Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
30 years on the Road To Progressively Better Data
Tire Pavement Interaction Noise and Correlation with Pavement
Texture ParametersPresentation by
Dr. Ricardo Burdisso
1
Lucas Spies1; Sterling McBride2; Ricardo Burdisso3; Corina Sandu4 and Vincent Bongioanni5
[email protected]; [email protected]; [email protected]; 4 [email protected]; [email protected]
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Outline• Introduction• Experiments• Experimental results
• Tread and Non-tread pattern noise• Relationship between pavement profile and noise
• Discussions
2
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Introduction
3
• Tire noise is the main contributor to vehicle noise at highway speeds.
• Typical mitigation is to implement acoustic barriers for main highways and roads.
• The main noise sources for tire-pavement noise (TPIN) have not been accurately modeled.
• An experimental TPIN campaign was undertaken at Virginia Tech for:
• Model development• Empirical and physically based predictions
• Uncover physical insight into TPIN
Donavan, P. (2008) - Exterior Noise of Vehicles
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Outline• Introduction• Experiments• Experimental results
• Tread and Non-tread pattern noise• Relationship between pavement profile and noise
• Discussions
4
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 5
Experiments: Pavements and Tires• US460 Road
• VT SMART road
• 42 tires
• 5 tiresDense graded hot mix asphalt (HMA)
26 pavement sections:• 14 mixes asphalt• 8 concrete• 3 bridges
• 1 Open Graded Friction• 1 concrete section with longitudinal grooves• 7 concrete sections with transverse grooves
SRTT Tire was tested in all pavements
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 6
Experiments: Test conditionsTest information
Number of tires 428 winter tires1 SRTT tire1 slick tire
Hardness range¥ 56 to 79 Shore A
Steady state speeds tested 45|50|55|60|65 mph
Acceleration test 45 to 65 mph
Tire pressure¥ 26|32|40 psi
Ambient temperature range 37°F to 86°F
For tires with D < 700 mm 2012 Chevrolet Impala (FWD)
For tires with D > 700 mm 2017 Chevrolet Tahoe (AWD)
¥ - Tire pressure, hardness, and temperature measured at start and end of each tire test.
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 7
• Noise: OBSI with optical sensor
Experiments: Noise Measurements
OBSI: On-Board Sound Intensity system
Sound intensity probes
Optical sensor and retroreflective tape
• Optical sensor produces a once per revolution signal. It is used to
• obtain vehicle speed accurately. • perform order tracking analysis.
Each peak represents the retroreflective tape going in front of the optical sensor.
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 8
• Pavement profile data was collected using a Sideway-Force Coefficient Routine Investigation Machine (SCRIM) equipped with a profiling laser.
Profiling Laser
Experiments: Pavement Measurements
Pavement profile measurement parameters.
Measurement resolution (after processing) 0.5 - 1 mm
Lowest velocity for noise measurements 45 mph – 20 m/s
Sampling period. ~ 49 microseconds
Sampling frequency 64 kHz
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Outline• Introduction• Experiments• Experimental results
• Tread and Non-tread pattern noise• Relationship between pavement profile and noise
• Discussions
9
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 10
• Noise spectrogram from acceleration test (Tire 12, 45 to 65 mph).
Experiments results: Tread and non-tread pattern Noise
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 11
• Noise spectrogram from acceleration test (Tire 12, 45 to 65 mph).
Experiments results: Tread and non-tread pattern Noise
Noise component a function of vehicle speed: amplitude and frequency increases with speed.
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 12
• Noise spectrogram from acceleration test (Tire 12, 45 to 65 mph).
Experiments results: Tread and non-tread pattern Noise
Noise component with frequency content independent of speed. However, amplitude increases with speed.
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 13
• TPIN can be separated into two components: tread (TPN) and non-tread pattern (NTPN) noise
Experiments results: Tread and non-tread pattern Noise
Order tracking analysis allows to extract the tread pattern noise from the total noise signal
Optical signal used to perform order tracking analysis.
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 14
• Extraction of TPN noise:
1 revolution of the tire (window).
Experiments results: Tread and non-tread pattern NoiseOrder tracking analysis:For each window• Noise signal resampled • Compute DFT• Average DFTs (TPN in
frequency domain)• Take inverse DFT of average
DFT (TPN in time domain)• Subtract TPN signal from total
signal (NTPN in time domain)
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency [Hz]
65
70
75
80
85
90
A-w
eigh
ted
SP
L [d
BA
re 2
0e-6
Pa]
A-weighted Sound Pressure Level (Frequency of Interest)
15
• Tire noise separation results: winter tire – US460 – 60 mph
12
Michelin X-ICE X13(Winter tire)215/60R16
TPN accounts for 23.4% of total acoustic energy.
Experiments results: Tread and non-tread pattern Noise103.4 97.1 102.3
Total noiseTread pattern noiseNon-tread pattern noise
103.4 dBA97.1 dBA
102.3 dBA
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 16
• Tire noise separation results: SRTT – US460 – 60 mph
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency [Hz]
65
70
75
80
85
90
A-w
eigh
ted
SP
L [d
BA
re 2
0e-6
Pa]
A-weighted Sound Pressure Level (Frequency of Interest)
Total noiseTread pattern noiseNon-tread pattern noise
SRTT - Standard Reference Test Tire
Experiments results: Tread and non-tread pattern Noise
TPN accounts for 3.8% of total acoustic energy (for the pavement tested).
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 17
SMART road data – 60 mph
Tread Pattern Noise spectrogram - SMART road - Run # 4 @60mph
10 20 30 40 50 60 70 80 90
Time in seconds
1280
2560
Freq
uenc
y (H
z)
0
5
10
15
20
25
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
Tread pattern noise spectrogram
Experiments results: Tread and non-tread pattern Noise
Non Tread Pattern Noise spectrogram - SMART road - Run # 4 @60mph
10 20 30 40 50 60 70 80 90
Time in seconds
1280
2560
Freq
uenc
y (H
z)
10
15
20
25
30
35
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
Non-Tread pattern noise spectrogram
Effect of different pavement surfaces on the TPN and NTPN
24
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 18
Effect of different pavement surfaces on the TPN and NTPN
OGFCDense graded
HMA PCC section
Sec_F RR_B PCC_1d Sec_K
Concrete with transverse grooves
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
40
45
50
55
60
65
70
75S
PL
(dB
A) -
Ref
pre
ssur
e 20
x 1
0- 6
Pa
TPN spectrum for different pavements - Tire 24 - 60 mph
Sec F
PCC 1d
OGFC
RR bridge
Experiments results: Tread and non-tread pattern Noise
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
65
70
75
80
85
90
SP
L (d
BA
) - R
ef p
ress
ure
20 x
10
- 6 P
a
NTPN spectrum for different pavements - Tire 24 - 60 mph
Sec F
PCC 1d
OGFC
RR bridge
24
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 19
Experiments results: Tread and non-tread pattern Noise
• TPN is produced only by the tread pattern.
• NTPN is mainly produced by the pavement (independent of tread pattern).
• These observations suggest that the characterization of pavement noise should be based only on the NTPN.
• The rest of the results will focus on NTPN component.
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Outline• Introduction• Experiments• Experimental results
• Tread and Non-tread pattern noise• Relationship between pavement profile and noise
• Discussions
20
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 21
Pavement Profile Spectrum
Railroad bridge
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
20
25
30
35
40
45
50
55
60
Dis
plac
emen
t (dB
) - R
ef 1
m
Pavement spectrum - dB scale
100 150 200 250 300 350 400 450 500 550 600
Wavenumber k (rad/m)
20
25
30
35
40
45
50
55
60
Dis
plac
emen
t (dB
) - R
ef 1
m
Pavement profile spectrum - dB scale
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Wavelength (m)
20
25
30
35
40
45
50
55
60
Am
plitu
de (d
B) -
Ref
1m
Pavement profile spectrum - dB scale
Wavelength
Frequency
2kπλ =
Experiments results: Pavement profile vs Noise
• Pavement profile spectrum is computed using the wavenumber transform (plotted vs k).
• It is also plotted vs wavelength (λ) to easily observe periodicity of wavenumber components.
• It is also plotted vs frequency (based on vehicle speed, V) to compare to noise spectrum.
( ) Vf Hzλ
=
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 22Dense graded HMA PCC section
Sec_F Sec_I PCC_1d
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
65
70
75
80
85
90
SP
L (d
BA
) - R
ef p
ress
ure
20 x
10
- 6 P
a
NTPN spectrum for different pavements - SRTT tire - 55 mph
Sec F
Sec I
PCC 1d
Experiments results: Pavement profile vs Noise
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
20
25
30
35
40
Dis
plac
emen
t (dB
) - R
ef 1
m
Pavement spectrum - dB scale
Sec F
Sec I
PCC 1d
Pavement Profile Spectrum (Non-porous)NTPN Spectrum
Non-porous pavements
SRTT
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 23Dense graded HMA PCC section
Sec_F Sec_I PCC_1d
Experiments results: Pavement profile vs NoisePavement Profile SpectrumNTPN Spectrum
Non-porous pavements
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
20
25
30
35
40
Dis
plac
emen
t (dB
) - R
ef 1
m
Pavement spectrum - dB scale
Sec F
Sec I
PCC 1d
OGFC
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
65
70
75
80
85
90
SP
L (d
BA
) - R
ef p
ress
ure
20 x
10
- 6 P
a
NTPN spectrum for different pavements - SRTT tire - 55 mph
Sec F
Sec I
PCC 1d
OGFC
OGFC
Sec_K
Porous pavement
SRTT
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
20
25
30
35
40
45
50
55
60
Dis
plac
emen
t (dB
) - R
ef 1
m
Pavement profile spectrum - dB scale
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
65
70
75
80
85
90
SP
L (d
BA
) - R
ef p
ress
ure
20 x
10
- 6 P
a
NTPN A weighted spectrum - Decibels scale
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
20
25
30
35
40
45
50
55
60
Dis
plac
emen
t (dB
) - R
ef 1
m
Pavement profile spectrum - dB scale
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
65
70
75
80
85
90
SP
L (d
BA
) - R
ef p
ress
ure
20 x
10
- 6 P
a
NTPN A weighted spectrum - Decibels scale
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
65
70
75
80
85
90
SP
L (d
BA
) - R
ef p
ress
ure
20 x
10
- 6 P
a
NTPN A weighted spectrum - Decibels scale
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
20
25
30
35
40
45
50
55
60
Dis
plac
emen
t (dB
) - R
ef 1
m
Pavement profile spectrum - dB scale
24
Tonal noise associated with transverse grooves.
HWY bridge PCC 1g RR Bridge
Experiments results: Pavement profile vs Noise
SRTT
Pave
men
t Spe
ctru
mN
TPN
Spe
ctru
m
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Outline• Introduction• Experiments• Experimental results
• Tread and Non-tread pattern noise• Relationship between pavement profile and noise
• Discussions
25
30 years on the Road To Progressively Better Data
Discussions
26
• A large number of tire noise data was collected using an OBSI system with an optical sensor (for order tracking analysis) under multiple testing conditions.
• Pavement profile data was acquired using a scanning laser.• Tire noise was separated into two main components: tread (TPN) and non-tread-
pattern (NTPN) noise• TPN is due only the tread pattern• NTPN is mainly a function of pavement.
• The NTPN spectrum is correlated to the pavement profile spectrum only over a limited frequency range (~ 200 to 900 Hz).
30 years on the Road To Progressively Better Data
Extras
27
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 28
US460 road data – 60 mph – Tonal noise associated with the pavement distress.
• On the US460 road test section, there was a segment (~300 m) where the non-tread pattern noise component showed tonal components.
• It was speculated that the tonal noise was due to corrugation of the pavement. However, visual inspection of the pavement didn’t revealed these corrugations.
TPIN vs pavement transverse corrugation distress
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency [Hz]
0
0.1
0.2
0.3
0.4
A-w
eigh
ted
p2 rm
s [P
a2
]
A-weighted Power Spectrum of Sound Pressure (Frequency of Interest)
Tone section
Normal section
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 29
US460 road data – 60 mph – Tonal noise associated with the pavement distress.Spectrograms shown are for tire 20 (i.e. SRTT tire). It is important to highlight that the similar behavior was observed in all 5 tested tires.
US460 road - Tire 20 - Total Noise Spectrogram - Tone section
1 1 2 2 3 3 4 4 5
Time in seconds
1280
Freq
uenc
y (H
z)
10
15
20
25
30
35
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
US460 road - Tire 20 - Total Noise Spectrogram - Normal section
1 2 3 4 5 6 7 8 9
Time in seconds
1280
Freq
uenc
y (H
z)
10
15
20
25
30
35
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
TPIN vs pavement transverse corrugation distress
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 30
The non-tread pattern noise can be used to estimate the wavelength of the corrugated pavement.
Non- Tread pattern noise
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency [Hz]
0
0.1
0.2
0.3
0.4
A-w
eigh
ted
p2 rm
s [P
a2
]
A-weighted Power Spectrum of Sound Pressure (Frequency of Interest)
Tone section
Normal section
120 Hz
13.1 2.1 230120
rot tire tirecorrugation
tone
f l Hz m mmf Hz
λ − ⋅ ⋅= = ≅
: Circumference of the tire
: Rotational speed of the tire in Hz
: Frequency interval between noise tones
tire
rot tire
tone
lff
−
TPIN vs pavement transverse corrugation distress
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Wavelength (m)
25
30
35
40
Am
plitu
de (d
B) -
Ref
1m
m
Spectrum - Decibels scale
Normal section
Tone section
The corrugated pavement wavelength was confirmed from direct measurements of the pavement profile
and the computation of the spectrum.
Tone with ~ 230 mm wavelength, which matches the wavelength predicted by the noise data
Non-tread pattern TIPN can be used for efficienlty monitoring pavement distress
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 31
US460 road data – 60 mph – Tonal noise associated with the pavement distress.
Results shown are for tire 20 (i.e. SRTT tire). It is important to highlight that the similar behavior was observed in all 5 tested tires.
Total tire noise Non- Tread pattern noiseTread pattern noise
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency [Hz]
0
0.1
0.2
0.3
0.4
A-w
eigh
ted
p2 rm
s [P
a2
]
A-weighted Power Spectrum of Sound Pressure (Frequency of Interest)
Tone section
Normal section
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency [Hz]
0
0.05
0.1
0.15
0.2
A-w
eigh
ted
p2 rm
s [P
a2
]
A-weighted Power Spectrum of Sound Pressure (Frequency of Interest)
Tone section
Normal section
200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency [Hz]
0
0.1
0.2
0.3
0.4
A-w
eigh
ted
p2 rm
s [P
a2
]
A-weighted Power Spectrum of Sound Pressure (Frequency of Interest)
Tone section
Normal section
120 Hz
TPIN vs pavement transverse corrugation distress
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 32
Experiments• VT SMART road test
Test information
Number of pavements 26
Among the tested pavement there were:• 14 surface mixes asphalt sections• 8 concrete sections• 3 bridges sections• 1 Open Graded Friction Course• 1 concrete section with longitudinal grooves• 7 concrete sections with transverse grooves
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
• There are 26 different pavements.
9/23/2018 33
Extend TPIN model to include pavement parameters.
Monthly Project Update – Confidential & Proprietary to CenTiRe
SMART Road pavement data.
S.R. Bridge
PCC 2
RR Bridge
PCC 1g
PCC 1f
PCC 1e
PCC 1d
PCC 1c
PCC 1b
PCC 1a
L1
KI
JHF
HWY. Bridge
D2
D1
C
B
A E1
E2 G
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 34
Experimental results• Tire noise separation procedure
Noise signal resampled to a fixed number of
points inside each window.Total tire
noise signal
Optical sensor signal The Fourier transform
of the resampled signal is computed for
each window.
Coherent average of the multiple Fourier
transforms computed for all the windows
Inverse Fourier transform to go back to the time
domain
The noise time signal obtained is denoted
as Tread-Pattern noise (i.e. TPN)
The noise time signal obtain is denoted as Non-Tread Pattern Noise (i.e. NTPN)
The Tread Pattern noise signal is
subtracted from the total noise signal
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data9/23/2018 INTER-NOISE 2016 IN HAMBURG, GERMANY 35
Tread Pattern Noise Contribution
No. Total Noise Level [dBA]
Tread Pattern Noise Level [dBA]
Non-Tread Pattern Noise
Level [dBA]
Tread Pattern Noise Contribution [%]
12 103.4 97.1 102.3 23.415 102.1 86.6 102.0 2.818 105.2 90.6 105.1 3.519 102.4 93.1 101.8 12.020 105.0 90.7 104.8 3.8
• Overall A-weighted sound pressure level for all tires (dBA)
• For the pavement tested, the tread pattern noise is not the dominant noise source.
• For a newer/smoother pavement (very limited data), the tread pattern noise component account for about 50% of the total noise (Tire 12).
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Non Tread Pattern Noise spectrogram - SMART road - Run # 4 @60mph
10 20 30 40 50 60 70 80 90
Time in seconds
1280
2560
Freq
uenc
y (H
z)
10
15
20
25
30
35
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
36
Experimental resultsSMART road data – 60 mph – Tonal noise associated with transverse grooves.
Non tread pattern noise spectrogram
HWY bridgePCC 1-a, PCC 1-c, PCC 1-e and PCC 1-gRR bridgeS.R. bridgeThe NTPN spectrogram shows tonal content at certain intervals, considered to be associated with the presence of transversal grooves on the pavement.
24
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data
Non Tread Pattern Noise spectrogram - SMART road - Run # 4 @60mph
10 20 30 40 50 60 70 80 90
Time in seconds
1280
2560
Freq
uenc
y (H
z)
10
15
20
25
30
35
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
37
Experimental resultsSMART road data – 60 mph – Tonal noise associated with transverse grooves.
Non tread pattern noise spectrogram
HWY bridgePCC 1-a, PCC 1-c, PCC 1-e and PCC 1-gRR bridgeS.R. bridgeThe NTPN spectrogram shows tonal content at certain intervals, considered to be associated with the presence of transversal grooves on the pavement.
13.1 2.1 201378
rot tire tiregrooves
tone
f l Hz m mmf Hz
λ − ⋅ ⋅= = ≅
24
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 38
Experimental resultsSMART road data – 60 mph – Tonal noise associated with transverse grooves.The pavement data spectrogram is also computed. The first tone appears at a similar frequency as in the NTPN spectrogram
SMART road pavement profile spectrogram
S.R
. Brid
ge PC
C 2
RR
Brid
ge
PC
C 1
g
PC
C 1
f
PC
C 1
eP
CC
1d
PC
C 1
c
PC
C 1
bP
CC
1a
L2 L1
K J I
H G F
E2
E1
HW
Y b
ridge
D2
D1 C B A
Distance in meters
0
1333
2667
4000
5333
Freq
uenc
y (H
z)
-5
0
5
10
15
20
25
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
30 years on the Road To Progressively Better Data30 years on the Road To Progressively Better Data 39
Experimental resultsSMART road data – 60 mph – Tonal noise associated with transverse grooves.The pavement data spectrogram is also computed. The first tone appears at a similar frequency as in the NTPN spectrogram.
SMART road pavement profile spectrogram
S.R
. Brid
ge PC
C 2
RR
Brid
ge
PC
C 1
g
PC
C 1
f
PC
C 1
eP
CC
1d
PC
C 1
c
PC
C 1
bP
CC
1a
L2 L1
K J I
H G F
E2
E1
HW
Y b
ridge
D2
D1 C B A
Distance in meters
0
1333
2667
4000
5333
Freq
uenc
y (H
z)
-5
0
5
10
15
20
25
Pow
er/fr
eque
ncy
(dB
/rad/
sam
ple)
HWY bridgePCC 1-a, PCC 1-c, PCC 1-e and PCC 1-gRR bridgeS.R. bridge