
its (0,3) mode. Again, sound power radiated by the source areas
partially sinks into the adjacent areas. Figure 7 shows the pressure
distribution which is fairly uniform and does not provide any
information on the actual source behavior. We can now show the
usefulness of reactive intensity d, which is proportional to the P2
29
gradient as revealed in Equation 6. Figure 8 shows that the vibrating 3
portions of the hard surface are sources of , because the soundI
pressure has maxima at these surfaces.
I
I
I
I
I
I
I
I
I
30 1
I
CHAPTER 3
EXPERIMENTAL PROCEDURES FOR THE NVAC SYSTEM
3.1 Acoustic Intensity Mapping

The acoustic measurements were performed in the ARL Penn State
Flow-Through Anechoic Chamber. Figure 9 shows the construction and
dimensions of this chamber; it has a low-frequency cutoff of 70 Hz. The
stand alone HVAC system was positioned near the floor at the center of
the chamber. It was attached to a simple wooden frame for support above
the floor. Acoustically absorbing foam was draped over support members
during the tests. Figure 10 shows a photograph of the setup in the
anechoic chamber. The metallic structure seen in this view is part of a
large 3-D microphone positioning apparatus (scanner). Figure 11 shows
an isometric drawing of the scanner which is typically 8 feet on a side.
A computer controlled stepping motor powers each of the three
independent carriages.
Two (2) pairs of phased-matched 1/4-inch microphones were remotely
positioned by the scanner. The microphone pairs were oriented 900 to
each other so that two orthogonal components of the intensity vector
could be measured at each measurement point in space. Thus, a
measurement plane was selected, measurement locations were programmed
into the computer (these required detailed physical measurements of the
HVAC system), and the scanner moved the intensity probes to the points
selected. Once set into operation the scanner moves the probes and then
pauses for a sufficiently long period of time so that mechanical
vibrations dampen out, and then the acoustic data acquisition is
performed. Once the signals are in computer memory, the computer
31
U
directs the scanner to move to the next position and repeat the data 3
acquisition. Based on the several hundred measurement points required,
it was not uncommon for a typical scan (over one plane) to take two I
hours. Data acquisition and post-processing of the acoustic intensity 3
quantities were performed by a MASSCOMP computer.
3.2 INCE Plenum Box 5

In 1985, the Institute of Noise Control Engineering (INCE) adopted
a standard practice for the measurement of sound power from small air- U
moving devices. This practice is commonly referred to as the INCE 3
Plenum Box Method and is documented as Recommended Practice 1-85 (INCE
Publication TG-CBE 1-85). The fundamental feature of the method is that 5

a simple frame box has walls covered with acoustically transparent mylar
sheets. The fan or blower under consideration is mounted in one wall as I

depicted in Figure 12; the box thus provides a pressure chamber that
allows the fan to be tested under various conditions of back pressure.
The back pressure (head rise) is varied by a sliding door opening in one
wall, and the static pressure is measured by a piezometer. Microphones
can be positioned around the box and are sensitive to the emissions from 3
the fan because the mylar is transparent; furthermore, the box prevents
extraneous flow noise from acting on the microphones. The box fabricated I
for the current project was made slightly smaller than the recommended 3
dimensions. A photograph of this box with the production blower mounted
is shown in Figure 13. The box was used to measure the sound power of 3
the blower over a range of pressures and wheel speeds.
I
CHAPTER 4
EXPERIMENTAL RESULTS
4.1 Nearfield Acoustic Intensity Maps of the HVAC System
The computer controlled, three axis traverser described in the
previous section positioned the intensity probe for nearfield acoustic
intensity mapping. The data were acquired with the HVAC system
operating in its maximum AC configuration with a volumetric flow rate
equivalent to the maximum AC throughput. Two dimensional, acoustic
intensity vector measurements were performed at numerous locations in
five planes for up to five one-third octave frequency bands. The
measurement planes are described with the data. In planes 1, 2, and 3,
data were taken both with the blower installed in the system and with
the blower located in an acoustically isolated box. The same air flow
rate was maintained to the system at all times. Data in Plane 4 were
acquired only with the blower installed and data in Plane 5 were
acquired only with the blower removed.
Figures 14 through 48 present the intensity maps in a format
designed to orient the reader so that he/she can easily relate the
intensity to features on the HVAC system. Each map consists of a three
dimensional figure of the HVAC system with the corresponding measurement
plane in the upper left corner of the page, one-third octave band active
intensity levels appear at top right, active intensity vectors at bottom
right, and reactive intensity vectors at bottom left. Photos of the
HVAC system were taken at the appropriate angles and were modified for
placewent in the blank spaces of the intensity maps representing the
surface in the measurement plane.
33
U
4.1.1 Measurement Plane 1
Intensity vector maps taken in measurement Plane 1 are presented in
Figures 14 through 22. For comparison purposes, at a given frequency, I

the data are first presented with the blower installed in the system, 3
and then with equivalent airflow being supplied by a fan in an
acoustically isolated box. 5

At 250 Hz, with the blower installed, Figure 14 clearly shows that
the blower dominates the field; radiating sound through the inlet and U
recirculating door. With the blower removed, the strongest source 3
region is, as shown in Figure 15, the top of the evaporator core; the
intensity vectors fan out radially from that location. Although the 3
source at the top of the evaporator core contributes locally to the
sound field, it is a secondary source in terms of the integrated I

intensity (i.e. sound power) radiated by the HVAC system in Plane 1 at 3
250 Hz; the blower being the primary source.
Figure 16 shows that at 500 Hz the blower is the dominant noise 3

source in Plane 1. The blower radiates sound predominantly through the
inlet although there is a substantial amount transmitted through the 3

recirculating door. Figure 17 reveals that, with the blower removed,
the exit of the evaporator is the strongest source region. However,
compared to the fan, the evaporator exit source contributes only a small 5
amount of acoustic energy to the field (comparisons of the maximum level
indicated on all of these figures aids in such conclusions). 3

At 1000 Hz in Plane 1, Figure 18 shows that the blower is again the
dominant noise source although there is a local contribution from the
area at the evaporator exit. With the blower removed, Figure 19 shows 5
that the area at the evaporator exit is the strongest source region,
341
and, compared to the fan, contributes equal energy to the sound field at
that location. However, in terms of the total sound power radiated by
the system, the contribution from the evaporator exit source is small.
Figure 20 also shows that the blower is the dominant noise source
for 2000 Hz in measurement Plane 1. The blower radiates sound primarily
through the inlet and recirculating doors. There is, however, a local
contribution from a source at the evaporator exit. Figure 21 is a
repeat of the active intensity shown in Figure 20, but color has been
applied to aid in hot spot identification. The red indicates an
acoustic hot spot. Figure 22 shows that, in measurement Plane 1 with
the fan removed, the area at the evaporator exit is the strongest source
region at 2000 Hz. Again, in terms of the total sound power radiated by
the system, the contribution from the evaporator exit source is small.
Figures 14 through 22 clearly show that the blower is by far the
dominant noise source measured in Plane 1 at frequencies ranging from
250 to 2000 Hz. Two additional flow noise source locations have been
identified though; the top of the evaporator at 250 Hz and the exit of
the evaporator at frequencies ranging from 500 to 2000 Hz. These flow
noise sources contribute to the sound field only locally. If the
intensity is measured a nhort distance away from either flow noise
source region, the sound field would be dominated by the fan. Locally,
flow noise can contribute equally to the sound field (relative to the
local contribution from the fan), but in terms of the overall sound
power radiated by the HVAC system (in Plane 1), the flow noise sources
are econdary contributors.
35
!
4.1.2 Measurement Plane 2 3
Figures 23 through 30 present intensity vector maps measured in
Plane 2. As in Plane 1, the data taken with the blower installed in the I
system are presented first, and then, for comparison purposes, the 3
intensity vector map taken with the air moving device located in an
acoustically isolated box is presented. 3

Figures 23 and 24 show that in Plane 2 at 250 Hz (even with the
blower installed in the HVAC system) the middle vents contribute I

significant energy to the field; the active intensities being high in 3
front of them. There is also a substantial, if not dominant,
contribution from the blower in that area of the measurement plane n
ranging from the middle vents to the edge. The passenger's side vent
ranks third out of the four vents in terms of sound generation with the m
driver's side vent contributing the least amount of energy to the field.
Figure 25 shows that with the blower removed the middle vents contribute
most of the acoustic energy to the field. The passenger's side vent n
ranks third out of the four vents in terms of noise radiation with the
driver's side vent contributing the least amount of energy to the field.
With the blower removed, the active intensity is reduced only
approximately 3 dB in front of the middle vents and passenger's side

vent. This indicates that flow noise contributes a significant amount
of energy to the field. Locally, register flow noise contributes up to
50% of the acoustic energy with the remaining 50% coming from the fan. 3
Near the blower, however, the sound field is dominated by the blower.
At 500 Hz in Plane 2, Figure 26 shows the active intensity to be I

greatest between the middle vents and the passenger's side vent and to
the side of the passenger's side vent. The active intensity is also
I
fairly high in front of the middle vents. The intensity to the side of
the passenger's side vent is thought to be high due to the proximity of
the measurement location to the fan. The fan also contributes to the
measured intensities between the middle vents and the passenger's side
vent and in front of the middle vents, but to a lesser extent. As one
would expect, the influence of the fan is inversely proportional to
distance from the fan and decreases as other sound sources are
approached. With the blower removed from the system, Figure 27, like
Figure 25, shows that the active intensity is greatest over the middle
vents. The passenger's side vent contributes a lesser amount of energy
to the field and the driver's side vent contributes the least amoun, of
energy. Locally, the middle vents contribute most of the acoustic
energy to the field compared to the fan. Overall, flow noise is a
significant contributor to the measured sound field.
Figure 28 shows that in Plane 2 at 1000 Hz with the blower
installed in the system, the active intensity is greatest in the area
between the middle vents and the passenger side vent. This is due to
the proximity of the measurement location to the fan. The middle vents
contribute to the sound field as well but to a lesser extent. The
driver's side vent contributes only a small amount of energy to the
field.
In Plane 2 at 2000 Hz with the blower installed, the active
intensity is seen in Figure 29 to be greatest in that portion of the
measurement plane located to the side of the passenger's side vent. The
active intensity is also fairly high between the middle vents and the
passenger's side vent and over the middle vents. The contribution of
the blower to the local sound field appears to decrease with distance
37
I
away from it, with flow noise sources dominating locally. Figure 30 I
shows that with the blower removed from the system the middle vents are
the dominant sources of sound. The passenger's side vent and the
driver's side vent contribute approximately equal amounts of energy to 3

the field. Comparing the active intensity levels measured with the
blower in place to the levels measured with the blower removed indicates 3
that flow noise is a very important source of sound in Plane 2 at this
frequency. U
Figures 23 through 30 show that both the fan and register flow
noise are significant contributors to the intensities measured in Plane
2. With the blower installed, the active intensities tend to be highest
between the middle vents and the passenger's side vent and in that area
of the measurement plane located to the side of the passenger's side I

vent. This is undoubtedly due to the proximity of the measurement 3
location to the fan. With the blower installed, the middle vents are
also primary contributors to the sound field, but, in general, to a 3

lesser extent. The driver's side vent contributes little energy to the
field. I
With the blower removed from the system, the high active intensity 3
levels between the middle vents and the passenger's side vent disappear
as do the high activa intensity levels in that area of the measurement 3

plane located to the side of the passenger's side vent. In this
configuration, the middle vents are the dominant sound source. The I
passenger's side vent contributes to the sound field as well, but to a g
lesser extent. Of all four vents, the driver's side vent contributes
the least amount of energy to the field. 3

38
I
Figures 31 through 38 present intensity vector maps taken in
Plane 3. As in measurement planes I ard 2, maps taken with the blower
in place are presented first and then, for comparison purposes, maps
taken with the airflow supplied by an acoustically isolated source are
shown.
In Plane 3 at 250 Hz, Figure 31 shows the intensities to be
greatest at the fan inlet. There is a substantial contribution also
from the area at the back of the fan. Fan noise may be leaking through
the gap at the juncture of the two sections of the casing. With the
blower removed, Figure 32 fails to reveal a single dominant sound source
region. Sound is radiating from the surface enclosed by Plane 3 in a
fairly uniform fashion. Slightly higher active intensity levels were
measured at the inlet, at the back of the fan, and between the blower
and the duct. In general, however, the flow noise measured in this
plane at this frequency is negligible and by comparing the maximum
active intensity level measured with the blower in place to the level
measured with the blower removed, it is obvious that the blower is
totally dominant at 250 Hz in Plane 3.
Figure 33 shows that at 500 Hz the fan inlet is the dominant sound
generating region. As at a frequency of 250 Hz, fan noise appears to be
leaking through the gap at the juncture of the two sections of the
casing. Figure 34 shows the fan inlet to be the dominant noise source
region even though the blower is removed from the system. However, its
maximum value is 54.3 dB; some 12.5 dB lower than the maximum intensity
value measured with the blower in place. Obviously, the fan is the
dominant noise source at 500 Hz in Plane 3.
39
I
At 1000 Hz, the fan inlet is again identified in Figure 35 as the 3
dominant source region. As at 250 and 500 Hz, souad leaks are thought
to be present at the casing juncture. With the blower removed,
Figure 36 still shows the most intense sound to be radiating from the 3
fan inlet. The magnitude of the intensity at the inlet is 13.2 dB lower
than with the blower installed. Obviously, at this location and 3

frequency, flow noise is negligible compared to the noise generated by
the fan.
Figure 37 shows that at 2000 Hz the intensity is highest at the 3

lower part of the fan inlet and at a corresponding location behind the
fan. Figure 38 also shows that with the blower removed the intensities 3
are greatest at the fan inlet and at a corresponding location behind the
fan. However, with the blower removed, the active intensity is reduced I
by 5.5 dB. 3
Figures 31 through 38 show that the blower is the dominant noise
source in Plane 3. Unlike in Planes I and 2, distinct secondary flow
noise sources were not identified. Ihe source at the fan inlet which
appears in the measurements performed with the blower removed is I

probably due to the delivery of air to the system. 3
Figures 39 through 43 are intensity vector maps of measuremenL 3
Plane 4. All data were taken with the blower installed in the system.
At 125 Hz (Figure 39) the intensities are largest at the entrance I

to the plenum. The active intensity fans out radially from that
location. The measurements performed at the back (opposite the area
where the flow enters the plenum) indicate that the noise emanates from 3
40 |I
the area where the flow enters the plenum. It is noted that the
intensity values measured in this plane probably contain contributions
from the fan.
Figure 40 (250 Hz) indicates that the intensities are greatest at
the area where the flow enters the plenum. The measurements taken at
the back of the plenum also indicate that this area is a source region.
Figure 41 (500 Hz) reveals that the active intensities are fairly
uniform over the measurement plane. The sound from the fan is affecting
the measured intensities considerably.
At 1000 Hz the intensities are shown in Figure 42. They are
greatest over a relatively large area at the back of the plenum at the
location of the door used to control the flow to the floor vents. The
fan appears to contribute less to the measured intensities at this
frequency than at 500 Hz.
The intensity vectors farthest from the HVAC system in Figure 43
indicate that at 2000 Hz sound is emanating from the lack of the plenum
at the location of the floor vent door. Some blower noise is
contributing to this frequency band, however.
The intensity maps presented in Figures 39 through 43 appear to be
influenced by fan noise, particularly at 500 Hz and above. At the lower
measurement frequencies, the dominant source region is the area where
the flow enters the plenum. At 500 Hz, the area where the flow enters
the plenum is contributing approximately as much energy to the field as
is the area at the ba,k of the plenum. Above 500 Hz, the dominant
source region is the area at the back of the plenum at the location of
the floor vent door.
41
I
4.1.5 Measurement Plane 5 3
Figures 44 through 48 present intensity vector maps taken in Plane
5. All measurements were performed with the blower removed from the
system; the air being supplied by a fan in an acoustically isolated box. 3

At 250 Hz (Figure 44), the intensities are greatest in the area
between the plenum and the exit of the evaporator. The active 3
intensities fan out radially from that location.
Figure 45 (500 Hz) reveals the notch between the plenum and the I
exit of the evaporator to be the dominant source region. The active 3
intensities are greatest at this location and fan out radially from it.
Figure 46 again shows the notch between the plenum and the exit of 3
the evaporator to be the dominant source regior with the 1000 Hz active
intensity vectors fanning out radially from it. An additional source I

near the location of the air entrance contributes to the acoustic field 3
as well. The color representation of Figure 47 shows this second source
as a yellow region at the top of the figure. 3

At 2000 Hz, Figure 48 shows the active intensity to fan out
radially from the area between the plenum and the exit of the 3
evaporator; this is a source region.
Figures 44 through 48 indicate that the area between the plenum and
the exit of the evaporator is the dominant source region in Plane 5 with 5
the blower removed. A source at the location where the air is delivered
to the system may contribute at some frequencies, however, this source 3

is probably due to the experimental set-up (i.e. the flow from the
ducting impinging on the motor backplate). In any case, the
plenum/evaporator exit source dominates te field. 3

42
U
4.2 Sound Power Measurements of the HVAC System
The sound power radiated by the HVAC system was measured by the
method described in a previous section on applications of the acoustic
intensity technique. The system was operating in its maximum AC mode
during these measurements. This operating condition has been defined as
the "standard" condition. For comparison purposes, the sound power was
also measured with airflow only through the system. The mass flow rate
jof air through the system was identical in both cases since the static
pressures at the evaporator exit were matched. Air was supplied to the
system through a 30 ft. long section of dryer hose connected to a quiet
airflow facility that is used in related studies of the aeroacoustics of
flow separations. Initially, the discharge from the dryer hose was
perpendicular to the motor backplate (the wheel was removed). This
resulted in a "jet impinging on a solid surface" and was diagnosed as
being a possible noise source. As a precaution against generating
unwanted noise, the discharge from the dryer hose was subsequently
turned gradually into the scroll using a flexible duct. Broadband five
point sound power measurements confirmed that the resulting noise was
reduced by approximately 2 dB.
Figure 49 shows the sound power spectrum radiated by the HVAC
system operating in the "standard" configuration and with "airflow only"
through the system. The sound power results confirm that the blower is
the douinant noise source, although the data also show that the sound
power radiated by the blower decreases with increasing frequency and
that secondary flow noise sources become more important as the frequency
increases. These results were verified through five point farfield
sound power measurements performed using an octave band sound pressure
43
I
level meter and 20 point sound power measurements performed using the i
intensity measurement system.
4.3 INCE Plenum Box Measurements

Five point sound power measurements were performed for the standard 3
blower mounted in the INCE plenum box, see Figure 50 for a schematic.
The pressure rise across the blower, AP, ranged between 0 and 1.72 3
inches of water in these tests, while the blower speed ranged from
approximately 1700 to 2900 rpm. !
Figures 51 through 54 summarize the INCE box sound power
measurements performed. Each figure is for a constant blowe°- speed and
for various total pressure rises across the unit. It is clear that a 3
line component at the blade passage frequency (BPF) appears in the
lightly-loaded cases. A means of reducing this line component is i

discussed in Appendix 1. As the pressure rise increases, wheel blades 3
are prone to separation (stall) and turbulence noise and turbulence-
turbulence interaction mechanisms begin to mask the time-invariant 3

mechanisms that govern BPF noise generation. When the blower is
installed in the HVAC system, pressure rises in excess of 1.5 inches of i

water are present and as a result, the BPF noise is reduced somewhat i
compared to the lightly-loaded cases.
One may use the data shown in the previous figures to establish a 5
blower speed scaling law. As an example, and because the spectra are
broadband in nature, we choose 500 and 1500 Hz for the analysis. The U
pressure coefficient is defined as:
* = 2p~2/V (17)
I
44 U
where Ptot - total pressure rise across the fan,
Vt impeller tip speed,
and p -air density.
Figure 55 is a cross plot of sound power level versus wheel speed
at constant t. These may or may not correspond to the HVAC system
operating points, but that is irielevant because the sound power
increases as the fifth power of wheel speed, regardless of the pressure
coefficient.
45
CHAPTER 5
CONCLUSIONS
The blower is the dominant noise source in the HVAC system,
particularly at lower frequencies. In general, the sound power radiated
by the blower decreases with increasing frequency and the secondary flow
noise sources (which are broadband in nature) become more dominant at
higher frequencies. At 250 Hz, the blower is contributing approximately
10 times more energy to the acoustic field than is flow noise; however,
at 3 kHz, the blower and flow noise contribute nearly equally to the
acoustic field.
Specific flow (or flow-induced) noise source locations identified
by the intensity vector maps include the region between the evaporator
core and plenum, the inlet to the evaporator core, and the dash board
registers, particularly the middle ones. Denger et al. (1990), document
dye injection flow visualization through all of these regions. Figure
56 shows the quantitative flow path in the blower exit region. Because
the diffuser angle is large (greater than 70), separated flow regions
exist and these are believed to contribute to the secondary flow noise.
Perhaps more important is the contracted flow on the downstream side of
the evaporator core. Figure 57 shows the detailed flow path, the
dynamics of which are lost here, but real-time flow visualization
confirms that there is a great deal of chaotic flow in this contraction.
A gusset is in the flow path and it is capable of creating significant
edge noise. The flow visualization shows considerable unsteady
vorticity at this edge; such vorticity is an efficient generator of
46
aerodynamic sound. Furthermore, a pair of freon supply pipes cross the I
flow path as indicated in Figure 58. The vorticity in the wakes of
these pipes is another acoustic mechanism.
Flows impinging on non-rigid surfaces can also result in sound
radiation due to structural vibration and subsequent re-radiation. The
diagnosis of this mechanism was beyond the scope of the present 3

investigation, but will be the subject of future work.
The dash board registers also were found to be sources of sound I

with the middle registers radiating the most noise, followed by the 3
passenger's side vent and then the driver's side vent. The strength of
the acoustic emission from a register appears to be proportional to the 3

airflow velocity through it. Also, in general, the active intensity
level measured in front of the middle registers decreased by less than3 I

dB when the fan was removed from the system and equivalent airflow was 3
supplied by the acoustically isolated blower. This indicates that
blower noise is not being transmitted through the ductwork and radiating 3
from the vents. It appears that high-velocity airflow over the louvers
in these registers is the primary cause for the sound. A preliminary I

investigation of this type of noise was conducted and the results are 3
given in Appendix 2. The noise appears to depend on the trailing edge
configuration of the louvers since the levels are observed to increase 3

as the louver angle of attack is increased. Although the blower is the
dominant contributor to the sound power radiated by the HVAC system, the I
registers are important local contributors to the acoustic energy field.
Due to their proximity to the vehicle occupants, and due to the fact
that the dash helps shield the occupants from the blower, register noise 3
warrants further attention.
47
I
Identifying particular noise sources within the blower itself was
not within the capability or scope of the current project. In general,
I however, Denger's results indicate that inlet distortions and the
circumferential variation in blade stall may contribute to sound
generation by the blower. Local separation zones on the volute wall may
also produce significant broadband noise, as will the convection of
boundary layer turbulence off the trailing edges of the fan blades.
Basically, there are several noise generating mechanisms that will
contribute to the overall sound produced by the standard HVAC system
blower and only additional research will enable one to establish a rank
ordering of these sources and implement effective noise control
measures.
48
CHAPTER 6
RECOMMENDATIONS FOR QUIETING
It is obvious that in order to quiet the subject HVAC system, the
noise generated by the blower must be reduced. However, as shown by the
results for equivalent airflow being supplied by a quiet external
source, blower noise reductions of 10 dB at 250 Hz and only 3 dB at
3 kHz will result in a total system noise emission dominated by flow
noise created by other system components and ducts. Flow noise in the
present system results from poor flow quality, i.e., flow separation,
flow impingement on solid surfaces, obstructions jutting into the flow
path, high fluid velocity over edges, etc. The poor flow quality
results in severe pressure losses necessitating a given blower to run
faster to overcome these losses and achieve a given mass flow rate.
Figure 55 shows that the faster the blower runs, the more noise it will
produce. Thus, the quality of the flow through the HVAC system and the
noise generated by a given blower are coupled.
In general, improved flow quality and reduced noise can be achieved
if the following features are avoided:
1) sharp corners and bends (to avoid flow separation and
pressure losses)
2) edges conducive to the generation of trailing edge noise
3) sudden area expansions (to avoid flow separation)
4) sudden area contractions (to avoid fluid impingement and
I accelerating flow)
- 5) small cross sectional area ducts (to reduce flow velocity)
6) obstructions jutting into the flow path (to avoid edge tones)
-- 49
U
UI
7) structures prone to vibration (loose doors, etc.) 3

8) air leaks (to avoid edge tones)
Following the above recommendations will reduce flow noise and U

pressure losses. Reduced pressure losses will enable a given blower to
run slower. A 25% reduction in blower speed alone will result in more
than a 75% reduction in the sound power radiated by the system. 3

Obviously, a quieter blower design will reduce the noise radiatcd by the
system even further. I

I
REFEPR-CES
[1] Denger, G. R., M. W. McBride, G. C. Lauchle. An Experimental 3

Evaluation of the Internal Flow Field of an Automotive Heating,
Ventilation, and Air Conditioning System. ARL Penn State Technical I

Report TR 90-011 (July 1990). 3
[2] Neise, W., G. H. Koopmann. Reduction of Blower Noise by
Resonators. J. Sound Vib. 73:297 (1980). 3

5
I
I
I
I
I
50 1
U
IA ERNS-
IOS p
CONSAERNTS 1r)Q(r) 2w P1 r V r
15
G.11(I
IIQ
NTENITIE
f PG
ST
22 W) LNEARFIEL
Ar FTPROESI ,Q IQUANSITIES
G2V, ()4 T ENERGY DENSITIES
Figre .BockDiaramof
easremnt echiqu
52I
30.I i
E 2)0.1 1 Y Y i AI .... / , I I I , I
r r 1 11 1.11 - - 'kI I t I
010. 1 T 7 P ' I T iI 17
T I N II
T11
o 0. 17 77P '- -- '
-24. -14. -4 6. 16.
Distance (cm)
Figure 3. Intensity Field and Wavefronts of Three

Sources Separated by 1/10 Wavelength
53
I
I
I
I
I
I
30.-2 o o o o o o o o o oo o o
3. 0 a0 0 0 0 0 0 0 0 0 000 0 0 0 0 0 00 000 1
S00000000000
00000000000000000000O0000I 00 00000
0 00 00 000 0 00 00 0 00 0 0000 00 0
20. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 n0 0 0 0 0 0 0 0 0 0 0 0
o0 0 0 0 0 o 0 o o o o o o o o o o o o o 0 0 0 0 o
.0 00 00000 0 o o
0 0 0 0 0 0 0
o 00 00 0 000
O O Q ooo oooooI00 0
8 10.-0000000000000000000000000
000000000 0 0000000 0000000
0-00000oo00000oo00000
o 000000 000000 0
0. O0 _ _ 00ooQ
_0oo0_ a_ OC
-24. -14. -4. 6. 16.
Distance (cm)
I
I
I
Figure 4. Sound Pressure Field of Three Sources

Separated by 1/10 Wavelength
54I
C
0. - ..- 44 0
-24. -12. 0 191 E-05 12. 24.

Distance (cm)
Figure 5. Intensity Field and Wavefronts of Three

Sources Separated by 1.35 Wavelength
55
U
I
I
12. 1 r iI i ir i i I
I
I
I
12.0-' ,\ \ '. I I I I 1 1 1 1 1 1 r ? /
11.0 \ " 1 1 1 / / I I I I I I I I 1 1 / /
8o.0 IT /"-\-\ I
t 1-,,', I I
I4~t50 .x ,\I.,... j~b.\ J N.J.... 7// 3

Nodal Lines
-20.0-16.0 -12.0 -8.0 -4.0 -0.0 4.0 8.0 12.0 16.0
I
I
I
I
Figure 6. Active Intensity in Nearfield of Circular
Membrane Excited to (0,3) Mode
56 I
i
I
I
I
I
I
12.0, 000C,000-000ooo
i11.00000000000000000000
0000000000000000000
o.o
9.0"00000000000000000
8.0 OO00000000000000000
m 8.0 0000000000000000000
67.o 0000000000000000000
4.0 000000000000000001
3.0-0000000000 000000
Nodal Unes
-20.0-1 6.0-12.0 -8.0 -4.0 -0.0 4.0 8.0 12 .0'16.0

Distance (cm)
I
I
I
I
Figure 7. Potential Energy for Configuration of
IFigure 6
U 57
I
I
11.0. .,., ,, , , ,, ,
12.0 - ' - ' '- '- t I I - - I1. I / /
U8.-.%l.1I-I,
C
0 .0 - 1/% -1 %,-.1I
Nodal Unes
, ,, . *L " F. I-
-20.0-16.0 -12.0 -8.0-40-0.0 4.0 8.0 12.0 16.0
Distance (cm) i
I
I
~I
I
Figure 8. Reactive Intensity for Configuration of
6
Figure
58
I
TURNING
VANE
AIR RETURN
PLENUM
PLENUM
INLETS
FACILITY
MOUNTING
ACCESS
CUT-OFF FREQUENCY
70 Hz
WEDGE STRUCTURE
36.5' h x 24' w x 29' d
WORKING VOLUME
30.6 i x 18' w x 22.5' d
Figure 9. The ARL Penn State Flow-Through Anechoic

Chamber
59
cc
I4
to
I1
7L
In
4J
C.,
60U
I'entl
Cotole
Moo1rv
Ix
Ien
InestUrb
Fiur 11 _chmti_ igrmofte3- cosi
Scne
U6
EXI TPORTASSMB
ADJUTABE
BOTTOM.* EACEPT UOM06NNT

INGSLDRONG
PANEL AND EXIT PORT I

. .......
PIEZOMETER PRESSURE 82
SEHIND PANEL U3TADVMIO

3LTO
MOUNTING PANEL ASSY.-
FAN RTAINER
Figure 12. Construction Notes for INCE Plenum Box3
62I
Ic
I0
633
z
U.'
z
:5m a
53.
CL 2 E 3 "g9a
f C
1.IJLJ\\W-.'-4
>V IfI, -f~~. -Fu- 0 C
w 'aU c
P. 0
z -4
I9r
e
/4 e % *~' e., j III I i I I I
I In
In
0 Uo
-4
5 E
64
LU
u
z
j5 Naa
7 7~~
ii4- T f7Y-
I E
N
0
'0 10
I w#1.1
U
, N..
In,. N
I~ eouUloSp
sO E *o oOl~
N
b
I 3
Il
Iw
I, cc
I~c
E ZI
i 5
(f1h W.o
0
4 40. 1
fa
Ln
-
U)4
4 Cd~ I
W>
IE
5o
E
_ _
of
_
C-4 W4
-Iog~
Io
--
Cd
4)
CD. 2 j *-
~d0 d'
IDI
inIn
663
Iu
Iz
MN~~~t T T T : !~ N.T
N
LUU
qww
I LU
CD
In
'a
-- SO-DIBu U
xE N t - 4
k N-II 9
3 I/
IIs
I I.
x $4
0 0
N cc
u*j soU-n 10
z
67$ .
in 0 InU
LUU
ulI
PII
K0
i
:?
j
-W
-2 Wj 0~OS~
,,
L>LI I
41
41
'C4)
UI-I
_ _ I _ _ T_
"II
68~
uj z
z 4
5 - ,''~ IL t,,,
Z~ E
7 7 04. 10 0
LUU
-rI
>V
Ij z.
690
'U Z
z5
L( ZN i4__
ZE -
-C N-
1I
9(- M I' c -J
.. T I - > . 0
40 40
to
'U InI
~
E~~~ ~~ ~ w ac
-l2 .
ID
L~ > E U
N _______________________ ->
I-. 4)I
_ -4
70I
44
A)
in
4
L4)
w
bo
Btal
I
7. 1VIABECP
wI
I I
__ _ __ _ __ _ __ _ _ Z Cm
o E ;; I
7~ T
C14 x
j 0U I
zZ C
0
b- n
E > E
- 4j
INI
a. ca
> C)
0 -O
wo 0z~ll
72~
UL
IL
I~ ~~~C E
Iz 9~lt~~*9*~*i9~*99
!5 C4
.
l ,
m ddd~i.dil
f'diilt
4l&
ItI
__________________________________________
soi
0f ~0 0
N N N - Na
~~~....
P.
n"
oilU~B
li
* U '
SU........9,-
0'12
z0
E E -W
A.4C
I cc
~ -r4
0'
* 73
0
*1 "4
04
'-4
BEST
74 AVAIBL -4COP
w z
5 z
0
to o
Nuco .. 010qN~N 0
40
.~z E w~*~o~I
E 0
IE
4N4
I ii
0
~4
r.
I.... it
ini 0W;
.. ........
.....
0
....
C
Ou
;14
* 75
CCd
Cdd
I- (4
0 flY N
K U
..... ....
Ida'$
" a.1%
....... .......
C4N3 4) I
4
U~x 0 -
NN
E3
in NO P.0O4
76 10
tkI
Uo
IC
Iw
3 0
00
in 0 B
In00 ~
P V0 0 an 0 0
in.UO In LnN
s1pc
so *Do
r, N 0 P, N
N N
CI 4
II / .. *~ **UO)S~9
p
/E
I 77
51
ILei
1
-x I-
7t vi .. If
. I
(4 I C 11 0 N aC)
wZ s oa CNDUII
5 U)
E E -.
C)4
0......................--- --
~1, _
v N
N' N N 34~
78
z
55
x~0
0~ I .....
Sidd
N N N a
-E
- z E
0
UU
Ir -s4)
=y -
C1
4
IL C
I~ 79
I.a 1
5 1
w~c
z zl
I
x x 0
g o Co .... .. 0II" 1 11%%11 C
0 0 n'g 0 C 0 In 0 n 0 n~ 0
V,~~~.'i in N 0
z C N(
SE 0
>E E '-. U
U
>)
'S
s~e-*,*Jlss~.~
St
0 in3
0 0
0 -4)
80I
I
I
I -
I~ 7 7 '7 77'?7
7''''
77 7 7-'
!So4 ;-; , . rcmTT

9.2 EE 7% 7 7 N 7 T7
In 9- ~2
I-. C 0i
x S0 !w~- 1 N
Ev UjU
o
ri -
o • o
w
C4 w
/L Cc ~ - \I /
!z
L 0. '\ iW w'i! 0 In
-0 0~
v
\. Iz>
z- . .. N..- )
!a
~d ~
U41
IfI
LIU
41& . . .. T , t/!/-0
x n0 '. 4
*u 101
v.
0
' no
> 11J'0
N 7 7 17 7 C4______
a T T 7 N N-___
C1o
it
Qj*U~! ~ ~ ~*Ul!
w~) 3c
z TTJTr/
ItT I??4-U
C T
46 .
82I
7 7
CM~~ I TT I1
It7* - 0 (9- in uc
C 0
0 7 ~ 1 0
, :!
- V
oou
7no70 T*7 .O).#.
E E ' .
I.4
ILI
, -/I
cm
o
. ij.\.
'
.
04o
INI\N
'1 0
I~~~~ ~ uj~\N LO
hi L v I \t
I I I I 1 4\
. 0 *O.O)I*!P
I 83
i II
I,-
~ '~ I T I1 I t - o
7 7
m m, I
: 9 02 . .- . . ' /
IFi
IFI
P 0 P a)
n E) >
T I
a..c
7Z0'-
>- E
z
>- )
F 7083
I
I
II
7 77 Z
- N
CL . . . . . , , .. . . . - v vC4/ 1 T \I \
e0
I > V 7. T
U
I 7r
1 V V =I= ( I J 0
E -W
in 'o \ \ \ \ n
:4,0 >
0
n N >
Zl n
0 '0
I 85
z 5
ONC . - Z04
r L,- - 4 - - S - E N TTI
I Io TJ T
IrII /
7 7 . . . . . .
r~tit
4t I
7 T -n /-- u -4
* Co ~ . c -
7im 7 7
-U ..- 02---
!?T- T -P Tc - 7/ 7 c -E / 1 1
LN 7
4
IE
. ~t ItI r
It I I IIr c
--/I \/\- T~-
~~10
I1
ca , -
IEI
86I
T T~
z '- 72 7 IT t
LL Z- .Nc .v~
x
I aa 02aa a~. ~
x
\I I T f
f/m
- 1 ~ " ~ a - i I/ l
IJ
-7 -)
T~ T_ CC _ _ _ _ __ _ _ _ _
o .j
to*uj.pa *u1
> c 7I 0 W
, I 1;4 Es,
E >0
Ul 2
z
EN
z
E
2 J
0
in-.--.
fn/A. -
X E 02
/0-.-- - 0
:sS CL
i£4 ~ ~~ 'I. '\ o
3 87
I
I
~7 77 7 7 7 77 7 7 77 7 IL x , x r t t ¢ t
; i v t :1 tIII
7- - TI
Cc T 1 tV v1212Z
___________0 III I t I7 I\\
I
•
x
7
*
-j - - TF 7 \
-, - T, 1/7. T -J 11
o
I ,, /I ,
I I I II
0J C4.
P
0 0 n
ev, //
ItI 0
N IIl ~ 77 /'
x.
E ft
88j:i#'
- -
T =T0S. 2. -. z NNN 'ti
7 ~ 7~o~~ 7.O~~ 7~ C
w CN
Co7 (Ud
r- 77 7 7
7 ' . -
#. . ; I
zz
UU 89
InI
uj .2 .
z T5r.?T
a\ IL.. % N. NI
Ir TI
xx
*Y m7
L a T ; 2VV L UC
0*
. T-* * 4
~* ~ ~2 ~2(A/1 1 0
T T T , , * * * *
n 4r.4
In
r 4.)
U
E v
C4. N- C4. t
xx.
CT,
90I
I w w
e4~ TTT/r /
- j
co 1 .U 07
-20L
>e T v Tp 17 17 1? T T- I U 0
r * ,*T17L .- p. - ,I
*7 T p 1 17 T e T~e
If IF T I CTZ
w~. J44
PPp.
'o*Uoqp 0M oaosp
0-..--~.'11,in 4)
21 IT 0J0\\~
E 00i
91
uj z
5 iTTTITT//1i?
1. T\\' I - I
ON r\~ TI T T 7/to ZN
7 7
T TiTi1\'\
7-
a* P
ta-. U ODDI -.-.--. 02
I
- 1 4
0 0 004 u
1-4
\zN
92~
ZZ
ZN 7 r '
- 7" 01 cc
>ir T 7 T-r
07TTT
7
007-.7 1_____ 01
z 'rT T 7 -7 7 7 I
P in 0 41
WD oouolg~p
0.
wjz
-4
coOI!
I 4L
4)
E 1-1 I - 0
UL
93
In
z
zz
LU k
N
N
9 493 L .
L;C- U U4
>0r
0U n :Sq n333.3:.
'333 0
-)
I3n1 0
5 n~,/d
n 0. 0IS
I
z w&auojm~p U' a
DUOoP 1
r:-E
p- 1
orz I-I4
zL
Cqn
U) x
C -4
941
IL
IA
r ITv ' -- - I
Ii M
.....
UU
U-
t
00
0
0 on ~ n
so *Duo ag so *DOC1SIp
>E
-E
> E E4.)
z1d LU,
4)
'aa
EZ +
/t
* 95
I
I
" i I
in w
UJ z
z5
ZNE
i-N . i1..
.. . .
* n 0 ini
-
No eoa a
- 4- --
.j-a.
I"l
'' *eWD
SOI a 9 . 7
9 S N
- E -E -4St01~
Iin
z 0a
CLi
E6 E I
\4. )
961
44
441-
01
I4
BES
97AAkARLCP
ZE
4 ,C4
ReJrv.TV.,e
0 v
In In0 - i
n -
'A
0 , n I0 1
Z) *
3 E E >-
LIn
ui)
41
z .
SI
98
0 In I
70.00 I
Airflow Ony
Standard Configuration I
3 60.00-
0
-D 50.00-
0
IL
:3
0
40.00-
C/)
30.00-i
0.00 0.50 1.00 1.50 2.00 2.50 3.00
Frequency (kHz)
Figure 49. Sound Power Radiated by the HVAC System

Operating in its "Standard" Condition
and With Airflow Only
99
2m :
I3'
4 10
JI
(CO N CRETE)
FIVE POINT POWER MEASUREMENT

INTENSITY MEASURED AT 5 PTS.
AS =4 M 2 I
Figure 50. Measurement Positions Used in INCE Plenum

Box Sound Power MeasurementsI
100I
100 Spi vs. Static Pressure at 1700 RPM
0
8n
- 40
05
0.5
~20
ci)
S 0
0
2 500 1000 1500 2000 2500

Frequency, Hz
Figure 51. Sound Power Measurements of the Blower

at ±hIti RPh
101
100 SPL vs. Static Pressure at 2000 RPM3
rH8OA P =0 IN./H 2 0
460
ai)
a4 0 20
0 500 1000 150020050

Frequency, Hz3
Fixure 59. ZCUzT rw- 11,aS I~ik b -Cf t~he Blover
at 2000 RPM
1023
10 SPL vs. Static Pressure at 2500 RPM
:Rz90
80 ______
ai)
70A P = 0.28 IN./H 2 0
~60 05
320
0
a410
00 500 1000 1500 2000 2500

Frequency, Hz
Figure 53. Sound Power Measuremencs of the Blower

at 2500 RPM
103
100 SPL vs. Static Pressure at 2900 RPM3
~903
~80
5605
~30
20
~10 .
0
C1 0_ __
10 0 500' 1000 1500 2000 2500 3000

Frequency, Hz3
Figure 54. Sound Power Measurements of the Blower

at 2900 RPMI
1041
I OPEN SYMBOLS: 500 HZ
SOLID SYMBOLS: 1500 HZ
0 * =0.0
I A ~i=0.93
1 0 '=1.25
I70 -W- NSI@ CONST. '
-j
0
0. 40
1 0
I1000 2000 3000

I SPEED, N(RPM)
Figure 55. Sound Power Level as a Function of Wheel

I Speed at Constant Pressure Coefficient
I 105
I
I
I
I
SEPARATED FLOW AT INLET OF
EVAPORATOR CORE ARE IMPORTANT I
SECONDARY SOURCES OF NOISE
EVAPORATOR CORE
BLOWER IS PRIMARY
SOURCE OF NOISE
I
Figure 56. Flow Visualization in the Exit Region

of the BlowerI
106
II I I I I- - -
CONTRACTED FLOW IN
TRANSITION REGION CREATES
HIGH FLOW NOISE
EvApORATOR CORE / EDGE NOISE
Figure 57. Flow Visualization in the Transition Region

Between the Evaporator Core and Plenum
Section
107
VORTEX SHEDDING FROM PIPESI
CAUSES HIGH FLOW NOISE
EATE
EVAPOATORCORE
Fiur lwVsaiainUptemo etrCr 38

and_ DitiuioIut
N108
APPENDIX .
HVAC System Noise Reduction Using Resonators

Measurements of the sound power radiated by the HVAC system
operating in its maximum AC condition showed a tone at the blade passage
frequency (BPF) to be several dB above the surroundin bioadband noise.
The BPF tone is also one of the most irritating components of the
overall spectrum and becomes increasingly noticeable as the total
pressu;re rise across the blower is reduced. Therefore, if the system
losses and wheel speed are decrcased, the BPF tone may become an even
more significant problem. Following Neise and Koopmann (1983), the use
of a quarter-wavelength resonator tc reduce the BPF noise was
investigated. Figure 1.1 shows the concept employed. The body of the
resonator was fabricatec from sheet stainless steel and the piston was
teflon. The perforated area at the scroll cutoff was very fine metal
screen. ine polypropylene throat on a production blower was cut cut and
replaced with the resonator. Tests were conducted in the INCE plenum
box, but only the sourd pressure level at a few spatial locations was
measured. The results were found to be independent of microphone
position. Shown in Figure 1.2 are typical experimeLtaL results for the
blower operating under free-delivery conditions. The resonator was set
(tuned) to cancel pure tone energy at 1460 Hz. The fan speed was varied
such that the BPF occurs at frequencies both above and below this tuned
frequency. The results show, quite dramatically, the effectiveness of
the reso-tator. The tonal is reduced by up to 10 dB, almost to the point
109
3
where it is of the same level as the broadband noise. Further testing i
with the resonator has not been pursued because system losses must be
reduced before such noise control devices warrant serious consideration. I
i
i
i
I
i
I
I
I
I
I
I
I
I
U
110 I
i
X/4 RESONATORS
RESONANCE CHAMBER
PISTON ADJUST
PERFORATED
AREA
Figure 1.1 The 1/4 Wavelength Resonator Concept
111
I
I
I
I
I
EXPERIMENTAL DATA FOR RESONATOR
TUNED TO 1460 HZ
35 I
2100 RPM (1540 Hz)
30 _ _
,,1854 RPM (1360 Hz)
I
_ _ _I 1 _1
25- 1995 RPM (1460 Hz)

co i
0
I
0 0 500 1000 1500 2000 2500 3000i
FREQUENCY, HZ
10 _____
Figure 1.2 Experimental Data Showing the Effectiveness

of the 1/4 Wavelength ResonatorI
112
APPENDIX 2
Register Noise
Acoustic intensity measurements performed in Plane 2 showed the
HVAC system registers to be significant local sources of noise.
Although their contribution to the total sound power radiated by the
system is small, their proximity to the vehicle's occupants justifies
giving them additional attention. A simple experiment was conducted
which gives further insight into the noise generating mechanisms of
registers. Noise measurements were performed on the HVAC system
J operating with all registers open, and with the middle two registers
closed. Closing the middle two registers results in an approximate 40%
decrease in the mass flow rate through the system and an approximate 20%
increase in the flow velocity from the two remaining registers. Both
the acoustic intensity and sound pressure level (SPL) were measured
approximately 18 inches in front of the middle registers. The SPL
measurements were performed in addition to the intensity measurements
because of interest in frequencies higher than the 3 kHz limit of the
intensity system and in order to help verify the results. Airflow
I equivalent to the throughput when operating in its maximum AC condition
(when all vents are open) was supplied to the HVAC system through a
dryer hose connected to the ARL Penn State quiet airflow facility.
Figure 2.1 shows relative intensity levels measured with all vents
open and with the middle vents closed. Table 2.1 presents the SPL
results. The SPL data show a 13 dB noise decrease at 125 Hz and a 5 dB
noise increase at 8 kHz when the two middle vents are closed. When the
middle vents are closed, the high-frequency noise radiated from the
I113
I
I
outer vents increases very noticeably. This is interesting in that
although the mass flow rate through the system decreases by
approximately 40%; the higher frequency noise still increases. The I

velocity through the outer vents is increased under these conditions and u
a dipole source mechanism is known to exist at the trailing edges of the
louvers. Such sources have a radiation efficiency that increases as the 3

sixth power of velocity. Thus, the high-velocity airflow over the
louvers appears to be an important noise source with the magnitude and I

location of spectral peaks dependent on the flow velocity. The louvers
act as little airfoils in a high-velocity airflow and radiate sound from
their trailing edges. 3

In order to help verify the register noise hypothesis, a 30 ft.
section of dryer hose was connected to a blower located in an 3

acoustically isolated box, and a stock register and insert (to hold the
register) were placed on the exit end. Sound pressure level I

measurements were performed when the register directed the airflow 3
straight out, and when the airflow was directed at a rather sharp angle
(louvers set at approximately 45°). The airflow rate was approximately 3

equal to that through a given register when the HVAC system is operated
in its maximum AC condition. For comparison purposes, SPL measurements I

were also obtained for the noise emitted by the airflow exiting the hose 3
without a register or insert present. The measurements were performed
approximately I ft. to the side of the hose exit and are depicted 3
schematically in Figure 2.2. The measurement results are summarized in
Table 2.2 and show that, compared to the open jet, the louvers are very I
significant high-frequency noise sources, particularly when the airflow 3
is directed at a sharp angle.
114 3
I
Table 2.1 Octave Band SPL for HVAC System
with Air Supplied by an Auxiliary
Quiet Airflow Source
Octave Band Measured SPL Measured SPL

Frequency (kHz) All Vents Middle Vents
Open (dB Closed (dB
re. 20 pPa) re. 20 pPa)
0.125 65.0 52.0

0.250 62.8 57.5
0.500 60.2 57.2
1.000 58.3 58.0

2.000 56.4 57.7
4.000 55.8 58.3
8.000 53.3 58.3
Table 2.2 Octave Band SPL Measurements of

Register Noise
Octave Band Open Jet Register Register

Frequency No Register Louvers at 00 Louvers at 450
0.125 46.5 48.0 47.0
0.250 39.5 43.0 43.5
0.500 34.0 41.5 44.5
1.000 29.0 43.0 47.0
2.000 24.0 42.0 45.5

4.000 <20.0 40.0 44.0
115
all vents open
x middle vents closed
CIO
r
CE
0.00 0.50 1.00 1.50 2.00 2.50 3.003

Frequency (kHz)
Figure 2.1 Relative Intensity LevelstheMeasured

MiddleI With
All Vents Open and With
Vents Closed
116I
BASELINE MEASUREMENT - OPEN JET, NO REGISTER OR INSERT
TO ACOUSTICALLY
ISOLATED BLOWER
DRYER HOSE (30 ft.)
OPEN JET
(HOSE EXIT) \
AIR FLOW
MEASUREMENT WITH REGISTER AND INSERT
TO ACOUSTICALLY
ISOLATED BLOWER
DRYER HOSE (30 ft.)
REGISTSPL METER
AIR FLOW
Figure 2.2 Schematic of the Register Noise Measurements
117

A231288 PDF

Uploaded by

Copyright:

Available Formats

A231288 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A231288 PDF

Uploaded by

Copyright:

Available Formats

00 4 4

The Pennsylvania State University

Technical Report No. TR 91-001

Supported by: L.R. Hettche, Director

Approved for public release; distibution unlimited

The heating, ventilation, and air conditioning (HVAC) system of an

automobile can be a source of acoustic annoyance, particularly when it

is operated under maximum airflow conditions. In order to suggest

possible design changes for the purpose of noise control, a systematic

characterization of the acoustic sources and mechanisms must be

conducted initially. This report addresses such a characterization for

a typical automotive system. The approach is based on acoustic

intensity measurements of the st.nd-alone HVAC system operating under

maximum airflow conditions in the ARL Penn State flow-through anechoic

chamber. The experimental data indicate that the centrifugal blower is

the dominant low-frequency source of noise, while separation zones and

sources of noise that become increasingly dominant as the frequency

exceeds 3 kHz. Qualification of these identified sources of sound is

aided by detailed flow visualizations of the subject system.

Recommendations for acoustic improvements to the system are provided.

LIST OF TABLES ........................... 9

LIST OF SYMBOLS ......................... 10

CHAPTER 1. INTRODUCTION. .................... 12

CHAPTER 2. ACOUSTIC INTENSITY TECHNIQUE .. ........... 15

2.1 Background. ...................... 15

CHAPTER 3. EXPERIMENTAL PROCEDURES FOR THE HVAC SYSTEM .31

3.1 Acoustic Intensity Mapping ................ 31

CHAPTER 4. EXPERIMENTAL RESULTS .................. 33

4.1 Nearfield Acoustic Intensity Maps of

4.1.1 Measurement Plane 1 ................ 34

4.2 Sound Power Measurements of the HVAC System . .43

CHAPTER 5. CONCLUSIONS ..................... 46

CHAPTER 6. RECOMMENDATIONS FOR QUIETING .. ........... 49

APPENDIX 1. HVAC System Noise Reduction

APPENDIX 2. Register Nnise ................... 113

Figure Number Title Page Number

1 Intensity Vectors, Wavefronts arid Contours

2 Block Diagram of Measurement Technique 52

3 Intensity Field and Wavefronts of Three

4 Sound Pressure Field of Three Sources

5 Intensity Field and Wavefronts of Three

6 Active Intensity in Nearfield of Circular

7 Potential Energy for Configuration of

8 Reactive Intensity for Configuration of

9 The ARL Penn State Flow-Through Anechoic

10 The HVAC System Situated in the Anechoic

11 Schematic Diagram of the 3-D Acoustic

12 Construction Notes for INCE Plenum Box . . 62

13 INCE Plenum Box Used in Production Blower

14 Measurement Plane 1 Intensity Vector Map,

15 Measurement Plane 1 Intensity Vector Map,

16 Measurement Plane 1 Intensity Vector Map,

17 Measurement Plane 1 Intensity Vector Map,

18 Measurement Plane 1 Intensity Vector Map,

19 Measurement Plane 1 Intensity Vector Map, I

20 Measurement Plane 1 Intensity Vector Map,

24 Active Intensity Data of Figure 23 with Color

26 Measurement Plane 2 Intensity Vector Map,

27 Measurement Plane 2 Intensity Vector Map,