May 2010
Reprint Report: UCPRC- RP- 2010- 01
Accoouussttiicc Aggiinngg ooff
Asspphhaalltt Paavveemeennttss::
A Caalliiffoorrnniiaann// Daanniisshh
Coomppaarriissoonn
Authors:
Hans Bendtsen, Danish Road Institute— Road Directorate;
Qing Lu, UC Pavement Research Center; and
Erwin Kohler, Dynatest Consulting, Inc.
This report is based on research performed by the Danish Road Institute- Road Directorate
on behalf of the University of California Pavement Research Center for the California
Department of Transportation, and is reprinted here in its original form.
Work Conducted as Part of the “ Supplementary Studies for Caltrans QPR Program” Contract
PREPARED FOR:
California Department of Transportation
( Caltrans)
Division of Research and Innovation
PREPARED BY:
The Danish Road Institute—
Road Directorate and
University of California
Pavement Research Center
Danish Road Institute
ii UCPRC- RP- 2010- 01
DOCUMENT RETRIEVAL PAGE Reprint Report
UCPRC- RP- 2010- 01
Title: Acoustic Aging of Asphalt Pavements: A Californian/ Danish Comparison
Author: H. Bendtsen, Q. Lu, and E. Kohler
Prepared for:
Caltrans
FHWA No.:
CA101735A
Date Work Submitted:
July 2009
Date:
May 2010
Contract/ Subcontract Nos.:
Caltrans Contract: 65A0293
UC DRI- DK Subcontract:
08- 001779- 01
Status:
Final
Version No:
1
Abstract:
The level of noise generated by tire/ pavement interaction of a pavement section changes over time. While the
general consensus is that the noise level tends to increase as the pavement ages, more scientific investigation is
necessary to better understand the process of acoustic aging of pavements.
For more than a decade, independent studies by Caltrans and the Danish Road Institute ( DRI- DK) have
included monitoring of tire/ pavement noise levels on selected pavements. Using data sets collected as part of those
studies, a comprehensive analysis was conducted in this study to characterize the acoustic aging properties of
different types of asphalt pavements. Pavement types considered in the analysis include dense- graded asphalt
concrete ( DGAC), open- graded asphalt concrete ( OGAC), thin open- graded asphalt layer, and porous asphalt
concrete ( PAC).
This report presents the results of the data analysis in terms of the relative changes of tire/ pavement noise over
time for the respective pavements. It also describes the development of an acoustic aging model for asphalt
pavements. The model predicts the increase in noise level as a function of pavement age, traffic volume, and
pavement type, primarily for highways with speeds over 50 mph. Further study is recommended to improve the
prediction model and to integrate the noise model in a Pavement Management System.
Keywords: asphalt pavements, noise, tire/ pavement interaction, acoustic aging
Proposals for implementation: No recommendations
Related documents:
• H. Bendtsen, H., Q. Lu, and E. Kohler. 2009. Temperature Influence on Road Traffic Noise: Californian OBSI
Measurement Study. Reprint report: UCPRC- RP- 2009- 02
• Bendtsen, H. 2009. Highway Noise Abatement: Planning Tools and Danish Examples. Reprint report: UCPRC-RP-
2010- 03
• Bendtsen, H. 2009. Noise Barrier Design: Danish and Some European Examples. Reprint report: UCPRC- RP-
2010- 04
• Q. Lu, E. Kohler, J. T. Harvey, and A. Ongel. 2009. Investigation of Noise and Durability Performance Trends
for Asphaltic Pavement Surface Types: Three- Year Results. Research report: UCPRC- RR- 2009- 01
Signatures:
Hans Bendtsen
1st Author
DRI- DK
John T. Harvey
Technical Review
UCPRC
John T. Harvey
Principal Investigator
UCPRC
S. David Lim
Contract Manager
Caltrans
UCPRC- RR- 2010- 01 iii
DISCLAIMER
This report is based on a subcontract research study performed by the Danish Road Institute- Road Directorate
( DRI- DK) on behalf of the University of California Pavement Research Center ( UCPRC) for the California
Department of Transportation ( Caltrans). The contents of this report reflect the views of the authors and DRI-DK
who are responsible for the facts and accuracy of the data presented herein. The contents do not necessarily
reflect the official views or policies of the UCPRC, the State of California or the Federal Highway
Administration. This report does not constitute a standard, specification, or regulation. The content of the
original is unchanged in this version and has been reprinted with the consent of DRI- DK.
For more information:
University of California Pavement Research Center, Davis
One Shields Avenue, Davis, CA 95616
University of California Pavement Research Center, Berkeley
1353 S. 46th St., Bldg. 452, Richmond, CA 94804
www. ucprc. ucdavis. edu
Danish Road Institute
Report 171
2009
Hans Bendtsen
Qing Lu
Erwin Kohler
Acoustic aging of
asphalt pavements
A Californian / Danish comparison
xx
3
Contents
Executive summary ....................................................................................................... 5
Sammenfatning............................................................................................................. 11
Preface ......................................................................................................................... 16
Forord ......................................................................................................................... 18
1. Introduction and existing knowledge ....................................................................... 21
1.1 Pavement properties and noise generation ......................................................... 21
1.2 The “ Tyre/ Road Noise Reference Book”........................................................... 23
1.3 Literature survey ................................................................................................ 24
1.3.1 The SILENCE inventory........................................................................... 24
1.3.2 US studies.................................................................................................. 25
1.3.3 A Norwegian study.................................................................................... 28
1.4 Important factors affecting noise increase.......................................................... 29
1.4.1 Pavement aging and distress ..................................................................... 29
1.4.2 Acoustical aging........................................................................................ 32
1.5 Noise measurement methods.............................................................................. 34
1.6 Layout of study .................................................................................................. 35
2. Road sections included ............................................................................................. 37
2.1 I- 80 test section by Davis, CA ........................................................................... 37
2.2 LA138 test sections in Mojave, CA ................................................................... 38
2.3 Viskinge test sections, DK................................................................................. 39
2.4 M10 test sections, DK........................................................................................ 41
2.5 Californian investigation.................................................................................... 43
3. Analyses of noise over time...................................................................................... 45
3.1 I- 80 test section by Davis, CA ........................................................................... 45
3.2 LA138 test sections in Mojave, CA ................................................................... 48
3.3 Viskinge test sections, DK................................................................................. 52
3.4 M10 test sections, DK........................................................................................ 61
3.5 Californian investigation.................................................................................... 66
4. Analysis and comparison.......................................................................................... 73
4.1 Comparison of all seventeen pavements ............................................................ 73
4.2 Results divided in four pavement groups........................................................... 79
4.3 Average results for each pavement group .......................................................... 81
4.4 Model for noise increase .................................................................................... 85
5. . Conclusions and recommendations ......................................................................... 87
6. References ................................................................................................................ 91
4
5
Executive summary
The purpose of this project is to contribute to the ongoing international development
in the field of acoustical aging of tire/ pavement interactions by performing a compre-hensive
analysis of some Californian and Danish results from noise measurement se-ries
on asphalt pavements carried out over long time. The main focus is on asphalt
concrete pavements applied on highways. Cement concrete pavements have not been
included in this study.
An international literature survey has been conducted. It shows that the noise level
generally increases as the pavement gets older. For porous pavements ( built- in air
void content of more than 15 % or so) it is a known phenomenon that air voids tends
to clog and that this increases the noise significantly. But for other dense and open
graded ( but not really porous) pavement types there is not much knowledge on which
changes occur in the surface structure causing this increase in noise in the period from
when the bitumen film is worn off until the pavements begin to deteriorate with dis-tresses
like raveling, cracking, etc.
This project focuses on the trend in noise levels measured in the same way – the rela-tive
changes of noise over the years – and not on the actual absolute noise levels.
The objective is to analyze and compare trends in the development of noise over time.
Therefore, it is not so crucial if noise results have been measured by different methods
or by the same method applied by different measurement teams/ organizations. These
factors might influence the actual noise levels and can complicate direct comparison,
but when only trends are compared these differences in measurement methods are not
so important. The vehicle fleet of California and Denmark may differ, for example,
with smaller passenger cars in Denmark. This might influence the comparison of abso-lute
noise levels but will presumably be less important when comparing changes in
noise emission over the years measured at the same site. Only changes in the noise
levels happening over time are included in this document. Changes in other factors
relevant for the description of the development of the physical structure of the pave-ment
surface like texture, porosity, visual signs of wear and tear etc. have not been
considered.
Two well documented long- time noise measurement series from California and two
from Denmark have been analyzed in this project. The results have already been
documented in detail in separate national reports. The objective of the current report is
to perform a comparison study of the trends for acoustical aging found in these four
projects. The University of California Pavement Research Center ( UCPRC) finalized
in 2009 the third year report on annual On Board Sound Intensity noise measurements
on 65 to 76 pavement sections of different ages and mix types in California. Some re-sults
from this project are also included.
6
The two Californian measurement series and the UCPRC study have all been carried
out for California Department of Transportation ( Caltrans) and the two Danish meas-urement
series have been carried out for the Danish Road Directorate. The following
five measurement series are included:
1. Open graded pavement ( OGAC) on I- 80 near Davis, California ( 11 years),
one pavement type.
2. 5 test sections with dense and open graded pavements on LA138 in the Mojave
Desert, California ( 5 years).
3. Around 70 pavements in the UCPRC/ Caltrans monitoring project ( 3 years
and pavements in different age groups and of different mix types).
4. 3 single layer porous ( PAC), one dense graded ( DGAC) and one open graded
( OGAC) pavement at “ Viskinge”, Denmark ( 8 years).
5. 5 Thin Open graded ( OGAC) and one dense ( DGAC) pavement at M10
(“ Solrød”) near Copenhagen, Denmark ( 5 years).
The pavements included in this project were grouped in the following four main types:
• DGAC Dense Graded Asphalt Concrete.
• OGAC Open Graded Asphalt Concrete.
• PAC Porous Asphalt Concrete.
• Thin Open Asphalt Layers.
The following conclusions for highways can be made on the background of this study
on acoustical aging of asphalt pavements:
• The noise level on asphalt pavements normally increases with time.
• The increases occur continuously and before significant pavement deterioration
with raveling and cracks etc. begins.
• There are exceptions where the noise is reduced during the first year of porous
pavement lifetime.
• A linear regression gives a good fit of the relation between pavement age and noise
both for passenger cars and multi- axle heavy vehicles. This was also seen in the
European SILENCE study.
• The yearly noise increase is generally around 2 times higher for passenger cars
than for heavy vehicles.
Different parameters have been used to describe the increase of noise. The increase
in noise is often expressed as dB per year. Two main factors affect the changes on
the noise properties of a pavement. One relates to the physical/ chemical changes in
the materials caused by weather elements and time, and the other has to do with the
wear and tear caused by traffic. It can be argued that the combined effects of both the
physical age of a pavement as well as the wear and tear from traffic determine the
increase of noise.
7
The age reflects an accumulated effect of changing weather conditions like sun radia-tion,
rain, ice, freeze/ thaw, oxidation, etc. In order to try to define an indicator that
combines these two very different factors ( age and traffic load), two artificial indica-tors
called “ Mixed Indicator” ( ΔLMix50/ 50 and ΔLMix25/ 75) have been defined. The noise
increase has been analyzed for five different indicators:
1. ΔLAge: The change of noise level per year ( actual physical age of the pavement).
2. ΔLADT: The change of noise level per 1 million vehicles ( all types) passing per
lane.
3. ΔLHvy: The change in noise level per 100.000 heavy vehicles passing per lane.
4. ΔLMix50/ 50: The change of noise level predicted as a combination of actual
physical age and traffic load where the age counts for 50 % and the traffic
load counts for 50 % called “ Mixed Indicator 50/ 50”. This is a 50/ 50 %
combination of ΔLAge and ΔLADT.
5. ΔLMix25/ 75: The change of noise level predicted as a combination of actual
physical age and traffic load where the age counts for 25 % and the traffic
load counts for 75 % called “ Mixed Indicator 25/ 75”). This is a 25/ 75 % combi-nation
of ΔLAge and ΔLADT.
The results showed that the ΔLADT and the ΔLHvy resulted in the same ranking of the
pavements presumably because there was no big variation in the percentage of heavy
vehicles in the roads included. By using ΔLMix25/ 75 instead of ΔLMix50/ 50 the porous
pavements generally have a higher noise increase than the thin layers which is be-lieved
to be correct as the porous pavements have a tendency of clogging which is not
seen on non- porous thin layers. The selection of ΔLMix25/ 75 instead of ΔLMix50/ 50 carries
implicitly the assumption ( which lies beyond the scope of this study) that traffic ef-fects
dominate over climatic/ time aging effects. Therefore the following three indica-tors
ΔLAge, ΔLADT and ΔLMix25/ 75 are considered to be the most relevant for the descrip-tion
of the development of noise emission. The results for these three indicators for the
four pavement types can be seen for passenger cars in table below.
Table. The average noise level increase for passenger cars for the four pavement groups expressed
as the three indicators ΔL Age , ΔL ADT and ΔL Mix25/ 75 .
Pavement type ΔLAge
[ dB/ year]
ΔLADT
[ dB/ 1 mil vehicles]
ΔLMix25/ 75
[ dB/ mix]
All average 0.58 0.28 0.32
DGAC 0.40 0.21 0.26
OGAC 0.41 0.26 0.30
Thin Open 0.84 0.24 0.39
PAC 0.53 0.42 0.45
8
The average noise increase per year for passenger cars ( ΔLAge) is 0.58 dB/ year. The
DGAC pavements have the lowest increase of 0.40 dB/ year followed by OGAC with
0.41 dB/ year. The two pavement types with the highest increase are the PAC and the
Thin Open pavements with respectively 0.53 and 0.84 dB/ year.
When instead of ΔLAge the traffic volume ( ΔLADT) is taken into consideration as an in-dicator
for noise increase, the ranking of the pavement types changes significantly.
The increase for the porous pavements is almost twice the increase for the dense, thin,
or open graded asphalt concrete pavements.
The average ΔLMix25/ 75 for passenger cars is 0.32 dB. The ΔLMix25/ 75 indicator ranks the
four pavement types in the following way. DGAC has the lowest increase of 0.26 dB
followed by OGAC with 0.30 dB. The Thin Open pavements follow next with an in-crease
of 0.39 dB. The porous pavements have the highest increase of 0.45 dB/ using
the ΔLMix25/ 75 indicator.
Spectral analyses have been performed. The following very general tendencies are
observed for the four pavement types:
• For the Dense Graded Asphalt Concrete ( DGAC), the higher frequency noise
increases in the first years indicating that the pavement surface becomes denser
(“ post compaction”, increased air pumping noise). After some years there is also
an increase in the frequencies below 1600 Hz indicating that the pavement surface
becomes rougher with an increase in the tire vibration noise.
• For the Open Graded Asphalt Concrete ( OGAC), the tendencies for the different
pavements included in the investigation are not very clear. For some of the pave-ments
there is a tendency that the higher frequency noise increases in the first years
indicating that the pavement surface becomes denser (“ post compaction”), and af-ter
some years there is also an increase in the frequencies below 1600 Hz indicating
that the pavement surface becomes rougher with an increase in the tire vibration
noise. But for some of the pavements the increase at the lower frequencies hap-pened
before the increase at the higher frequencies.
• For the Thin Open pavements, the noise increases at the same time both at the
lower and at the higher frequencies. This indicates both that the pavement surface
becomes rougher with an increase in the tire vibration noise and that the pavement
surface becomes denser causing increased higher frequency air pumping noise.
• For the porous pavements ( PAC) the engine noise absorption effect at frequencies
between 400 and 1000 Hz is significantly reduced in the first two years. In the sec-ond
year clogging begins and this also increases the higher frequency noise over
1000 Hz presumably because of increased generation of air pumping noise. As the
porous pavements get older there is an increase in the low frequency noise ( below
1600 Hz) indicating increased tire vibration noise caused by a rougher pavement
surface structure.
• When heavy raveling occur the tire vibration generated low frequency noise
( below 1600 Hz) increases for all pavement types.
9
A first attempt to develop a noise emission performance model for pavements on
highways has been performed. Such a model can be used as a building stone if noise
is to be integrated in a Pavement Management System. The results from this project
indicate that it seems most relevant to develop such an aging model taking both the
physical age of a pavement as well as the traffic load into consideration. The model
has been developed using the ΔLmix25/ 75 indicator.
The model predicts the increase in noise level ΔLAging( A) that has to be added to the
noise emission of highway pavements when they are new ( not more than one year old)
as a function of the age of the pavement, the traffic load, and the pavement type.
The model is the following:
ΔLAging( A) = ( 0.25 · ΔLAge · A) + (( 0.75 · ΔLADT · ADT · 365 · A) / ( 106 · N))
Where:
ΔLAge is defined as the age component of noise level increase [ dB/ year].
ΔLADT is defined as the traffic component of noise level increase
[ dB/ 106 vehicle per lane].
A is the physical age of the pavement in years.
ADT = Average Daily Traffic ( total of both directions).
N = Number of lanes ( total of both directions).
This model is primarily useful for highways with speeds over 80 km/ h ( 50 mph). This
is a first version of a pavement noise performance model that can be improved and re-fined
as more knowledge is gained.
On the background of this project the following recommendations can be highlighted:
• The first version of a noise emission performance model for highway pavements
has been developed and can be used in Pavement Management Systems as a build-ing
stone to integrate noise as an important parameter in such systems.
• The noise emission performance model can be improved and refined like for
example by including analyses of pavements on urban roads with lower speed.
• There is a need for further research in order to give a better understanding on
which changes in the pavement surface structure cause the noise increase. Detailed
analyses of pavement structure and noise spectra etc. might be a lead to follow.
• More long- time measurement series are needed to get an even better understanding
of the noise increase as pavements get older.
10
• It is important whenever possible to follow the development of noise at existing
experimental road pavement test sections from the time they are new until they are
replaced.
• It is necessary to combine the results of noise measurements with results from
measurements of other pavement properties like surface texture, built- in air void
content, permeability, acoustical absorption, etc.
11
Sammenfatning
Formålet med dette projekt er at bidrage til den aktuelle internationale udvikling angå-ende
akustisk ældning af vejbelægninger ved at udføre en sammenlignende analyse af
nogle californiske og danske resultater fra langvarige støjmålinger på asfaltbelægnin-ger
på hovedlandeveje. Betonbelægninger er ikke inkluderet i denne undersøgelse.
Et internationalt litteraturstudie er blevet gennemført. Dette viser, at støjniveauet gene-relt
stiger som belægningen bliver ældre. Hvad angår drænasfalt ( porøse belægninger
med indbygget porevolumen på mere end 15 %), er det et kendt fænomen, at porerne i
nogle tilfælde tilstoppes, og at dette øger støj fra luftpumpning. Men for andre tætte og
åbne ( men egentlig ikke porøse) belægninger findes der ikke meget viden om eventu-elle
ændringer i overfladestrukturen i perioden fra bitumenfilm er nedslidt og til be-lægningerne
begynder at nedbrydes på grund af stentab, revnedannelser osv.
Dette projekt fokuserer på udviklingen af støjniveauet - de relative ændringer i støj
i årenes løb - og ikke de faktiske støjniveauer. Formålet er at analysere og sammenlig-ne
tendenser i udviklingen af støj over tid. Derfor er det ikke så afgørende, om støjre-sultaterne
er blevet målt ved hjælp af forskellige metoder eller af forskellige måle-hold/
organisationer. Disse faktorer kan påvirke de faktiske støjniveauer og kan van-skeliggøre
en direkte sammenligning, men når kun tendenser bliver sammenlignet
er eventuelle forskelle i målemetoder ikke så vigtige.
Den gennemsnitlige alder og størrelse af køretøjerne i Californien og Danmark kan
variere, for eksempel med mindre personbiler i Danmark. Dette kunne have indflydel-se
på sammenligningen af de faktiske støjniveauer, men vil formentlig være mindre
vigtigt, når man sammenligner udviklingen i støjemissionen over årene målt på samme
strækning. Kun ændringer i støjniveauet over tid er medtaget i dette dokument. Andre
faktorer som er relevante for beskrivelsen af udviklingen, som den fysiske struktur af
vejbelægningen såsom tekstur, porøsitet, visuelle tegn på slitage etc. er ikke blevet
taget i betragtning.
To veldokumenterede langtidsstøjmåleserier fra Californien og to fra Danmark er
blevet analyseret. Resultaterne er allerede blevet dokumenteret detaljeret i særskilte
nationale rapporter. Formålet med den aktuelle rapport er at foretage en sammenlig-ning
af tendenserne for akustisk ældning fundet i disse 4 projekter. University of Cali-fornia
Pavement Research Center ( UCPRC) afsluttede i 2009 den tredje års rapporte-ring
af årlige ” On Board Sound Intensity” støjmålinger på 65 til 76 belægninger af
forskellig type og alder i Californien. Nogle af resultaterne fra dette projekt er også
inkluderet.
12
De to californiske måleserier og UCPRC undersøgelsen er alle udført for California
Department of Transportation ( Caltrans) og de to danske måleserier er blevet udført
for det danske Vejdirektorat. Følgende fem måleserier er medtaget:
1. Åben asfaltbeton på I 80 nær Davis, Californien ( 11 år), 1 belægning.
2. 5 prøvestrækninger med tæt og åben asfaltbeton på LA138 i Mojave ørkenen,
Californien ( 5 år).
3. 65 til 76 belægninger i UCPRC/ Caltrans moniteringsprojekt ( 3 år og belægnin-ger
af forskellig alder og forskellige typer).
4. 3 belægninger med et lag drænasfalt samt en tæt og en åben asfaltbeton belæg-ning
i “ Viskinge”, Danmark ( 8 år).
5. 5 tynde åbne belægninger og en tæt asfaltbeton belægning på M10 (“ Solrød”)
nær København, Danmark ( 5 år).
Belægningerne medtaget i dette projekt blev grupperet i følgende 4 hovedgrupper:
• Tæt asfalt beton.
• Åben asfalt beton.
• Drænasfalt.
• Tynde åbne belægninger.
Følgende konklusioner for hovedlandeveje kan drages på baggrund af denne undersø-gelse
på akustisk ældning af asfaltbelægninger:
• Støjniveauet på asfaltbelægninger stiger normalt med tiden.
• Støjniveauet stiger kontinuerligt og inden betydelig belægningsnedbrydning med
stentab og revner osv. begynder.
• Der er undtagelser, hvor støjniveauet er reduceret i løbet af det første år af en porøs
belægnings levetid.
• En lineær regression giver en god beskrivelse af forholdet mellem belægningsalde-ren
og støj, både for personbiler og flereakslede køretøjer. Dette blev også fundet i
den europæiske SILENCE undersøgelse.
• Den årlige støjforøgelse er generelt ca. 2 gange højere for personbiler end for tunge
køretøjer.
Forskellige parametre har været brugt til at beskrive støjforøgelsen. Stigningen i støj-niveauet
er ofte udtrykt som dB/ år. To vigtige faktorer påvirker ændringer af støjegen-skaberne
af en belægning. Den ene faktor drejer sig om de fysiske/ kemiske ændringer
i materialet, forårsaget af vejrets påvirkning, og den anden har at gøre med slitage for-årsaget
af trafik. Det kan hævdes, at den kombinerede effekt af belægningens fysiske
alder samt slid fra trafik er afgørende for støjforøgelsen.
13
Alder afspejler en akkumuleret effekt af ændringer i vejrforholdene, som solstråling,
regn, is, frost/ tø osv. For at forsøge at definere en indikator, der kombinerer disse to
meget forskellige faktorer, alder og trafikbelastning er to kunstige indikatorer kaldet
" blandet indikator" ( ΔLMix50/ 50 og ΔLMix25/ 75) er blevet defineret. Støjforøgelsen er i alt
blevet analyseret for fem forskellige indikatorer:
1. ΔLAge: Ændring af støj pr. år ( faktiske fysiske alder af belægningen).
2. ΔLADT: Ændring af støj per 1 million køretøjer ( alle typer) som passerer pr
vognbane.
3. ΔLHvy: Ændring i støj per 100.000 tunge køretøjer som passerer pr vognbane.
4. ΔLMix50/ 50: En kunstig indikator for ændring af støj beskrevet som en kombinati-on
af den faktiske fysiske alder og trafikbelastningen, hvor alder tæller for 50%
og trafikbelastningen tæller for 50%. Kaldes " blandet indikator 50/ 50".
5. ΔLMix25/ 75: En kunstig indikator for ændring af støj beskrevet som en kombinati-on
af den faktiske fysiske alder og trafikbelastningen, hvor alder tæller for 25%
og trafikbelastningen tæller for 75%. Kaldes " blandet indikator 25/ 75".
Undersøgelserne viste, at ΔLADT og ΔLHvy resulterede i den samme rangordning af be-lægningerne,
formentlig fordi der ikke var nogen stor variation i andelen af tunge
køretøjer på de inkluderede veje. Ved at bruge ΔLMix25/ 75 i stedet for ΔLMix50/ 50 får
drænasfalt belægningerne generelt en højere støjstigning end de tynde åbne belægnin-ger;
dette vurderes at være korrekt, da dræmnasfalt har en tendens til tilstopning, som
ikke er set på ikke- porøse tynde belægninger. Ved at vælge ΔLMix25/ 75 i stedet for
ΔLMix50/ 50 indgår der en antagelse om at trafikkens indvirkninger dominerer over kli-matiske
ældningsvirkninger. De følgende tre indikatorer ΔLAge, ΔLADT og ΔLMix25/ 75
anses for at være de mest relevante til beskrivelse af udviklingen af støjemission.
Resultaterne for disse tre indikatorer for de fire belægningstyper kan ses i nedenståen-de
tabel ( for personbiler).
Tabel. Forøgelsen af gennemsnitsstøj for personbiler for 4 typer belægninger dom tre indikatorer
ΔL Age , ΔL ADT og ΔL Mix25/ 75 .
Belægningstype ΔLAge
[ dB/ år]
ΔLADT
[ dB/ 1 mio. køretøjer]
ΔLMix25/ 75
[ dB/ mix]
Gennemsnit af alle belægninger 0,58 0.28 0.32
Tæt asfaltbeton 0.40 0.21 0.26
Åben asfalt beton 0.41 0.26 0.30
Tynde åbne belægninger 0.84 0.24 0.39
Drænasfalt 0.53 0.42 0.45
14
Den gennemsnitlige støjstigning pr. år for personbiler ( ΔLAge) er 0,58 dB/ år. Tæt as-faltbeton
har den laveste stigning med 0,40 dB/ år efterfulgt af åben asfaltbeton med
0,41 dB/ år. De to belægningstyper med den højeste stigning er drænasfalt og tynde
åbne belægninger med henholdsvis 0,53 og 0,84 dB/ år.
Når der i stedet for ΔLAge tages trafikmængden ( ΔLADT) i betragtning som en indikator
for støjstigninger, rangordningen af belægningstyperne betydeligt.
Den gennemsnitlige ΔLMix25/ 75 for personbiler er 0,32 dB. ΔLMix25/ 75 indikatoren rang-ordner
de fire belægningstyper på følgende måde. Tæt asfaltbeton har den laveste
stigning på 0,26 dB, efterfulgt af åben asfaltbeton med 0,30 dB. Dernæst kommer
tynde åbne belægninger med en forøgelse på 0,39 dB. Drænasfalt har den højeste
stigning på 0,45 dB når ΔLMix25/ 75 indikatoren anvendes.
Spektrale analyser af støjen er blevet udført. Følgende, meget generelle, tendenser er
observeret for de fire belægningstyper:
• For tæt asfaltbeton stiger støjniveauet i de første år pga. øget luftpumpningen som
indikerer, at vejbelægningen bliver tættere ( efterkomprimering). Efter nogle år er
der også en stigning i de lavere frekvenser under 1600 Hz der indikerer, at vejbe-lægningen
bliver mere ujævn med en stigning i dækvibrationsstøj.
• For de tynde åbne belægninger, stiger støjen samtidig både på de lavere og højere
frekvenser. Dette indikerer både, at belægningen bliver mere ujævne med en stig-ning
i dækvibrationsstøj og at belægningens overflade bliver tættere ( efterkompri-mering),
hvilket forårsager en øget højfrekvent støj fra luftpumpning.
• For drænasfalt bliver absorptionen af motorstøj på frekvenser mellem 400 og 1000
Hz reduceret betydeligt i de første to år. I det andet år, begynder tilstopningen, og
dette øger støj i de højere frekvenser over 1000 Hz på grund af øget luftpumpestøj.
Efterhånden som drænasfalt bliver ældre er der en stigning i den lavfrekvente støj
ved mindre end 1.600 Hz som resultat af en mere ujævn belægningens overflade.
• Når der forekommer stentab ses stigninger i den lavfrekvente dækvibrationsstøj
under 1.600 Hz.
Et første forsøg på at udvikle en model for at beskrive støjens udvikling over tid for
belægninger på hovedlandeveje er blevet udført. En sådan model kan bruges som en
byggesten, såfremt støj skal integreres i et Pavement Management System. Resultater-ne
fra dette projekt viser, at det ser ud at være mest relevant at udvikle sådan en æld-ningsmodel
ved at tage både den fysiske alder af en belægning samt trafikmængde i
betragtning. Modellen er blevet udviklet ved hjælp af ΔLmix25/ 75 indikatoren.
15
Modellen beregner stigningen i støj ΔLaging( A), der skal adderes til støjen, når belæg-ninger
er nye ( ikke mere end et år gamle) som en funktion af alderen på belægningen
og trafikmængden på af en given vej såvel som en funktion af belægningstypen. Mo-dellen
er således:
ΔLAging( A) = ( 0.25 x ΔLAge x A) + (( 0.75 x ΔLADT x ADT x 365 x A) / ( 106 x N))
Hvor:
ΔLAge er defineret som alderskomponent af støjstigningen [ dB].
ΔLADT er defineret som trafikkomponent af støjstigning [ dB].
A er den fysiske alder af belægningen i antal år.
ADT = gennemsnitlig daglig trafik ( total i begge retninger).
N = antal vognbaner ( total i begge retninger).
Denne model er primært udviklet for hovedveje med hastigheder over 80 km/ t. Dette
er en første version af en model, der vil kunne videreudvikles og forbedres.
På baggrund af dette projekt, kan følgende anbefalinger fremhæves:
• Den første version af en støjemissionsmodel for hovedlandsvejsbelægninger, som
er udviklet i projektet, kan anvendes i Pavement Management Systemer som en
byggesten, der gør det muligt at integrere støj som en vigtig parameter i sådanne
systemer.
• Støjemissionsmodellen kan forbedres, for eksempel ved at inddrage analyser af be-lægninger
på byveje med lavere hastigheder mv.
• Der er behov for yderligere forskning med henblik på at give en bedre forståelse af
hvilke ændringer i belægningens overfladestruktur, der forårsager støjstigningerne.
Grundige analyser af belægningsstrukturen og støjspektre etc. kan være emner at
tage op!
• Flere langtidsmålingsserier er nødvendige for at få en endnu bedre forståelse for
hvorfor støjen stiger efterhånden som belægninger bliver ældre.
• Det er derfor vigtigt, hvis det er muligt, at følge eksisterende testtrækninger over så
lang tid som muligt for at få mere information om akustisk ældning.
• Det er nødvendigt at kombinere resultaterne af støjmålinger med resultater fra må-linger
af andre belægningsegenskaber som overfladetekstur, indbygget hulrum,
permeabilitet, akustisk absorption osv.
16
Preface
It is the experience of noise technicians that traffic noise emission of a given asphalt
pavement changes over time. Normally the noise level tends to increase over the
years. Knowledge on acoustical aging is important for road administrations in differ-ent
ways:
• When developing policies and strategies for noise abatement it is important to
know how noise- reducing pavements as well as standard pavements perform
over time.
• Acoustical aging is important information in order to achieve good accuracy
when noise is predicted with methods like the American TNM method, the
Nordic NORD2000 method, or the like.
• Noise performance models for road pavements are necessary if noise is to be
integrated as an active parameter in Pavement Management Systems.
The purpose of this project is to contribute to the ongoing international development
in the field of acoustical aging by performing a comprehensive analysis of some exist-ing
Californian and Danish results from long- time noise measurement series on as-phalt
pavements.
The project is carried out under the framework of the research technical agreement en-titled
“ Supplementary Studies for the Caltrans Quieter Pavement Research Program”
between California Department of Transportation ( Caltrans) and the University of
California Pavement Research Center ( UCPRC) as a part of the task: “ Policy docu-ments:
guidelines for Caltrans policy”. The Danish Road Institute ( DRI- DK) is sub-contracted
by UCPRC to work on the project. The work is carried out by a project
group with the following members:
• Hans Bendtsen, Danish Road Institute/ Road Directorate ( DRI- DK) working as a
guest researcher at UCPRC in 2008 and 2009.
• Qing Lu, University of California Pavement Research Center.
• Erwin Kohler, Dynatest Consulting Inc.
This report includes selected results from three Caltrans and two Danish projects.
From California the first site is the highway I- 80 project near Davis, where noise
measurements and analyses has been carried out by Illingworth & Rodkin, Inc.
Another site is the LA138 project where noise measurements and analyses have
been carried out by Volpe Center Acoustics Facility. The third site is actually a
large set of sections throughout the state of California that correspond to the UCPRC
project on field evaluation of tire/ pavement noise, being performed by the Partnered
Pavement Research Programs ( PPRC) Contract as Strategic Plan Element 4.19.
17
Results from two Danish projects are included. The “ Viskinge” project in which noise
measurements and analyses have been carried out by Delta Acoustics and the M10
project where DRI- DK has carried out noise measurements and analyses.
The authors would like to thank everybody involved in these projects. Without their
substantial work it would not have been possible to accomplish this aging study. Qing
Lu and Erwin Kohler were responsible for data analyses of OBSI measurements from
UCPRC collected in the field by Mark Hannum from UCPRC. Katrine Handberg from
the DRI- DK library carried out a literature search. The report is written by Hans
Bendtsen, DRI- DK. Jørgen Kragh DRI- DK has taken part in the evaluation and dis-cussion
of the results. Bent Andersen DRI- DK has performed Quality Assessment of
the report.
18
Forord
Erfaringen blandt støjteknikere er, at trafikstøj fra en asfaltbelægning ændrer sig med
tiden. Almindeligvis øges støjniveauet i løbet af en årrække. Kendskab til akustisk
ældning er vigtig for vejadministrationerne på forskellige måder:
• Når der udvikles politikker og strategier for støjreduktion, er det vigtigt at vide
hvorledes støjdæmpningen på støjreducerende belægninger såvel som på “ norma-le”
belægninger opfører sig i et givent tidsforløb.
• Akustisk ældning er en vigtig information for at opnå høj nøjagtighed, når støj
beregnes med metoder som den amerikanske TNM metode eller den nordiske
NORD2000 metode eller lign.
• Støjmodeller for vejbelægninger er nødvendige, hvis støj skal indgå som en aktiv
parameter i Pavement Management Systemer.
Formålet med dette projekt er at bidrage med den løbende internationale udvikling
angående akustisk ældning ved at gennemføre en omfattende analyse af nogle califor-niske
og danske resultater fra støjmålingsserier på asfaltbelægninger.
Projektet udføres inden for rammerne af den forskningstekniske aftale med titlen
" Supplerende Undersøgelser for Caltrans ” Quieter Pavement Research Program"”
mellem California Department of Transportation ( Caltrans) og UCPRC ( University
of California Pavement Research Center) som en del af opgaven: " Policy documents:
guidelines for Caltrans policy”. En del af dette arbejde er kontraheret til Vejdirektora-tet/
Vejteknisk Institut i Danmark
Arbejdet er udført af en projektgruppe bestående af følgende personer:
• Hans Bendtsen, Vejteknisk Institut/ Vejdirektoratet der arbejdede som gæsteforsker
på UCPRC i 2008 og 2009.
• Qing Lu, University of California Pavement Research Center.
• Erwin Kohler, Dynatest Consulting Inc.
Denne rapport indeholder udvalgte resultater fra Caltrans og fra danske projekter.
Fra Californien medtages I- 80 projektet, hvor støjmålinger and analyser er udført af
Illingworth & Rodkin, Inc; LA138 projektet, hvor støjmålinger og analyser er udført
af Volpe Center Acoustics Facility og UCPRC projektet på evaluering af dæk/ belæg-ningsstøj,
the Partnered Pavement Research Program ( PPRC) Contract Strategic Plan
Element 4.19.
19
Resultater fra to danske projekter er medtaget. “ Viskinge” projektet, hvor støjmålinger
og analyse er udført af Delta Acoustics og M10 projektet, hvor Vejteknisk Institut/
Vejdirektoratet har udført målinger og analyser. Forfatterne vil gerne takke alle som
har været involveret i disse projekter. Uden deres store indsats ville det ikke have
været muligt at gennemføre dette ældningsstudie.
Qing Lu og Erwin Kohler var ansvarlige for data analysen af OBSI målingerne fra
UCPRC som blev udført af Mark Hannum fra UCPRC. Katrine Handberg fra Vejtek-nisk
Institut, Danmark har gennemført en litteratur søgning. Rapporten er skrevet af
Hans Bendtsen, Vejteknisk Institut. Jørgen Kragh Vejteknisk Institut har deltaget i
evalueringen og diskussionen af resultaterne. Bent Andersen Vejteknisk
Institut har kvalitetssikret rapporten.
20
21
1. Introduction and existing knowledge
It is known from international literature [ 1] that noise emission from road traffic
changes over time as the road pavements age and are exposed to traffic and weather.
Normally an increase in noise level is seen. This report focuses on the change in noise
emission caused by changes in the pavement surface properties. Noise emission from
a given road can also change because of increase in traffic volume or percentage of
heavy vehicles as well as changes in average driving speed or driving pattern. The ef-fect
on noise of these traffic related changes is not the objective of this study which
focuses on pavements and the generation of tire- pavement noise.
There is a lack of well documented knowledge available today that can describe the
development of tire- pavement noise emission as a pavement gets older [ 1]. The main
purpose of this project is to contribute to the ongoing international development in the
field of acoustical aging of road surfaces. This is done in two ways:
1. By carrying out an international literature survey.
2. By performing a top down comprehensive analysis of existing Californian and
Danish results from long- time noise measurement series.
The results can be relevant for work on integrating noise as an active component in
Pavement Management Systems ( PM Systems) that are often used by road authorities
as a tool for planning the ongoing process of road and pavement maintenance and
renewal. Parameters like skid resistance etc are normally included in PM Systems in
order to optimize the traffic safety component, but noise might also be included in the
future [ 10]. An important part of integrating noise in PM Systems is to have perform-ance
models that can predict future noise emission from pavements as time goes by.
The results of this project will be relevant for developing such noise performance
models for road pavements.
The main focus of this project is asphalt concrete pavements. Cement concrete pave-ments
have not been included in this study.
The unit “ dB” is used in this report and it is considered in this document to be equal to
what is often denoted “ dB( A)” or “ dBA”.
1.1 Pavement properties and noise generation
In relation to pavement performance characteristics such as noise, skid resistance, roll-ing
resistance etc., the surface of a road pavement is often described by the following
three texture components [ 1] ( see Figure 1.1):
22
1. The microstructure is the deviation of a road surface from a completely plane
surface with characteristic dimensions along the surface less than 0.5 mm
( wavelengths less than 0.5 mm). The microtexture of a pavement is generally
created by the surface structure ( sharpness and harshness) of the individual
aggregates of the pavement surface. The microtexture is important for the skid
resistance of a pavement, but it is not considered to have an important influence
on the tire- noise generation.
2. The macrotexture is the deviation of a road surface from a completely plane
surface with characteristic dimensions along the surface between 0.5 mm and
50 mm ( wavelengths between 0.5 mm and 50 mm). The macrotexture is ob-tained
by the size and proportioning of aggregates and mortar, and by the
compaction of the pavement. The macrotexture is very important for the tire-pavement
noise generation.
3. The megatexture is the deviation of a road surface from a completely plane
surface with characteristic dimensions along the surface between 50 mm and
500 mm ( wavelengths between 50 mm and 500 mm). The megatexture can
be created by potholes, other larger discontinuities of the pavement surface, or
unevenness of the bottom layer. Megatexture is important for rolling resistance,
driving comfort etc., and can have some influence on the tire- pavement noise
generation. Surface roughness with larger wavelengths (> 500 mm) is normally
called unevenness.
( The vehicle)
" On a single stone,
barely visible"
Unevenness
Amplification ca. 50 times
Macrotexture
" Tyre/ road contact patch"
Microtexture
Reference length:
" Short stretch of road"
c:\ eget \ excel _ 9804\ textur e\ r oad_ textur e. xl s
Megatexture
" Size of a tyre"
Amplification ca. 5 times
Amplification ca. 5 times
Figure 1.1. Simplified examples of microtexture, macrotexture and megatexture [ 1] ( used with
permission from Ulf Sandberg, VTI).
23
The generation of noise when the tires are rolling on a road surface is mainly deter-mined
by the following different mechanisms [ 1] even though other mechanisms
might also play a minor role:
• Vibrations in the tires: The vibrations are generated by the contact between the
surface of the pavement and the rubber blocks of the tread pattern of the tire.
Tire vibrations generate noise in the frequencies from 500 to 1500 Hz. Macrotex-ture
is important for this noise. The noise level increases when the road surface
gets rougher. Therefore, an increase in the maximum aggregate size generally leads
to an increase in the noise level.
• The air pumping effect: When the rubber blocks on the tread pattern of the tire hit
the road surface, air is pressed out through the cavities between the rubber blocks
and the road surface. When the rubber blocks leave the road surface air is sucked
back into the cavities. This air pumping generates noise at frequencies over 1000
Hz. The macrotexture is important for the generation of this noise. If the road sur-face
is open or porous, the air will instead be pumped down into the pavement
structure and the noise level will be reduced.
• The horn effect: The curved belt of the tires and the road surface act as an acousti-cal
horn which amplifies the road noise generated around the contact point between
the tire and the road surface. This effect is highly directional and most important
for high frequencies. If the road surface is porous ( and therefore sound absorbing),
this amplification effect will be reduced.
• Absorption during propagation: The engine and road- tire noise are propagated from
the vehicle to the receiver. Under this propagation, the noise might be reflected on
the road surface. If the road surface is porous and therefore sound absorbing, the
noise at some frequency bands will be reduced.
• The effect of stiffness: The stiffness of the pavement is important for the noise
generated by the contact between the surface of the pavement and the rubber
blocks of the tread pattern of the tire. If the pavement is much less stiff, less noise
will be generated.
1.2 The “ Tyre/ Road Noise Reference Book”
The “ Tyre/ Road Noise Reference Book” by Sandberg and Ejsmont from 2002 [ 1]
includes an international status on current knowledge at that time on the influence of
pavement age etc. on tire/ road noise generation. According to this book, a road surface
changes its characteristics influencing noise over its lifetime. For some surface types
the influence is small and for others it can be large. It is summarized that the age and
wear influence on noise emission are caused by the following phenomena [ 1]:
• “ Mega- and macrotexture are changed, as particles and other materials are worn
away.
• Mega- and macrotexture, as well as stiffness, are changed due to the pavement
structure being compacted by traffic.
• Microtexture is changed, mainly by a polishing effect of many tires passing over
the surface ( studs on tires may counteract this effectively).
24
• The chemical effects of the weather, maybe assisted by road salt, creates a weath-ering
and crumbling of the surface ( loss of fine material), affecting both microtex-ture
and macrotexture. Rain may also play a role in changing the microstructure.
• Cracks may be created.
• If the surface is porous, its pores will become clogged by accumulated dirt.”
The Reference Book concludes that for smooth and medium textured dense asphalt
concrete ( DGAC) and SMA, the noise levels normally increase 1 to 2 dB the first 1 to
2 years of the pavement’s lifetime and then stabilizes until raveling and cracking occur
at the end of the lifetime of the pavement. The increase generally occurs at the higher
frequencies indicating an increase in the air pumping noise. According to [ 1], although
macrotexture will change over time there are cases where this change cannot explain
the observed increase in noise. The reason for this is ( according to [ 1]) still unknown
and the matter requires further research.
1.3 Literature survey
The library at the Danish Road Institute/ Road Directorate ( DRI- DK) has carried out
an international literature search for references to studies on the aging effects of as-phalt
road pavement in relation to noise. The general result is that not very many stud-ies
have been carried out worldwide with this specific focus. As a part of the European
EU funded project SILENCE [ 4], a large comprehensive inventory of currently avail-able
European results was performed [ 5]. The main results from this European study
are presented in the following pages together with the results of some recent studies
carried out in Arizona [ 20] and Washington State [ 21 to 24] in the US as well as in
Norway [ 26].
1.3.1 The SILENCE inventory
The SILENCE project included a task to provide models for the effect of pavement
aging on the noise reducing effect of road pavements [ 5]. This task was carried out by
DRI- DK. Existing historical measurement data on long- time noise performance of
pavements was collected from Belgium, Denmark, France, Germany, the Netherlands,
Poland, Sweden, and United Kingdom. A series of data from the noise study at the
UCPRC as a part of Caltrans Quiet Pavement Research ( QPR) program in California
was also included [ 11].
The scatter of all these results was quite large. The results were analyzed with the
main focus of finding slopes or trends for the increase of noise level every year for
different pavement types. No indication could be found that any model ( polynomial,
logarithmic or exponential) would yield better fit to the data than a simple linear rela-tion
between vehicle noise level and pavement age. This may be due to large scatter in
measurement results. Therefore a simple linear model was selected. This study did not
include spectral analyses for the noise data in relation to aging.
25
In Table 1.1, a summary is given of the average slopes to be expected for the linear
time history of vehicle noise levels. For both light and heavy vehicles, the slope to be
expected at dense asphalt pavements is in the order of 0.1 dB per year of pavement
service time. This applies to high speed as well as low speed roads. For porous or open
graded asphalt pavements the time history slope for light vehicles can be expected to
be in the order of 0.4 dB per year at high speed roads and 0.9 dB per year at city
streets with low traffic speed. Heavy vehicle noise levels can be expected to increase
with 0.2 dB per year at high speed roads with open/ porous pavements. Data was not
available for heavy vehicles at low speed porous/ open pavements.
Table 1.1. Overall time history slopes, dB per year of pavement service time [ 5].
Surface type Light vehicles Heavy vehicles
High
speed
[ dB/ Year]
Low
speed
[ dB/ Year]
High
speed
[ dB/ Year]
Low
Speed
[ dB/ Year]
Dense asphalt
( DGAC)
0.1 0.1 0.1 0.1
Porous / Open graded asphalt
( PAC/ OGAC)
0.4 0.9 0.2 -
1.3.2 US studies
Figure 1.2. Asphalt- Rubber Friction Course ( ARFC) has been used as a tool to reduce highway noise
in Arizona.
26
94
95
96
97
98
99
100
101
0 2 4 6 8 10 12 14
Age [ years]
Noise CPX [ dB]
R2 = 0.58
Figure 1.3. Results of CPX noise measurements in Arizona on Asphalt- Rubber Friction Course at
different ages [ 20] performed at 96 km/ h ( 60 mph) [ 20].
The Arizona Department of Transportation ( ADOT) has used Asphalt- Rubber Friction
Course ( ARFC) as a tool to reduce highway noise for a longer period. In 2002, ADOT
conducted a survey where noise was measured at 20 ARFC pavements [ 20] with ages
ranging from 3 to 12 years using the CPX noise measurement method [ 2]. This meas-urement
series did not follow the same pavement over the years. The main result was
that the noise increased 0.55 dB/ year as can be seen in Figure 1.3.
Figure 1.4. Interstate I- 5 in Washington State where noise reducing pavements were constructed in
June/ July 2006 with a close up photo of a pavement after being exposed to studded tires ( Photo
Washington State Department of Transportation).
27
The Department of Transportation in Washington State is carrying out a project on
evaluating noise reducing pavements [ 23]. On Interstate I- 5 three different pavement
types have been constructed in June/ July 2006. The pavements were a DGAC and two
open graded noise reducing pavements, one polymer modified and the other rubber
modified. The maximum aggregate size was 12.5 mm for the DGAC and 9.5 mm for
the two open graded pavements [ 24]. On Board Sound Intensity ( OBSI) [ 14] noise
measurements with the SRTT tire have been performed nearly every month on these
pavements since they were one month old. Studded tires are frequently used in the
winter period on the test road.
The results of the OBSI noise measurements can be seen in Figure 1.5. There is a sig-nificant
increase in noise level especially for the open graded polymer and rubberized
pavements. The increase mainly happens in the winter month where studded tires are
used. The average noise increase over the two and a half year period is 1.8 dB/ year for
the DGAC and 2.4 and 3.6 dB/ year for the two open graded pavements; the highest
increase for the rubberized pavement.
The OBSI noise measurements are not corrected for the influence of temperature on
the noise levels. A recent UCPRC and DRI- DK study [ 27] has resulted in a suggestion
of an air temperature correction factor for the SRTT tire of - 0.027 dB/° C. This means
that the noise level increases as the air ( and the tires and pavement) gets colder. But
the increase is not at all so large that it can explain the increase seen in the winter pe-riods
in these measurements. But there is a tendency that the noise level decreases a
little in the summer periods. This could to some extent be caused by higher tempera-ture
in the summer.
I- 5 Lynnwood, WA
95
96
97
98
99
100
101
102
103
104
105
aug- 06
okt- 06
dec- 06
feb- 07
apr- 07
jun- 07
aug- 07
okt- 07
dec- 07
feb- 08
apr- 08
jun- 08
aug- 08
okt- 08
dec- 08
feb- 09
Date
Noise OBSI [ dB]
DGAC
Polymer
Rubberrized
Figure 1.5. Results of OBSI CPX noise measurements on test pavements on interstate I- 5 in Wash-ington
State using SRTT tire [ 21].
28
The Colorado Department of Transportation has started a study of tire- pavement noise
on different pavements. The results of the first year of OBSI measurements performed
in 2006 are reported in [ 25]. Time series are not yet available.
1.3.3 A Norwegian study
A research and development project called “ Environmentally friendly pavements”
has been initiated by the Norwegian Public Roads Administration in 2005. As a part
of this project, the Norwegian research institution SINTEF has performed yearly CPX
noise measurements on 37 asphalt pavements [ 26] in the period from 2005 to 2008.
The measurements were carried out at 50 km/ h ( 31 mph) and/ or 80 km/ h ( 50 mph).
The CPX “ A” tire was used for these measurements [ 2]. In Norway, studded tires are
commonly used in the long winter season causing wear and tear of the pavements.
Some results from measurements performed over a four- year period are shown in Fig-ures
1.7 and 1.8. The main results show that the noise level on a newly laid pavement
increased 2- 4 dB after the first winter with exposure to studded tires. There is a trend
that the noise level increase was higher for pavements with a small maximum aggre-gate
size. In the following years, the increase was around 0.5 to 1.0 dB/ year. The two
porous pavements are exceptions with yearly increase of 1.6 and 2.0 dB.
Figure 1.6. Road included in the CPX measurements carried out as a part of the Norwegian “ Envi-ronmentally
friendly pavements” ( Photo Truls Berge, SINTEF ICT Acoustics, Norway).
29
50 km/ h
84
85
86
87
88
89
90
91
92
93
94
0 1 Age [ years] 2 3
CPX - level [ dB]
1- DGAC6
4- SMA6
2- DGAC8
5- SMA8
3- DGAC11
6- SMA11
Figure 1.7. CPX measurements performed at 50 km/ h ( 31 mph) ( using the CPX “ A” tire) on differ-ent
pavement types in Norway [ 26].
80 km/ h
92
93
94
95
96
97
98
99
100
101
102
0 1 2 3
Age [ years]
CPX - level [ dB]
9- SMA6
16- DGAC6
10- SMA8
7- SMA11
8- SMA11
11- SMA11
12- SMA16
18- Porous8
19- Porous
Figure 1.8. CPX measurements performed at 80 km/ h ( 50 mph) ( using the CPX “ A” tire) on differ-ent
pavement types in Norway [ 26].
1.4 Important factors affecting noise increase
1.4.1 Pavement aging and distress
The structure and condition of the surface of a given asphalt pavement changes over
time! When the pavement is newly laid, a black bitumen film covers the aggregate at
the surface ( see Figure 1.9). This bitumen film is worn off by the traffic ( the tires)
driving on the road within a shorter period of up to some months ( see Figure 1.10).
The effect on noise of this initial bitumen film is not clearly documented as most noise
measurements are first conducted when this film has been worn off.
30
Figure 1.9. A one day old porous asphalt concrete ( PAC) with a black bitumen film covering the
aggregate prior to opening to traffic.
Figure 1.10. A three years old dense- graded asphalt concrete ( DGAC) where the bitumen film has
been worn off and where no significant signs of wear and tear and distress can be observed.
The mix design ( aggregate gradation, binder type and content, use of additives etc.) as
well as the aging of the binder are important for the structural performance of a pave-ment
as time goes by. When asphalt pavements get older, significant signs of wear and
tear as well as distress on the pavement surface can be observed like for example:
• Polishing of aggregate.
• Raveling where aggregates are lost creating small holes in the surface
( see Figure 1.11).
• Cracks in the pavement ( see Figure 1.11).
• Development of potholes where material is missing over a larger area
( see Figure 1.12).
• Fatigue cracking ( see Figure 1.12).
• Bleeding where bitumen covers parts of the surface ( see Figure 1.13).
31
Besides polishing of aggregate that generally influences the microtexture of a pave-ment,
these types of distress generally increase the macrotexture/ roughness of the sur-face
structure and will result in an increase in vibration generated noise depending on
the severity of the distress. The impact on noise is very often rather localized; a few
ten meters back and in front of the area with distress, the noise level is not affected
[ 13]. On some occasions “ bleeding” occurs. This might also lead to changes in the
noise emission. The level of pavement distress is reflected in the overall pavement
condition index that is established on the background of visual inspection. An over-view
of the main influence on pavement texture on the different types of distress can
be seen in Table 1.2.
In the European SILENCE project [ 4], a series of CPX noise measurements were per-formed
on asphalt pavements with different kinds of distress. An increase of 2 - 3 dB
was measured with the CPX method at a speed of around 50 km/ h ( 31 mph) on a 6 m
long section with severe alligator cracking relative to a section of the same pavement
with no distress [ 13].
Figure 1.11. Raveling, loss of aggregate ( to the left) and cracking ( to the right).
Figure 1.12. Pothole ( to the left) and alligator cracking ( to the right).
32
Figure 1.13. Bleeding where bitumen covers part of the pavement surface.
Table 1.2. Pavement distress types and their main influence on pavement texture.
Distress Polishing of
aggregate
Raveling Cracks Potholes Alligator
cracking
Bleeding
Microtexture XX XX
Macrotexture XX XX X XX XX
Megatexture XX X
1.4.2 Acoustical aging
The literature survey shows that the noise level generally increases as the pavement
gets older. For porous pavement ( built in air void more than 15 % or so) it is a known
phenomenon that voids tends to clog and that this increases the noise generated from
air pumping [ 12]. But for other dense and open graded ( but not really porous) pave-ment
types there is not much knowledge on which changes in the surface structure
causes this increase in noise level in the period from when the bitumen film is worn
off until the pavements begins to deteriorate with distresses like raveling, cracking etc.
An example of the development of the pavement texture over a long period of time
can be seen in Figure 1.14 which shows the Mean Profile Depth ( MPD) on a dense
graded asphalt pavement ( DGAC) with 8 mm maximum aggregate size over ten years.
The pavement is located on an urban road in Copenhagen with an average daily traffic
( ADT) of around 7000 vehicles ( 7 percent heavy) and a speed limit of 50 km/ h ( 31
mph). The figure shows a remarkable decrease of the MPD when an old worn down
dense asphalt concrete pavement with distress was replaced by a new DGAC. Over the
years MPD gradually increase and after 8 years the MPD has nearly reached the level
of the former old pavement. No significant signs of pavement distress have been ob-served
over the years on the “ new” DGAC pavement.
33
0,0
0,2
0,4
0,6
0,8
1,0
1,2
Old
pavement
1999 2000 2001 2002 2003 2004 2005 2006 2007
MPD [ mm]
Figure 1.14. Development of MPD of a dense graded asphalt concrete ( DGAC) in the Danish Øster
Søgade experiment [ 12]. Bars show the standard deviation of the MPD measurements along the
100 m long road section.
The following are suggestions for factors that might influence the acoustical aging
phenomena before significant pavement distress occurs:
• Extra “ post” compaction because of traffic load.
• Aggregate is pressed further down in the mortar and the openness of the surface
structure is reduced.
• Change in the orientation of the aggregate because of traffic load.
• Clogging where the open structure with communicating pores in the upper part of a
porous pavement layer is more or less clogged.
• The average driving speed might have an influence on the clogging process.
• Ordinary pavement maintenance and cleaning such as sweeping etc.
• If utility works are carried out, patchwork pavement repair can change the surface
structure.
• In areas with “ white” winter periods, snow and ice removal procedures, salting,
and plowing etc. might influence the surface structure.
• In areas where studded tires and/ or snow chains are used this definitely has an ef-fect
on the surface structure creating rutting and raveling.
• Meteorological conditions like rain water, sun, snow, freeze- thaw, oxidation etc.
• Ultraviolet radiation from the sun.
34
On the background of the list of factors shown above which might influence the
acoustic aging of a pavement, it can be discussed which indicator or indicators for
aging are relevant to use. In previous investigations the pavement age has normally
been used ( see Section 1.3). Besides age the following indicators could be relevant:
• Accumulated traffic load on relevant driving lane ( total vehicles/ lane/ year).
• Accumulated light vehicles on relevant driving lane ( light vehicles/ lane/ year).
• Accumulated heavy vehicles on relevant driving lane ( heavy vehicles/ lane/ year).
• Traffic speed and acceleration/ deceleration ( including curvature, gradients etc.).
Together these traffic load related indicators, the climate zone ( effect of weather), the
winter maintenance procedures and the use of studded tires and/ or snow chains might
define some basic parameters that influence the tire pavement noise emission over
time.
In the current project, the indicators pavement age as well as the accumulated traffic
on the driving lane will be taken into consideration in the analyses. The road sections
included in the survey are all highways with a speed limit of 80 to 110 km/ h ( 50 to
69 mph) and they are located in areas where studded tires and/ or snow chains are not
used ( see Section 1.5). The sections in California are located in a subtropical climate
zone with very little frost and long warm and sunny summers, while the Danish sec-tions
are located in a temperate coastland climate zone with many thaw/ frost periods
and colder and cloudier summer periods than in California.
1.5 Noise measurement methods
Some factors might “ contaminate” long- time measurement series of noise emission
from pavements. If wayside measurements including a large population of vehicles
like the Statistical Pass- By method ( SPB) [ 3] or LAeq measurements are used, the noise
measured reflects the noise produced by the average vehicle fleet with the average
tires used at the time and in the region where the measurements are performed.
The noise emission related to vehicles and especially tires might change over the
years because of introduction of new types. The tire tread pattern, the rubber hardness,
and the tire width etc. has an influence on this.
If “ close to source” methods like the On Board Sound Intensity method [ 14] ( OBSI)
or the Close Proximity method ( CPX) [ 2] are used, there will be no dependencies of
the development of noise emission from vehicles or tires, since a standardized set of
test tires are used. But over the years the specifications for the methods might be
changed or new corrections applied and this might complicate comparison of results
measured over long time periods. Different reference tires have also been used in US
and in Europe.
35
This study will focus on wayside measurements because it turns out that the longest
comparable time series measured in California and Denmark have been performed
using these methods. The use of close by ( OBSI and CPX) methods has been more
common in the last 5 years in California and in Denmark. So in the future also results
from long- time measurement series using these methods may become available.
This project focuses on the trend in noise levels measured in the same way – the
relative changes of noise levels over the years – and not on the actual absolute noise
levels. Therefore it is not so crucial if noise results have been measured by different
methods or by the same method applied by different measurement teams/ organiza-tions.
These factors might influence the absolute noise levels and can complicate
direct comparison, but when only trends are compared, these differences in measure-ment
methods are not that important.
The average vehicle fleet of California and Denmark may differ, for example with
smaller passenger cars in Denmark. This might influence the comparison of absolute
noise levels but will presumably be less important when comparing trends in noise
emission over the years measured at the same site.
1.6 Layout of study
As already mentioned, this study focuses on the acoustical aging of asphalt concrete
pavements on highways. The objective is to analyze and compare trends in the devel-opment
of noise over time. A comparison of the actual nominal noise levels is not the
main objective of this study. When possible, noise emission from light and heavy ve-hicles
is analyzed separately. Frequency spectra will be used for detailed analyses of
the changes in noise emission.
This project focuses on the development of noise levels with time. It is outside the
scope of the project to include parameters describing the development of the pavement
surface structure with measurements like:
• Surface profiles
• Mean Profile Depth
• Built in air- void content
• Permeability
• Acoustical absorption
• Etc.
Examples of such comprehensive measurements and analyses are the Danish “ Øster
Søgade” experiment on two- layer porous pavements for urban roads [ 12] and the
UCPRC experiment in 2006 on approximately 70 pavement sections that span over
various climate regions, pavement age, traffic conditions and pavement types [ 17].
36
Two well documented long- time noise measurement series from California [ 15 and
16] and two from Denmark [ 18 and 19] have been selected for this project. The results
have already been documented in detail in separate national reports. The objective of
the current report is to perform a comparison study of the trends for acoustical aging
found in these four projects. UCPRC has finalized the third year report on yearly
OBSI measurements on 65 to 76 pavements at different ages in California [ 17]. Some
results of trends of noise emission and spectras from this project are also included in
this current report. The two Californian measurement series and the UCPRC study
have all been carried out for Caltrans and the two Danish measurement series have
been carried out for the Danish Road Directorate. The following measurement series
are included:
• California:
1. Open graded pavement ( OGAC) on I- 80 near Davis ( 10 years) [ 15].
2. Five test sections on LA138 in the Mojave Desert ( 5 years) [ 16].
3. Around 70 pavements in the UCPRC/ Caltrans monitoring project
( 3 years, pavements in different age groups and mix types) [ 17].
• Denmark:
1. Three single layer porous ( PAC), one dense graded ( DGAC) and
one open graded ( OGAC) asphalt concrete pavements at “ Viskinge”
( 8 years) [ 18].
2. Five Thin Open graded and one dense ( DGAC) pavement at M10
(“ Solrød”) near Copenhagen ( 5 years) [ 19].
The test sections and the pavements will be briefly described in Chapter 2 and the
trends of noise emission will be analyzed in Chapter 3 with focus on both the trends of
the development of the noise emission and on the development of the noise spectra
over the years in order to be able to evaluate what might cause the acoustical changes
of the pavements. A comparison of the results will be performed in Chapter 4, where
also a noise performance model in relation to noise is developed.
37
2. Road sections included
Table 2.1 gives an overview of the measurement series on the five test roads/ projects
that are included in this study. Presentations of the projects as well as more detailed
descriptions and data on the pavements at these five test roads/ projects can be found
in the following publications [ 15, 16, 17, 18 and 19]. Studded tires and/ or snow chains
are not used in the regions where the test roads are located though there might be a
few exceptions on some of the 65 to 76 pavements in [ 17].
Table 2.1. Overview of the five measurement series/ projects included in this comparison study.
Test road/ project
Number of
pavements
Year of con-struction
Noise measure-ment
method
Years of meas-urements
Number of lanes
Speed limit
km/ h / mph
Yearly Aver-ageDaily
Traffic
ADT
Percentage
heavy vehicles
Average ADT
per lane
I- 80 Cal. 1 1998 LAeq 11 6 104 / 65 146000 7.6% 24300
LA138 Cal. 5 2002 SPB 5 2 88 / 55 4300 17% 2150
Viskinge DK 5 1990 SPB 8- 9 2 80 / 50 7000 12.5% 3500
M10 DK 6 2004 SPB 5 6 110 / 69 90000 13.2% 15000
UCPRC 70 Cal. 65- 76 1989 - 2005 OBSI 3 variable variable variable variable variable
2.1 I- 80 test section by Davis, CA
The test section with Open Graded Asphalt Concrete with a nominal maximum aggre-gate
size of 9.5 mm was constructed in 1998 [ 15]. Wayside LAeq noise measurements
have been performed every year over an eleven year period in the summer season.
Figure 2.1. The Open Graded Asphalt Concrete on the I- 80 test section. The diameter of the US
quarter dollar coin is 24 mm.
38
Figure 2.2. The test section on Highway I- 80 east of Davis California with an OGAC pavement.
2.2 LA138 test sections in Mojave, CA
Figure 2.3. The test section on Highway LA138 in the Mojave Desert California with three open
graded and two dense graded pavements.
The test sections on LA138 were constructed in 2002 [ 16]. The purpose was to per-form
full scale testing of noise properties of a series of open graded pavements. All
the pavements have a nominal maximum aggregate size of 12.5 mm. The following
pavements were constructed on the test road:
• A Dense Graded Asphalt Concrete ( DGAC) with a specified thickness of
30 mm used as a noise reference pavement ( air void around 9 %).
• An Open Graded Asphalt Concrete ( OGAC 30) with a specified thickness of
30 mm ( air void around 15 %).
• An Open Graded Asphalt Concrete ( OGAC 75) with a specified thickness of
75 mm ( air void around 12 %).
39
• An Open Graded Asphalt Concrete with rubber powder added to the bitumen
( RAC- O) and a specified thickness of 30 mm ( air void around 12 %).
• A Bonded Wearing Course ( BWC). A propriety product used in California
( air void around 7 %). ( Noise measurements have only been performed on this
pavement when it was 10 and 16 months old so it is not included in the further
analyses in this project).
DGAC OGAC75
OGAC30 RAC- O
BWC
Figure 2.4. The five test pavements on the Highway LA138 test section when they were 6 years
old. All the pavements have a nominal maximum aggregate size of 12.5 mm. The size of the black
and white squares on the photos is 10 mm by 10 mm.
2.3 Viskinge test sections, DK
The test sections at Viskinge were built as the first Danish full scale noise experiment
with porous and open graded asphalt concrete [ 18]. The sections were constructed in
1990 and yearly SPB noise measurements were conducted over a period of 8 to 9
years. There were five test pavements:
40
• PAC8 type A: Single layer porous pavement with 8 mm nominal maximum aggre-gate
size and a built in air void between 18 and 22 %.
• PAC8 type B: Single layer porous pavement with 8 mm nominal maximum aggre-gate
size and a built in air void of more than 22 %.
• PAC12: Single layer porous pavement with 12 mm nominal maximum aggregate
size and a built in air void of more than 22 %.
• DGAC12: Dense graded asphalt concrete with 12 mm nominal maximum aggre-gate
size.
• OGAC12: Open graded asphalt concrete with 12 mm nominal maximum aggregate
size.
In order to be able to carry out a full lifecycle experiment of these porous pavements
in the shortest time period possible, the five pavements were deliberately built to break
down faster than would normally be the case. Therefore it was decided not to use bi-tumen
modifiers. With this choice it was expected that the lifetime of these porous
pavements would be minimized. Heavy raveling occurred when the pavements were
seven years old. Then the experiment was ended and the road was repaved with other
non- porous pavement types.
From recent Dutch experiences such porous pavements with bitumen modifiers have
a lifetime around eleven years on highways [ 28]. In the further analyses in Chapter 4
of this report an estimated lifetime of eleven years is assumed for these pavements,
because this is considered more realistic for such porous pavements if they were con-structed
today using modified bitumen. The yearly noise increase is reduced to reflect
a lifetime of eleven and not seven years. By doing so it becomes more relevant to
compare the results of this study started nearly 20 years ago with results from more
recent studies.
Figure 2.5. The test section at a highway near Viskinge in Denmark.
41
PAC8 type A PAC8 type B
PAC12 DGAC12
OGAC12
Figure 2.6. The five test pavements on the Viskinge test sections [ 18] when they were 3 years old.
The aggregate size varies between 8 and 12 mm. The red knife has a length of 8.5 cm.
2.4 M10 test sections, DK
The test sections on highway M10 near Solrød in Denmark were established in 2004
[ 19]. The purpose was to test different types of noise reducing thin open graded pave-ments
on a motorway. Yearly SPB noise measurements have now been conducted
over a 5 year period. There are six test pavements. A dense graded pavement and five
noise reducing thin layers were laid ( see Figure 2.8):
• DGAC11: Dense asphalt concrete with 11 mm nominal maximum aggregate size
( air void when constructed 2.8 %).
• SMA8: Stone Mastic Asphalt with a nominal maximum aggregate size of 8 mm
constructed as an open graded pavement ( air void when constructed 12.4 %).
42
• OGAC8: A very open graded asphalt concrete with a nominal maximum aggregate
size of 8 mm ( air void when constructed 15.3 %).
• UTLAC8: An ultra thin layer pavement with a nominal maximum aggregate size
of 8 mm. On the existing road surface, a thick layer of polymer modified bitumen
emulsion is spread. On the top of this unbroken bitumen emulsion a very open
graded mix is paved ( like porous asphalt) with a built- in air void of approx. 14 %
or even more. The unbroken bitumen emulsion “ boils up” in the air voids of the
pavement leaving only the upper part of the structure open. This reduces the built-in
air voids of the pavement because the pores of the pavement are almost filled
with bitumen.
• SMA6+: Stone Mastic Asphalt with a nominal maximum aggregate size of 6 mm.
A small amount of 5/ 8 mm aggregate is added ( air void when constructed 3.0 %).
• SMA8+: Stone Mastic Asphalt with a basic nominal maximum aggregate size of
6 mm. A very small amount of 6/ 8 aggregate and a larger amount of 8/ 11 mm ag-gregate
are added ( air void when constructed 5.7 %).
Figure 2.7. The Highway M10 test section near Solrød in Denmark with five open graded and one
dense pavement.
UTLAC8 DGAC11
43
Figure 2.8. The six test pavements on the on the Highway M10 test section when they were 2
years old. The maximum aggregate size varies between 6 and 11 mm [ 19]. The size of the black
and white squares on the photos is 10 mm by 10 mm.
2.5 Californian investigation
In 2006, UCPRC started a project where noise, durability and other pavement proper-ties
were measured on 76 selected pavements on the Californian highway network [ 8].
The pavements were selected in order to include both new, some years old and older
pavements of the same type. The following pavement types were included ( see Figure
2.10):
• DGAC: Dense Graded Asphalt Concrete.
• OGAC: Open Graded Asphalt Concrete.
• RAC- O: Open Graded Asphalt Concrete with rubber.
• RAC- G: Dense graded Asphalt Concrete with rubber.
Each of the four main pavement groups were divided into the following three age
categories in 2006 when the project started: less than one year old, one to less than
four years old and more than four years old. The nominal maximum aggregate size
of the pavements in these three groups is 9.5, 12.5 and 19 mm. The influence on noise
caused by different aggregate size is generally assumed to be an increase of noise level
of around 0.25 dB per 1 mm increase in aggregate size. In this data set, the variation in
aggregate size can be expected to influence the noise levels in a range of 2.5 dB. This
will cause some spread in the results.
SMA6+ SMA8
SMA8+ OGAC8
44
The noise has been measured by the OBSI method on all these pavements in 2006,
2007 and 2008 [ 17] and the measurements will be continued in 2009. Generally, the
OBSI measurements were carried out at a speed of 96 km/ h ( 60 mph). Where this for
practical reasons was not possible, the noise levels have been normalized to a refer-ence
speed of 96 km/ h. Due to pavement rehabilitation etc. the amount of pavements
included every year has decreased. In 2007 the number was 71 and in 2008 65. An
analysis of the noise from pavements belonging to the same pavement type of differ-ent
ages is included in Section 3.5 of this report. Generally each of the pavements is
represented three times in the data set, each time being one year older.
Figure 2.9. One of the 76 roads included in the UCPRC pavement noise study, started in 2006 in
California.
Figure 2.10. Typical pavements included in the Californian project. The diameter of the US quarter
dollar coin is 24 mm.
DGAC OGAC
RAC- O RAC- G
45
3. Analyses of noise over time
The noise level trends from each of the five projects are presented in this chapter.
The Residual Standard Error is used to describe the spread of the actual measurement
data around the regression lines. Frequency spectra are presented for the frequency
range from 400 to 4000 Hz important for the A- weighted noise level. The OBSI
measurements used at the 76 Californian pavement sections do not include the 400 Hz
frequency band. The noise trends generally range between 0.1 and 1.0 dB/ year. In or-der
to show differences between the pavements two decimals are used for noise trends
even though the normal accuracy of noise measurements only calls for using one
decimal.
3.1 I- 80 test section by Davis, CA
The I- 80 test section with an Open Graded Asphalt Concrete was constructed in 1998.
Yearly wayside LAeq noise measurements have been performed for mixed traffic using
the American Continuous Flow Traffic Time Integrated Method ( CTIM). Noise has
been measured by Illingworth & Rodkin Inc., both for the eastbound and for the west-bound
lanes [ 15]. The noise is normalized to a fixed traffic volume. The speed limit is
104 km/ h ( 65 mph). Temperature correction has not been performed, but the meas-urements
have all been carried out in the summer period with temperatures generally
ranging between 21 and 29° C ( 70 to 85° F), so the temperature should not have a sig-nificant
influence on the noise trend [ 27].
The results can be seen in Figures 3.1 and 3.2 and in Table 3.1. A linear regression
gives a reasonable fit with Residual Standard Error of respectively 0.3 and 0.5 dB.
The yearly increase of noise on this OGAC pavement is 0.11 dB/ year in the eastbound
direction and 0.19 dB in the westbound direction.
46
OGAC12 - Mixed traffic Eastbound direction
y = 0,11x + 73,30
R2 = 0,64
70,0
72,0
74,0
76,0
78,0
80,0
0 1 2 3 4 5 6 7 8 9 10
Age [ years]
LAeq [ dB]
Figure 3.1. Measured wayside eastbound noise as L Aeq for mixed traffic ( speed limit 104 km/ h
( 65 mph)) [ 15].
OGAC12 - Mixed traffic Westbound direction
y = 0,19x + 72,07
R2 = 0,65
70,0
72,0
74,0
76,0
78,0
80,0
0 1 2 3 4 5 6 7 8 9 10
Age [ years]
LAeq [ dB]
Figure 3.2. Measured wayside westbound noise as L Aeq for mixed traffic ( speed limit 104 km/ h
( 65 mph)) [ 15].
Table 3.1. Average noise increase per year and the Residual Standard Error for mixed traffic at the
I- 80 test section ( speed limit 104 km/ h ( 65 mph)).
Section Increase Residual Standard Error
Eastbound 0.11 dB/ year 0.3 dB
Westbound 0.19 dB/ year 0.5 dB
47
The development of the noise spectra for mixed traffic over the years can be seen in
Figures 3.3 and 3.4. The trend shows that the noise is generally slightly increased over
the 10 year period in all frequency bands.
I 80 Eastbound
50
55
60
65
70
75
400 500 630 800 1000 1250 1600 2000 2500 3150 4000
Frequency [ Hz]
LAeq [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Year 8
Year 9
Year 10
Figure 3.3. The spectra of L Aeq for mixed traffic at the different ages for the OGAC12 pavement
at I- 80 in the eastbound direction [ 15].
I 80 Westbound
50
55
60
65
70
75
400 500 630 800 1000 1250 1600 2000 2500 3150 4000
Frequency ( Hz)
LAeq [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Year 8
Year 9
Year 10
Figure 3.4. The spectra of L Aeq for mixed traffic at the different ages for the OGAC12 pavement
at I- 80 in the westbound direction [ 15].
48
3.2 LA138 test sections in Mojave, CA
A test section with five test pavements was constructed on LA138 in 2002. Detailed
wayside SPB measurements have been carried out by Volpe Center for Acoustics four
times in a five year period [ 16] ( except for the BWC pavement where data are only
available from month 10 and 16). Therefore, the BWC pavement has not been in-cluded
in the following analyses. The measurements were carried out at a temperature
range of 16 to 28° C ( 61 to 83° F). Temperature corrections have not been applied to
these data, but with the relatively narrow temperature range this will not have a sig-nificant
influence on the trends for noise increase. The microphone position used was
7.5 m ( 25 feet) from the centerline of the lane and at a height of 1.5 m ( 5 feet). The
SPB measurements have been carried out when the pavements were 4, 10, 16 and 52
months old. Therefore, the time scale used in the analyses is months after construction
and not years as in the other measurement series included in this report. The noise in-crease
per month therefore has to be multiplied by 12 to get the yearly noise increase.
The trends for the development of the noise for passenger cars as well as for multi axle
vehicles and the spectra can be seen in the below figures and in Table 3.2 where also
the Residual Standard Error is given.
DGAC - Passenger cars
y = 0,020x + 78,572
R2 = 0,952
72
74
76
78
80
82
0 20 Age [ month4s0] 60
LAmax [ dB]
DGAC - Multi Axle
y = 0,024x + 86,006
R2 = 0,899
80
82
84
86
88
90
0 20 Age [ month40s] 60
LAmax [ dB]
Figure 3.5. Maximum SPB noise level for passenger cars to the left ( reference speed 96 km/ h
( 60 mph) and for multi axle vehicles to the right ( reference speed 88 km/ h ( 55 mph)) for the
DGAC pavement) [ 16].
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Month 4
Month 10
Month 16
Month 52
DGAC
Passenger cars
Figure 3.6. SPB noise spectra for passenger cars for the DGAC pavement ( reference speed 96 km/ h
( 60 mph)) [ 16].
49
In Figure 3.5 the development of noise for the DCAC pavement at the LA138 test sec-tion
can be seen. The noise increase fits quite well with a linear regression with Resid-ual
Standard Errors of 0.1 and 0.2 dB ( see Table 3.2). The yearly increase for passen-ger
cars was 0.24 dB/ year and for multi axle vehicles it was 0.29 dB/ year. According
to Figure 3.6 the increases take place in the whole range of the frequency spectrum.
OGAC75 - Passenger cars
y = 0,026x + 74,594
R2 = 0,878
72
74
76
78
80
82
0 20 Age [ months4] 0 60
LAmax [ dB]
OGAC75 - Multi Axle
y = 0,008x + 83,418
R2 = 0,551
80
82
84
86
88
90
0 20 Age [ mont4h0s] 60
LAmax [ dB]
Figure 3.7. Maximum SPB noise level for passenger cars to the left ( reference speed 96 km/ h
( 60 mph) and for multi axle vehicles to the right ( reference speed 88 km/ h ( 55 mph)) for the
OGAC75 pavement) [ 16].
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz] SPB [ dB]
Month 4
Month 10
Month 16
Month 52
OGAC75
Passenger cars
Figure 3.8. SPB noise spectra for passenger cars for the OGAC75 pavement ( reference speed 96
km/ h ( 60 mph)) [ 16].
The development of noise over the years for passenger cars for the open graded
OGAC75 pavement was 0.31 dB/ year ( see Figure 3.7) whereas it was lower for multi
axle vehicles ( 0.10 dB/ year). The noise increase fits quite well with a linear regression
with Residual Standard Error of 0.3 and 0.2 dB ( see Table 3.2). Figure 3.8 show that
the increase basically happens in the low frequencies indicating that the pavement sur-face
becomes rougher.
50
OGAC30 - Passenger cars
y = 0,017x + 77,070
R2 = 0,894
72
74
76
78
80
82
0 20 Age [ month4s0] 60
LAmax [ dB]
OGAC30 - Multi Axle
y = 0,010x + 85,366
R2 = 0,640
80
82
84
86
88
90
0 20 Age [ mont4h0s] 60
LAmax [ dB]
Figure 3.9. Maximum SPB noise level for passenger cars to the left ( reference speed 96 km/ h ( 60
mph) and for multi axle vehicles to the right ( reference speed 88 km/ h ( 55 mph)) for the OGAC30
pavement) [ 16].
Figure 3.9 shows the trends for the open graded OGAC30 pavement. The Residual
Standard Error is 0.2 dB for both vehicle categories ( see Table 3.2). The increase is
0.20 dB/ year for passenger cars and 0.12 dB/ year for multi axle vehicles. Figure 3.10
shows that the increase like for the OGAC75 pavement basically happens in the low
frequencies indicating that the pavement surface becomes rougher.
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Month 4
Month 10
Month 16
Month 52
OGAC30
Passenger cars
Figure 3.10. SPB noise spectra for passenger cars for the OGAC30 pavement ( reference speed
96 km/ h ( 60 mph)) [ 16].
RAC- O - Passenger cars
y = 0,033x + 76,030
R2 = 0,898
72
74
76
78
80
82
0 20 Age [ mont4h0s] 60
LAmax [ dB]
RAC- O - Multi Axle
y = 0,030x + 84,351
R2 = 0,959
80
82
84
86
88
90
0 20 Age [ mont4h0s] 60
LAmax [ dB]
Figure 3.11. Maximum SPB noise level for passenger cars to the left ( reference speed 96 km/ h
( 60 mph) and for multi axle vehicles to the right ( reference speed 88 km/ h ( 55 mph)) for the RAC- O
pavement) [ 16].
51
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Month 4
Month 10
Month 16
Month 52
RAC- O
Passenger cars
Figure 3.12. SPB noise spectra for passenger cars for the RAC- O pavement ( reference speed 96
km/ h ( 60 mph)) [ 16].
Finally, the trends for the open graded RAC- O can be seen in Figure 3.11. This pave-ment
has the highest increase of the LA138 test pavements with a yearly increase of
0.40 dB for passenger cars and 0.36 dB for multi axle vehicles ( Residual Standard Er-ror
of 0.3 and 0.2 dB [ see Table 3.2]). The spectra in Figure 3.12 show that the highest
increase happens at frequencies below 1000 Hz but there is also some minor increase
at the higher frequencies.
Table 3.2. Average noise level increase per year for passenger cars and multi axle vehicles and the
Residual Standard Error for the four test pavements on LA138 ( reference speed 96 km/ h ( 60 mph)
for passenger cars and 88 km/ h ( 55 mph) for multi axle vehicles).
Pavement Passenger
cars
Residual
Standard
Error
Passenger
Multi axle
vehicles
Residual
Standard
Error
Multi axle
DGAC 0.24 dB/ year 0.1 dB 0.29 dB/ year 0.2 dB
OGAC75 0.31 dB/ year 0.3 dB 0.10 dB/ year 0.2 dB
OGAC30 0.20 dB/ year 0.2 dB 0.12 dB/ year 0.2 dB
RAC- O 0.40 dB/ year 0.3 dB 0.36 dB/ year 0.2 dB
Table 3.2 gives an overview of the noise trends for the four LA138 pavements. For
passenger cars the dense graded OGAC30 pavement has the lowest increase of 0.20
dB/ year. For the other three more open graded pavements, the increases vary between
0.24 and 0.40 dB/ year. The increases for multi axle vehicles are generally less than for
passenger cars with the DGAC pavement as an exception; here the trend for multi axle
vehicles is slightly higher than for passenger cars.
52
3.3 Viskinge test sections, DK
The five test pavements at Viskinge were constructed in 1990. SPB noise measure-ments
have been conducted over a period of eight to nine years [ 18]. The microphone
position used was 7.5 m ( 25 feet) from the centerline of the lane and at a height of
1.2 m ( 4 feet). The results are reported as LAE values and not LAmax which is now
common for SPB measurements. There is a linear correlation between LAE and LAmax
for SPB traffic noise measurements [ 29]. Temperature corrections have not been per-formed,
but the measurements have all been made in periods not covering the colder
winter months in a temperature range of generally 15 to 25° C ( 59 to 77° F) so this will
not have a significant influence on the trends for noise increase. The trends for the de-velopment
of the noise level for passenger cars as well as for multi axle vehicles and
the spectra can be seen in the below figures. The Residual Standard Error can be seen
in Table 3.3.
DGAC12 - Passenger cars
y = 0,40x + 75,94
R2 = 0,92
70
72
74
76
78
80
82
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
DGAC12 - Multi axle
y = 0,21x + 88,12
R2 = 0,60
80
82
84
86
88
90
92
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
Figure 3.13. L AE SPB noise level from passenger cars ( reference speed 80 km/ h ( 50 mph)) to the left
and for multi axle vehicles to the right for the DGAC12 pavement [ 18].
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Year 8
DGAC12
Figure 3.14a. The SPB spectra for passenger cars at the different ages for the DGAC12 pavement
( reference speed 80 km/ h ( 50 mph)) [ 18].
Figure 3.13 shows the development of the noise level over the years for the dense
graded DGAC12 pavement. The noise level increase fits quite well with a linear re-gression
with a Residual Standard Error of 0.3 and 0.5 dB ( see Table 3.3).
53
The yearly increase for passenger cars was 0.40 dB/ year and for multi axle vehicles
0.21 dB/ year. The noise for passenger cars generally increased over time at all fre-quencies
above 630 Hz ( see Figure 3.14a). In the first one to two years the noise in-creased
1 to 2 dB at frequencies above 1000 Hz. This could indicate that the dense
surface structure of the pavement has become even denser causing an increase in the
high frequency air pumping generated noise! This might be caused by the pavement
being “ post compacted” by the tires driving on the pavement. Heavy raveling occurred
on the DGAC pavement in year 8. The results show a noise increase of 1.0 to 1.5 dB
in the frequency range from 800 to 1600 Hz. The spectra for multi axle vehicles
( Figure 3.14b) generally show the same trends. From the third year the spectra is
nearly unchanged to year 8. This could indicate that truck tires are not as sensitive
to changes in the openness of the pavement surface structure as passenger car tires.
The raveling in year 8 does not have any significant effect on the noise emission
from the truck tires.
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Year 8
DGAC12
Figure 3.14b. The SPB spectra for multi axle vehicles at the different ages for the DGAC12 pave-ment
( reference speed 80 km/ h ( 50 mph)) [ 18].
OGAC12 - Passenger cars
y = 0,51x + 76,99
R2 = 0,94
70
72
74
76
78
80
82
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
OGAC12 - Multi axle
y = 0,27x + 87,96
R2 = 0,93
80
82
84
86
88
90
92
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
Figure 3.15. L AE SPB noise level from passenger cars to the left and for multi axle vehicles to the
right for the OGAC12 pavement ( reference speed 80 km/ h ( 50 mph)) [ 18].
54
The development of noise level at the open graded OGAC12 is shown in Figure 3.15.
Also for this pavement the noise increase fits quite well with a linear regression ( Re-sidual
Standard Error of 0.3 and 0.2 dB ( see Table 3.3)). The yearly increase for pas-senger
cars was 0.51 dB/ year and for multi axle vehicles 0.27 dB/ year. The noise for
passenger cars generally increased at all frequencies above 630 Hz ( see Figure 3.16a).
Like for the DGAC12 pavement the noise increased 1 to 2 dB at frequencies above
1000 Hz in the first one to two years. Also for the OGAC12 pavement this might be
caused by the pavement being “ post compacted” reducing the open structure of the
pavement surface. For the truck tires there is an increase of 1- 2 dB over 1000 Hz from
year 0 to year 1. After this the changes of the spectra are quite small.
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
OGAC12
Figure 3.16a. The SPB spectra for passenger cars at the different ages for the OGAC12 pavement
( reference speed 80 km/ h ( 50 mph)) [ 18].
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
OGAC12
Figure 3.16b. The SPB spectra for multi axle vehicles at the different ages for the OGAC12 pave-ment
( reference speed 80 km/ h ( 50 mph)) [ 18].
55
The first of the three porous pavements at the Viskinge test site is the PAC8 Type A.
The development of noise can be seen in Figure 3.17. The yearly increase for passen-ger
cars was 0.87 dB/ year and for multi axle vehicles the increase was 0.37 dB/ year.
These increases are twice as high as for the dense DGAC12 and the open but not po-rous
OGAC12 pavement. The Residual Standard Error was greater for this porous
pavement and was 0.6 dB for both vehicle categories ( see Table 3.3).
PAC8 Type A - Passenger cars
y = 0,87x + 70,45
R2 = 0,93
70
72
74
76
78
80
82
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
PAC8 Type A - Multi axle
y = 0,37x + 84,13
R2 = 0,69
80
82
84
86
88
90
92
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
Figure 3.17. L AE SPB noise level for passenger cars to the left and for multi axle vehicles to the right
for the PAC8 type A pavement ( reference speed 80 km/ h ( 50 mph)) [ 18].
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
PAC8 type A
Figure 3.18a. The SPB spectra for passenger cars at the different ages for the PAC8 type A pave-ment
( reference speed 80 km/ h ( 50 mph)) [ 18].
In the first year, a decrease of noise level of 0.3 dB for passenger cars and 0.7 dB for
multi axle vehicles was observed. The frequency spectra for passenger cars in Figure
3.18a give an indication of what might be happening. This spectrum is significantly
different from the spectra at the two previous not porous pavements.
56
The noise level decreased by 2 dB in the frequency range 800 to 1000 Hz, which is
important for the total A- weighted noise level. A new open porous pavement absorbs
noise reflected on the pavement at frequencies typically around 400 to 1000 Hz ( en-gine
noise) depending on the thickness of the porous layer. It seems that this absorp-tion
effect improved over the first year. But at the same time, the noise over 1250 Hz
increased indicating an increase in the noise from air pumping. This might be caused
by post compaction of the pavement.
From the first to the second year the noise level increased by 2 to 3 dB at frequencies
above 1000 Hz giving an increase in air pumping noise indicating that the open pores
of the pavement were beginning to clog! The noise level also increased at 800 to 1000
Hz indicating that the noise absorption effect was reduced significantly! This is also
an indication of clogging.
In year 7, heavy raveling was observed on this pavement. In this year there was a sig-nificant
increase of around 2 to 3 dB of low frequency noise ( below 1600 Hz). This
indicates that the pavement has become rougher because of the raveling. Figure 3.18b
shows the spectra for multi axle vehicles. The trends are generally the same as for pas-senger
cars. Here the noise level increased around 2 dB in the frequencies below 1600
Hz from year 6 to 7 when raveling occurs. This is different than for the dense DGAC
pavement where the raveling did not increase the truck tire noise.
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
PAC8 type A
Figure 3.18b. The SPB spectra for multi axle vehicles at the different ages for the PAC8 type A
pavement ( reference speed 80 km/ h ( 50 mph)) [ 18].
57
PAC8 Type B - Passenger cars
y = 0,81x + 70,39
R2 = 0,89
70
72
74
76
78
80
82
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
PAC8 Type B - Multi axle
y = 0,20x + 84,47
R2 = 0,27
80
82
84
86
88
90
92
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
Figure 3.19. L AE SPB noise level from passenger cars to the left and for multi axle vehicles to the
right for the PAC8 type B pavement ( reference speed 80 km/ h ( 50 mph)) [ 18].
Figure 3.19 shows the development of noise level of the porous PAC8 Type B pave-ment
( Residual Standard Error of 0.8 and 0.9 dB [ see Table 3.3]). The yearly increase
for passenger cars was 0.81 dB/ year similar to the increase at the PAC8 Type A
pavement. For multi axle vehicles the increase was only 0.20 dB/ year. But in the first
year a decrease of noise of 0.9 dB for both passenger cars and multi axle vehicles were
observed. The frequency spectra for passenger cars in Figure 3.20a give an indication
of what might be happening. The same trends as described for the PAC Type A pave-ment
with increased absorption over the first year at 800 to 1000 Hz can be observed.
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
PAC8 type B
Figure 3.20a. The SPB spectra for passenger cars at the different ages for the PAC8 type B pave-ment
( reference speed 80 km/ h ( 50 mph)) [ 18].
From the first to the second year, the noise level increased 2 to 5 dB at frequencies
above 1000 Hz indicating an increase in air pumping noise because the open pores of
the pavement were beginning to clog. The noise also increased at 800 to 1000 Hz indi-cating
that the noise absorption effect was significantly reduced. This is also an indica-tion
of clogging.
58
Also for the PAC8 type B heavy raveling was observed in year 7. In this year there
was a significant increase of low frequency noise level ( around 2 dB) at below 1600
Hz. This reflects that the pavement has become more uneven because of the raveling.
Figure 3.20b shows the spectra for multi axle vehicles. The trends are similar to the
trends seen for passenger cars.
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
PAC8 type B
Figure 3.20b. The SPB spectra for multi axle vehicles at the different ages for the PAC8 type B
pavement ( reference speed 80 km/ h ( 50 mph)) [ 18].
PAC12 - Passenger cars
y = 0,83x + 72,51
R2 = 0,81
70
72
74
76
78
80
82
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
PAC12 - Multi axle
y = 0,44x + 84,38
R2 = 0,69
80
82
84
86
88
90
92
0 1 2 3 4 5 6 7 8
Age [ years]
LAE [ dB]
Figure 3.21. L AE SPB noise level from passenger cars to the left and for multi axle vehicles to the
right for the PAC12 pavement ( reference speed 80 km/ h ( 50 mph)) [ 18].
The results from the third and last porous pavement on the Viskinge test site can be
seen in Figure 3.21( Residual Standard Error of 1.1 and 0.8 dB [ see Table 3.3]). The
yearly increase for passenger cars was 0.83 dB/ year and for multi axle vehicles 0.44
dB/ year, similar to the PAC8 Type A and B pavement. But in the first year a decrease
of noise level of as much as 2.3 dB for passenger cars and 1.9 dB for multi axle vehi-cles
was observed. The frequency spectra for passenger cars can be seen in Figure
3.22a. The same trends as described for the PAC Type A and Type B pavements with
increased absorption at 800 to 1000 Hz can be observed over the first year.
59
50
55
60
65
70
75
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
PAC12
Figure 3.22a. The SPB spectra for passenger cars at the different ages for the PAC12 pavement
( reference speed 80 km/ h ( 50 mph)) [ 18].
From the first to the second year the noise level increased 1 to 4 dB at frequencies
above 1000 Hz indicating an increase in air pumping noise, because the open pores of
the pavement were beginning to clog! The noise level also increased at 800 to 1000
Hz indicating that the noise absorption effect was significantly reduced. This is also an
indication of clogging.
Heavy raveling was also observed on the PAC12 pavement in year 7. In this year there
was a significant increase of low frequency noise level ( below 1600 Hz) of around 2
to 3 dB. This indicates that the pavement has become rougher because of the raveling.
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
PAC12
Figure 3.22b. The SPB spectra for multi axle vehicles at the different ages for the PAC12 pavement
( reference speed 80 km/ h ( 50 mph)) [ 18].
60
Also for this PAC12 pavement the same spectral tendencies as for passenger cars can
be seen for multi axle vehicles ( Figure 3.22b).
Table 3.3. Average noise level increase per year for passenger cars and multi axle vehicles and the
Residual Standard Error for the five test pavements measured at Viskinge ( reference speed 80 km/ h
( 50 mph)).
Pavement Passenger
cars
Residual
Standard
Error
Passenger
Multi axle ve-hicles
Residual
Standard
Error
Multi axle
DGAC12 0.40 dB/ year 0.3 dB 0.21 dB/ year 0.5 dB
OGAC12 0.51 dB/ year 0.3 dB 0.27 dB/ year 0.2 dB
PAC8 type A 0.87 dB/ year 0.6 dB 0.37 dB/ year 0.6 dB
PAC8 type B 0.81 dB/ year 0.8 dB 0.20 dB/ year 0.9 dB
PAC12 0.83 dB/ year 1.1 dB 0.44 dB/ year 0.8 dB
Table 3.3 gives an overview of the noise level trends on the five Viskinge pavements.
For passenger cars, the dense graded DGAC12 pavement had the lowest increase of
0.40 dB/ year, followed by the open graded OGAC12 pavement with 0.51 dB/ year.
For the three porous pavements the increase was around twice as high with 0.81 and
0.87 dB/ year. The increase for multi axle vehicles were generally around 50% of the
increase for passenger cars, with the PAC8 Type B pavement as an exception; here the
trend for multi axle vehicles were only a fourth of the trend for passenger cars.
Table 3.4. Predicted average noise level increase per year for passenger cars and multi axle vehicles
for five pavements like the Viskinge test sections, but constructed for long structural lifetime with
modified bitumen assuming a lifetime of eleven years ( reference speed 80 km/ h ( 50 mph)).
Pavement Passenger cars Multi axle vehicles
DGAC12 0.25 dB/ year 0.13 dB/ year
OGAC12 0.32 dB/ year 0.17 dB/ year
PAC8 type A 0.55 dB/ year 0.24 dB/ year
PAC8 type B 0.52 dB/ year 0.13 dB/ year
PAC12 0.53 dB/ year 0.28 dB/ year
As mentioned in section 2.3, the intention of the Viskinge experiment was to perform
a “ fast” life cycle testing of porous pavements. For this reason the five pavements
were deliberately built to break down faster than would normally be the case. Modifi-ers
were not added to the bitumen. New Dutch results show that porous pavements
built for long structural lifetime ( with modified bitumen) can be constructed, so they
have a lifetime of around eleven years [ 28]. The acoustical performance of the five
pavements is “ stretched” to a lifetime of eleven years by multiplying the yearly in-creases
by 7/ 11.
61
Table 3.4 shows the expected noise increases of new durable porous pavements with
modified bitumen. The dense and the open graded asphalt concrete for passenger cars
then get a noise increase of respectively 0.25 and 0.32 dB/ year and the porous pave-ments
an increase of 0.52 to 0.55 dB/ year. This “ stretching” of the noise increase as a
function of age makes it possible to compare the results with the results from the other
test sections included in this report.
3.4 M10 test sections, DK
The test sections on highway M10 in Denmark were constructed in 2004 [ 19].
SPB measurements have been conducted every year over a five year period and it is
planned to continue these measurements. The SPB results are here reported as LAmax
levels. The microphone position used was 7.5 m ( 25 feet) from the centerline of the
lane and at a height of 1.2 m ( 4 feet) according to ISO 11819- 1 [ 3]. The data are all
normalized to an air reference temperature of 20° C ( 68° F). The trends for the devel-opment
of the noise for passenger cars as well as for multi axle vehicles, and the
spectra can be seen in the below figures. The Residual Standard Error can be seen
in Table 3.5.
DGAC11 - Passenger cars
y = 0,72x + 80,82
R2 = 0,97
77
79
81
83
85
87
0 1Age [ yea2rs] 3 4
LAmax [ dB]
55
60
65
70
75
80
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
DGAC11
Figure 3.23. Maximum SPB noise level from passenger cars to the left and SPB noise spectra to the
right for the DGAC11 pavement ( reference speed 110 km/ h ( 69 mph)).
DGAC11 - Multi axle y = 0,28x + 88,29
R2 = 0,76
85,0
87,0
89,0
91,0
93,0
95,0
0 1 2 3 4
Age [ years]
LAmax [ dB]
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
DGAC11
Figure 3.24. Maximum SPB noise level from multi axle vehicles to the left and SPB noise spectra to
the right for the DGAC11 pavement ( reference speed 85 km/ h ( 53 mph)).
62
The results for the DGAC11 pavement on the M10 test sections can be seen in Figure
3.23 and 3.24. The noise level increase fits quite well with a linear regression ( Resid-ual
Standard Error of 0.2 and 0.3 dB ( see Table 3.5)). The yearly increase for passen-ger
cars is 0.72 dB/ year and a for multi axle trucks 0.28 dB/ year increase. The noise
level increases mainly at frequencies below 1600 Hz. This can indicate an increase in
the vibration generated noise caused by the pavement texture becoming rougher.
UTLAC8 - Passenger cars
y = 1,06x + 78,78
R2 = 0,97
77
79
81
83
85
87
0 1 Age [ 2years] 3 4
LAmax [ dB]
55
60
65
70
75
80
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
UTLAC8
Figure 3.25. Maximum SPB noise level from passenger cars to the left and SPB noise spectra to the
right for the UTLAC8 pavement ( reference speed 110 km/ h ( 69 mph)).
UTLAC 8 - Multi axle y = 0,35x + 86,32
R2 = 0,58
85,0
87,0
89,0
91,0
93,0
95,0
0 1 Age [ 2years] 3 4
LAmax [ dB]
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
UTLAC 8
Figure 3.26. Maximum SPB noise level from multi axle vehicles to the left and SPB noise spectra to
the right for the UTLAC8 pavement ( reference speed 85 km/ h ( 53 mph)).
The results for the UTLAC8 pavement can be seen in Figures 3.24 and 3.25. The noise
increase from this pavement fits well with a linear regression ( Residual Standard Error
of 0.3 and 0.5 dB ( see Table 3.5)). The yearly increase for passenger cars is 1.06 dB/
year and around three times the increase for multi axle trucks which is 0.35 dB/ year.
The yearly increase is higher for this open graded pavement than for the dense graded
DGAC11. For passenger cars, the noise level increases both in the frequencies above
and below 1000 Hz. This could indicate an increase in the vibration generated lower
frequency noise caused by the pavement texture becoming more rough as well as an
increase in the high frequency air pumping generated noise caused by the pavement
surface becoming denser.
63
OGAC8 Passenger cars
y = 0,80x + 78,50
R2 = 0,80
77
79
81
83
85
87
0 1 Age [ 2years] 3 4
LAmax [ dB]
55
60
65
70
75
80
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
OGAC8
Figure 3.27. Maximum SPB noise level from passenger cars to the left and SPB noise spectra to the
right for the OGAC8 pavement ( reference speed 110 km/ h ( 69 mph)).
OGAC8 - Multi axle y = 0,09x + 86,16
R2 = 0,04
85,0
87,0
89,0
91,0
93,0
95,0
0 1 Age [ 2years] 3 4
LAmax [ dB]
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
OGAC8
Figure 3.28. Maximum SPB noise level from multi axle vehicles to the left and SPB noise spectra to
the right for the OGAC8 pavement ( reference speed 85 km/ h ( 53 mph)).
The results for the OGAC8 pavement can be seen in Figures 3.27 and 3.28. The noise
level increase fits with a linear regression but with higher Residual Standard Error
than the other pavements in the M10 experiment ( 0.7 and 0.8 dB [ see Table 3.5]). The
yearly increase for passenger cars is 0.80 dB/ year whereas the increase for multi axle
trucks is just 0.09 dB/ year. Like for the UTLAC8 pavement the noise level for passen-ger
cars increases both in the frequencies above and below 1000 Hz, with the most
prominent increase at frequencies above 1000 Hz. This can indicate an increase in the
vibration generated lower frequency noise as well as an increase in the high frequency
air pumping generated noise caused by the pavement surface becoming denser espe-cially
during the first two years.
64
SMA8 - Passenger cars
y = 0,50x + 81,13
R2 = 0,73
77
79
81
83
85
87
0 1 Age [ 2years] 3 4
LAmax [ dB]
55
60
65
70
75
80
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
SMA8
Figure 3.29. Maximum SPB noise level from passenger cars to the left and SPB noise spectra to the
right for the SMA8 pavement ( reference speed 110 km/ h ( 69 mph)).
SMA8 - Multi axle y = 0,21x + 87,54
R2 = 0,92
85,0
87,0
89,0
91,0
93,0
95,0
0 1 Age [ 2years] 3 4
LAmax [ dB]
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
SMA8
Figure 3.30. Maximum SPB noise level from multi axle vehicles to the left and SPB noise spectra to
the right for the SMA8 pavement ( reference speed 85 km/ h ( 53 mph)).
The results for the SMA8 pavement can be seen in Figures 3.29 and 3.30 ( Residual
Standard Error of 0.5 and 0.1 dB [ see Table 3.5]). The yearly increase in noise level
for passenger cars is 0.50 dB/ year and corresponds to the lowest increase of the six
pavements on the M10 test section. For multi axle vehicles the increase is 0.21
dB/ year. The increase is mainly seen at the lower frequencies ( below 1600 Hz) indi-cating
that the pavement surface becomes rougher.
SMA6+ - Passenger cars
y = 0,93x + 78,98
R2 = 0,98
77
79
81
83
85
87
0 1 Age [ 2years] 3 4
LAmax [ dB]
55
60
65
70
75
80
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
SMA6+
Figure 3.31. Maximum SPB noise level from passenger cars to the left and SPB noise spectra to the
right for the SMA6+ pavement ( reference speed 110 km/ h ( 69 mph)).
65
SMA6+ - Multi axle y = 0,63x + 86,68
R2 = 0,81
85,0
87,0
89,0
91,0
93,0
95,0
0 1 Age [ 2years] 3 4
LAmax [ dB]
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
Year 4
SMA6+
Figure 3.32. Maximum SPB noise level from multi axle vehicles to the left and SPB noise spectra to
the right for the SMA6+ pavement ( reference speed 85 km/ h ( 53 mph)).
The SMA6+ pavement in Figures 3.31 and 3.32 has a noise level increase for passen-ger
cars and multi axle vehicles of respectively 0.93 and 0.63 dB/ year ( Residual Stan-dard
Error of 0.3 and 0.6 dB [ see Table 3.5]). The increase primarily occurs for fre-quencies
below 1600 Hz indicating that the pavement structure becomes rougher over
the years.
SMA8+ - Passenger cars
y = 1,32x + 77,62
R2 = 0,96
77
79
81
83
85
87
0 1 Age [ 2years] 3 4
LAmax [ dB]
55
60
65
70
75
80
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
SMA8+
Figure 3.33. Maximum SPB noise level from passenger cars to the left and SPB noise spectra to the
right for the SMA8+ pavement ( reference speed 110 km/ h ( 69 mph)).
SMA8+ - Multi axle y = 0,67x + 86,19
R2 = 0,83
85,0
87,0
89,0
91,0
93,0
95,0
0 1 2 3 4
Age [ years]
LAmax [ dB]
60
65
70
75
80
85
400 630 1000 1600 2500 4000
Frequency [ Hz]
SPB [ dB]
Year 0
Year 1
Year 2
Year 3
SMA8+
Figure 3.34. Maximum SPB noise level from multi axle vehicles to the left and SPB noise spectra to
the right for the SMA8+ pavement ( reference speed 85 km/ h ( 53 mph)).
66
The SMA8+ pavement was constructed one year after the other pavements on M10.
Therefore Figures 3.33 and 3.34 only include results from four years. This SMA8+ has
the highest noise level increase of the six pavements on M10 with increase for passen-ger
cars and multi axle vehicle of 1.32 and 0.67 dB/ year respectively. The spectra
show increase over the whole frequency range.
Table 3.5. Average noise level increase per year for passenger cars and multi axle vehicles and Re-sidual
Standard Error for the six test pavements on M10 ( reference speed 110 km/ h ( 69 mph) for
passenger cars and 85 km/ h ( 53 mph) for multi axle vehicles).
Pavement Passenger
cars
Residual
Standard
Error
Passenger
Multi axle
vehicles
Residual
Standard
Error
Multi axle
DGAC11 0.72 dB/ year 0.2 dB 0.28 dB/ year 0.3 dB
UTLAC8 1.06 dB/ year 0.3 dB 0.35 dB/ year 0.5 dB
OGAC8 0.80 dB/ year 0.7 dB 0.09 dB/ year 0.8 dB
SMA8 0.50 dB/ year 0.5 dB 0.21 dB/ year 0.1 dB
SMA6+ 0.93 dB/ year 0.3 dB 0.63 dB/ year 0.6 dB
SMA8+ 1.32 dB/ year 0.4 dB 0.67 dB/ year 0.5 dB
Table 3.5 gives an overview of the slope of the trend lines found at each of the six
M10 pavements. For passenger cars the DGAC and the SMA8 pavements have the
lowest noise level increase of 0.50 to 0.72 dB/ year. For the more open graded pave-ments,
the increase varies between 0.8 and 1.32 dB/ year. The increase for multi axle
trucks is generally much lower.
3.5 Californian investigation
On the 65 to 76 Californian test pavements the noise was measured by UCPRC using
the OBSI measurement method. All the results have been converted, so they represent
measurements performed by the SRTT Standard Reference Tire [ 17] at a reference
speed of 96 km/ h ( 60 mph). Temperature corrections are not applied to the results, but
according to [ 27] the SRTT tire is not very sensitive to variations in temperature. The
results of noise over the years are grouped for four pavements types ( irrespective of
maximum aggregate size):
• DGAC
• OGAC
• RAC- O
• RAC- G
The results are presented as the trend of OBSI noise level for each pavement type as
well as by an average spectrum for each pavement type in years with data available.
67
It must be remarked that the layout of this study is much different from the previous
four studies where the noise emission from exactly the same pavements has been
monitored over the years. In this project a population of pavements ( belonging to each
pavement type) with different ages are monitored. Each of the pavements has been
monitored for three consecutive years ( the research continues with the fourth year
evaluation in 2009). In addition of tracking each individual section for three years, the
study allows for combining sections of nominally the same mix type to obtain a longer
time span for pavement of the same type. Combining sections to derive noise trends
has some difficulties. Variation in the noise emission can be caused by differences in
the design and production of the pavements belonging to the same type for example by
the use of different nominal maximum aggregate size ( 9.5, 12.5 and 19.0 mm) etc.
Therefore, a larger spread in the data than in the previous four studies must be ex-pected.
Table 3.6 shows Residual Standard Errors in the order of magnitude of 1.3 to
1.5 dB. Due to the spread in the data it will be more difficult to draw firm conclusions
on the development of the noise spectra from year to year. The spectra are presented
anyway in the following.
y = 0,16x + 102,00
R2 = 0,14
96
98
100
102
104
106
108
0 2 4 6 8 10 12 14 16
Age [ Year]
OBSI [ dB]
DGAC
Figure 3.35. OBSI noise level measured for DGAC pavements in the Californian investigation [ 17].
The yearly increase in noise level for DGAC pavements can be seen in Figure 3.35.
The increase is 0.16 dB/ year. The frequency spectra in Figure 3.36 shows a “ dip”
around 1250 Hz when the pavements are 0 to 1 year old. This dip disappears in the
second year. This might indicate that the pavements become less open in the surface
structure during this period resulting in increased air pumping noise. From the second
year, the noise level increases at all frequencies.
68
70
75
80
85
90
95
100
500 630 800 1000 1250 1600 2000 2500 3150 4000
Frequency [ Hz]
OBSI [ dB]
Year 0 ( 3)
Year 1 ( 5)
Year 2 ( 5)
Year 3 ( 4)
Year 4 ( 3)
Year 5 ( 4)
Year 6 ( 4)
Year 7 ( 3)
Year 8 ( 3)
Year 9 ( 1)
Year 14 ( 1)
Year 16 ( 1)
DGAC
Figure 3.36. Noise spectra for OBSI measurements for DGAC pavements in the Californian investi-gation.
The number in the brackets indicates the number of measurement results that are included
in the actual spectrum [ 17].
y = 0,31x + 99,31
R2 = 0,28
96
98
100
102
104
106
108
0 2 4 6 8 10 12 14 16
Age [ Year]
OBSI [ dB]
OGAC
Figure 3.37. OBSI noise level measured for OGAC pavements in the Californian investigation [ 17].
69
70
75
80
85
90
95
100
500 630 800 1000 1250 1600 2000 2500 3150 4000
Frequency [ Hz]
OBSI [ dB]
Year 0 ( 3)
Year 1 ( 3)
Year 2 ( 3)
Year 3 ( 4)
Year 4 ( 6)
Year 5 ( 9)
Year 6 ( 8)
Year 7 ( 6)
Year 8 ( 3)
Year 9 ( 2)
Year 10 ( 2)
OGAC
Figure 3.38. Noise spectra for OBSI measurements for OGAC pavements in the Californian investi-gation.
The number in the brackets indicates the number of measurement results that are included
in the actual spectrum [ 17].
The yearly increase in noise level for OGAC pavements can be seen in Figure 3.37.
The increase is 0.31 dB/ year. The frequency spectra in Figure 3.38 show a “ dip”
around 1600 Hz when the pavements are 0 to 1 year old, like for the DGAC pave-ments.
This dip disappears in the second year. This might indicate that the pavements
become less open in the surface structure during this period, resulting in increased air
pumping noise. From the second year, the noise level increases at all frequencies like
for the DGAC pavements.
y = 0,18x + 99,36
R2 = 0,10
96
98
100
102
104
106
108
0 2 4 6 8 10 12 14 16
Age Year]
OBSI [ dB]
RAC- O
Figure 3.39. OBSI noise level measured for RAC- O pavements in the Californian investigation [ 17].
70
70
75
80
85
90
95
100
500 630 800 1000 1250 1600 2000 2500 3150 4000
Frequency [ Hz]
OBSI [ dB]
Year 0 ( 5)
Year 1 ( 7)
Year 2 ( 7)
Year 3 ( 10)
Year 4 ( 6)
Year 5 ( 8)
Year 6 ( 2)
Year 7 ( 2)
Year 8 ( 2)
Year 9 ( 2)
Year 10 ( 2)
RAC- O
Figure 3.40. Noise spectra for OBSI measurements for RAC- O pavements in the Californian investi-gation.
The number in the brackets indicates the number of measurement results that are included
in the actual spectrum [ 17].
The yearly increase in noise level for the RAC- O pavements can be seen in Figure
3.39. The increase is 0.18 dB/ year, similar to that of the DGAC pavements. The fre-quency
spectra in Figure 3.40 also show a “ dip” around 1600 Hz when the pavements
are new like for the DGAC and OGAC pavements. This might also here indicate that
the pavements become less open in the surface structure during this period resulting in
increased air pumping noise. From the second year,, the noise level mainly increases in
the frequency range between 800 and 2000 Hz.
y = 0,31x + 100,10
R2 = 0,33
96
98
100
102
104
106
108
0 2 4 6 8 10 12 14 16
Age [ Year]
OBSI [ dB]
RAC- G
Figure 3.41. OBSI noise level measured for RAC- G pavements in the Californian investigation [ 17].
71
70
75
80
85
90
95
100
500 630 800 1000 1250 1600 2000 2500 3150 4000
Frequency [ Hz]
OBSI [ dB]
Year 0 ( 3)
Year 1 ( 5)
Year 2 ( 5)
Year 3 ( 2)
Year 4 ( 2)
Year 5 ( 4)
Year 6 ( 6)
Year 7 ( 4)
Year 8 ( 1)
Year 9 ( 1)
Year 10 ( 1)
Year 11 ( 1)
RAC- G
Figure 3.42. Noise spectra for OBSI measurements for RAC- G pavements in the Californian investi-gation.
The number in the brackets indicates the number of measurement results that are included
in the actual spectrum [ 17].
Figure 3.41 shows that the increase in noise level for RAC- G pavements is 0.31
dB/ year. The frequency spectra in Figure 3.42 show a “ dip” around 1250- 1600 Hz
when the pavements are new ( 0 years old) like for the other three pavements types, but
this dip disappears in the first year. From the first year, the noise level increases at all
frequencies.
Table 3.6. Average noise level increase per year for OBSI noise levels and Residual Standard Error
for the four Californian pavement types ( reference speed 96 km/ h ( 60 mph).
Pavement OBSI noise increase Residual Standard
Error
DGAC 0.16 dB/ year 1.4 dB
OGAC 0.31 dB/ year 1.3 dB
RAC- O 0.18 dB/ year 1.5 dB
RAC- G 0.31 dB/ year 1.3 dB
Table 3.6 shows a summary of the results for the four pavement types. The noise in-crease
is between 0.16 and 0.31 dB/ year.
72
73
4. Analysis and comparison
The results from the different test sections in Chapter 3 are compared and evaluated in
the following. The comparison is structured in three main parts:
1. The data from all seventeen pavements plus the four pavement types in the Cali-fornian
investigation are compared ( Section 4.1).
2. The data are subdivided into four pavement groups and the results are compared
for the pavements in each pavement group ( Section 4.2).
3. The average results for each pavement group are compared ( Section 4.3).
In Section 4.4, a first attempt to develop a model describing the increase of noise
emission for pavements at highways is presented. Such a model could be integrated in
a Pavement Management System ( PM System) taking noise into consideration.
The results from the I- 80 study on OGAC pavements were performed as LAeq meas-urements
of mixed traffic. Therefore it is not possible to separate the increase of noise
from passenger cars and heavy