null  null
Acoustics 08 Paris
Noise source mapping for trucks, part 1: development
and design
Kenneth Plotkina , Yuriy Gurovicha , William Blakeb and Paul Donavanc
a
b
Wyle Laboratories Inc., 241 18th Street S., Suite 701, Arlington, VA 22202, USA
Naval Surface Warfare Center (Ret.), 6905 Hillmead Road, Bethesda, MD 20817, USA
c
Illingworth & Rodkin, Inc., 505 Petaluma Blvd. South, Petaluma, CA 94952, USA
[email protected]
1869
Acoustics 08 Paris
Mapping and quantifying noise sources on trucks under actual operating conditions on the road are important for
traffic noise modeling and mitigation. The purpose of this study is to develop a practical truck noise source
localization technique using acoustic beam-forming. An experimental 70+ microphone elliptical array was
designed and fabricated for truck testing. Beam-forming software was developed and implemented using a
computerized data acquisition system. Proof-of-concept tests were performed at low-speed and high-speed truck
testing facilities for a representative sample of trucks with widely different characteristics to validate the
measurement system performance. The measurement system design parameters were verified experimentally,
and certain improvements to the system were recommended for future implementation based on the field
experience. The developed beam-forming measurement system provided adequate noise mapping and
localization for various noise sources on trucks, stationary and moving with the speed up to 50 mph. The results
of the proof-of-concept testing presented in an accompanying paper (Part 2) confirm that the developed
microphone array, data acquisition system and beam-forming software performed generally as expected and
required no major adjustments. This ongoing project is funded by the National Cooperative Highway Research
Program of the Transportation Research Board of the National Academies, USA.
1
observed in typical passby measurements. The objective of
this study is to use noise-source mapping techniques to
accurately localize, identify, and quantify the noise sources
on typical commercial trucks operating in the actual
roadway environment. This paper demonstrates a practical
technology developed for truck noise source localization
and describes the experimental measurements performed
during the proof-of-concept truck testing of the technique.
Introduction
Heavy trucks are significant contributors to overall traffic
noise levels. At highway speeds, the noise level produced
by heavy trucks is about 10 dB greater than that of light
vehicles. As a result, every one truck in the traffic flow
contributes the same amount to the average noise level
values as 10 light vehicles. Because of their contribution, a
thorough understanding of the trucks as a noise source is
crucial to the prediction and mitigation of traffic noise.
Noise from heavy trucks originates from a variety of
sources, which include exhaust stack outlet, muffler shell,
exhaust pipes, engine block, air intake, cooling fan, tires,
and aerodynamics. The relative contributions of these
sources vary with vehicle type, operating condition, and
(for tire noise) the type of pavement.
2
Noise mapping techniques
Acoustic beam forming is considered the most suitable
noise mapping method for this application. It is capable of
mapping both vertical and horizontal distributions, and
implicitly conveys spectral information about sources
under actual operating conditions. Traditional methods
such as near field measurements, component wrapping,
removal and substitution of components during stationary
and moving tests are time- and labor-intensive, performed
on a relatively small number of vehicles, and are not
applicable
in
uncontrolled
roadside
conditions.
Substantially higher productivity could be achieved with
remote sensing methods, such as acoustic holography or
acoustic
intensity
measurements. However, the
microphone array used for acoustic holography must be
physically as large as the source region of interest. It is
also not suitable for passby applications when the array is
stationary and the source is moving. As with traditional
methods or sound intensity, this method could not be
employed for uncontrolled passby testing.
For modeling and abatement of traffic noise, the barrier
performance of sound walls depends on the assumed
distribution of noise source heights. Since trucks contribute
the tallest noise sources, highway noise walls are typically
designed so the top of the exhaust stack was obscured from
the receiver sight under the assumption that exhaust noise
is a major source. The current treatment of truck noise for
highway conditions is often simplistic, placing about 50%
of the source strength at a height of 3.7 m (12 ft) and the
other half at ground level, independent of vehicle speed or
pavement type.
A number of recent observations, however, challenge the
current treatment of trucks and lead to the need for new
research. First, as a result of the federal regulations in the
U.S., truck noise levels have been incrementally lowered
over the past few decades. In achieving this lower level of
noise performance, engine and exhaust noise have been
effectively addressed.
The more compelling method of localizing sources has
been the application of acoustic beam forming, as
described in [1]. Beam forming techniques in a horizontal
direction have also been employed in French research
specifically on truck sources [2]. Acoustic beam forming
uses an array of microphones to focus measurement on a
specific point on an imaginary source plane near the actual
source. This focus point is electronically swept across this
plane, and the sound pressure level is determined.
From recent studies performed in Europe, a much larger
contribution of tire/pavement noise, 63% to 84%, has been
found for highway speeds. A few studies have also shown
that there is a strong dependence on truck tire type. From
research work related to the application of quieter
pavements, reductions in traffic noise have been measured
consistent with the reduction of tire/pavement source levels
when pavement modifications have been made, which goes
beyond what would be predicted based on the current 5050 split of tire/pavement and elevated sources on trucks.
In more advanced approaches, which should be used in the
case of moving vehicles, the source plane moves with the
vehicle and points are scanned over the time the vehicle
goes by. This requires the algorithm to account for both the
moving source plane and Doppler shifting of the sound
during the passby. In this manner, several slices in time
can be evaluated near the position of interest, such as the
For all these reasons, it is essential to have more
information about truck noise sources than what can be
1870
Acoustics 08 Paris
time when the maximum passby noise level occurs, which
improves the accuracy of the resulting beam forming.
3
Because of the large physical size of the source plane for
trucks, spherical beam forming is the only appropriate
approach for the current application. It assumes that path
lengths for different points on the source plane are
different and are accounted for by including spherical
divergence.
Performance of the notional array that is illustrated in Fig. 1
was the baseline for the current design study. It represents
one of the possible approximations to the array used in the
2005 demonstration tests mentioned above. The side lobe
Microphone array design
Approx.
Spiral
Angle
The algorithm used in this study incorporates a crossspectral density (CSD) method, which was initially
developed for resolving locations of stationary sources in
wind tunnels [3]. This method has an advantage over the
classical “delay and sum” method, e.g. [1], by providing a
better reduction of background contamination over the
frequencies of interest. In our case wind noise at
microphones could be better reduced by employing a
complete two-dimensional CSD matrix for all 70
microphones. The array response is otherwise analogous to
that of a parabolic radio antenna for which the main lobe
can be computationally directed to the source. The signal
processing used here included spherical divergence, source
tracking, and Doppler shift correction.
(a)
(b)
Fig.1 Notional 90-element circular spiral array with 15
spokes of 6 elements each and inner and outer radii of 0.5
and 1 m (20 and 50 in), respectively: (a) array pattern,
(b) directivity pattern of array.
Several issues had to be resolved before the beam forming
could be used for the truck application. The first was a
spatial resolution of the technique for the frequencies of
interest for truck noise. In several successful applications,
the revealing “pictures” of sound are typically higher in
frequency, above 1.5 kHz. In part because of the longer
wavelengths and the limitations of the array, at frequencies
below 1 kHz the source regions may appear quite large and
source identification uncertain. The second was the sourceto-array distance. For controlled tests, distance can be
optimized to be relatively close to the vehicle. For roadside
measurements, practical issues of safe access and not
distracting drivers limit how close the array can be
positioned to the lane of travel. Thirdly, practical issues
such as the effect of large vehicle wakes, random
turbulence and other background noise were of concern.
suppression method used here was developed in [6] to
provide minimum number of redundant microphone
separations to effectively optimize the required number of
microphones per unit area needed for a given tolerated side
lobe level. The resulting spiral angle and the spoke
configuration introduce spatial irregularity into the circular
array, thus enabling required side lobe suppression. The
circular array for this application, however, provides
unnecessary localization in the horizontal direction and
insufficient localization in the vertical direction. Given the
intrinsic horizontal localization afforded by the pass-by
itself, a deformed array was considered to optimize the
effective localization vertically and horizontally for
enhanced performance at low frequencies. Deformation of
the parent circular array into an ellipse provides a means to
tailor these arrays for truck pass-by. This facet of the array
design represents one of the new products of the current project.
The design of the microphone array was built on lessons
learned during the array-based demonstration tests of
truck-noise source localization conducted under Caltrans
sponsorship in 2005 [4]. That study utilized a 90microphone commercial wheel array WA0890 developed
by Bruel & Kjaer [5] with software that was an early
prototype of that used here. Useful results were obtained,
but there were limits associated with the system being
general-purpose, not optimized for the specific type of
measurements for moving trucks. The data from that
project was used to optimize the methodology with regard
to the number of microphones needed, the configuration of
the array, and software enhancements to utilize the
complete CSD matrix of all microphones.
All of the arrays shown in Fig. 2 have the same area, i.e.
AB=D2, and all have the same area density of elements.
D= 7 ft
A= 9.1 ft
B= 5.37 ft
The primary objectives for the new array design in this
study were: (1) extend the directivity gain to frequencies
below 400-500 Hz attained with the Bruel & Kjaer array in
the Caltrans demonstration of 2005; (2) meet the Bruel &
Kjaer array gain performance (side lobe suppression from
-10 to -15 dB); (3) optimize array performance for an
affordable 70+ element array; and (4) optimize array
geometry to provide for (a) vertical directivity through
vertical array aperture, (b) horizontal directivity through
combined array aperture and cross-range spreading loss
during vehicle pass-by, and (c) array dimensions for
assembly/disassembly in the field.
A= 11.8 ft
B= 4.1 ft
(a)
(b)
(c)
Fig.2 Patterns for various arrays of 70 elements distributed
around 14 spokes of 5 elements each: (a) Circular spiral
array, (b) Ellipse-1 array, (c) Ellipse-2 array. A – major
axis, B – minor axis. Spiral angle = 0.5°.
Fig. 3 shows the similar lobe structures of these arrays at
1 kHz. The projection of the lobe structure has a shape
outline which follows that of the array, but is rotated 90°
relative to it, i.e. increased array length in X-axis results in
1871
Acoustics 08 Paris
(at 250 Hz) for the three arrays that are shown in Fig. 2. In
all cases the closest point of approach (CPA) is 6 m (20 ft).
This sketch shows that spreading loss provides added
discrimination as the horizontal directivity is reduced to
provide for an increased vertical dimension of the array.
One cannot arbitrarily increase the horizontal dimension
since this would provide degradation of the side lobe
structure.
narrow beam in X; conversely, reduced array length in Yaxis results in wider beam in Y.
of Array Alternatives at 250 Hz
5
(b)
Variation in SPL (dB)
(a)
0
-5
Range Limited
Ellipse 2
-10
Ellipse 1
Circle
-15
-80
0
20
40
60
80
The following conclusions were drawn from the design
analysis for the frequency range of the highest A-weighted
sound levels of emissions (during cruise):
Fig. 4 provides a chart that gives the approximate lobe
width, here called “spot width”, as it refers to the
localization in the truck side plane 6 m (20 ft) from the
array plane. It is seen that lengths on the order of 3.7 m (12
ft) are required to localize to within +/- 1.5 m (5 ft). Given
the design requirement of constant area for all arrays, the
degradation in horizontal beam width for the benefit of
vertical discrimination can also be seen.
• A 70-element elliptical array provides adequate
aperture with acceptable side lobe suppression;
• The aspect ratio 1.7 of an elliptic aperture array
provides beam patterns that are geometrically similar to
the array shape at all frequencies of interest;
• Vertical directivity provides a beam focus spot 1.2 m
(4 ft) wide (-6 dB-down) at 465 Hz;
Array Length or
Diameter , ft
• Excellent side lobe suppression of approximately
-14 dB over the range of 250 to 2250 Hz and
approximately -11 dB at 8000 Hz;
La=4
La=6
La=8
• A minimum spiral angle is needed to suppress side
lobes so element supports may be radial;
La=10
La=12
Spot Width, ft
-20
Fig. 5 Approximate profile of sound pressure level for
simple source passing by the array at 250 Hz.
Fig.3 Directivity pattern projections in front of the array
plane shown in Figure 2 at 1 kHz: (a) Circular spiral array,
(b) Ellipse-1 array, (c) Ellipse-2 array.
10
-40
Cross-range, ft
(c)
100
-60
• Side lobes at high frequencies seem well distributed,
i.e. appear amorphous, which minimizes the likelihood
of unwanted highlights that may lead to false source
images.
1
With regard to low frequency performance, it was
concluded that:
0.1
100
1000
10000
Frequency, Hz
Fig.4 Width of the focus spot of arrays of various lengths
or diameters as function of frequency at a distance of 6 m
(20 ft).
• With a 3.7 m (12 ft) major (vertical) axis of the
ellipse, the vertical beam half height (-6 dB) is about
1.4 m (4.5 ft) at 250 Hz. This will allow imaging
resolution to the upper half of a large truck cab.
• The horizontal effective beam width during pass-by,
including both beam width and spherical source
spreading loss, would be about 2.7 m (9 ft) for the minor
axis (horizontal dimension of 1.2 m (4 ft)).
As the truck sources pass by the microphone array, there is
a sound level change at the array microphones that is due
just to the spherical spreading loss resulting from the
varying distance. Ignoring possible effects of source
directivity, which are probably of little concern in the lowfrequency range of interest, this variation in the sound
level provides some localization along the truck. Fig. 5
illustrates this combined effect of spreading loss and
directivity gain. The -6 dB – down points for the pass-by
are indicated along with the -6 dB – down points
The elliptical array described herein represents a new
result of this study. Based on the design analysis for
Ellipse-2 array, a 70-element array schematically shown in
Fig. 6a was selected for experimental engineering and
implementation through the proof-of-concept testing, as
described in the following sections.
1872
Acoustics 08 Paris
80
60
40
20
0
-80
-60
-40
-20
0
20
40
(a)
60
(b)
-20
Fig. 7 (a) Data acquisition system and (b) equipment setup.
-40
signals from the vehicle tracking system for determining
the truck speed.
-60
-80
(a)
(b)
5
Fig. 6 Experimental elliptical microphone array:
(a) 70-element design; (b) array assembly (note 7
additional microphones in the array center).
4
Proof-of-concept testing
The proof-of-concept testing was conducted at the
International Truck and Engine Corp. (IT) test facilities in
Fort Wayne, Indiana. This company has two test tracks.
One is a low-speed pass-by sound pad used for standard
truck pass-by noise emission measurements at speeds up to
56 km/h (35 mph). Fig. 8 shows how the array was
calibrated at this track using an omni-directional
(spherical) loudspeaker on a tripod. The speaker was used
for initial evaluation of the fully assembled system as
known stationary “point” source placed at several on- and
off-axis locations with different distances and heights in
front of the array.
Experimental array engineering
As the result of the development described above, a 70microphone elliptical array was designed and constructed,
with an aspect ratio of 1.7, a width of 1.2 m (4 ft), and a
height of 3.7 m (12 ft). The assembly is shown in Fig. 6b.
The mechanical design of the array, due to a significant
size of the array aperture, contains a metal frame
composed of three separate sections mounted together
vertically and installed on a four-wheel metal base. The
sections, each of approximately 1.2 by 0.9 m (4 by 3 ft) in
dimensions, could be easily disassembled for shipping. 14
PVC pipe spokes, each holding 5 microphones, were
mounted on the frame sections, providing the 70microphone elliptical pattern designed. The identical lower
and upper frame sections hold five spokes each, with five
microphones mounted equidistantly on each spoke. The
middle frame section holds four spokes with five
microphones each. During the field tests of the array,
additional microphones were mounted along the central
vertical axis of the middle frame section, raising the total
number of microphones to 73 or 77 for some tests.
Fig. 8 Array calibration using a spherical loudspeaker.
The array was equipped with the PCB ¼-inch electret
microphones with ICP® preamplifiers. The microphones/
preamplifiers were inserted in holes predrilled in the
spokes of the array, each provided with a windscreen.
Fig. 9 shows calculated and measured images of the
spherical source at a series of frequencies. The color bar
legend indicates approximately equivalent one-third octave
band sound levels in decibels. All images of a single
acoustic source show a single “hot spot” above ground
with its mirror image below ground. At 922 Hz, for
example, Fig. 9 shows elliptical spots whose major and
minor axes are complementary to those of the array. The
vertical -6 dB width of the spot is about 0.38 m (1.25 ft),
while the horizontal width is about 0.67 m (2.2 ft) at a
distance of 6 m (20 ft) and an elevation of 2 m (6 ft). Fig. 9
also shows examples at lower and higher frequencies,
respectively, for the same source at this distance and
elevation. At higher frequencies, the existence of side
lobes in the array directivity due to certain phase
relationships (frequency and spatial aliasing) between the
array microphones creates unreal ‘ghost’ images.
The data acquisition system, shown in Fig. 7 (a), was
completed using a National Instruments Model PXI-1044
embedded controller chassis with twelve data acquisition
cards providing analog-to-digital conversion for total of 80
data channels providing simultaneous signal recording.
The measurement signals from the array microphones were
fed into the PXI channels through 17-meter (50 ft) long
microphone cables. The software for running the system in
real time and transferring data from the PXI to a laptop
computer for post-processing was also developed. The
measurement equipment setup is shown in Fig. 7 (b).
During the proof-of-concept testing described below, one
of the remaining PXI channels was used for recording
the time signal from a GPS-based time code generator
for data synchronization. Another available channel
received a signal from a pair of photo cells installed on
tripods near the microphone array to register truck
passbys. Another PXI channel was used for recording
The lower and higher frequencies define the approximate
limits of the array performance. It can be seen from the
figure that the array shows adequate performance between
approximately 250 and 2000 Hz. As the result of these and
similar tests, it was determined that the array reliably images
1873
Acoustics 08 Paris
Measured Equivalent Ls, OTO at F=276.8555Hz
Theoretical Equivalent Ls, OTO at F=276.8555Hz and GRC= 0.65
6
6
97
96
4
93
0
sy/m
sy/m
95
2
94
93
0
92
91
91
-2
90
-4
-6
-6
90
-4
89
88
-4
-2
0
sx/m
2
4
6
-6
-6
277 Hz
89
88
-4
-2
0
sx/m
2
4
6
Theoretical Equivalent Ls, OTO at F=922.8516Hz and GRC= 0.7
Measured Equivalent Ls, OTO at F=922.8516Hz
6
91
6
91
90
90
4
4
89
89
88
2
88
2
87
sy/m
sy/m
87
0
0
86
86
85
-2
-4
83
84
-4
83
82
82
-4
-2
0
sx/m
2
4
6
-6
-6
922 Hz
-4
-2
0
sx/m
2
4
6
Theoretical Equivalent Ls, OTO at F=1707.2754Hz and GRC= 0.7
Measured Equivalent Ls, OTO at F=1707.2754Hz
6
6
90
90
89
4
89
4
88
88
2
87
86
0
sy/m
sy/m
2
87
86
0
85
85
-2
84
-2
84
Acknowledgments
83
83
-4
-4
82
82
-6
-6
Proof-of-concept tests were performed at low-speed and
high-speed testing facilities for a sample of trucks with
widely different characteristics to validate the system
performance. The results of the proof-of-concept testing
are presented in an accompanying paper [7] and confirm
that the developed beam-forming system provided adequate
noise mapping and localization for various noise sources on
trucks, stationary and moving at speeds up to 80 km/h.
85
-2
84
-6
-6
This paper describes the microphone array, data
acquisition system, and software developed and
implemented for mapping truck noise sources using
acoustic beam-forming. The measurement system design
parameters
were
verified
experimentally.
The
measurement system with the 70+ elliptical microphone
array showed adequate performance at frequencies
between 250 and 2000 Hz with side lobe suppression of
-14 dB, optimized vertical and horizontal directivities, and
suitable handling in the field application.
94
92
-2
81
-4
-2
0
sx/m
2
4
(a) Measured
6
Conclusion
96
95
2
6
97
4
-6
-6
1707 Hz
81
-4
-2
0
sx/m
2
4
6
This ongoing research is funded by the National
Cooperative Highway Research Program (NCHRP) of the
Transportation Research Board of the National Academies,
USA. The authors are grateful to Mr. Les A. Grundman,
Mr. Lee E. Schroeder and other IT employees for their
invaluable active support during the proof-of-concept testing.
(b) Calculated
Fig.9 Images of (a) measured and (b) calculated spherical
source emission at three frequencies; source elevation 2 m
(6 ft), array stand-off at road side 6 m (20 ft).
an omni-directional source at steering angles up to 45
degrees off axis and stand-off distances up to at least 20 m
(65 ft) at road side.
References
A number of tests were performed with the same speaker
mounted on a rear frame of a truck in stationary, idle, and
moving operations. Those tests determined the array’s
ability to localize the truck noise sources (engine, tires and
exhaust) in comparison with the known calibrated source,
as the truck passed by the array at speeds up to 56 km/h
(35 mph).
[1] A. Crewe, F. Perrin, V. Benoit, and K. Haddad, "RealTime Pass-by Noise Source Identification Using a
Beam-Forming Approach", Proc. The Society of
Automotive Engineers Noise and Vibration Conf.,
Paper No. 2003-01-1537, Traverse City, MI (2003)
[2] M. A. Pallas, "Acoustic Behavior of the Noise Sources
of a Truck", Proc., Inter-Noise 2004, Prague, Czech
Republic, International INCE (2004)
For high-speed tests, the microphone array and data
acquisition system were transported to the second IT track.
It consists of a one-mile long multiple-lane loop designed
for conducting endurance truck testing with the maximum
speed limit of 80 km/h (50 mph). The microphone array
was placed at a distance of 6 m (20 ft) from the edge of the
nearest asphalt driving lane. A number of the speaker and
truck pass-by tests were carried out at this location,
including three widely different types of trucks (all by
International®) in various configurations. A medium utility
truck, a long-haul heavy-duty truck and a “severe” truck
used for very heavy payloads were equipped with different
engines, a variety of tires, diverse exhaust configurations
(horizontal, vertical, muffled or “straight-through” exhaust
with no muffler); some contained trailers, fuel tank skirts,
or an aerodynamic wind fairing over the cabin. The trucks
were tested at several speeds, engine RPM's and gear
settings, in cruse, coast down, acceleration and
compression brake modes. This testing was supplemented
with conventional single microphone and sound intensity
measurements. The experimental results of the proof-of
concept testing are presented and analyzed in an
accompanying paper [7].
[3] R. P. Dougherty, “Beamforming in Acoustic Testing”
in Aeroacoustic Measurements, Chapter 2, T .J.
Mueller, ed., Springer Verlag (2002)
[4] W. Blake, P. Donovan, "A New Road-Side ArrayBased Method for Characterization of Truck Noise
During Passby", Paper NCAD2008-73056, Proc.,
NCAD2008, NoiseCon2008-ASME NCAD, July 28-30,
Dearborn, Michigan, USA (2008)
[5] PULSE Beamforming – Type 7768. Product Data.
Bruel & Kjaer Sound & Vibration Measurement A/S,
Naerum, Denmark (2004)
[6] J. R. Underbrink, “Aeroacoustic Phased Array Testing
in Low Speed Wind Tunnels", in Aeroacoustic
Measurements, Chapter 3, T. J. Mueller, ed., Springer
Verlag (2002)
[7] W. Blake, K. Plotkin, Y. Gurovich, P. Donavan,
"Noise source mapping for trucks, Part 2:
Experimental results", Proc., Acoustics'08 Paris,
France (2008)
1874
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement