THE IMPACT OF MATERIAL CHARACTERISTICS ON TIRE PAVEMENT
INTERACTION NOISE FOR FLEXIBLE PAVEMENTS
By
Salih Kocak

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Civil Engineering
2011

ABSTRACT
THE IMPACT OF MATERIAL CHARACTERISTICS ON TIRE PAVEMENT
INTERACTION NOISE FOR FLEXIBLE PAVEMENTS
By
Salih Kocak
Noise pollution has recently been one of the growing problems all over the world. While there
are many sources of the noise, traffic noise is the main contributor to the total environmental
noise. Although there are different sources for traffic noise, the tire pavement interaction noise is
the most dominant component within most city and highway limits. One of the ways to reduce
the tire pavement noise is to improve the material characteristics of the pavements such that they
produce less noise. In this study, the relationship between basic material characteristics (e.g., Hot
Mix Asphalt (HMA) volumetrics) and sound generation and absorption characteristics of flexible
pavements was investigated. In addition, the effect of linear visco-elastic properties (e.g.,
dynamic modulus (|E*|) and phase angle (δ)) on sound absorption was studied. In order to focus
only on impact of material characteristics and overshadow the effect of surface texture, a novel
laboratory tire pavement noise measurement simulator (TIPANOS) was developed. The
statistical analysis results showed that although the individual material characteristics do not
have appreciable influence on sound absorption, there is a significant correlation between sound
pressure levels (SPL) and combination of several material and linear visco-elastic parameters.

To My Family

iii

ACKNOWLEDGEMENTS
I would like to thank my advisor, Dr. Muhammed Emin Kutay, for his endless support, guidance
and friendship throughout my graduate career. He always encouraged and gave me back my
confidence whenever I lost it. I also would like to thank to the Civil and Environmental
Department at Michigan State University for providing me with the opportunity to pursue my
graduate degree. The efforts and inputs of Dr. Syed Waqar Haider and members of the advisory
committee, Dr. Karim Chatti, and Dr. Neeraj Buch, are highly appreciated.
I would like to extend my appreciations to my colleagues, Ashvini Kumar Thottempudi, Hande
Isik Ozturk, and Abrar Siddiqui for their invaluable discussions and supports.
Thanks to my family and friends for their continuous emotional support and encouragement.
The research presented in this thesis was supported by the MSU’s Intramural Research Grants
Program (IRGP).

iv

TABLE OF CONTENTS
LIST OF TABLES .................................................................................................................... VII
LIST OF FIGURES .................................................................................................................... IX
CHAPTER 1
INTRODUCTION......................................................................................................................... 1
1.1 BACKGROUND ................................................................................................................... 1
1.2 PROBLEM STATEMENT .................................................................................................... 3
1.3 OBJECTIVES ........................................................................................................................ 3
1.4 RESEARCH PLAN ............................................................................................................... 3
CHAPTER 2
LITERATURE REVIEW ............................................................................................................ 6
2.1 INTRODUCTION ................................................................................................................. 6
2.2 TIRE-PAVEMENT NOISE MECHANISM.......................................................................... 7
2.2.1 Air Resonant Mechanism ................................................................................................ 8
2.2.2 Radial Vibration Mechanism .......................................................................................... 8
2.2.3 Adhesion Mechanism ...................................................................................................... 8
2.3 QUITE PAVEMENTS........................................................................................................... 9
2.4 PARAMETERS AFFECTING NOISE................................................................................ 10
2.5 IMPROVING MATERIAL CHARACTERISTICS TO REDUCE NOISE......................... 12
2.6 SENTHESIS OF THE PREVIOUS WORK AND MOTIVATION FOR THE CURRENT
STUDY ...................................................................................................................................... 13
CHAPTER 3
MATERIAL AND MIXTURE PROPERTIES ........................................................................ 15
3.1 PHYSICAL PROPERTIES OF AGGREGATES ................................................................ 15
3.2 PHYSICAL PROPERTIES AND ENGINEERING CHARACTERISTICS OF ASPHALT
BINDERS .................................................................................................................................. 16
3.3 PHYSICAL PROPERTIES OF HMA........................................................................................... 19
3.3.1 Gradation of Aggregates and Sieve Analysis ................................................................ 19
3.3.2 Control Mix Design Parameters .................................................................................... 21
3.3.3 Preparation of Specimens for |E*| and Laboratory Noise Testing Using TIPANOS .... 27
CHAPTER 4
LABORATORY INVESTIGATIONS AND INSTRUMENTATION ................................... 31
4.1 TIRE PAVEMENT INTERACTION NOISE MEASUREMENTS .................................... 31
4.1.1 Tire Pavement Noise Measurement Simulator (TIPANOS) ......................................... 33
4.1.2 Dynamic Signal Analyzer (DP-QUATTRO) ................................................................ 35
4.1.3 Free Field Microphone (G.R.A.S- Type 46AE) ............................................................ 35
4.1.4 Sound Calibrator (LD-CAL200) ................................................................................... 36
4.2 DYNAMIC MODULUS MEASUREMENTS .................................................................... 37
v

4.2.1 Asphalt Mixture Performance Tester (IPC-AMPT) ...................................................... 43
4.2.2 Environmental Chamber ............................................................................................... 44
CHAPTER 5
DATA ANALYSIS ...................................................................................................................... 45
5.1 CORRELATION OF INDIVIDUAL MATERIAL CHARACTERISTICS AND SOUND
PRESSURE LEVEL (SPL) ........................................................................................................ 45
5.1.1 Relation between Linear Visco-Elastic Parameters of Asphalt and SPL ...................... 45
5.1.2 Relation between Mixture Volumetric Parameters and SPL ........................................ 51
5.1.3 Relation between Gradation Parameters, Material Types and SPL .............................. 55
5.2 MULTIVARIATE REGRESSION ANALYSIS ................................................................. 61
CHAPTER 6
SUMMARY AND CONCLUSIONS ......................................................................................... 69
6.1 SUMMARY ......................................................................................................................... 69
6.2 CONCLUSIONS ................................................................................................................. 69
APPENDICES ............................................................................................................................. 71
APPENDIX A
(JOB MIX FORMULAS) ........................................................................................................... 72
APPENDIX B
(TIPANOS SPL MEASUREMENT RESULTS)...................................................................... 77
APPENDIX C
(CORRELATIONS BETWEEN LINEAR VISCO-ELASTIC PARAMETERS AND SPL
AT DIFFERENT FREQUENCY AND TEMPERATURE COMBINATIONS) .................. 82
REFERENCES ............................................................................................................................ 89

vi

LIST OF TABLES
Table 3.1Particle size and physical properties of aggregates ....................................................... 16
Table 3.2 Properties of binders used during the research ............................................................ 17
Table 3.3 Mixing and compaction temperatures for the binders used in the research ................. 18
Table 3.4 Aggregate gradations used in the research mix designs............................................... 19
Table 3.5 Initial gradations and control points for control mix design ........................................ 21
Table 3.6 Superpave design criteria for control mix and obtained results ................................... 24
Table 3.7 Superpave mix design criteria for Nmas 9.5 and 25 mm (AASHTO R-35) ................ 25
Table 3.8 Mix designs and basic mixture-volumetric values ....................................................... 25
Table 3.9 Material notation and parameters controlled................................................................ 28
Table 4.1 Properties of the laboratory prepared specimens ......................................................... 31
Table 4.2 SPL at 250 Hz and 23OC .............................................................................................. 32
Table 4.3 Averaging dynamic modulus and phase angles for sample 125SA4CR5 .................... 39
Table 4.4 Master curve construction for sample 125SA4CR5..................................................... 40
Table 4.5 Shift and sigmoid coefficients for sample 125SA4CR5 .............................................. 42
Table 5.1 Specimens linear visco-elastic properties at 212 Hz and 23OC ................................... 47
Table 5.2 Volumetric parameters of the specimens ..................................................................... 52
Table 5.3 Gradation parameters and control parameters of the asphalt mixtures ........................ 57
Table 5.4 Gradation parameters for dense and uniform graded aggregates ................................. 58
Table 5.5 SPSS results for SPL and linear visco-elastic parameters ........................................... 62
Table 5.6 SPSS results for SPL and gradation parameters .......................................................... 64
Table 5.7 SPSS results for SPL and significant parameters 1 ...................................................... 65
Table 5.8 SPSS results for SPL and significant parameters 2 ...................................................... 68
Table B.1 Sound pressure level at 1/3 octave frequencies ........................................................... 78

vii

Table B.2 Sound pressure level at 1/3 octave frequencies (continued) ....................................... 79
Table B.3 Sound pressure level at 1/3 octave frequencies (continued) ....................................... 80
Table B.4 Sound pressure level at 1/3 octave frequencies (continued) ....................................... 81

viii

LIST OF FIGURES
Figure 1.1 On-Board Sound Intensity (OBSI) levels versus the maximum texture height
relationships for a variety of pavement surface (Rasmussen et al. 2006) ............................... 2
Figure 3.1 Compaction and mixing temperature determination for original binder .................... 18
Figure 3.2 Aggregate gradation curves and maximum density lines (MDL) .............................. 20
Figure 3.3 Superpave control mix design gradations on 0.45 power graph................................. 22
Figure 3.4 Gmb values for 3 gradations and 1 binder content...................................................... 23
Figure 3.5 Va values for 3 gradations and 1 binder content......................................................... 23
Figure 3.6 VMA values for 3 gradations and 1 binder content.................................................... 23
Figure 3.7 VFA values for 3 gradations and 1 binder content ..................................................... 24
Figure 3.8 Job mix formula for Superpave mix design NMAS 12.5 mm (control mix).............. 27
Figure 3.9 Flow chart showing the material matrix ..................................................................... 30
Figure 4.1 SPL versus 1/3 octave frequency of the specimens. ................................................... 32
Figure 4.2 (a) TIPANOS set-up, (b) TIPANOS plan view ........................................................ 34
Figure 4.3 QUATTRO data acquisition hardware (www.dataphysics.com) ............................... 35
Figure 4.4 GRAS type 40 AE free field microphone and CCP 26 CA preamplifier combination
(http://www.gras.dk/)............................................................................................................. 36
Figure 4.5 Larson Davis L 200 portable sound calibrator (www.larsondavis.com) .................... 37
Figure 4.6 Sigmoidal master curve for sample 125SA4CR5 ....................................................... 41
Figure 4.7 Phase angle fitting for sample 125SA4CR5 ............................................................... 41
o

Figure 4.8 |E*| versus frequency for all samples at 21 C ........................................................... 42
o

Figure 4.9 Phase angle versus frequency for all samples at 21 C ............................................... 43
Figure 4.10 IPC Global AMPT .................................................................................................... 44
Figure 5.1 |E*| versus reduced frequency curves for specimens ................................................. 48
Figure 5.2 Change in SPL with change in |E*| ............................................................................ 48
ix

Figure 5.3 Change in SPL with change in |E*|/sin δ.................................................................... 49
Figure 5.4 Change in SPL with change in δ ................................................................................ 49
Figure 5.5 Change in SPL with change in E’ ............................................................................... 49
Figure 5.6 Change in SPL with change in E” .............................................................................. 50
Figure 5.7 Change in SPL with change in air void content ......................................................... 53
Figure 5.8 Change in SPL with change in voids in mineral aggregate ........................................ 53
Figure 5.9 Change in SPL with change in voids filled with asphalt ........................................... 54
Figure 5.10 Change in SPL with change in binder content ........................................................ 54
Figure 5.11 Change in SPL with change in Gmm and Gmb......................................................... 55
o

Figure 5.12 Peak SPL values at 250 Hz and 23 C....................................................................... 56
Figure 5.13 Change in SPL with change in Cu ............................................................................ 57
Figure 5.14 Dense and uniform gradation for maximum aggregate size 12.5 mm ..................... 58
Figure 5.15 Change in SPL with change in Cu, only one for each gradation .............................. 60
Figure 5.16 Change in SPL with change in Cc ............................................................................ 60
Figure 5.17 Measured and predicted SPL values for general prediction model 1 ....................... 66
Figure 5.18 Measured and predicted SPL values for general prediction model 2 ....................... 68
Figure A.1 Job Mix Formula for Superpave NMAS 9.5 mm ...................................................... 73
Figure A.2 Job Mix Formula for Superpave NMAS 25 mm ....................................................... 74
Figure A.3 Job Mix Formula for Stone Matrix Asphalt .............................................................. 75
Figure A.4 Job Mix Formula for Open Graded Friction Course ................................................. 76
o

Figure C.1 SPL vs (a) Phase angle (b) Dynamic Modulus at 4 C and 10 Hz ............................. 83
o

Figure C.2 SPL vs (a) |E*| (b) |E*|/ sin δ at 4 C and 10 Hz ........................................................ 83
o

Figure C.3 SPL vs (a) Phase angle (b) Dynamic Modulus at 4 C and 5 Hz ............................... 84

x

o

Figure C.4 SPL vs (a) |E*| (b) |E*|/ sin δ at 4 C and 5 Hz .......................................................... 84
o

Figure C.5 SPL vs (a) Phase angle (b) Dynamic Modulus at 4 C and 0.1 Hz ............................ 85
o

Figure C.6 SPL vs (a) |E*| (b) |E*|/ sin δ at 4 C and 0.1 Hz ....................................................... 85
o

Figure C.7 SPL vs (a) Phase angle (b) Dynamic Modulus at 21 C and 10 Hz ........................... 86
o

Figure C.8 SPL vs (a) |E*| (b) |E*|/ sin δ at 21 C and 10 Hz ...................................................... 86
o

Figure C.9 SPL vs (a) Phase angle (b) Dynamic Modulus at 21 C and 5 Hz ............................. 87
o

Figure C.10 SPL vs (a) |E*| (b) |E*|/ sin δ at 21 C and 5 Hz ...................................................... 87
o

Figure C.11 SPL vs (a) Phase angle (b) Dynamic Modulus at 21 C and 0.1 Hz ........................ 88
o

Figure C.12 SPL vs (a) |E*| (b) |E*|/ sin δ at 21 C and 0.1 Hz ................................................... 88

xi

CHAPTER 1
INTRODUCTION
1.1 BACKGROUND
The noise pollution has been one of the most obtrusive problems all over the world recently.
Traffic noise is one of the major contributors to total environmental noise. People living at a
closer proximity to the highways, who are constantly exposed to the traffic noise, complain about
physical and psychological health related problems (Kropp et al. 2007). While there are different
sources of traffic noise, noise due to the tire/pavement interaction is dominant at high speeds
greater than 30 miles/hr (Sandberg 2001, Kutay et al. 2010). To minimize the impact of
tire/pavement interaction noise, tire manufacturers have been working on producing tires that
generate less noise. On the other hand, constructing sound walls, to absorb the traffic noise, was
the most commonly used sound mitigating solution by State Departments of Transportation
(DOTs). However, the noise walls absorb a small percent of the noise and reflect most of the
sound waves. This leads to driver discomfort. Moreover, the noise walls are expensive to
construct (~$2.1 million per mile) (Hanson et al. 2004). Therefore, to solve the problem at the
source, DOTs have been spending great effort to produce pavements having good sound
absorption capacity.
One of the major contributors to the tire/pavement noise generation is the surface texture. Recent
studies have primarily focused on measurement of texture using both 2D and 3D laser-based
techniques and developed correlations between tire/pavement noise and certain parameters
representing texture such as the mean profile depth. However, it was observed that the
correlation between the texture characteristics and pavement noise is not always consistent
(Rasmussen et al. 2006). This is possibly due to the material characteristics of the pavements that

1

cause damping or amplification of the sound. Depending on the design of the pavement material,
these characteristics may overshadow the effects of texture. Certain pavements with surface
textures designed for low levels of tire/pavement noise were reported in the literature to lose
their characteristics quickly over the years. Moreover, both microtexture and macrotexture result
in the deformations on the tread of the tire and escorting to the vibrations of tire and all
suspension system (Wayson 1998).

Figure 1.1 On-Board Sound Intensity (OBSI) levels versus the maximum texture height
relationships for a variety of pavement surface (Rasmussen et al. 2006).
“For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this thesis.”

2

Pavement surface can be textured to minimize tire/pavement noise. However, the degradation of
the texture is dependent on the material properties such as the aggregate type and shape used in
the mixture, asphalt binder characteristics...etc. Even though pavement material properties can
have such a significant influence on the highway noise, there is a lack of literature on the
relationship between the material properties and the generation of noise. Furthermore, the
durability of surface texture calls for development of such relationships to better understand the
tire/pavement noise generation and for a better design of long-lasting quiet pavements.
1.2 PROBLEM STATEMENT
As mentioned earlier, the noise generated by the pavement is dependent on the surface texture of
the pavement, which is further dependent on the material properties such as aggregate type and
shape, asphalt binder characteristics…etc. The primary aim of this research is to identify the
fundamental material characteristics of asphalt pavements that affect the tire-pavement noise
generation and propagation to lead to an improved pavement design for sustainable green
highways in the future.
1.3 OBJECTIVES
The main objective of this research is to investigate relationship between tire/pavement noise
generation/absorption and material characteristics of flexible pavements. It also presents the
impact of material mix design characteristics as well as linear visco-elastic properties on sound
absorption.
To accomplish the objective, a research plan consisting of three major tasks was developed and
is presented in the next section.
1.4 RESEARCH PLAN
As stated above, the research plan for this study consists of the three tasks detailed below.

3

Task 1 – Information and Material Supply – In this task, mix design types, specifications and
standards utilized by most of the state DOT’s were obtained and analyzed. The most suitable mix
designs for the research were adopted. AASHTO standards were followed in both testing of the
materials and constructing mix designs. The following steps can summarize the task 1,
•

Getting familiar with the standards, specifications and DOT’s practices. All tests
performed on aggregates, binders and HMAs follow AASHTO standards.

•

Obtaining different mix designs typically used by DOTs and choosing the most suitable
designs for the research.

•

Acquiring ample aggregate and binder samples for preparation of HMAs.

•

Gathering the information and ideas to construct a system capable of measuring the
tire/pavement interaction noise in the laboratory environment without taking into account
the effect of surface texture.

•

Getting familiar with the laboratory equipment for tests and specimen preparation.

Task 2 — Evaluation of Materials and Preparation of Specimens- All material supplies used
in the study were tested according to the corresponding AASHTO standards. Only those
materials complying with the AASHTO standards are used in this research. The assessment and
preparation of specimens include:
•

Testing as-received supplies according to the AASHTO standards. These tests include all
performance grading (PG) tests for binders and physical property tests of aggregates. The
materials were assessed if they conformed to the specifications after testing part was
performed.

•

Performing mix designs. Superpave (SP) 9.5mm, SP 12.5 mm, SP 25 mm, open graded
friction course (OGFC) and stone matrix asphalt (SMA) are the mix designs performed.

4

•

Preparation of specimens. All specimens were prepared by using a laboratory slab
compactor with shearing capabilities.

Task 3 – Laboratory Testing and Data Analysis – The cylindrical specimens prepared during
the previous task were tested for dynamic modulus measurements and sound pressure levels.
Since both the tests are non-destructive in nature, same specimens were used. Statistical analyses
were performed on the measured dynamic modulus and sound pressure levels. This task
includes:
•

Dynamic modulus measurement by using Asphalt Mixture Performance Tester (AMPT).

•

Tire/pavement interaction noise measurements via the laboratory device developed in this
research (Tire-Pavement Noise Measurement Simulator (TIPANOS)).

•

Statistical analysis of relationship between individual material characteristics and sound
pressure levels.

5

CHAPTER 2
LITERATURE REVIEW
2.1 INTRODUCTION
Noise pollution has recently been one of the most important problems in the United States and
other countries all over the world. Loss of hearing, anxiety, sleeplessness, aggression, speech
interference, increase in heart rate and stress are just few common examples of physical and
psychological health related problems caused by noise pollution (Shatanawi et al. 2008,
SCDREA 1999). While there are many sources of the noise, traffic noise is a major contributor
to the total environmental noise. Traffic noise mainly composed of three different sources, which
can be classified as propulsion, tire-pavement interaction, and aerodynamic noises. Sounds
generated from the engine, exhaust, intake and other power-train components constitute the
propulsion noise type. This kind of noise governs the total noise at low speeds (Rasmussen et al.
2007). The sound generated by the interaction of tire and pavement surface is called tirepavement noise, which becomes dominant when a crossover speed is reached. Crossover speed is
between 10-25 mph for cars and 35-50 mph for trucks (Rasmussen et al. 2007). With today’s
technology in manufacturing quieter engines, crossover speed has been decreasing significantly.
The third kind of noise is the aerodynamic noise and it is generated as a result of the air
turbulence around a vehicle. This type of noise becomes dominant at very high speeds. In city
and highway speed limits, tire pavement interaction noise controls the overall noise generated by
moving vehicles. To reduce the noise generated by running traffic, both the vehicle and tire
manufacturers have shown significant accomplishments in this multidisciplinary problem. As the
third party in the solution of the traffic noise problem, the road owners also have been sharing
the responsibility by conducting or supporting research on quite pavements. In order to meet the

6

public demand, Federal Highway Administration (FHWA) recommends six traffic noise
abatement methods (FHWA 1997). The most commonly used method is construction of noise
barriers. Besides being costly (~$2.1 million per mile), the noise walls are not very effective and
desirable because of the driveway access, height limitation and their aesthetical view. Typical
concrete noise barriers have high acoustic reflectivity with 95% or above and very limited sound
absorption (Campbell 2000, Zhu et al. 2008). Therefore, they can only alter the direction of
propagation. Redirected noise from the barriers adversely affects the passengers and drivers
using the roads. Moreover, Campbell (2000) showed the inefficiency of concrete noise barriers
in controlling and reducing the impact of traffic noise by both field measurements and modeling
studies. Constructing quieter pavements is anticipated to be a more economic and sustainable
way in reducing the tire pavement interaction noise. The low noise road surface is defined as the
“road surface which, when interacting with a rolling tire, influences the vehicle noise in such a
way as to cause at least a 3 dB (A) lower vehicle noise than that obtained on conventional and
most common road surfaces” (Sandberg et al. 2002). The definition of common road surface
differs from country to country. In the ISO standards, it corresponds to the reference surface with
dense, smooth- textured asphalt concrete surface with a maximum aggregate size between 11
mm. to 16 mm (ISO 1997).
2.2 TIRE-PAVEMENT NOISE MECHANISM
The interaction between tire and pavement results in generation of noise. Depending on the type
of the tire and pavement surface, the noise level can vary greatly. There are several mechanisms
that explain the generation of the sound at the interface between the tire and pavement.
Moreover, there are some other factors that can contribute to the amplification of the
mechanisms.

7

All mechanisms can be gathered under three categories. These are air resonant, radial vibration
and adhesion mechanisms.
2.2.1 Air Resonant Mechanism
Air resonant mechanism includes three main components and becomes dominant after 1000 Hz
frequency level. The first one is the pipe resonance which amplifies the sound generated in
another place inside the grooves of the tire tread and on channels along the surface the tire
(Sandberg 1992, Rasmussen 2007). Another component is the Helmholtz resonance that occurs
when the air in the tire tread cavity behaves as a spring which resonates with the mass of air in
between the cavity and the atmospheric air while the tire rotates. The last one is the air pumping.
It forms in between the tire tread and pavement surface texture as the gaps in between filled with
air. While the tire rolls over the pavement, either air is squeezed out or trapped and compressed.
When tire loses the contact with pavement at a point, trapped air is forced to out. This process is
repeated hundreds of times in a second and results in a large amount of air turbulence and as a
result noise (Leasure et al. 1975, Rasmussen 2007).
2.2.2 Radial Vibration Mechanism
The radial vibration mechanism occurs as the tire rolls over the pavement. It is more pronounced
at frequencies below 1000 Hz. In this mechanism, the vibrations (noise) are induced by small
deflections due to the interactions between pavement texture and the tread of the tire and
propagate to the air. It can be described by a hammer physical analogy. It can be visualized by
assuming each tread as a hammer stroking to the pavement thousands of times a second.
2.2.3 Adhesion Mechanism
The adhesion mechanism includes two components. As in the case of air resonant mechanism,
adhesion mechanism is more pronounced at 1000 Hz and higher frequencies. Stick-slip is the

8

first component and it occurs as a result of the vibrations due to tangential slippage of the tire
tread between tire and the road surface. The second constituent is stick-snap. It happens when the
rubber adheres and is released vertically from the road surface as the tire rotates. The physical
analogy for stick-snap can be seen as suction cup. There are some other components of the tire
pavement interaction noise that amplify the mechanisms explained as well.
2.3 QUITE PAVEMENTS
The research on quite pavements first started in Europe in 1970’s, and one-decade later, Japan
researchers began to implement low noise pavements. A few decades ago, the importance of the
quite pavements was realized by FHWA and research projects were initiated in this field in
United States (Meiarashi 1999, Rasmussen et al. 2007).
There are three typical types of hot mix asphalt (HMA) designs used in the high volume
highways: Open Graded Friction Course (OGFC), Dense Graded HMA Mix (DGA), and Stone
Matrix Asphalt (SMA) mix (Hanson et al. 2004). In literature, porous pavements also known as
OGFC, gap graded asphalt or drainage asphalt, are reported to be one of the quietest pavement
types. Studies have shown that an OGFC can reduce the noise level 3 to 5 dB (A), when
compared to a dense HMA pavement. This is because the air voids in the pavement provide a
means for air trapped between the tire and the pavement surface to escape, and lead to increased
sound absorption. To be able to damp the noise successfully, the pores need to be interconnected
(Sandberg et al. 2002). Furthermore, porous surfaces have an advantage of efficient drainage of
water and reduce the splash and spray behind vehicles during rainfalls (Malcom et al. 2003).
However, one of the important challenges associated with the porous pavements is their
durability and effectiveness over years. Fine particulate on the roadways can quickly clog the
voids reducing pavement’s capability to absorb noise. The recent researches to solve the

9

clogging of surface with dirt and dust from environment and snow removal operations and
durability issues related with OGFC wearing surfaces have suggested the use of two layer system
(Hanson et al. 2004). Otherwise, clogging can become a serious problem especially in urban
areas and periodic cleaning operations are needed, which creates additional costs. A surface
named as Twinlay has been optimized to have a long acoustical lifetime for urban applications at
speeds around 50 kilometer per hour (km/h). In the applications of porous pavements where the
traffic speed is 90-130 km/h, there is a self-cleaning of the pavement surface and the acoustic
lifetime can be acceptable even without cleaning (Sandberg 1999).
2.4 PARAMETERS AFFECTING NOISE
There are several geometrical and road pavement parameters affecting the sound absorption of
porous pavement surfaces. Porosity, pore size distribution, air void, tortuosity, coarseness of the
aggregate mix, thickness of the porous layer and the airflow resistance per unit length are
parameters typically related to noise (Malcolm et al. 2003, Nelson et al. 2008).
Typical in-place air voids in dense HMA is 5-7%, whereas, the percent air void in porous mixes
ranges from 15% to 30%. While the tortuosity or shape factor is a measure of the shape of air
voids passages, the airflow resistance is the resistance experienced by air when it passes through
open pores in the pavement (Malcolm et al. 2003). It has been shown that airflow resistance and
air void content play an important role on peak sound absorption coefficient of porous surfaces
with 40 mm thick and tortuosity value of 5 (Von Meier 1998, Von Meier et al. 1990).
The frequency range of the traffic noise, which is undesirable for the public, lies between 250
Hz. and 4000 Hz (Lapcik 1998). The thickness of the porous mixes changes both the value and
the shape of the absorption peak. Hamet et al. (1990) performed the measurement of sound
absorption of various surfaces ranging from 50 mm to 400 mm thick. While 50 mm thick porous

10

surface had a sharp peak with almost unity at about 900 Hz, 100 mm surface had first peak in
450 Hz with a smooth peak and second sharp peak at 1,350 Hz. Moreover, the 150 mm thick
pavement had three different peak values at 300, 900 and 1,500 Hz (Hamet et al. 1990). The
thickness of the porous surface has a significant effect on both sharpness and peak frequency. In
general, as the thickness of the surface layer increases, peak frequency decreases. It is worth to
note that the most objectionable noise to human ear occurs between frequencies of 800 and 1200
Hz. Hence, the frequency at which the maximum acoustic absorption occurs could be
manipulated by changing the thickness of the surfaces (Narayanan et al. 2004). When the peak
absorption is desired at a frequency approximately 1000 Hz, which is the interstate highway
noise frequency for vehicles with most of the tire and road combinations, a porous surface
between 1.5 and 2.0 inch thick can be used effectively (Malcolm et al. 2003, Sandberg 2003). To
shift the absorption maxima to low frequency region when the traffic speed is low as in the case
of in city conditions and/or the percentage of heavy vehicles of the traffic volume is high, the use
of thicker asphalt pavement layer is suggested (Kropp et al. 2007). In particular, for the
frequency ranges below 1000 Hz sound waves show less attenuation inside the material, the
sound absorption coefficient peaks at a frequency at which the anti phase condition is satisfied
between the multi reflective waves in the material and the sound wave reflected from the front
surface of the material. Moreover, for the frequencies above 1000 Hz, attenuation gradually
increases while the interference decreases (Yamaguchi et al. 1999). Since the thickness, air void
and the interconnectivity of the air voids play an important role in sound absorption, most of the
researchers have been working on these characteristics to increase the efficiency. While dense
asphalt mixes’ air voids vary between 4% and 8%, with absorption coefficients ranging from 0.1
to 0.2, open graded mixes with air void content 15% absorption coefficients of 0.4 to 0.7 are

11

easily achieved (Hanson et al. 2004). In order to obtain reasonable noise reduction, which is
around 8 dB(A), it is advised to construct a pavement with an air void content 20% and
minimum thickness of 40-50 mm by using small chippings in the top layer. However, the
pavements thicker than 100 mm do not show significant sound absorption (Sandberg 1999).
2.5 IMPROVING MATERIAL CHARACTERISTICS TO REDUCE NOISE
Besides trying to increase the efficiency and durability of porous surfaces, some other
researchers have been working on different ways of reducing the tire pavement interaction noise.
One of the study areas is to use different kinds of bitumen and mixes. Use of crumb rubber or
polymer-modified binder in asphalt to increase the sound absorption has been one of the mostly
funded areas in this field recently. Especially the use of crumb rubber modified binder has more
than one advantage. While resulting in a decrease in noise level generated by tire pavement
interaction, it also helps environment by recycling the old tires inside asphalt. Crumb rubber is
obtained by re-processing (shredding) the disposed automobile tires into small pieces after
removing the fiber and the metal present inside (Zhu et al. 1999). Even though the idea of using
old tires to make asphalt was started in United States in 1940’s, the idea has not gained much
momentum. However, nowadays it has been gaining more popularity both in the States and all
over the world and it is used with confidence (SCDERA 1999). Using crumb rubber inside the
asphalt pavement also helps solving environmental problems caused by disposal of automobile
tires (Zhu et al. 1999). Furthermore, the presence of crumb rubber particles inside the binder
affected the volumetric properties of the mixtures such as permeability and the binder content
that in return have effect on the sound absorption (Shatanawi et al. 2008). A study in University
of Waterloo, Ontario revealed that crumb rubber modified open friction course absorbed 9% of
the sound generated (Ahammed et al. 2010). In another research performed, the surface of the

12

aged and cracked Portland Cement Concrete Pavement was overlaid by crumb rubber modified
asphalt (CRM). The noise measurements just before and after placement of the CRM asphalt
layer showed a decrease of 6.1 dB (A) in residential areas which are almost 40 ft away from the
roadway. Also, the noise level inside the car was measured and in average the noise reduction
was 5.2 dB (A) compared to the previously existing pavement surface. Upon the extremely
positive public opinion in the area, the City of Phoenix allocated new funds for rehabilitation of
roads with CRM asphalt (Carlson et al. 2005). The similar analysis was performed in the eight
streets in the City of Thousand Oaks, Sacramento. Before and after laying the asphalt rubber
overlay, the noise measurements were taken from 50 ft or more from the roadway centerlines.
The results indicated that traffic noise level reductions were between 3 and 7 dB (A) on freshly
resurfaced roadways compared to old one and 2 to 5 dB (A) of total noise reduction was
attributed by asphalt rubber overlay. Noise reduction was even more on the sites where traffic
speed is higher (Bucka 2002). As the time passed and pavements aged under temperature and
pressure applied by vehicles, the reduction in noise level decreased; however, it still stayed
noticeable around 1 to 3 dB (A). In his research, Meiarashi (1999) measured the annual noise
absorption degradation in drainage asphalt is about 1 dB (A)/year. From the literature review on
acoustic properties of pavements, using drainage surface with two layer system and crumb
rubber modified binder with anticipated thickness by considering the frequency of traffic
according to the flow speed will yield the maximum reduction in traffic noise. The reduction in
noise level degrades as the time passes but above-mentioned pavement surface never loses all its
ability to absorb noise compared to dense graded asphalt mixes (Meiarashi 1999).
2.6 SENTHESIS OF THE PREVIOUS WORK AND MOTIVATION FOR THE
CURRENT STUDY

13

Since one of the major contributors to the tire/pavement noise generation is the surface texture,
most of the recent studies focused on measurement of surface texture by using newly developed
techniques. The researchers worked on developing correlations between tire/pavement noise and
certain parameters representing texture such as the mean profile depth. However, it was observed
that the correlation between the texture characteristics and pavement noise is not always
consistent (Rasmussen et al. 2006). This might be because of the effect of pavements’ material
characteristics (causing damping or amplification) overshadowing the effects of surface texture
on the tire/pavement noise. Some other researchers studied the relation between individual
material parameters and tire pavement interaction noise. In their research, Kaloush et al. (2006)
analyzed the relation between the dynamic modulus (|E*|) test parameters for conventional and
asphalt rubber mixtures and tire/road noise characteristics. However, they focused only on the
individual correlations between the |E*| and noise levels and did not investigate combined effect
of other materials characteristics such as HMA volumetrics and aggregate properties. The
primary motivation for this research was that there was very limited or no information on the
relation between the material characteristics of HMA pavements and their ability to absorb
tire/pavement noise. These characteristics include (i) HMA mix design parameters (e.g., binder
and air void content, voids in mineral aggregate (VMA), voids filled with asphalt (VFA)…etc.),
(ii) aggregate properties (e.g., nominal maximum aggregate size (NMAS), coefficient of
uniformity (Cu) …etc.) and (iii) linear viscoelastic properties (e.g., |E*|, phase angle, storage
modulus (E’), loss modulus(E”) …etc.). Also, there was no realistic and practical sound
absorption measurement system that can be used on the laboratory-size asphalt samples. Such
system is important for evaluating different HMA types at the mix design stage so that the future
pavements can be designed by taking into account the HMA’s ability to absorb noise.

14

CHAPTER 3
MATERIAL AND MIXTURE PROPERTIES
This chapter covers the physical properties of aggregates, binders and mixtures along with their
sources. The aggregates were obtained from different locations in the State of Michigan (MI).
Michigan Paving and Materials Company (a.k.a. Spartan Asphalt) and Rieth-Riley Construction
Co., Inc. were the two companies providing the aggregates in Lansing, MI for the research. Light
weight aggregates used to investigate the effect of aggregates in a wider range were supplied by
Buildex Incorporated in Ottawa, Kansas.
The main supplier of the asphalt binder was Mathy Technology and Engineering (MTE) Services
in Onalaska, Wisconsin. MTE provided original and polymer modified binders used during the
research. Crumb rubber modified binders were supplied by Seneca Petroleum Co. Inc located at
Crestwood, Illinois.
All material tests were performed according to the corresponding AASHTO and ASTM
standards at Michigan State University (MSU) Advanced Asphalt Characterization Laboratory
(AACL).
3.1 PHYSICAL PROPERTIES OF AGGREGATES
The aggregates used during the research were mainly river gravels with at least one crush face.
Limestone was another mostly used aggregate type to produce specimens. Besides the regular
aggregates, some of the specimens were prepared by using commercially available light weight
aggregates which is also known as hydite. The more detailed information about the physical
properties of the aggregates is presented in job mix formula (JMF) tables in Appendix A. Table
3.1 shows the basic properties of the coarse, fine and light weight aggregates used throughout the
research.

15

Table 3.1Particle size and physical properties of aggregates
Sieve Opening
in mm

River Gravel
Specific Gravity

River Gravel
Absorption

Light Weight Agg.
Specific Gravity

3/4

19.5

2.706

NA

1/2

12.5

2.702

3/8

9.5

2.684

1.208

#4

4.75

2.633

1.212

#8

2.36

2.814

1.351

#16

1.18

2.821

1.501

#30

0.6

2.802

1.711

#50

0.3

2.632

#100

0.15

2.695

2.101

#200

0.075

2.695

2.303

Dust

<0.075

2.604

NA

1.66%

0.73%

1.202

1.902

As it can be seen from the table above, the specific gravities of the coarser light weight
aggregates are less than half of the specific gravities of river gravels. However, as the particle
size gets smaller, the difference between the specific gravities reduces. Since researchers and
manufactures of the light weight aggregates have been conducting tests for potential use in
asphalt and concrete mixtures, this research also aimed to study the effect of light weight
aggregates on tire pavement interaction noise.
3.2 PHYSICAL PROPERTIES AND ENGINEERING CHARACTERISTICS OF
ASPHALT BINDERS
Binders used throughout the research were evaluated according to AASHTO performance grade
standards. Since the binder determines the visco-elastic properties and climatic condition that
asphalt pavements can resist, it needs to be investigated deeply. In this research, other than
original binder, polymer and crumb rubber modified binders were used to analyze the effect of
modified binders on sound absorption capacity of the asphalts as well. Performance grade (PG)
16

of the binders is performed according to AASHTO R-29 and M-320 protocols. PG tests include;
•

Flash Point by Cleveland Open Cup, AASHTO-48

•

Viscosity Determination of Asphalt Binder by Using Rotational Viscometer (RV),
AASHTO T-316

•

Determining the Rheological Properties of Asphalt Binder Using a Dynamic Shear
Rheometer (DSR), AASHTO T-315

•

Effect of Heat and Air on a Moving Film of Asphalt, Rolling Thin Film Oven (RTFO),
AASHTO T-240

•

Accelerated Aging of Asphalt Binder Using a Pressurized Aging Vessel (PAV),
AASHTO R-28

•

Determining the Flexural Creep Stiffness of Asphalt Binder Using the Bending Beam
Rheometer (BBR), AASHTO T-313

The Table 3.2 summarizes the performance grade (PG), flash point, viscosity and specific gravity
of asphalt binders used in the research.
Table 3.2 Properties of binders used during the research
Binder Type

PG

Flash
Point
o

Viscosity
o
(at 135 C)

C

Specific
Gravity

Pa.s

Original Binder

58-28

272

0.34

1.023

Polymer Modified (PM)

70-28

314

1.26

1.029

5% Crumb Rubber Modified (CRM)

70-28

298

1.29

1.031

10% Crumb Rubber Modified (CRM)

76-28

335

2.20

1.033

Moreover, the mixing temperature (MT) and compaction temperature (CT) of the hot mix
asphalts can be obtained by using RV. According to AASHTO T312- Preparing and Determining
the Density of Hot Mix Asphalt Specimens by Means of the Superpave Gyratory Compactor,
17

MT and CT can be obtained by running the RV more than one temperature level and
demonstrating them on semi-log paper. Temperatures corresponding to 170 ± 20 milipascalsseconds (mPa.s) and 280 ± 30 mPa.s kinematic viscosities will give MT and CT of HMA
prepared by that asphalt binder (AASHTO T-312). Table 3.3 summarizes mixing and
compaction temperatures of the HMAs prepared by using each binder type. It is noted that the
mixing and compaction temperatures of CRM binders were recommended by the manufacturer.
Table 3.3 Mixing and compaction temperatures for the binders used in the research
Binder Type

Mixing Temp.

Compaction Temp.

o

o

C
149
175
180
180

Original
Polymer Modified
5% CRM
10% CRM

C
138
165
170
170

Since HMA containing modified binders follows manufacturer’s recommendations, Figure 3.1
shows how to obtain MT and CT of the original binder used in the research.
500

Viscosity (mPa.s)

Compaction and Mixing Temperatures

50
130

134

138

142

146

150

154
oC)
Temperature (

158

162

Figure 3.1 Compaction and mixing temperature determination for original binder
18

3.3 Physical Properties of HMA
After the physical properties of the aggregates and binders used in HMA preparation are
explored, the next step is to prepare asphalt specimens and check their properties whether they
conform to standards. The most important criteria to be checked is the air void content.
This section covers the aggregate gradations, control mix design parameters and process and
preparation of the specimens.
3.3.1 Gradation of Aggregates and Sieve Analysis
The effect of mix design parameters on sound absorption characteristics of pavements has been
investigated by performing five different mix designs: (i) Superpave (SP) 12.5 mm, (ii) SP 9.5
mm, (iii) SP 25.0mm, (iv) stone matrix asphalt (SMA) and (v) open graded friction course
(OGFC). Table 3.4 and Figure 3.2 provide the aggregate gradations for each type of the mix
designs.
Table 3.4 Aggregate gradations used in the research mix designs
Sieve Size
mm
25.000
19.000
12.500
9.500
4.750
2.360
1.180
0.600
0.300
0.150
0.075
(dust)
TOTAL

SP-9.5 mm
0.0%
0.0%
0.0%
9.0%
35.0%
18.0%
12.0%
10.0%
6.0%
3.0%
2.0%
5.0%
100.00%

Percent Retained in Each Sieve
SP-12.5 mm SP-25 mm
SMA
0.0%
3.0%
0.0%
0.0%
8.0%
0.0%
3.1%
14.0%
6.0%
7.0%
8.0%
29.0%
18.0%
13.0%
35.0%
19.0%
13.0%
10.0%
18.0%
11.0%
5.0%
15.4%
8.0%
3.0%
7.3%
7.0%
2.0%
4.7%
5.0%
4.0%
3.6%
7.0%
2.0%
4.0%
3.0%
4.0%
100.00%
100.00%
100.00%

19

OGFC
0.0%
0.0%
0.0%
0.0%
62.0%
28.0%
4.0%
2.0%
1.0%
0.0%
0.4%
2.6%
100.00%

1"

3/4"

1/2"

3/8"

#4

#8

#16

#30

#200
#100
#50

100.0
90.0
80.0

SP 9.5
SP 12.5

60.0
50.0

SMA 12.5
SP 25

40.0

OGFC
MDL _SMA

30.0
20.0
10.0

1/2"

Sieve Size

3/8"

#4

#8

#16

#30

#200
#100
#50

0.0

1"

MDL_SP 12.5
MDL_SP 25
MDL_SP 9.5
3/4"

% Passing

70.0

Figure 3.2 Aggregate gradation curves and maximum density lines (MDL)
Superpave mix design with maximum aggregate size 12.5 mm was chosen as the control mix
design since it has been very commonly used in asphalt pavement projects in Michigan. SMA
mixture design uses superpave mix design with some modifications (AASHTO M-325 & R-46,
2009). National Pavement Asphalt Association’s (NAPA 1999) designing and constructing SMA
mixtures and NCHRP report 425 (Brown et al. 1999) were two other sources studied for SMA
mixture design. Since the amount of binder was considerably high in SMA and OGFC mix
designs, cellulose fibers were introduced to the mixtures to prevent the draindown. The amount
of fiber added was 0.2% and 0.3% by weight of the mixture, in SMA and OGFC, respectively
(Kumar et al. 2005). Fibers were not kept in the oven with aggregates since they might catch fire
easily. They were introduced mixture during mixing period. Furthermore, 1% hydrated lime was
added to the OGFC mix design as an antistripping agent. Since OGFC mixtures have high air

20

void content compared to other mix designs, water can freely infiltrate to entire depth of the
pavement. To prevent the moisture damage to the asphalt, using hydrated lime as an antistripping
agent on OGFC mixtures is the most commonly used technique (Cooley et al. 2004).
3.3.2 Control Mix Design Parameters
The aggregate gradation for the control mix design can be seen in Table 3.4 as SP -12.5.
Furthermore, in this section detailed information about the findings of control mix design is
presented. AASHTO M323- Superpave Volumetric Mix Design and R35- Superpave Volumetric
Design for Hot Mix Asphalt were two main sources used for mix designs. Table 3.5 shows the
standard sieve sizes, maximum and minimum control points specified in standards and three
different gradations chosen to conduct the test. Figure 3.3 demonstrates the three trial gradations
on 0.45 power curve.
Table 3.5 Initial gradations and control points for control mix design
Sieve
Size (SS)
mm
19.5
12.5
9.5
4.75
2.36
1.18
0.6
0.3
0.15
0.075
0

SS

0.45

3.81
3.12
2.75
2.02
1.47
1.08
0.79
0.58
0.43
0.31
0.00

Gradation 1
% retained

Gradation 2
% retained

0.0
10.0
8.5
22.0
21.0
20.0
8.0
4.0
2.0
1.3
3.2

0.0
5.5
8.1
19.0
21.0
19.0
11.0
6.0
3.4
3.0
4.0

21

12.5mm
Superpave mixture
Gradation 3
gradation limits
% retained
Min.
Max.
(%)
(%)
0.0
100
3.4
90
100
7.0
90
18.0
19.0
28
58
18.0
15.3
7.3
4.7
3.6
2
10
4.0

120.00
100.00

% Passing

80.00
60.00

0.45 power
Gradation 1 % passsing
Gradation 2 % passsing
Gradation 3 % passsing
Min Ctrl pts
Max Ctrl pts

40.00
20.00
0.00
0.00

1.00

2.00

3.00

4.00

Sieve Size Power to 0.45
Figure 3.3 Superpave control mix design gradations on 0.45 power graph
After determining three aggregate gradations, the next step was to choose one binder content and
it was taken as 5% by using the previous experience with aggregates. At the end of the first step
of Superpave mix design, third gradation with 5.1 % estimated optimum binder content was the
closest one to the Superpave mix design requirements. During the second step of the design
process, specimens with third gradation and having optimum, optimum +/-0.5% and optimum
+1.0% binder contents were prepared and analyzed. The results of the analysis include the
change in theoretical maximum specific gravity (Gmm), bulk specific gravity (Gmb) and air void
content (Va). Figures 3.4 through 3.7 show the results of values obtained during the first step of
control mix design.

22

2.500

Gmb vs Gradation
G3, 2.485

2.450

Gmb

G2, 2.438
2.400
2.350
G1, 2.332
2.300

Gradations

Figure 3.4 Gmb values for 3 gradations and 1 binder content

Va vs Gradation
9.00
8.00

Va

7.00

G1, 7.91

6.00
5.00

G2, 4.79

4.00
3.00

G3, 3.98

Gradations

Figure 3.5 Va values for 3 gradations and 1 binder content

VMA vs Gradation
20.00
G1, 19.09

VMA

18.00
16.00
G2, 15.46
14.00
G3, 14.05
12.00

Gradations

Figure 3.6 VMA values for 3 gradations and 1 binder content

23

20.0

VFA vs Gradation
G1, 19.09

VFA

18.0
16.0
G2, 15.46
14.0
G3, 14.05
12.0

Gradations

Figure 3.7 VFA values for 3 gradations and 1 binder content
6

Equivalent single axle load (ESAL) for the design life was assumed a value in between 3x10 to
6

30x10 . According to this value, minimum voids in mineral aggregate (VMA), target voids filled
with asphalt binder (VFA) and number of gyrations to be applied to compact the specimens to
target air content (Va) were obtained from AASHTO design tables. The design criteria and
results obtained for the Superpave mix design NMAS 12.5 mm are given in the Table 3.6.
Table 3.6 Superpave design criteria for control mix and obtained results
Design Criteria
Va target
VMA (min)
VFA target
ESAL (20 years)
Ndes (gyration)

Results

4.00%
14.00%
65-75
6

>3x10 and <30x10

6

100

4.00%
14.41%
72.28
6

>3x10 and <30x10

6

100

Superpave mix design criteria for other NMAS values are important to mention. It is important to
note how the volumetric values change with NMAS although all other design requirements are
kept the same for the research purposes. SP mix design criteria for NMAS 9.5 mm and 25 mm
are also provided in the Table 3.7.

24

Table 3.7 Superpave mix design criteria for NMAS 9.5 and 25 mm (AASHTO R-35)
Superpave Mix Design Criteria
Requirements
NMAS 9.5 mm
NMAS 25 mm
Va target
VMA (min)

4.00%
15.00%

VFA target

4.00%
12.00%

65-75

ESAL (20 years)

6

>3x10 and <30x10

Ndes (gyration)

6

65-75
6

>3x10 and <30x10

100

6

100

The common point for all SP mix designs is the anticipated traffic loading. Both the aggregate
and mixture volumetric properties change with amount of ESALs. It is also important to
understand that the numbers of ESALs are based upon the expected traffic on the design lane
over 20 - year period. The actual roadway design life is most probably different than the design
lane assumptions. Table 3.8 shows the mix design calculations for all the mixes performed
during the tire/pavement interaction noise study.
Table 3.8 Mix designs and basic mixture-volumetric values
Mix Design Type
Superpave
SMA
OGFC
NMAS
9.5 mm
12.5 mm
25 mm
12.5 mm
4.75 mm
Gmm

2.648

2.583

2.624

2.587

2.497

Gmb

2.542

2.480

2.519

2.484

2.122

Gb

1.023

1.023

1.023

1.023

1.023

Gse

2.876

2.819

2.840

2.867

2.912

Gsb

2.849

2.747

2.802

2.819

2.664

Ps

95.24

94.80

95.37

94.00

91.00

Pb

4.76

5.20

4.63

6.00

9.00

Pba

0.34

0.95

0.49

0.61

3.27

Pbe

4.44

4.30

4.16

5.43

6.02

P200

0.05

0.04

0.03

0.04

0.03

P200/Pbe
VMA
VFA

1.13
15.02
73.35

0.93
14.41
72.28

0.72
14.26
71.94

0.74
17.19
76.72

NA
NA
NA

Va (%)

4.00

4.00

4.00

4.00

15.02

25

It is important to note the parameters provided in Tables 3.6 and 3.7 are only for Superpave mix
designs. AASHTO specifies the dust to binder ratio (P200/Pbe) for mix designs as 0.6-1.2
(AASHTO M-323). Moreover, VMA for SMA mix design should not be less than 17 and voids
in coarse aggregate of the compacted mix (VCAMIX) should be less than dry-rodded voids in
coarse aggregates (VCADRC) of the coarse fraction (AASHTO R-46 and AASHTO M-325).
These conditions are satisfied for SMA mix design.
Figure 3.8 demonstrates the job mix formula (JMF) for control mix design. The JMFs for other
type of mix designs are provided in the Appendix A. Since all the aggregate sources were graded
to each sieve size, pit number and percent values on the right hand side of the JMF are either
blank or assigned as not available (NA). Furthermore, JMF were prepared according to the
original virgin binder with PG 58-28. Although the specific gravities (Gb) of the binders used
were so close to each other, volumetric parameters were checked by using new binder Gb. All of
the values conformed to the standards.
All specimens, while applying the mix design procedures, were prepared by using gyratory
compactor which was capable of recording the properties at initial, design and maximum
gyrations. The volumetric properties measured on prepared specimens. The same process was
performed for each mix design. Once the gradation and optimum binder content was obtained for
each mix design type, the specimens were produced by using presbox shear compactor (a.k.a
slab compactor). The preparation of the specimens by using slab compactor is explained under
section 3.3.3.

26

JOB MIX FORMULA
HMA FIELD COMMUNICATION
CONTROL SECTION
CONTROL MIX DESIGN

JOB NO.
0.01

PROJECT ENGINEER
Salih KOCAK

DATE EFFECTIVE
07/21/2009

MIXTURE TYPE
MIX DESIGN NO.
PLANT LOCATION
SUPERPAVE
NMAS 12.5
EAST LANSING
ANGULARITY
% AIR VOIDS
VMA
VFA
COMP.TEMP
MIX. TEMP
47.0
4.00
14.41
72.28
138 C
149 C
Gmm
Gmb
Gb
Gse
Gsb
P200/Pbe
% AIR VOIDS
2.583
2.480
1.023 2.819 2.747
0.93
4.00
MIX/AGG. GRADATION, %
MIX/AGG. PROPORTION, %
ITEM

PERCENT

MATERIAL/PRODUCER

PIT NO.

ASPHALT,%

5.20%

SPARTAN & RIETH-RILEY

N/A

P 1-1/2" (37.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1" (25.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 3/4" (19.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1/2" (12.5 mm)

96.95%

SPARTAN & RIETH-RILEY

N/A

P 3/8" (9.5 mm)

89.95%

SPARTAN & RIETH-RILEY

N/A

P No. 4 (4.75 mm)

71.95%

SPARTAN & RIETH-RILEY

N/A

P No. 8 (2.36 mm)

52.95%

SPARTAN & RIETH-RILEY

N/A

P No. 16 (1.18 mm)

34.95%

SPARTAN & RIETH-RILEY

N/A

P No. 30 (600 µm)

16.90%

SPARTAN & RIETH-RILEY

N/A

P No. 50 (300 µm)

12.30%

RECLAIMED

NO RAP

P No. 100 (150 µm)

7.60%

Supplier: MTE

P No. 200 (75 µm)

4.00%

ASPHALT
BINDER

CRUSHED 1 FACE

100%

PRODUCER LOCATION: MSU-CEE-AACL

CRUSHED 2 FACES

100%

REGULAR TESTING

GRADE: PG 58-28

Figure 3.8 Job mix formula for Superpave mix design NMAS 12.5 mm (control mix)
3.3.3 Preparation of Specimens for |E*| and Laboratory Noise Testing Using TIPANOS
Specimens for both |E*| and TIPANOS testing were compacted by using slab compactor. The
advantage of using slab compactor is to be able to obtain three (3) 100 mm diameter 150 mm tall
27

cylindrical HMA replicates from a single slab. This reduced the preparation time, labor and the
amount of the material wasted. Table 3.9 shows the material test matrix with corresponding air
voids, aging condition and control parameter for each slab.
Table 3.9 Material notation and parameters controlled
Gradation
Aging
Air
Slab
Control Parameter
Sample Code
(NMAS)
Condition Voids *
No
#1
125SA4
12.5 mm
4 hours
7.35% Control Mixture
#2
125SA48
12.5 mm
48 hours
6.98% Aging Performed
#3
125SA4PM
12.5 mm
4 hours
7.38% PM** Binder
#4
125SA4CR5
12.5 mm
4 hours
7.20% 5% CRM*** Binder.
#5
125SA4CR10
12.5 mm
4 hours
6.92% 10% CRM Binder
#6
125SA4LW
12.5 mm
4 hours
7.05% Lightweight Aggregate
#7
095SA4F
9.5 mm
4 hours
6.94% Fine Gradation
#8
25SA4C
25.0 mm
4 hours
6.74% Coarse Gradation
#9
0475OA4
4.75 mm
4 hours
15.87% OGFC Mix Design
#10 0475OA4CR10
4.75 mm
4 hours
13.20% OGFC-10%CRM Binder
#11 125SMAA4
12.5 mm
4 hours
4.90% SMA Mix Design
#12 125SMAA4CR10 12.5 mm
4 hours
4.32% SMA-10% CRM Binder
*Air void content is the average of three replicates for each slab, **PM- Polymer modified,
***CRM- Crumb rubber modified.

Total twelve slabs were prepared for this research. Each slab yielded three (3) 100 mm diameter,
150 mm tall replicates, which resulted in total 36 specimens. First slab is the control mixture
with NMAS of 12.5 mm (125SA4 in Table 3.9). Second slab is prepared by using the same
Superpave mix design however the binder was aged for 48 more hours (125SA48 in Table 3.9).
Similar to the second one, the other consecutive three slabs were prepared by just altering the
binder types. They include polymer modified (125SA4PM in Table 3.9), 5% and 10% crumb
rubber modified binders (125SA4CR5, 125SA4CR10 in Table 3.9), respectively. The last SP12.5 mm slab is prepared by changing the aggregate type. Instead of gravel, it includes light
weight aggregates (125SA4LW in Table 3.9). Slab #7 is prepared by SP mix design with NMAS

28

of 9.5 mm (095SA4F in Table 3.9) and slab#8 with NMAS of 25 mm (25SA4C in Table 3.9).
The following two slabs are constructed according to OGFC mix design procedure with original
and 10% CRM binders (0475OA4, 0475OA4CR10), respectively. The last two slabs follow the
SMA mix design procedure. Slab#11 includes original binder (125SMAA4) whereas slab#12 has
10% CRM binder (125SMAA4CR10).
Figure 3.9 illustrates the schematic representation of the material matrix tested during the
research.

29

Mix Design Type
Superpave Mix Design
(SMD)

Stone Matrix Asphalt
(SMA)
Virgin Binder PG 58-28
28
4 hours Aging

NMAS 9.5 mm

NMAS 25 mm
Virgin Binder PG 58
58-28
4 Hours Aging

5 % Crumb Rubber
Modfied Binder
4hours Aging

Virgin Binder PG 58-28
4 Hours Aging

Open Graded Friction
Course (OGFC)

Virgin Binder PG 58
58-28
4 Hours Aging

5 % Crumb Rubber
Modfied Binder
4 Hours Aging

NMAS 12.5 mm
Virgin PG 58
58-28
Virgin Binder PG 58-28

4 Hours Aging

4 Hours Aging

Light Weight Aggregate
Virgin Binder PG 58-28

Polymer Modified Binder

48 Hours Aging

4 Hours Aging

10 % Crumb Rubber
Modfied Binder

5 % Crumb Rubber
Modfied Binder
4 Hours Aging

4 Hours Aging

Figure 3.9 Flow chart showing the material matrix
30

CHAPTER 4
LABORATORY INVESTIGATIONS AND INSTRUMENTATION
Laboratory investigations included three different measurements on the specimens during and
after the specimen preparation periods. They were aggregate consensus properties and HMA
volumetric measurements/calculations, tire pavement interaction noise measurement using
TIPANOS and dynamic modulus (|E*|) measurement by means of asphalt mixture performance
tester (AMPT). Aggregate properties can be found in chapter 3 and on the JMF tables provided
in the Appendix A. Table 4.1 provides the gradation and volumetric properties for the laboratory
prepared specimens.
Table 4.1 Properties of the laboratory prepared specimens
Design Type
SP- Control

Sample Code

Va

VMA

VFA

Cu

Cc

Pb

7.35%

17.08 57.0% 14.27 1.35 5.20%

SP- Aged

125SA4
125SA48

6.98%

16.75 58.3% 14.27 1.35 5.20%

SP- PM Binder

125SA4PM

7.38%

17.11 56.9% 14.27 1.35 5.20%

SP- CRM 5%

125SA4CR5

7.20%

16.95 57.5% 14.27 1.35 5.20%

SP- CRM 10%

125SA4CR10

6.92%

16.68 58.6% 14.27 1.35 5.20%

SP- 9.5 mm

095SA4F

6.94%

14.63 52.5% 17.64 1.54 4.76%

SP- 25 mm

25SA4C

6.74%

14.88 54.8% 46.28 1.38 4.63%

OGFC-Virgin Bin.

0475OA4

15.87%

27.08 42.0%

2.73 1.09 9.00%

OGFC-CRM 10%

0475OA4CR10

13.19%

24.88 47.1%

2.73 1.09 9.00%

SMA-Virgin Bind.

125SMAA4

4.90%

14.77 66.8% 29.40 8.53 6.00%

SMA-CRM 10%

125SMAA4CR10

4.32%

14.25 69.6% 29.40 8.53 6.00%

4.1 TIRE PAVEMENT INTERACTION NOISE MEASUREMENTS
To simulate the tire/pavement interaction noise generation and the measurement on laboratory
size specimens, tire/pavement noise simulator (TIPANOS) was developed (Figure 4.2). The
simulator is explained in detail in the subsequent parts of the chapter.Tire pavement interaction
noise measurement was the first test performed by using non destructive laboratory measurement
31

device TIPANOS. Tests were performed at room temperature. Each design type had three
replicates and the resulting sound pressure was recorded as the average of all three
measurements. Each specimen was tested three times for repeatability concerns and satisfactory
results were obtained. Coefficient of variation (COV) of repeatability test was around 4.2%
overall. The results given in linear SPL scale were then converted into A-weighted scale. Table
4.2 demonstrates the averaged and A-scale converted SPL.
o

Table 4.2 SPL at 250 Hz and 23 C
Sample Code

125SA4

125SA48

125SA4PM

125SA4CR5

SPL dB(A)

91.4

91.8

89.8

90.7

Sample Code

125SA4CR10

095SA4F

125SA4LW

25SA4C

SPL dB(A)

89.1

89.8

92.0

94.0

Sample Code

0475OA4

SPL dB(A)

86.7

0475OA4CR10 125SMAA4 125SMAA4CR10
85.2

92.9

90.0

The peak SPL at interstate roads is experienced at around 1000 Hz. (Sandberg 2003). However,
since TIPANOS tire and specimen’s sizes are smaller, the peak SPL in laboratory conditions was
acquired at 250 Hz. Figure 4.1 shows the SPL at 1/3 octave frequency band.

Sound Pressure Level

100
90
80
70
60
50
40
30

125SA4
125SA48
125SA4PM
125SA4CR5
125SA4CR10
125SA4LW
095SA4F
25SA4C
0475OA4
0475OA4CR10
125SMAA4
125SMAA4CR10
40 63 100 160 250 400 630 1k 1.6k 2.5k 4k 6.3k 10k 16k
25

1/3 Octave Band Frequency (Hz)
Figure 4.1 SPL versus 1/3 octave frequency of the specimens.
32

Moreover, the speed of the TIPANOS can be calculated at its highest rotation. For this
calculation, a tachometer was used to record the rotational speed of the tire. It was measured as
810 revolutions per minute (rpm). Equations 4.1 through 4.5 show the computation steps for the
speed of TIPANOS.

‫ ݓ‬ൌ 810 ‫݉݌ݎ‬

4.1

ܿ ൌ2ൈ ߨ ൈ‫ ݎ‬ൌ2 ൈ ߨ ൈ2ൌ4ߨ
‫ ݔ‬ൌ 810 ൈ 4 ߨ ൎ 10179

௜௡௖௛

ଵ

଺ଷଷ଺଴

4.3

௠௜௡௨௧௘

ܺ ൌ 60 ݉݅݊.ൈ 10179 ൌ 610740
ܺ ൌ 610740 ൈ

4.2

ൎ 9.64

௜௡௖௛

௛௢௨௥
௠௜௟௘

௛௢௨௥

4.4
4.5

4.1.1 Tire Pavement Noise Measurement Simulator (TIPANOS)
The set-up consists of a treaded tire rotating over the cored specimens. Dimensions of the
specimens (100 mm diameter and 150 mm tall) used in TIPANOS have been chosen
intentionally so that they can be used in dynamic modulus (|E*|) tests as well. Since the sound
measurement test performed by TIPANOS is non-destructive, the same samples can be used in
dynamic modulus tests conveniently by eliminating the cost, time and effort necessary for
preparation of extra samples. As shown in the Figure 4.2, TIPANOS has two specimen holders to
maintain the cored and sawed cylindrical specimens in place. They have 101 mm interior
diameter and 15 mm lip thickness. To be able to simulate the sound generated on the roadways
as closely as possible, a small scale threaded tire was used. The other two unthreaded tires
support the specimen from bending at the bottom. The diameter of the tires is 100 mm and the

33

width is 50 mm. Treaded tire has 2.5 mm width and 2.5 mm depth grooves on the surface and
they are spaced at 6 mm. Both specimen holders and tires are kept in place by means of metal
rods and these rods are supported by other metal arms attached to the main frame. Rotation is
provided by an electric motor connected to the main frame via belt. Distance between motor and
the pulley is 250 mm. The speed of the rotation can be easily adjusted to any level by using the
tunable switch available on the motor. All system sits on a metal block having dimensions 50 cm
length, 50 cm width and 5 cm thickness. To record the real time sound generated between tire
and the specimen, an intensity microphone, dynamic signal analyzer and a computer are utilized.
Microphone is pointed to the interface between threaded tire and the specimen from 250 mm
distance and hold in place with a metal bar. Dynamic signal analyzer, which is connected to the
microphone in one end and to the computational station on the other end, is capable of
measuring, analyzing and recording the sound pressure level (SPL) values to the computer
between 0 and 20000 Hz.

(a)

(b)

Figure 4.2 (a) TIPANOS set-up, (b) TIPANOS plan view
34

4.1.2 Dynamic Signal Analyzer (DP-QUATTRO)
Data physics Quattro Dynamic Signal Analyzer (simply QUATTRO) hardware and SignalCalc
Ace 240 software were used for recording and analyzing the real time sound data. Both the
analyzer and the software are commercially available. Data acquisition system saves the data to a
computer via USB connection. The capability of the software to record sound pressure level data
is determined by the capability of microphone. After recording is performed, the data can be
manipulated by converting other weighting scales. Traffic noise is generally analyzed and
published according to A-weighing scheme. The A-scale converted data for all recorded
frequencies can be reached in the Appendix B. Figure 4.3 illustrates the QUAATRO data
acquisition system.

Figure 4.3 QUATTRO data acquisition hardware (www.dataphysics.com)
4.1.3 Free Field Microphone (G.R.A.S- Type 46AE)
GRAS Type 46 AE is the combination of ½-inch pre-polarized free field microphone 40 AE and
the same size constant current power type 26CA. It is highly sensitive to sound pressure and can

35

cover a frequency range of 3.15 Hz to 20000 Hz with its large dynamic range. As a free field
microphone system, it can measure sound pressure and compensate for the influence of its
presence in the sound field. One of the most important properties of the free field microphones is
the need to point them towards the sound source. The incidence to the source other than 0 degree
angle can affect the results, the effect is drastic at higher frequencies. Moreover, the presence of
the microphone in the sound field has a small impact at low frequencies, however the effects of
reflections and diffractions at higher frequencies can cause and increase in sound pressure level
measurements. The connection between QUATTRO and amplifier is provided by using BNC coaxial connector. Figure 4.4 shows GRAS type 46 AE free field microphone unit.

Figure 4.4 GRAS type 40 AE free field microphone and CCP 26 CA preamplifier combination
(http://www.gras.dk/)
4.1.4 Sound Calibrator (LD-CAL200)
Larson Davis CAL 200 (Figure 4.5) is used as a sound calibrator for ½-inch free field
microphone. It has a calibration frequency of 1000 Hz and it provides an output level of 94 and
114 dB with National Institute of Standards and Technology (NIST). Figure 4.5 demonstrates the
sound calibrator.

36

Figure 4.5 Larson Davis L 200 portable sound calibrator (www.larsondavis.com)
4.2 DYNAMIC MODULUS MEASUREMENTS
Dynamic modulus measurement was one of the most important parts of this study. Since the
noise measurements are time consuming and expensive compared to dynamic modulus (|E*|)
measurements, any relation between |E*| and the SPL could help the researchers and the asphalt
producers understand the noise level that specific pavement causes.
Since noise measurement by using TIPANOS was non-destructive, the same specimens were
tested in AMPT to obtain the |E*| master curve. The specimens were tested at 4, 21, 37 and 54
o

C temperatures and 0.1, 0.5, 1, 5, 10 and 25 Hz frequencies. After the |E*| and phase angle (δ)

values were averaged for all three replicates, |E*| master curves were constructed. Sigmoidal and
Gaussian curves were fitted for |E*| and phase angle, respectively. Phase angle fit was performed
by using commercially available curve fitting software “Curve Expert”. AASHTO TP 62 –
Determining Dynamic Modulus of Hot Mix Asphalt (HMA) was the applied standard for |E*|
measurements. It should be recalled that complex modulus (E*) is actually a complex number

37

with real and imaginary parts (i.e., E*=E’+E”i). Real part (E’=|E*|cosδ) is known as storage or
elastic modulus, whereas the imaginary part (E” =|E*|sinδ) is called as loss or viscous modulus.
The magnitude of complex modulus is defined as dynamic modulus and denoted by |E*|. The
phase angle “δ” in degrees is also measured for each and every frequency and temperature levels.
Since the asphalt mixtures exhibit linear visco-elastic behavior (at small deformations), when the
stress is applied to the specimen, strain will be obtained after some time lag. This lag between
applied stress and resultant strain is the phase angle and it ranges from 0 to 90 degrees. When
phase angle is 0°, the material shows only elastic behavior. As δ increases, the viscous behavior
of the material increases as well and when δ = 90°, the material behaves purely viscous. Table
4.3 shows the averaging process for sample 125SA4CR5 replicates.

38

Table 4.3 Averaging dynamic modulus and phase angles for sample 125SA4CR5
Measured E* (MPa)
Replicate #1
Specimen
Code
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5

Temp
(°C)
4.0
4.0
4.0
4.0
4.0
4.0

f
(Hz)
25
10
5
1
0.5
0.1

125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5

21.0
21.0
21.0
21.0
21.0
21.0
37.0

125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5
125SA4CR5

Replicate #2

Replicate #3

|E*|

|E*|

|E*|

11614
10481
9593
7488
6546
4451

δ
10.89
12.01
13.02
15.68
17.13
21.68

25
10
5
1
0.5
0.1
25

4965
3919
3243
1975
1572
863.3
1560

37.0
37.0
37.0
37.0
37.0
54.0
54.0
54.0
54.0

10
5
1
0.5
0.1
25
10
5
1

125SA4CR5

54.0

125SA4CR5

54.0

Avg.
|E*|
12370
11101
10104
7873
6909
4770

Avg.
δ
10.6
11.8
12.7
15.4
16.9
21.3

12486
11137
10083
7831
6876
4837

δ
11.09
12.12
13.04
15.74
17.12
21.49

13010
11684
10635
8300
7304
5023

δ
9.83
11.22
12.18
14.83
16.33
20.58

22.89
25.37
27.24
31.81
33.26
36.68
37.13

5286
4162
3413
2036
1601
863.7
1555

23.67
26.15
27.97
32.44
33.59
36.47
36.68

5496
4364
3625
2214
1775
983.5
1713

22.62
25.1
26.88
31.3
32.53
35.81
36.37

5249
4148
3427
2075
1649
904
1609

23.1
25.5
27.4
31.9
33.1
36.3
36.7

1087
819.8
402.5
301.4
151.9
442.2
288.9
211.1
106

38.6
39.04
39.65
38.7
36.54
38.82
37.47
35.85
32.5

1087
817.6
402.5
299.5
150.3
412.7
271.6
196.1
97.9

37.91
38.34
39.03
38.12
36.28
39.16
37.77
36.35
32.97

1215
920.4
464.5
347.7
175.2
409.7
267.9
196.9
98.2

37.53
37.92
38.41
37.41
35.51
38.24
37.03
35.53
32.1

1130
853
423
316
159
422
276
201
101

38.0
38.4
39.0
38.1
36.1
38.7
37.4
35.9
32.5

0.5

83.7

30.44

77.9

30.85

80.3

30.08

81

30.5

0.1

51.9

26.71

38.6

25.29

51.1

26.65

47

26.2

(MPa)

(MPa)

(MPa)

In order to obtain quality data, data quality indicators (DQI) were maintained in the limits
(AASHTO PP-62). Only high quality data is used in the calculations. Load standard error,
deformation standard error, deformation uniformity and phase uniformity were the parameters
checked each and every temperature-frequency combinations. Table 4.4 shows the basic
calculations to obtain the master curve for sample 125SA4CR5.

39

T
(C)
4.0
4.0
4.0
4.0
4.0
4.0
21.0
21.0
21.0
21.0
21.0
21.0
37.0
37.0
37.0
37.0
37.0
37.0
54.0
54.0
54.0
54.0
54.0
54.0

f
(Hz)
25
10
5
1
0.5
0.1
25
10
5
1
0.5
0.1
25
10
5
1
0.5
0.1
25
10
5
1
0.5
0.1

Table 4.4 Master curve construction for sample 125SA4CR5
Phase
Sigmoid
fR
Log fR
Avg. |E*| Avg.
Angle
Fit,
(Hz)
(Hz)
(MPa)
log(aT)
Fit
|E*| MPa
δ
3.2E+03 3.5E+00 1.1E+01
12187
12370
10.6
2.10
11022
1.3E+03 3.1E+00 1.1E+01
11101
11.8
2.10
6.3E+02 2.8E+00 1.2E+01
10096
10104
12.7
2.10
1.3E+02 2.1E+00 1.6E+01
7868
7873
15.4
2.10
6912
6.3E+01 1.8E+00 1.8E+01
6909
16.9
2.10
1.3E+01 1.1E+00 2.3E+01
4828
4770
21.3
2.10
5333
1.9E+01 1.3E+00 2.2E+01
5249
23.1
-0.12
4236
7.6E+00 8.8E-01 2.5E+01
4148
25.5
-0.12
3490
3.8E+00 5.8E-01 2.7E+01
3427
27.4
-0.12
2088
7.6E-01 -1.2E-01 3.3E+01
2075
31.9
-0.12
3.8E-01 -4.2E-01 3.5E+01
1630
1649
33.1
-0.12
871
7.6E-02 -1.1E+00 3.7E+01
904
36.3
-0.12
3.6E-01 -4.5E-01 3.5E+01
1591
1609
36.7
-1.85
1120
1.4E-01 -8.5E-01 3.7E+01
1130
38.0
-1.85
7.1E-02 -1.1E+00 3.8E+01
848
853
38.4
-1.85
1.4E-02 -1.8E+00 3.8E+01
430
423
39.0
-1.85
319
7.1E-03 -2.1E+00 3.6E+01
316
38.1
-1.85
1.4E-03 -2.8E+00 3.1E+01
161
159
36.1
-1.85
408
1.3E-02 -1.9E+00 3.7E+01
422
38.7
-3.30
275
5.0E-03 -2.3E+00 3.6E+01
276
37.4
-3.30
204
2.5E-03 -2.6E+00 3.3E+01
201
35.9
-3.30
105
5.0E-04 -3.3E+00 2.4E+01
101
32.5
-3.30
2.5E-04 -3.6E+00 1.9E+01
81
81
30.5
-3.30
46
5.0E-05 -4.3E+00 1.5E+00
47
26.2
-3.30
Error sum
40

2

Error
4.1903E-05
9.497E-06
1.1641E-07
8.1833E-08
5.2095E-08
2.7369E-05
4.7154E-05
8.1581E-05
6.2791E-05
7.355E-06
2.6236E-05
0.00025502
2.565E-05
1.3096E-05
6.4792E-06
4.9656E-05
1.5752E-05
2.2402E-05
0.00020199
3.466E-06
4.4047E-05
0.00040226
7.2687E-08
0.0001827
0.00152672

Figure 4. 6 and Figure 4.7 show the master curve for dynamic modulus and Gaussian fitting for
phase angle after the data in Table 4.4 is analyzed and studied.

Sigmoidal Master Curve

100000

|E*| (MPa)

10000
1000
E* Average
(Mpa)

100
10
1.E-06

1.E-04

1.E-02
1.E+00
1.E+02
Reduced Frequency (logfR) (Hz)

1.E+04

Figure 4.6 Sigmoidal master curve for sample 125SA4CR5

Phase Angle Fitting
Phase angle (degree)

phase angle (measured)

50
phase angle (fitted)

40
30
20
10
0
-6

-4

-2
0
2
Reduced Frequency (logfR) (Hz)

4

Figure 4.7 Phase angle fitting for sample 125SA4CR5
Master curve for |E*| is constructed by using the data provided in Table 4.4. AASHTO PP 61 Developing Dynamic Modulus master curves for Hot Mix Asphalt (HMA) Using the Asphalt
Mixture Performance Tester (AMPT) and AASHTO PP 62- Developing Dynamic Modulus
master curves for Hot Mix Asphalt (HMA) were the two standards used to construct the master

41

curves for all samples. Equations 4.6 and 4.7 demonstrate the formulations for calculations used
during construction of master curves.

log ( a ( T )) 4.6 a
=

1

T2 + a

2

T + a

4.6
3

α

log(| Ε* |) = δ +
1+ e







β + γ log f

R

4.7







where;
a1, a2, a3 are the shift coefficient factors,

α,β, γ, δ are sigmoid coefficients,
fR is reduced frequency given in Table 4.3
The values of the shift and sigmoid coefficients are provided in Table 4.5. These coefficients
were obtained by using the solver option available in excel spread sheets.
Table 4.5 Shift and sigmoid coefficients for sample 125SA4CR5
Shift Factor
Coefficients
Sigmoid
Coefficients

a1

a2

a3

6.85E-04

-1.48E-01

2.68E+00

20.0

α

β

γ

δ

3.55

1.01

0.49

0.76

|E*| (Mpa)

10000

Tref (C) aT at Tref
2.35E-06

125SA4
125SA48
125SA4PM
125SA4CR5
125SA4CR10
125SA4LW
095SA4F
25SA4C
0475OA4
0475OA4CR10
125SMAA4
125SMAA4CR10

1000

100
0.1

1

Frequency (Hz)

10

100

O

Figure 4.8 |E*| versus frequency for all samples at 21 C
42

125SA4
125SA48
125SA4PM
125SA4CR5
125SA4CR10
125SA4LW
095SA4F
25SA4C
0475OA4
0475OA4CR10
125SMAA4
125SMAA4CR10

Phase Angle (Degree)

45
40
35
30
25
20
15
0.1

1

Frequency (Hz)

10

100

O

Figure 4.9 Phase angle versus frequency for all samples at 21 C
O

Figure 4.8 and 4.9 illustrate |E*| and δ versus frequency change for all the samples at 21 C,
respectively.
4.2.1 Asphalt Mixture Performance Tester (IPC-AMPT)
Dynamic modulus tests were performed by using asphalt mixture performance tester (AMPT)
(previously known as Simple Performance Tester (SPT)). AMPT was developed after the
adaptation of Superpave mix design process almost two decades ago. First studies started in
1996 in University of Maryland at College Park (UMCP). Three years later National Cooperative
Highway Research Program (NCHRP) task C took the effort under project 9-19 “Superpave
Support and Performance Models Management” (NCHRP 2002). The definition of the simple
performance test is described in NCHRP (2002) report as follows:
“A test method(s) that accurately and reliable measures a mixture response characteristics or
parameter that is highly correlated to the occurrence of pavement distress (e.g., cracking and
rutting) over a diverse range of traffic and climatic condition.”
Being one of the oldest triaxial compression tests, dynamic modulus test includes the application
of haversine compressive stress to cylindrical test specimens in confined or unconfined

43

condition. The relationship between stress and strain of a linear viscoelastic material under
continuously applied sinusoidal loading is defined by a complex number which is called as
complex modulus and denote as E*.
In the NCHRP project 9-29 phase 2, the SPT was constructed having the properties investigated
and published in project 9-19. At the end of the research, small bottom loading, servo-hydraulic
equipment with testing chamber serving both as an environmental chamber and confining
pressure cell was constructed. Today’s IPC AMPT uses circulating conditioned air through the
test chamber to maintain the desired temperature. Figure 4.10 demonstrates the IPC’s AMPT.

Figure 4.10 IPC Global AMPT
4.2.2 Environmental Chamber
Since the dynamic modulus measurements were performed at 4 different temperatures, Russell
environmental chamber was used to control the temperature of the samples. Specimens were
conditioned in the environmental chamber according to AASHTO specifications.

44

CHAPTER 5
DATA ANALYSIS
5.1 CORRELATION OF INDIVIDUAL MATERIAL CHARACTERISTICS AND
SOUND PRESSURE LEVEL (SPL)
Aggregates and binder are the two main ingredients in asphalt pavements. Although the behavior
of aggregates is well established, the behavior of the binders is very complex and has not been
completely discovered yet. The same complexity of asphalt pavements is valid for their sound
absorption capacities as well. The investigation of the contribution of each component to the
noise level is crucial to understand and manipulate the sound absorption of pavements. The
effect of individual material characteristics on sound pressure level was studied throughout the
chapter. The analyses include the mutual correlations between SPL and linear visco-elastic
parameters, mixture volumetric properties and aggregate gradations of HMA specimens.
5.1.1 Relation between Linear Visco-Elastic Parameters of Asphalt and SPL
Asphalt mixtures have complex mechanical behavior which changes with temperature and rate of
loading. This behavior can be characterized in the linear visco-elastic domain by different
dynamic material functions such as dynamic (complex) modulus and phase angle. New
mechanistic- empirical approach for asphalt pavement design and analysis makes the use of
complex modulus test data essential. Any relation between linear visco-elastic parameters and
SPL would simplify the noise studies by saving time, budget and effort since SPL measurements
of the pavements are more complex and expensive and time consuming compared to dynamic
modulus test.
In order to investigate the relation between linear visco-elastic parameters and SPL, novel
laboratory tire pavement noise measurement system (TIPANOS) was constructed. TIPANOS

45

focuses on the noise due to asphalt material characteristics by overshadowing the impact of
surface texture.
o

Dynamic modulus test was run at 4, 21, 37 and 54 C with 25, 10, 5, 1, 0.5and 0.1 hertz (Hz)
frequencies at each temperature level. The correlations for each combination have been studied
for a trend and they can be reached in the Appendix C. However, the relation between linear
0

visco-elastic parameters and SPL is presented at 23 C and 212 Hz which are the temperature and
frequency of the TIPANOS measurements according to the load impulse approach. It is
important to calculate the frequency of impact when a point on the tire makes the contact with
the pavement surface. Since the frequency that pavement experience is important to compute the
dynamic modulus and phase angle, the equation 5.6 shows the calculation of the impact
frequency on pavement. The revolution of TIPANOS is obtained by using a digital hand
tachometer. At the highest speed, the reading was 810 revolutions per meter (rpm). Moreover,
the contact length between the tire and the asphalt specimen surface was measured around 10
mm. As explained in the chapter 4, the diameter of the tire is 100 mm (with r =0.05m.). The
following computations show how to calculate the frequency at which TIPANOS runs at its
maximum speed.

∆‫ ݔ‬ൌ ∆‫ ݐ‬ൈ ‫ݒ‬
‫ ݒ‬ൌ‫ݓ‬ൈ‫ݎ‬
‫ ݓ‬ൌ 810

௥௘௩௢௟௨௧௜௢௡௦
௠௜௡௨௧௘

5.1
5.2

ൈ ሺ2ߨ ⁄ ‫݊݋݅ݐݑ݈݋ݒ݁ݎ‬ሻ/ሺ60 ‫ݏ݀݊݋ܿ݁ݏ‬ሻ

‫ ݓ‬ൎ 84.82 ‫ݏ݀݊݋ܿ݁ݏ⁄ݏ݊ܽ݅݀ܽݎ‬

5.3
5.4

∆‫ ݐ‬ൌ ∆‫/ݔ‬ሺ‫ ݓ‬ൈ ‫ݎ‬ሻ ൌ ሺ10 ሻ/ሺ84.82 ൈ 50ሻ ൌ 0.0023 ‫ ݏ݀݊݋ܿ݁ݏ‬ሺ‫ܿ݁ݏ‬ሻ 5.5
46

݂ ൌ 1/ሺ2 ൈ ∆‫ݐ‬ሻ ൌ 1/ሺ2 ൈ 0.0023ሻ ൎ 212 ‫ݖܪ‬

5.6

where;
r = radius of the tire (mm)
w= angular velocity (rpm)
∆x= contact length of tire and HMA specimen (mm)
∆t= contact time (sec)
f= frequency (Hz), calculated by diving 2 ∆‫ ݐ‬since it is impulse frequency.
Table 5.1 summarizes the fundamental linear visco-elastic properties of the each HMA specimen
o

at 212 Hz frequency at 23 C as well as the sound pressure level (SPL) at 250 Hz and at the same
temperature.
o

Table 5.1 Specimens linear visco-elastic properties at 212 Hz and 23 C

Sample Code

|E*|
MPa

125SA4
5949.5
125SA48
6096.5
125SA4PM
5466.2
125SA4CR5
7456.4
125SA4CR10
9523.3
095SA4F
7816.2
25SA4C
8012.7
0475OA4
2056.3
0475OA4CR10
4306.2
125SMAA4
8386.6
125SMAA4CR10 11389.0

sin δ |E*| / sin δ
δ
Degree
MPa
23.85
21.68
25.32
17.27
13.92
23.95
23.78
30.22
17.03
23.47
14.76

0.404
0.369
0.428
0.297
0.241
0.406
0.403
0.503
0.293
0.398
0.255

14715.1
16501.4
12779.3
25113.8
39581.4
19253.3
19869.4
4085.0
14700.3
21054.2
44693.0

E'
MPa

E"
MPa

SPL
dB(A)

5441.6
5665.2
4940.9
7120.2
9243.6
7143.2
7332.3
1776.8
4117.3
7692.5
11013.0

2405.5
2252.4
2338.1
2213.8
2291.3
3173.1
3231.3
1035.1
1261.4
3340.6
2902.3

91.4
91.8
89.8
90.7
89.1
89.8
94.0
86.7
85.2
92.9
90.0

In the Figure 5.1, resultant sigmoidal curves for frequency-temperature combinations are shown
for all asphalt samples tested.

47

100000

1000
125SA4
125SA4PM
125SA4CR10
095SA4F
0475OA4
125SMAA4

100

10
1.E-05

1.E-03

125SA48
125SA4CR5
125SA4LW
25SA4C
0475OA4CR10
125SMAA4CR10

1.E-01
1.E+01
Reduced Freq. (Hz)

1.E+03

Figure 5.1 |E*| versus reduced frequency curves for specimens

SPL vs |E*|
94
SPL (dB(A))

|E*| (MPa)

10000

92
90
88
y = 0.0005x + 86.675
R² = 0.2492

86
84
2,000

4,000

6,000
8,000
|E*| (MPa)

10,000

Figure 5.2 Change in SPL with change in |E*|

48

12,000

SPL vs |E*|/sin δ
SPL (dB(A))

94
92
90
88

y = 4E-05x + 89.32
R² = 0.0312

86
84
0

10,000

20,000

30,000

|E*|/ sin δ (MPa)

40,000

50,000

Figure 5.3 Change in SPL with change in |E*|/sin δ

SPL vs δ
SPL (dB(A))

94
92
90
88

y = 0.0551x + 88.949
R² = 0.0118

86
84
14.0

16.0

18.0

20.0

22.0

δ

24.0

26.0

Figure 5.4 Change in SPL with change in δ

SPL vs E'
SPL (dB(A))

94
92
90
88
y = 0.0005x + 87.14
R² = 0.2057

86
84
0

2,000

4,000

6,000

8,000 10,000 12,000

E'
Figure 5.5 Change in SPL with change in E’

49

SPL vs E"

SPL (dB(A))

94
92
90
88

y = 0.0028x + 83.505
R² = 0.6631

86
84
500

1,000

1,500

2,000

2,500

3,000

3,500

E"
Figure 5.6 Change in SPL with change in E”
o

The correlations between SPL and linear visco-elastic parameters at 212 Hz and 23 C are shown
on the Figures 5.2 through 5.6. Figure 5.2 demonstrates the relation between SPL and dynamic
2

modulus |E*|. Positive correlation between the variables with R =0.2492 is achieved. However,
it should be observed that the relation becomes very weak if the lowest two values are ignored.
Hence, it can be concluded that the correlation between |E*| and SPL is almost negligible. The
same rationale applies for |E*|/sin δ, sin δ and E’ versus SPL relations where the relation is very
weak. The last correlation for linear visco-elastic parameters is performed between SPL and E”.
As explained and formulated in chapter 4, E” is the viscous, or loss, modulus. The relation
2

between E” and SPL is positive with R =0.6631. With an increase in E”, SPL values increases as
well. Another important finding is the sound absorption behavior of the asphalt mixture by
changing time and frequency. Although all the volumetric parameters are somehow correlated
positively with SPL at the mentioned temperature and frequency combination, the relation
demonstrates negative correlation at some other combinations. Thus, it is not always proper to
make generalization for the relation between the linear visco-elastic parameters and SPL. More
analyses have been performed at different temperature – frequency combinations for each
50

specimen to disclose the relationship and some of them are presented in the Appendix C.
It is important to declare that specimens prepared by using light weight aggregates become
outliers for the data analysis. They absorbed unrealistic amount of asphalt binder. Since they
yield the misinterpretation of the data when included, they are assigned as outliers for all data
analysis part. However, the measurements performed on light weight aggregate specimens are
presented in SPL measurements for comparison reasons.
5.1.2 Relation between Mixture Volumetric Parameters and SPL
Superpave is a volumetric mix design. There are certain values that need to be accomplished to
perform the mix design according to the design criteria of AASHTO standards. Since SMA is a
modified version of the Superpave, it also conforms to the volumetric mix design requirements.
Mixture volumetric parameters studied in this research are air void content (Va), voids filled with
asphalt (VFA), voids in mineral aggregates (VMA), binder content (Pb) along with theoretical
maximum specific gravity (Gmm) and compacted bulk specific gravity (Gmb) of a mixture. The
equations from 5.7 to 5.12 show how to calculate the volumetric properties for a mix.

‫ ݉݉ܩ‬ൌ

ܹ‫ ݏ‬൅ ܹܽܿ
ܸ‫ ݏ‬൅ ܸܽܿ

5.7

‫ ܾ݉ܩ‬ൌ

ܹ‫ ݏ‬൅ ܹܽܿ
ܸ‫ ݏ‬൅ ܸܽܿ ൅ ܸܽ

5.8

ܸܽ ൌ ൬1 െ

‫ܾ݉ܩ‬
൰ ൈ 100
‫݉݉ܩ‬

ܸ‫ ܣܯ‬ൌ 100 െ

‫ ܾ݉ܩ‬ൈ ܲ‫ݏ‬
‫ܾݏܩ‬

51

5.9
5.10

ܸ‫ ܣܨ‬ൌ 100 ൈ
ܾܲ ൌ

ܸ‫ ܣܯ‬െ ܸܽ
ܸ‫ܣܯ‬

ܹܽܿ
ܹ݉݅‫ݔ‬

5.11
5.12

where;
ܹ‫ = ݏ‬Weight of aggregates
ܹܽܿ = Weight of asphalt cement (binder)
ܹ݉݅‫ = ݔ‬Weight of the asphalt mixture
ܸ‫ = ݏ‬Volume of aggregates
ܸܽܿ = Volume of asphalt cement
ܸܽ = Volume of air voids
ܲ‫ = ݏ‬Aggregate content ሺ 1 െ ܾܲሻ
‫ = ܾݏܩ‬Bulk specific gravity of aggregates
Table 5.2 gives the volumetric properties of the asphalt specimens prepared for the research. It
illustrates the parameters for all Superpave, SMA and OGFC mix designs.
Table 5.2 Volumetric parameters of the specimens
Sample Code
125SA4
125SA48
125SA4PM
125SA4CR5
125SA4CR10
095SA4F
25SA4C
0475OA4
0475OA4CR10
125SMAA4
125SMAA4CR10

Va
7.35
6.98
7.38
7.20
6.92
6.94
6.74
15.87
13.19
4.90
4.32

Gmm
Gmb
Pb
VMA
VFA
17.08
57.01
2.583
2.393 0.0520
16.75
58.33
2.583
2.403 0.0520
17.11
56.87
2.583
2.392 0.0520
16.95
57.52
2.583
2.397 0.0520
16.68
58.62
2.583
2.405 0.0520
14.63
52.50
2.648
2.464 0.0476
14.88
54.77
2.634
2.457 0.0463
27.08
41.98
2.497
2.105 0.0900
24.88
47.05
2.497
2.168 0.0900
14.77
66.76
2.587
2.460 0.0600
14.25
69.64
2.587
2.475 0.0600

52

Figures 5.7 to 5.11 illustrate the relation between individual volumetric properties of the asphalt
mixtures with SPL. In figure 5.7, the correlation between air void content and sound pressure
level is analyzed. There is an inverse; however, strong enough relation between the variables. It
is the same for VMA versus SPL as well. This phenomenon can be explained with the travel of
the sound waves. When there are more voids on the surface and internal structure of the asphalt,
sound waves can freely move through them and be both absorbed and refracted. This results in
less sound wave reflection and thus less SPL. The same logic applies for Figure 5.9. There is a
mild relation between VFA and SPL. The more the voids are filled with asphalt, the more sound
pressure level increases.

SPL vs Va
SPL (dB(A))

94.0
92.0

y = -0.565x + 94.636
R² = 0.5837

90.0
88.0
86.0
84.0
4.0

6.0

8.0

10.0

12.0

14.0

16.0

Va
Figure 5.7 Change in SPL with change in air void content

SPL (dB(A))

94.0

SPL vs VMA
y = -0.48x + 98.638
R² = 0.6427

92.0
90.0
88.0
86.0
84.0

12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0

VMA
Figure 5.8 Change in SPL with change in voids in mineral aggregate
Figure 5.10 shows the relationship between Pb and SPL, where a negative correlation is visible.
53

Visco-elastic behavior of the binder and the mastic (binder + fine aggregate) may be the reason
for such relationship. As the amount of binder is increased, the asphalt mixture becomes more in
the viscous behavior side. Materials showing viscous behavior are quieter than the materials with
elastic performance. However, it is noted that increase in asphalt content may yield the decrease
in overall void content which is inversely correlated with SPL as well. There should be an
optimum content for binder in which the highest sound reduction might be achieved. In order to
discover this type of relation, there is a need for more tests covering a large enough range. Since
the individual correlations with SPL do not contribute to the overall knowledge because of the
material characteristics interactions, there is a need for multivariate data analysis and it will be
presented in the section 5.2.

SPL vs VFA
SPL (dB(A))

94.0
92.0
90.0
88.0

y = 0.1974x + 78.98
R² = 0.3651

86.0
84.0
40.0

45.0

50.0

55.0

60.0

65.0

70.0

VFA
Figure 5.9 Change in SPL with change in voids filled with asphalt

SPL vs Pb

SPL (dB(A))

94.0

y = -128.94x + 97.792
R² = 0.6312

92.0
90.0
88.0
86.0
84.0
0.04

0.05

0.06

0.07

0.08

0.09

0.1

Pb
Figure 5.10 Change in SPL with change in binder content
54

SPL vs Gmb

SPL vs Gmm
95.0

93.0
91.0
89.0

y = 42.224x - 18.753

87.0

R² = 0.5904

85.0
2.450

SPL (dB(A))

SPL (dB(A))

95.0

2.500

2.550

2.600

2.650

93.0
91.0

y = 16.64x + 50.615
R² = 0.6433

89.0
87.0
85.0
2.100

2.700

2.200

2.300

2.400

2.500

Gmm
Gmb
Figure 5.11 Change in SPL with change in Gmm and Gmb
For the figure 5.11, the relations are almost negligible if the lowest two values are ignored.
Although the trend is positive for correlation between specific gravities and SPL, more data
points are required to scrutinize and establish the relation acceptably.
5.1.3 Relation between Gradation Parameters, Material Types and SPL
Gradation of the aggregates is another important parameter analyzed for sound
absorption\reflection capacities of the asphalt pavements. Gradation can be characterized
according to the nominal maximum aggregate size (NMAS), coefficient of curvature (Cc) and
coefficient of uniformity (Cu). The effect of NMAS, mix design type, binder aging, binder type,
aggregate type on SPL can indirectly be seen in the figure 5.12. The decrease in NMAS for
Superpave (SP) mix design results in the reduction on SPL as well. The drop in SPL between
25SA4C and 095SA4F specimens is approximately 4.2 dB (A). By choosing NMAS 12.5 mm
instead of 25 mm for SP mix design yields around 3dB (A) more noise damping.

55

SPL vs Mix Design and Material Type
96

SPL dB(A)

94
92
90
88
86
84

Mix Design and Material Type
o

Figure 5.12 Peak SPL values at 250 Hz and 23 C
Hence, as the HMA aggregate gradation gets finer, the sound pressure level decreases. The other
obvious information that can be obtained from the figure is the impact of binder type on SPL.
Modified binders work better for sound reduction on pavements. As the amount of crumb rubber
increases in the binder, the SPL level decreases. Another deduction that can be seen is the effect
of binder aging on SPL. The only difference between specimens 125SA4 and 125SA48 is the
aging of the binder used. The variation is 0.4 dB (A) higher on aged specimen side. This can be
explained by the visco-elastic behavior. Since the aged binder gets stiffer, it tends to become
closer on elastic behavior side. It is known that elastic materials are not as good as viscous
material on sound damping. The last conclusion is drawn from the figure is the type of material
mix design. OGFC mix design has the best sound absorption capacity compared to SP and SMA.
This can be attributed to the high air void content and connected void structure of the OGFC
mixes.

56

Table 5.3 Gradation parameters and control parameters of the asphalt mixtures
Gradation Aging
Air
Cu
Cc
(NMAS)
Condition Voids *
Control Parameter
14.27 1.35 Control Mixture
12.5 mm
4 hours
7.35%
125SA4
14.27 1.35 Aging Performed
12.5 mm
48 hours
6.98%
125SA48
14.27 1.35 PM*** Binder
12.5 mm
4 hours
7.38%
125SA4PM
14.27 1.35 5% CRM** Binder.
12.5 mm
4 hours
7.20%
125SA4CR5
14.27 1.35 10% CRM Binder
12.5 mm
4 hours
6.92%
125SA4CR10
14.27 1.35 Lightweight Aggregate
12.5 mm
4 hours
7.05%
125SA4LW
17.64 1.54 Fine Gradation
9.5 mm
4 hours
6.94%
095SA4F
46.28 1.38 Coarse Gradation
25.0 mm
4 hours
6.74%
25SA4C
2.73 1.09 OGFC Mix Design
4.75 mm
4 hours
15.87%
0475OA4
2.73 1.09 OGFC-10%CRM Binder
4.75 mm
4 hours
13.20%
0475OA4CR10
29.40 8.53 SMA Mix Design
12.5 mm
4 hours
4.90%
125SMAA4
29.40 8.53 SMA-10% CRM Binder
4 hours
4.32%
125SMAA4CR10 12.5 mm
*Air voids values are the average of three (3) replicates. CRM**-Crumb rubber modified.
PM***-Polymer modified
Sample Code

SPL vs Cu

SPL (dB(A))

94.0
92.0
90.0
88.0

y = -0.004x2 + 0.3519x + 85.643
R² = 0.7375

86.0
84.0
0.0

10.0

20.0

30.0

40.0

50.0

Cu
Figure 5.13 Change in SPL with change in Cu
The coefficient of uniformity is a measure of how well or poorly the aggregates are sorted. It can
be calculated by using equation 5.13. Figure 5.13 shows the correlation between SPL and Cu.

‫ ݑܥ‬ൌ ‫଺ܦ‬଴ /‫ܦ‬ଵ଴

5.13

‫ ܿܥ‬ൌ ሺ‫ܦ‬ଷ଴ ሻ^2/ሺ‫ܦ‬ଵ଴ ൈ ‫଺ܦ‬଴ ሻ

5.14

57

where;

‫଺ܦ‬଴ = Grain diameter at 60% of the aggregates passing
‫ܦ‬ଵ଴ = Grain diameter at 10% of the aggregates passing

1/2"

3/8"

#4

#8

#16

#200
#100
#50
#30

Dense
Gradation

1/2"

3/8"

#4

#8

Uniform
Gradation

#16

100
90
80
70
60
50
40
30
20
10
0

#200
#100
#50
#30

% Passing

‫ܦ‬ଷ଴ = Grain diameter at 30% of the aggregates passing

Sieve Size

Figure 5.14 Dense and uniform gradation for maximum aggregate size 12.5 mm
According to Unified Soil Classification System (USCS), well graded (dense (non-uniformly)
graded) aggregates are classified with ‫ ݑܥ‬൐ 4 ܽ݊݀ 1 ൏ ‫ ܿܥ‬൏ 3. If it does not satisfy the
criteria, it is classified as poorly graded (uniform graded). Table 5.4 shows grain sizes and the
gradation parameters Cu and Cc for the dense and uniform gradation examples demonstrated on
figure 5.14.
Table 5.4 Gradation parameters for dense and uniform graded aggregates

Uniform Gradation
Dense Gradation

Cc
D10 D30 D60 Cu
3.06 5.27 7.23 2.36 1.25
0.35 1.59 3.64 10.40 1.99
58

It can be inferred from the USCS that increase in Cu results in well, i.e., non-uniform gradation.
Hence, the uniformity coefficient is misnamed since as Cu gets smaller, the aggregate gradation
becomes more uniform. It should actually be called the coefficient of non-uniformity. For
instance, if Cu is equal to one (1), it means that there is only one grain size.
The relation between coefficient of uniformity and SPL is presented in figure 5.13, where a
2

strong correlation between the variables (with R =0.74) was observed. This analysis can be
further extended to discover the impact of Cu purely by overshadowing the interactions with
other material characteristics. In order to better illustrate the effect of the Cu, only the specimens
with unmodified binders were compared (i.e., mixtures with polymer modified and crumb rubber
modified binders are not included). Also, the mixture with lightweight aggregates was not
included in the analysis. Figure 5.15 illustrates the correlation between SPL and Cu. In this case,
2
the relation becomes stronger with approximately R = 0.96. This phenomenon can be explained

with the compaction and void content. If Cu increases, the gradation becomes well graded having
less air voids. When there are less air voids, more sound waves are reflected yielding higher SPL
values. It still needs to be remembered that there are always other material interactions with SPL
and multivariate analysis is required for better understanding.

59

SPL vs Cu (for each gradation)
95.0
SMA

SPL (dB(A))

93.0
91.0

SP-25 mm

SP-12.5mm

89.0

SP-9.5mm

87.0

y = -0.0033x2 + 0.3295x + 85.896
R² = 0.9565

OGFC

85.0
0.0

10.0

20.0

30.0

40.0

50.0

Cu
Figure 5.15 Change in SPL with change in Cu, only one for each gradation
The other gradation parameter is coefficient of curvature (Cc). The studies show that it almost
2

has no impact on SPL. The correlation between Cc and SPL is given in the figure 5.16. R =
0.0815 is also another indication very poor effect of Cc on SPL. Figure 5.16 demonstrates the
correlations between SPL and Cc

SPL vs Cc
SPL (dB(A))

94.0
92.0
90.0
88.0

y = 0.2488x + 89.473
R² = 0.0815

86.0
84.0
0.0

2.0

4.0

6.0

8.0

10.0

Cc
Figure 5.16 Change in SPL with change in Cc
In order to better understand the effect of gradation parameters on SPL, there is a need for an
intensive study. Further research on this area is recommended.
60

5.2 MULTIVARIATE REGRESSION ANALYSIS
The individual relations between material characteristics of the asphalt pavements and the SPL
reveal that there is a need for multivariate analyses for a better understanding about sound
absorption behavior of asphalt. There are certain interactions between visco-elastic, volumetric
and gradation parameters. As discussed in the previous sections, there is more than one
parameter affecting the void content of the asphalt such as aggregate gradation, binder content
and voids filled with asphalt. This occurrence can be elucidated by studying the interactions
between parameters. Multivariate statistics by using commercially available software SPSS
(statistical package for the social sciences) was performed to explore the relations. Multivariate
regression analysis between linear visco-elastic, volumetric and gradation parameters of asphalt
specimens and the SPL is studied to determine a formula which can be used to predict the SPL of
the different asphalt mixtures by only using the data obtained at laboratory conditions. Such
formulation might be helpful to fill the gap for future’s quieter pavements by predicting the noise
level of the asphalt pavements using basic mix design data.
The analyses performed show only the resultant significant parameters and results. SPSS were
run on by using one of the backward, stepwise or enter methods on linear regression analyze
mode. In any of the modes, the parameters were excluded from the model because of either
collinearity between them or insignificancy according to setup criterion.
The first empirical prediction model included only linear visco-elastic parameters of the asphalt
mixtures. The general denotation of the model utilized;
SPL = f ( |E*|, sin δ, |E*|/sinδ, E’, E”)

5.15

All the combinations between parameters were analyzed and the most suitable outcome was
obtained as;

61

SPL = f (|E*|, sin δ)

5.16

SPL = C1 + (X1 sin δ) + (X2 |E*|)

5.17

where C1, X1 and X2 are the unstandardized coefficients obtained at the end of analysis and
given in Table 5.5 which illustrates the results acquired from SPSS at the end of the run. The
table consists only the basic values need to be reported.
Table 5.5 SPSS results for SPL and linear visco-elastic parameters
VARIABLES ENTERED/REMOVED
Model
1
Model
1

Var. Entered
|E*|, Sin δ
R
0.762

Var. Removed
.

Method
Enter

MODEL SUMMARY
R Square
Adjusted R Square
.581
.476
ANOVA
Sum of Squares
37.613
27.128
64.742

Model

1

Model
Regression
Residual
Total

df
2
8
10

COEFFICIENTS
Unstandardized Coefficients
B
Std. Error

F
5.546

Sig.
0.031

t

Sig.

(Constant)

5.029

14.862

.000

Sin δ

X1=23.594655

9.375

2.517

.036

|E*|

1

C1=74.742577
X2=.0009799

.000

3.290

.011

2

The relation between SPL and |E*|, sinδ is not strong with R = 0.58. The significance (sig) of
the parameters to predict the model is good enough with sig value less than 0.05. However, the
correlation between the independent variables is in an appreciable amount. This makes the model
extremely weak to predict the SPL.
The second relation investigated was between SPL and gradation parameters. Cu and Cc were the
62

two parameters studied for the effect of aggregate gradation on SPL. The prediction model was
in the form of;
SPL = f (Cu, Cc)

5.18

Although Cu has a strong relation with SPL, it is weak for Cc. There is a need for a wider study
to better understand the correlation between parameters studied. In the final prediction function,
only Cu values are included since they seem to represent the impact of gradation on SPL better.
Table 5.6 demonstrates the result obtained by SPSS. Backward criterion yielded for removal of
Cc from the prediction since it does not make any contribution to predict the SPL in the model.
The prediction model for gradation parameters is;
SPL = C2+X3.Cu

5.19

where X2 and A2 are the unstandardized coefficients. As in the individual correlations part,
increase in Cu will result in an increase in SPL since the mixture becomes well graded.
Another prediction model was setup between mixture volumetric parameters and SPL.
SPL = f (Va, VMA, VFA, Gmm, Gmb, Pb)

5.20

Since volumetric parameters were highly correlated, there was a problem of collinearity in the
prediction model. SPSS excludes highly correlated variables by setting up the criteria and using
the backward method. The strongest variable in the model above yielded to be Pb, Va was the
second significant parameter. They were used in the general prediction model along with other
considerable gradation and visco-elastic parameters.

63

Table 5.6 SPSS results for SPL and gradation parameters
Model
1
2

Model
1
2
Model
1

2

Model
1

VARIABLES ENTERED/REMOVED
Var. Entered
Variables Removed
Method
Cc, Cu
.
Enter
Backward (criterion: Probability of F. Cc
to-remove >= .100)
MODEL SUMMARY
Std. Error of
R
R Square
Adjusted R Square
the Est.
.813
.661
.576
1.657
.651
.612
1.585
.807
ANOVA
Sum of Squares
df
Mean Square
F
Sig.
Regression
42.768
2
21.384
7.785 0.01
Residual
21.974
8
2.747
Total
64.742
10
Regression
42.141
1
42.141 16.782
0
Residual
22.601
9
2.511
Total
64.742
10
COEFFICIENTS
Unstandardized Coefficients Stand. Coef.
Std.
B
Error
Beta
t
Sig.
(Constant)
87.2460188 .914
95.441 .000
Cu

.1728447

.047

.858

3.694 .006

Cc

-.0966415

.202

-.111

-.478 .646

(Constant)

C2=87.1793222

Cu

2

X3=.1625192
.807
4.097 .003
.040
EXCLUDED VARIABLES
Partial Correlation
Collinearity Statistics (Tolerance)

Model
2

Cc

-.167

.864

100.926 .000

.786

The general prediction model includes all the variables.
SPL = f (|E*|, sin δ, |E*|/sin δ, E’, E”, Cu, Cc, Va, VMA, VFA, Gmm, Gmb, Pb)
Although all the variables measured are shown in the model above, only the significant ones
obtained in the previous analyses are included.

64

5.21

First considerable relation analyzed was
SPL = (|E*|, Cu, Pb, VFA)

5.22

Runs were performed by using backward criterion of the SPSS so that the significant or highly
collinear parameters were excluded from the model. Table 5.7 illustrates the results obtained by
using important parameters at enter criterion of SPSS.
Table 5.7 SPSS results for SPL and significant parameters 1
VARIABLES ENTERED/REMOVED
Variables
Removed
Method
Model Variables Entered
1

Pb, VFA, Cu, |E*|

Model
1

R
0.964

Model
1

Regression
Residual
Total

Model

.
MODEL SUMMARY
Adjusted
R Square
R Square
.930
.883
ANOVA
Sum of Squares
df
60.212
4
4.530
6
64.742
10
COEFFICIENTS
Unstandardized Coefficients
B
Std. Error

(Constant)
|E*|

X4=-.0006990

.000

VFA

X5=.1617178

Cu

X6=.1319413

Pb

1

C3=88.9664421

X7=-92.5891333

Enter

Std. Error of the Estimate
.869
Mean
Square
15.053
.755

Std. Coef.
Beta

F
19.937

t

Sig.
0.001

Sig.

25.730

.000

-.703

-3.172

.019

.065

.495

2.476

.048

.030

.655

4.367

.005

23.493

-.571

-3.941

.008

3.458

In the model all the parameters are significant with Sig. < 0.05 (very right column). The signs of
the unstandardized parameter coefficients reveal the relation between parameter and SPL
individually. VFA and Cu are positively correlated with SPL. It means that an increase in VFA
or Cu or both will yield SPL to ascend as well. The reason for this kind of relation was explained
65

under each individual parameter relation with SPL studies. On the other hand, rise in |E*| and/or
Pb will cause decrease in SPL. The fact for Pb was clarified previously and the phenomenon for
|E*| can be enlightened with following interpretations. This phenomenon can be enlightened with
two interpretations. The first one is the range of |E*| data used. Since only one combination of
o

temperature and frequency was utilized for visco-elastic parameter (212 Hz and 23 C), the data
used most probably does not cover the all range. The second reason might be the interaction
between parameters in the model. The relations between independent parameters can cause
inverse effect on certain parameters. In order to better understand the situation, more
2

measurements of |E*| data and analysis need to be performed. The model has R =0.930. This
tells how close the predicted and measured SPL’s are the ability of model to predict the SPL by
using the engineering characteristics of the asphalt mixtures.
In this case the general prediction model yields the following SPL formula,
SPL = C3 + (X4.|E*|) + (X5.VFA) + (X6. Cu) + (X7. Pb)

5.23

where C3, X4, X5, X6 and X7 are the coefficients given in Table 5.7.

SPL Measured vs SPL Predicted (Model-1)

SPL Measured (dB(A))

96.0
94.0

y = 0.93x + 6.3064
R² = 0.93

92.0
90.0
88.0
86.0
84.0
84.0

86.0

88.0

90.0

92.0

94.0

96.0

SPL Predicted (dB(A))

Figure 5.17 Measured and predicted SPL values for general prediction model 1
66

Figure 5.17 illustrated the relation between measured and predicted sound pressure levels which
is obtained as a result of analysis of first general prediction model.
The second considerable relation performed was
SPL = (|E*|, Cu, Va)

5.24

Runs were performed by using backward criterion of the SPSS. All the parameters yielded to
become significant according to the inherent removal criteria of SPSS. The aim of running a
second general prediction model is to decrease the amount of variables to predict SPL. In this
model, volumetric parameters VFA and Pb are replaced with another volumetric parameter Va.
Using only Va as an independent variable in the model has some advantages. One of them is
easier determination of Va compared to VFA and Pb. The second advantage is to use less
parameter to predict the model which in return results more efficient and time saving
computations.
Table 5.8 demonstrates the SPSS results of general model 2. Although the relation is not as
2

2

strong as in the model 1 (model1 R 1= 0.93 and model 2 R 2= 0.897), the significance of the
parameters are better (Table 5.7 and Table 5.8).
Prediction formula in this model becomes,
SPL = C4 + (X8.|E*|) + (X9.Cu) + (X10.Va)

5.25

Figure 5.18 demonstrates the measured and predicted SPL values with trend line and equation of
the line for general model 2.

67

Table 5.8 SPSS results for SPL and significant parameters 2

(Constant)

VARIABLES ENTERED/REMOVED
Variables
Method
Removed
.
Enter
MODEL SUMMARY
Adjusted R
R Square
Std. Error of the Estimate
Square
.897
.853
.9768393255844
ANOVA
Mean
Sum of Squares
df
F
Sig.
Square
58.062
3
19.354
20.283 0.001
6.680
7
.954
64.742
10
COEFFICIENTS
Unstandardized Coefficients
Stand. Coef.
t
Sig.
B
Std. Error
Beta
32.194 .000
3.089
C4=99.4611990

|E*|

X8=-0.0008062

.000

-.810

-3.300

.013

Cu

X9=0.1288282

.034

.640

3.784

.007

Va

X10=-0.7601706

.187

-1.028

-4.069

.005

Variables
Entered

Model
1

Va, Cu, |E*|

Model

R

1

0.947
Model

1

Regression
Residual
Total
Model

SPL Measured vs SPL Predicted (Model-2)
96.0

SPL Measured (dB(A))

1

94.0
92.0

y = 0.8968x + 9.2986
R² = 0.8968

90.0
88.0
86.0
84.0
84.0

86.0

88.0

90.0

92.0

94.0

96.0

SPL Predicted (dB(A))
Figure 5.18 Measured and predicted SPL values for general prediction model 2
68

CHAPTER 6
SUMMARY AND CONCLUSIONS
6.1 SUMMARY
The impact of material properties on sound absorption of flexible pavements have been studied
in this thesis. These material properties include asphalt linear visco-elastic properties, aggregate
gradation parameters and mixture volumetrics. Analyses performed include both individual
correlations and multivariate regressions between parameters and SPL. In order to focus on
difference between sounds generated from the material properties, a novel laboratory tire
pavement noise simulator (TIPANOS) was constructed. TIPANOS measurements were not
influenced by the surface texture of pavement since tests were performed on laboratory AMPT
specimens.
6.2 CONCLUSIONS
Based on the laboratory measurements of the samples and data analyses of the obtained results,
the following conclusions were drawn:
•

There is negative correlation between VMA and Va, and SPL. As these parameters
increase, there is possibility of more interconnected voids where sound waves can freely
move and can be absorbed and refracted.

•

A negative correlation between Pb and SPL was observed. This is logical because as the
amount of binder increases, the asphalt mixture becomes more viscous (rather than
elastic). Materials showing viscous behavior are typically quieter than the materials with
elastic characteristics (because of damping). However, it should be noted that increase in
asphalt content may yield to the decrease in void content, which may increase SPL.

69

Therefore, it can be suggested that, for sound absorption purposes, the binder content
should be optimized such that the highest amount of binder is used, without excessively
blocking the interconnected voids
•

2

A very good correlation (with an R of 0.89) between the Cu of aggregates used in the
HMA and SPL was observed (SPL increased with increasing Cu). High Cu means well
graded (more densely packed) aggregates, having less air void space between the
aggregates. It is hypothesized that when there are less air voids, more sound waves are
reflected yielding higher SPL values. However, there is a need for an intense study to
discover the effect of Cc on SPL.

•

Statistical analysis revealed that, among all the material parameters, the influence of |E*|,
VFA, Cu and Pb on SPL is statistically significant. A predictive multivariate regression
2

equation was developed. This regression equation revealed an R of 0.93, which shows
the significance of the combined effect of the parameters on the SPL.

70

APPENDICES

71

Appendix A
(Job Mix Formulas)

72

JOB MIX FORMULA
HMA FIELD COMMUNICATION
CONTROL SECTION
CONTROL MIX DESIGN

JOB NO.
0.02

PROJECT ENGINEER
Salih KOCAK

DATE EFFECTIVE
07/21/2009

MIXTURE TYPE
MIX DESIGN NO.
PLANT LOCATION
SUPERPAVE
NMAS 9.5
EAST LANSING
ANGULARITY
% AIR VOIDS
VMA
VFA
COMP.TEMP
MIX. TEMP
47.4
4.00
15.02
73.35
138 C
149 C
Gmm
Gmb
Gb
Gse
Gsb
P200/Pbe
% AIR VOIDS
2.648
2.542
1.023 2.876 2.849
1.13
4.00
MIX/AGG. GRADATION, %
MIX/AGG. PROPORTION, %
ITEM

PERCENT

MATERIAL/PRODUCER

PIT NO.

ASPHALT,%

4.76%

SPARTAN & RIETH-RILEY

N/A

P 1-1/2" (37.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1" (25.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 3/4" (19.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1/2" (12.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 3/8" (9.5 mm)

91.00%

SPARTAN & RIETH-RILEY

N/A

P No. 4 (4.75 mm)

56.00%

SPARTAN & RIETH-RILEY

N/A

P No. 8 (2.36 mm)

38.00%

SPARTAN & RIETH-RILEY

N/A

P No. 16 (1.18 mm)

26.00%

SPARTAN & RIETH-RILEY

N/A

P No. 30 (600 µm)

16.00%

SPARTAN & RIETH-RILEY

N/A

P No. 50 (300 µm)

10.00%

RECLAIMED

NO RAP

P No. 100 (150 µm)

7.00%

Supplier: MTE

P No. 200 (75 µm)

5.00%

ASPHALT
BINDER

CRUSHED 1 FACE

100%

PRODUCER LOCATION: MSU-CEE-AACL

CRUSHED 2 FACES

100%

REGULAR TESTING

GRADE: PG 58-28

Figure A.1 Job Mix Formula for Superpave NMAS 9.5 mm

73

JOB MIX FORMULA
HMA FIELD COMMUNICATION
CONTROL SECTION
CONTROL MIX DESIGN

JOB NO.
0.03

PROJECT ENGINEER
Salih KOCAK

DATE EFFECTIVE
07/21/2009

MIXTURE TYPE
MIX DESIGN NO.
PLANT LOCATION
SUPERPAVE
NMAS 25.00
EAST LANSING
ANGULARITY
% AIR VOIDS
VMA
VFA
COMP.TEMP
MIX. TEMP
46.1
4.00
14.26
71.94
138 C
149 C
Gmm
Gmb
Gb
Gse
Gsb
P200/Pbe
% AIR VOIDS
2.624
2.513
1.023 2.840 2.802
0.72
4.00
MIX/AGG. GRADATION, %
MIX/AGG. PROPORTION, %
ITEM

PERCENT

MATERIAL/PRODUCER

PIT NO.

ASPHALT,%

4.63%

SPARTAN & RIETH-RILEY

N/A

P 1-1/2" (37.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1" (25.0 mm)

97.00%

SPARTAN & RIETH-RILEY

N/A

P 3/4" (19.0 mm)

89.00%

SPARTAN & RIETH-RILEY

N/A

P 1/2" (12.5 mm)

75.00%

SPARTAN & RIETH-RILEY

N/A

P 3/8" (9.5 mm)

67.00%

SPARTAN & RIETH-RILEY

N/A

P No. 4 (4.75 mm)

54.00%

SPARTAN & RIETH-RILEY

N/A

P No. 8 (2.36 mm)

41.00%

SPARTAN & RIETH-RILEY

N/A

P No. 16 (1.18 mm)

30.00%

SPARTAN & RIETH-RILEY

N/A

P No. 30 (600 µm)

22.00%

SPARTAN & RIETH-RILEY

N/A

P No. 50 (300 µm)

15.00%

RECLAIMED

NO RAP

P No. 100 (150 µm)

10.00%

Supplier: MTE

P No. 200 (75 µm)

3.00%

ASPHALT
BINDER

CRUSHED 1 FACE

100%

PRODUCER LOCATION: MSU-CEE-AACL

CRUSHED 2 FACES

100%

REGULAR TESTING

GRADE: PG 58-28

Figure A.2 Job Mix Formula for Superpave NMAS 25 mm

74

JOB MIX FORMULA
HMA FIELD COMMUNICATION
CONTROL SECTION
CONTROL MIX DESIGN

JOB NO.
0.04

PROJECT ENGINEER
Salih KOCAK

DATE EFFECTIVE
07/21/2009

MIXTURE TYPE
MIX DESIGN NO.
PLANT LOCATION
SMA
12.5
EAST LANSING
ANGULARITY
% AIR VOIDS
VMA
VFA
COMP.TEMP
MIX. TEMP
45.0
4.00
17.19
76.72
138 C
149 C
Gmm
Gmb
Gb
Gse
Gsb
P200/Pbe
% AIR VOIDS
2.587
2.484
1.023 2.867 2.819
0.74
4.00
MIX/AGG. GRADATION, %
MIX/AGG. PROPORTION, %
ITEM

PERCENT

MATERIAL/PRODUCER

PIT NO.

ASPHALT,%

6.00%

SPARTAN & RIETH-RILEY

N/A

P 1-1/2" (37.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1" (25.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 3/4" (19.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1/2" (12.5 mm)

94.00%

SPARTAN & RIETH-RILEY

N/A

P 3/8" (9.5 mm)

65.00%

SPARTAN & RIETH-RILEY

N/A

P No. 4 (4.75 mm)

30.00%

SPARTAN & RIETH-RILEY

N/A

P No. 8 (2.36 mm)

20.00%

SPARTAN & RIETH-RILEY

N/A

P No. 16 (1.18 mm)

15.00%

SPARTAN & RIETH-RILEY

N/A

P No. 30 (600 µm)

12.00%

SPARTAN & RIETH-RILEY

N/A

P No. 50 (300 µm)

10.00%

RECLAIMED

NO RAP

P No. 100 (150 µm)

6.00%

Supplier: MTE

P No. 200 (75 µm)

4.00%

ASPHALT
BINDER

CRUSHED 1 FACE

100%

PRODUCER LOCATION: MSU-CEE-AACL

CRUSHED 2 FACES

100%

REGULAR TESTING

GRADE: PG 58-28

VCA (mix) & VCA (dry) conditions are satisfied, 0.2%fider was added.
Figure A.3 Job Mix Formula for Stone Matrix Asphalt

75

JOB MIX FORMULA
HMA FIELD COMMUNICATION
CONTROL SECTION
CONTROL MIX DESIGN

JOB NO.
0.05

PROJECT ENGINEER
Salih KOCAK

DATE EFFECTIVE
07/21/2009

MIXTURE TYPE
MIX DESIGN NO.
PLANT LOCATION
OGFC
4.75
EAST LANSING
ANGULARITY
% AIR VOIDS
VMA
VFA
COMP.TEMP
MIX. TEMP
47.0
4.00
NA
NA
138 C
149 C
Gmm
Gmb
Gb
Gse
Gsb
P200/Pbe
% AIR VOIDS
2.497
2.122
1.023 2.912 2.664
NA
15.00
MIX/AGG. GRADATION, %
MIX/AGG. PROPORTION, %
ITEM

PERCENT

MATERIAL/PRODUCER

PIT NO.

ASPHALT,%

9.00%

SPARTAN & RIETH-RILEY

N/A

P 1-1/2" (37.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1" (25.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 3/4" (19.0 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 1/2" (12.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P 3/8" (9.5 mm)

100.00%

SPARTAN & RIETH-RILEY

N/A

P No. 4 (4.75 mm)

38.00%

SPARTAN & RIETH-RILEY

N/A

P No. 8 (2.36 mm)

10.00%

SPARTAN & RIETH-RILEY

N/A

P No. 16 (1.18 mm)

6.00%

SPARTAN & RIETH-RILEY

N/A

P No. 30 (600 µm)

4.00%

SPARTAN & RIETH-RILEY

N/A

P No. 50 (300 µm)

3.00%

RECLAIMED

NO RAP

P No. 100 (150 µm)

3.00%

Supplier: MTE

P No. 200 (75 µm)

2.60%

ASPHALT
BINDER

CRUSHED 1 FACE

100%

PRODUCER LOCATION: MSU-CEE-AACL

CRUSHED 2 FACES

100%

REGULAR TESTING

GRADE: PG 58-28

1% hydrated lime & 0.3% fiber by weight were added
Figure A.4 Job Mix Formula for Open Graded Friction Course

76

Appendix B
(TIPANOS SPL Measurement Results)

77

Table B.1 Sound pressure level at 1/3 octave band frequencies
Frequency
Hz
25.0
40.0
63.0
100.0
160.0
250.0
400.0
630.0
1000.0
1600.0
2500.0
4000.0
6300.0
10000.0
16000.0

Sample Code 125SA4
1-1
1-2
1-3 Average
76.9
81.5
80.0
79.5
74.0
75.7
77.9
75.9
73.3
72.7
76.7
74.3
78.5
74.0
79.8
77.4
75.1
72.1
76.0
74.4
90.6
92.0
91.8
91.4
67.0
75.7
74.2
72.3
67.6
75.3
76.9
73.3
66.6
66.7
67.7
67.0
55.0
58.8
63.9
59.2
53.5
51.9
54.0
53.1
58.3
51.4
54.7
54.8
47.7
49.2
50.1
49.0
41.7
41.6
42.3
41.9
32.2
32.1
32.2
32.2

Sound Pressure Level of Samples dB(A)
Sample Code 125SA48
Sample Code
2-1
2-2
2-3
Average
3-1
3-2
91.9
78.4
80.5
83.6
79.3
81.4
80.2
75.4
78.8
78.1
76.3
75.8
72.2
75.9
77.1
75.1
76.1
72.5
76.4
82.1
81.3
79.9
80.4
78.8
76.8
76.7
77.8
77.1
77.3
77.6
92.5
91.4
91.6
91.8
90.8
88.5
75.2
72.3
74.1
73.9
71.1
69.4
73.7
70.1
70.1
71.3
67.6
68.3
67.9
67.7
68.5
68.0
66.7
66.3
62.4
64.5
63.2
63.4
57.1
59.4
52.2
61.2
59.1
57.5
54.1
52.6
55.2
59.4
60.3
58.3
56.4
55.6
49.4
47.2
50.2
48.9
51.9
49.5
44.9
47.2
45.2
45.8
36.8
39.9
29.3
30.0
32.7
30.7
28.7
30.3

78

125SA4PM
3-3
Average
78.4
79.7
76.6
76.2
77.3
75.3
79.0
79.4
74.4
76.4
90.0
89.8
71.4
70.6
72.7
69.5
71.3
68.1
59.7
58.7
56.8
54.5
59.2
57.1
49.2
50.2
41.0
39.2
29.7
29.6

Table B.2 Sound pressure level at 1/3 octave band frequencies (continued)
Frequency
Hz
25.0
40.0
63.0
100.0
160.0
250.0
400.0
625.0
1000.0
1600.0
2500.0
4000.0
6300.0
10000.0
16000.0

Sound Pressure Level of Samples dB(A)
Sample Code 125SA4CR5
Sample Code 125SA4CR10
Sample Code
4-1
4-2
4-3 Average
5-1
5-2
5-3 Average 6-1
6-2
76.7
76.7
73.7
75.7
81.8
75.5 77.1
78.1
79.0
78.8
74.6
75.1
71.6
73.8
77.2
72.7 75.4
75.1
76.3
75.7
75.4
76.6
72.4
74.8
75.2
71.2 74.0
73.5
75.3
73.6
77.6
77.8
76.0
77.1
77.7
74.2 77.7
76.5
77.7
79.0
74.5
77.9
75.5
76.0
77.7
72.5 72.7
74.3
75.2
76.3
91.8
90.9
89.6
90.7
88.7
89.4 89.0
89.1
92.2
92.6
70.9
73.3
70.3
71.5
71.3
68.6 68.2
69.4
73.9
75.4
75.5
72.3
73.5
73.8
70.1
71.3 69.1
70.2
73.7
74.4
71.3
66.8
66.8
68.3
68.0
68.0 65.9
67.3
69.5
71.9
59.1
58.5
60.6
59.4
64.6
58.5 57.7
60.3
59.6
64.1
52.4
52.8
53.3
52.8
54.8
57.3 52.2
54.8
57.7
56.2
57.0
56.5
57.2
56.9
53.8
58.1 57.5
56.5
64.0
64.2
47.7
48.5
50.0
48.8
48.4
49.8 47.2
48.5
52.6
54.5
40.5
40.3
42.8
41.2
38.5
42.4 39.8
40.2
45.4
48.8
30.1
31.5
29.8
30.5
29.7
29.5 34.1
31.1
36.1
36.6

79

125SA4LW
6-3 Average
78.1
78.6
76.5
76.1
74.9
74.6
76.7
77.8
76.6
76.0
91.2
92.0
76.2
75.2
76.2
74.7
71.5
71.0
65.6
63.1
62.8
58.9
62.9
63.7
56.1
54.4
46.5
46.9
35.2
36.0

Table B.3 Sound pressure level at 1/3 octave band frequencies (continued)
Frequency
Hz
25.0
40.0
63.0
100.0
160.0
250.0
400.0
625.0
1000.0
1600.0
2500.0
4000.0
6300.0
10000.0
16000.0

Sound Pressure Level of Samples dB(A)
Sample Code 095SA4F
Sample Code 255SA4C
Sample Code 0475OA4
7-1
7-2
7-3 Average 8-1
8-2
8-3
Average
9-1
9-2
9-3
Average
72.6
74.3
77.7
74.9
77.4
73.7
76.5
75.9
76.6
78.3
72.9
75.9
72.4
73.6
76.8
74.3
75.6
72.9
74.8
74.4
75.4
75.7
70.6
73.9
73.1
72.4
74.9
73.5
75.4
74.8
74.3
74.8
74.7
75.3
71.6
73.9
77.6
75.6
79.5
77.6
77.4
81.1
75.9
78.1
77.9
80.5
78.7
79.0
75.5
72.8
74.9
74.4
73.2
72.5
75.0
73.6
76.9
77.4
79.3
77.9
89.3
90.0
90.0
89.8
93.9
93.8
94.2
94.0
85.1
87.2
87.9
86.7
72.5
72.2
71.3
72.0
76.2
73.8
74.8
74.9
76.0
72.9
72.5
73.8
72.0
69.3
74.4
71.9
72.4
76.7
75.9
75.0
72.3
72.4
69.5
71.4
66.3
63.5
66.0
65.2
68.4
70.0
73.5
70.6
64.7
65.4
65.2
65.1
62.8
56.7
59.1
59.5
60.0
63.0
64.2
62.4
57.7
58.4
54.6
56.9
52.8
51.1
56.4
53.4
53.3
55.1
61.1
56.5
52.2
54.1
53.7
53.3
56.3
54.1
57.1
55.8
57.3
54.9
58.3
56.9
57.9
54.7
58.2
56.9
52.2
46.9
51.4
50.2
48.6
51.4
50.6
50.2
52.3
50.9
52.5
51.9
37.3
41.3
36.8
38.5
42.0
45.8
44.6
44.1
45.2
46.2
43.7
45.0
31.3
30.4
31.9
31.2
31.6
29.7
32.3
31.2
39.3
38.6
41.9
39.9

80

Table B.4 Sound pressure level at 1/3 octave band frequencies (continued)
Frequency
Hz
25.0
40.0
63.0
100.0
160.0
250.0
400.0
625.0
1000.0
1600.0
2500.0
4000.0
6300.0
10000.0
16000.0

Sound Pressure Level of Samples dB(A)
Sample Code 0475OA4CR10
Sample Code 125SMAA4
Sample Code 125SMAA4CR10
10-1
10-2
10-3 Average 11-1
11-2
11-3 Average
12-1
12-2
12-3 Average
76.0
75.0
77.1
77.1
76.3
79.7
79.2
78.4
74.7
76.6
77.5
76.3
75.1
75.2
75.3
75.3
73.1
76.1
74.3
74.5
73.5
73.9
74.8
74.0
74.3
75.0
75.2
75.2
72.9
73.5
72.2
72.9
73.7
74.3
75.3
74.4
82.4
76.5
78.6
78.6
76.9
74.3
75.8
75.6
76.4
79.6
77.3
77.8
77.5
74.0
74.8
74.8
74.3
73.0
74.6
74.0
73.9
74.8
74.0
74.2
88.5
86.7
85.2
85.2
92.8
92.4
93.5
92.9
90.8
89.6
89.4
90.0
74.2
67.3
69.8
69.8
75.1
73.6
72.2
73.6
70.3
72.0
74.4
72.2
68.9
69.5
69.2
69.2
74.1
72.8
72.2
73.0
71.6
74.7
71.2
72.5
64.8
68.4
64.1
64.1
69.9
68.9
71.4
70.0
70.1
68.2
69.8
69.4
56.7
55.9
57.2
57.2
59.4
61.2
57.9
59.5
57.7
60.7
60.6
59.7
49.0
48.1
48.8
48.8
56.2
56.2
55.2
55.9
59.9
53.9
58.8
57.5
53.1
58.1
56.9
56.9
62.6
59.6
60.5
60.9
60.3
61.7
62.8
61.6
49.8
49.8
46.2
46.2
49.0
47.9
50.0
49.0
50.1
48.5
50.1
49.6
40.8
43.2
43.1
43.1
40.4
38.8
41.5
40.2
43.9
40.6
40.2
41.6
34.6
38.9
33.7
33.7
36.1
30.5
34.3
33.6
31.9
28.2
28.6
29.6

81

Appendix C
(Correlations between Linear Visco-Elastic Parameters and SPL at Different
Frequency and Temperature Combinations)

82

SPL vs δ

95.0
93.0

93.0

SPL (dB(A))

SPL (dB(A))

SPL vs |E*|

95.0

91.0
89.0

y = -0.0007x + 90.294

87.0

R² = 1E-06

85.0

91.0
89.0

y = 0.0003x + 87.034
R² = 0.2114

87.0
85.0

5.00

10.00

15.00

20.00

0

5000

10000

15000

20000

Phase Angle (Deg)

|E*| (MPa)
(a)
(b)
o
Figure C.1 SPL versus (a) Phase angle (b) Dynamic modulus at 4 C and 10 Hz

SPL vs |E"|

95.0

93.0

SPL (dB(A))

93.0

SPL (dB(A))

SPL vs |E*|/ sin δ

95.0

91.0
89.0

y = 0.0019x + 85.371
R² = 0.4285

87.0
85.0
0

1000

2000

3000

91.0
89.0

y = 0.0003x + 87.034
R² = 0.2114

87.0
85.0
0

4000

5000

10000

15000

|E"| (MPa)
|E*| / sin δ (MPa)
(a)
(b)
o
Figure C.2 SPL versus (a) |E”| (b) |E*|/sin δ at 4 C and 10 Hz
83

20000

SPL vs δ

95.0
93.0

93.0

SPL (dB(A))

SPL (dB(A))

SPL vs |E*|

95.0

91.0
89.0

y = 0.0196x + 89.964
R² = 0.0011

87.0
85.0

91.0
89.0

y = 0.0003x + 87.338
R² = 0.1843

87.0
85.0

10.00

15.00

20.00

25.00

0.00

5000.00

10000.00

15000.00

20000.00

Phase Angle (Deg)

|E*| (MPa)
(a)
(b)
o
Figure C.3 SPL versus (a) Phase angle (b) Dynamic modulus at 4 C and 5 Hz

SPL vs |E"|

95.0

93.0

SPL (dB(A))

93.0

SPL (dB(A))

SPL vs |E*|/ sin δ

95.0

91.0

91.0

89.0

89.0

y = 0.0021x + 85.154
R² = 0.4569

87.0
85.0
0

1000

2000

3000

y = 0.0003x + 87.338
R² = 0.1843

87.0
85.0
0.00

4000

5000.00

10000.00

15000.00

|E"| (MPa)
|E*| / sin δ (MPa)
(a)
(b)
o
Figure C.4 SPL versus (a) |E”| (b) |E*|/sin δ at 4 C and 5 Hz
84

20000.00

SPL vs δ

95.0

93.0

SPL (dB(A))

93.0

SPL (dB(A))

SPL vs |E*|

95.0

91.0
89.0

y = 0.0813x + 88.082
R² = 0.0493

87.0

91.0
89.0

y = 0.0002x + 89.593
R² = 0.0215

87.0

85.0

85.0

15.00

20.00

25.00

30.00

35.00

40.00

0.00

2000.00 4000.00 6000.00 8000.00 10000.00

Phase Angle (Deg)

|E*| (MPa)
(a)
(b)
o
Figure C.5 SPL versus (a) Phase angle (b) Dynamic modulus at 4 C and 0.1 Hz

SPL vs |E"|

SPL vs |E*|/ sin δ

93.0

SPL (dB(A))

95.0

93.0

SPL (dB(A))

95.0

91.0

91.0

89.0

89.0

y = 0.0024x + 86.41
R² = 0.3012

87.0

y = 0.0002x + 89.593
R² = 0.0215

87.0
85.0

85.0
0

1000

2000

3000

0.00

2000.00 4000.00 6000.00 8000.00 10000.00

|E"| (MPa)
|E*| / sin δ (MPa)
(a)
(b)
o
Figure C.6 SPL versus (a) |E”| (b) |E*|/sin δ at 4 C and 0.1 Hz
85

SPL vs δ

95.0
93.0

93.0

SPL (dB(A))

SPL (dB(A))

SPL vs |E*|

95.0

91.0
89.0

y = 0.0871x + 87.663
R² = 0.0464

87.0
85.0

91.0
89.0

y = 0.0003x + 89.352
R² = 0.0353

87.0
85.0

20.00

25.00

30.00

35.00

40.00

0.00

2000.00

4000.00

6000.00

8000.00

Phase Angle (Deg)

|E*| (MPa)
(a)
(b)
o
Figure C.7 SPL versus (a) Phase angle (b) Dynamic modulus at 21 C and 10 Hz

SPL vs |E"|

95.0
93.0

93.0

SPL (dB(A))

SPL (dB(A))

SPL vs |E*|/ sin δ

95.0

91.0
89.0

y = 0.0021x + 86.832
R² = 0.2475

87.0
85.0

91.0
89.0

y = 0.0003x + 89.352
R² = 0.0353

87.0
85.0

0

1000

2000

3000

0.00

2000.00

4000.00

6000.00

|E"| (MPa)
|E*| / sin δ (MPa)
(a)
(b)
o
Figure C.8 SPL versus (a) |E”| (b) |E*|/sin δ at 21 C and 10 Hz
86

8000.00

SPL vs δ

95.0
93.0

93.0

SPL (dB(A))

SPL (dB(A))

SPL vs |E*|

95.0

91.0
89.0

y = 0.0969x + 87.221
R² = 0.0541

87.0
85.0

91.0
89.0

y = 0.0002x + 89.69
R² = 0.0172

87.0
85.0

20.00

25.00

30.00

35.00

40.00

0.00

2000.00

4000.00

6000.00

8000.00

Phase Angle (Deg)

|E*| (MPa)
(a)
(b)
o
Figure C.9 SPL versus (a) Phase angle (b) Dynamic modulus at 21 C and 5 Hz

SPL vs |E"|

95.0

93.0

SPL (dB(A))

93.0

SPL (dB(A))

SPL vs |E*|/ sin δ

95.0

91.0
89.0

y = 0.0018x + 87.801
R² = 0.1523

87.0
85.0
0

500

1000

1500

2000

91.0
89.0

y = 0.0002x + 89.69
R² = 0.0172

87.0
85.0

2500

0.00

2000.00

4000.00

6000.00

|E"| (MPa)
|E*| / sin δ (MPa)
(a)
(b)
o
Figure C.10 SPL versus (a) |E”| (b) |E*|/sin δ at 21 C and 5 Hz
87

8000.00

SPL vs δ

95.0
93.0

93.0

SPL (dB(A))

SPL (dB(A))

SPL vs |E*|

95.0

91.0
89.0

y = 0.2756x + 80.382
R² = 0.1089

87.0
85.0

91.0
89.0

y = -0.0003x + 90.503
R² = 0.0058

87.0
85.0

25.00

30.00

35.00

40.00

45.00

0.00

500.00

1000.00 1500.00 2000.00 2500.00

Phase Angle (Deg)

|E*| (MPa)
(a)
(b)
o
Figure C.11 SPL versus (a) Phase angle (b) Dynamic modulus at 21 C and 0.1 Hz

SPL vs |E"|

95.0

93.0

SPL (dB(A))

93.0

SPL (dB(A))

SPL vs |E*|/ sin δ

95.0

91.0
89.0

y = -0.0005x + 90.459
R² = 0.0029

87.0
85.0
0

500

1000

91.0
89.0

y = -0.0003x + 90.503
R² = 0.0058

87.0
85.0

1500

0.00

500.00

1000.00 1500.00 2000.00 2500.00

|E"| (MPa)
|E*| / sin δ (MPa)
(a)
(b)
o
Figure C.12 SPL versus (a) |E”| (b) |E*|/sin δ at 21 C and 0.1 Hz
88

REFERENCES

89

REFERENCES
AASHTO (1993). "American Association of State Highway and Transportation Officials, Guide
for Design of Pavement Structures." Washington, D.C.
AASHTO (2009). “Designing Stone Matrix Asphalt (SMA)”. American Association of State
Highway and Transportation Officials Standards, Specifications No: R-46, Volume 1B.
Washington, D.C.
AASHTO (2009). “Stone Matrix Asphalt (SMA)”. American Association of State Highway and
Transportation Officials Standards, Specifications No: M-325, Volume 1B. Washington,
D.C.
AASHTO (2009). “Standard Practice for Superpave Volumetric Desing for Hot-Mix Asphalt
(HMA)”. American Association of State Highway and Transportation Officials
Standards, Specifications No: R-35, Volume 1B. Washington, D.C.
AASHTO (2009). “Superpave Volumetric Mix Design”. American Association of State
Highway and Transportation Officials Standards, Specifications No: M-323, Volume 1B.
Washington, D.C.
AASHTO (2009). “Preparing and Determining the Density of Hot Mix Asphalt (HMA)
Specimens by Means of the Superpave Gyratory Compactor”. American Association of
State Highway and Transportation Officials Standards, Test No: T-312, Volume 2B.
Washington, D.C.
AASHTO (2009). “Preparation of Cylindrical Performance test Specimens Using the Superpave
Gyratory Compactor (SGC)”. American Association of State Highway and
Transportation Officials Standards, Test No: PP-60, Provisional Standards. Washington,
D.C.
AASHTO (2009). “Developing Dynamic Modulus Master Curves for Hot Mix Asphalt (HMA)
Using the Asphalt Mixture Performance Tester (AMPT).” American Association of State
Highway and Transportation Officials Standards, Test No: PP-61, Provisional Standards.
Washington, D.C.
AASHTO (2009). “Developing Dynamic Modulus Master Curves for Hot Mix Asphalt (HMA).”
American Association of State Highway and Transportation Officials Standards, Test No:
PP-62, Provisional Standards. Washington, D.C.
AASHTO (2009). “Determining Dynamic Modulus of Hot Mix Asphalt (HMA).” American
Association of State Highway and Transportation Officials Standards, Test No: TP-62,
Provisional Standards. Washington, D.C.

90

Ahammed, M.J., Tighe, S.L., (2010). “Sound Absorption characteristics of Typical Pavements:
An Ontario Stduy.” Transportation Research Board, 2010. 89th Annual Meeting CDROOM.
Ambroziak, M.A., (1999). “Effect of Pavement Type on Traffic Noise Levels.” Ohio Universty,
Master of Science Thesis.1999.
Brown, E.R., Cooley, L.A. Jr., (1999). “Designing Stone Matrix Asphalt Mixtures for RutResistant Pavements.” National Cooperative Highway Research Program (NCHRP).
Report 42, TRB, National Academy Press, Washington, D.C.
Bucka, M. P., (2002). “Asphalt Rubber Overlay Noise Study Update.” Acoustical Analysis
Associates Inc, County of Sacramento Public Works Agency Department of
Transportation, Report1272, Project 22110.
Campbell, K., (2000). “Propagation of Sound and Related Shadow Zones: the Reality of Field
Measurements versus Modeled Studies, The Wall Journal, Issue No 41.
Carlson D.D., Zhu, H., Xiao, C., (2005). “Analysis of Traffic Noise Before and After Paving
with Asphalt Rubber.” Third Joint UNCTAD/IRSG Workshop on Rubber and
Environment, International Rubber Forum, Veracruz, Mexico.
Cooley, L.A. Jr., Hurley, G.C. (2004). “Potential of Using Stone Matrix Asphalt (SMA) in
Mississippi.” National Center for Asphalt Technology. Auburn University.
Federal Highway Administration (FHWA) Legislation and Regulations, (1997). “Procedures for
Abatement of Highway Traffic Noise and Construction Noise.” Federal Aid Policy Guide
23-CFR-772.
Hamet, J.F., Deffayet, C., Palla, M.A., (1990). "Air Pumping Phenomena in Road Cavities.”
STU Info NO 794, Proceddings of International Tire/Road Noise Conference, Stockholm,
Sweden.
Hanson, D.I., James, R.S., NeSmith, C. (2004) “Tire/Pavement Noise Study.” National Center
for Asphalt Technology, Report 04-02.
International Organization for Standardization, (1997). “Measurement of the Influence of Road
Surfaces on Traffic Noise: Statistical Pass-By Method.” Part 1, ISO 11819-1.
Kaloush, K.E., Sotil, A., Biligiri, P.K., Golden, J.S. (2006) “The Dynamic Complex Modulus
(E*) Test as a Potential Indicator for an Asphalt Mixture’s Tire-Road Noise
Characteristics.” Transportation Research Board, Annual Meeting. Washington, DC.
2006.
Kropp, W., Kihlman, T., Forssen, J., and Ivarsson, L., (2007) “Reduction Potential of Road
Traffic Noise.” A Pilot Study of Applied Acoustics Chalmers University of Technology,
Alpha Print, Stockholm, ISBN 91 7082 770-2.
91

Kumar P., Chandra, S., Bose, S. (2005). “Comparative Study on Stone Matrix Asphalt Mixes
with Different Stabilizers.” Transportation Research Board, Annual Meeting.
Washington, DC. 2005.
Kutay, M.E., Kocak, S. (2010). “A New Numerical Modeling Approach for Sound Propagation
and Generation: the Lattice Boltzmann Method.” Transportation Research Board 89th
Annual Meeting. Washington, DC.
Lapcik, L., (1998). “Recycled Polymer Based Noise Absorbing Composite Materials.”
Proceedings of 5th International Conference on Composites Engineering, pp. 515, Las
Vegas, USA, ISBN 0-7333-0504-0.
Leasure, W.A. Jr, Bender, E.K., (1975). “Tyre-Road Interaction Noise.” Journal of the
Acoustical Society of America, 58, pp. 39-50.2.
Malcolm, J.C., Hanson, D., Li, Z., Karjatkar, R., Vissamraju, K.S., (2003). “Measurement of the
Acoustical and Mechanical Properties of Porous Road Surfaces and Tire/Pavement
Noise.” Transportation Research Record: Journal of the Transportation Research Board,
Volume 1981/2004, pp 16-22.
Meiarashi S., (1999). “ Researches on Low Noise Pavement in Japan.” Journal Acoustic Society
of Japan, Vol. 20, pp 19-27.
National Asphalt Pavement Association, (1999). “Designing and Constructing SMA MixturesState- of- the- Practice, Quality Improvement Series 122. NAPA. Landham, MD.
Narayanan, N., Jason, R.S., Jan, O., (2004). “Research on Enhanced Porosity Concrete.”
Cementing Future, Volume 15-1.
Nelson, T.J., Kohler E., Ongel, A., Rymer, B., (2008). “Acoustical Absorption of Porous
Pavement.” Transportation Research Board: Journal of the Transportation Research
Board, Volume 2058/2008, pp 125-132.
Rasmussen, R.O., (2006). “Surface Characteristics, Relating Pavement Texture to Tire-Pavement
Noise using RoboTex.” Presentation at the FHWA Technical Work Group Meeting, ISUFHWA-ACPA, 2006.
Rasmussen, R.O., Bernard, R.J., Sanberg, U, (2007). “The Little Book of Quieter Pavements.”
Technical Report No: FHWA-IF-08-004.
Sacramento County Department of Environmental Review and Assessment (SCDERA). (1999).
“Report on the Status of Rubberized Asphalt Traffic Noise Reduction in Sacramento
County.”
Sandberg, U., (1992). “Low Noise Road Surface. A State-of-the-Art Review.” Swedish Road and
Traffic Research institute, Linkoping. S-581/01.

92

Sandberg, U., (1999). “Low Noise Road Surfaces: A State-of-the-Art Review.” Journal Acoustic
Society of Japan.
Sandberg, U., (2001). “Tyre/road Noise – Myths and Realities.” International Congress and
Exhibition on Noise Control Engineering, Inter Noise 2001, The Hague, The
Netherlands.
Sandberg, U., Ejsmont, J.A., (2002). “Tyre/Road Noise Reference Book.” Informex, Sweden.
Sandberg, U., (2003). “The Multi-Coincidence Peak around 100 Hz in Tyre/Road Noise
Spectra.” Euronoise, Paper ID: 498, Naples.
Shatanawi K.M., Bradley, P.J., Thodesen, C., Amirkhanian, S.N., (2008). “The Effects of Crumb
Rubber Particles on Asphalt Pavement Traffic Noise Absorption.” TRIS, Transportation
Research Board 88th Annual Meeting, Report No 09-3197.
Von Meier, A., (1988). “Acoustically Porous Road Surfaces: Recent Experiences and New
Developments.” Proceedings of Inter-Noise 88, Avignon, France.
Von Meier, A., Van Blockland, G.J., Heerkens J.C.P., (1990). "Noise of Optimized Road
Surfaces and Further Improvements by Tire Choice.” Proceedings of INTEROC,
Gothenburg, Sweden.
Wayson, R. L. “Relationship Between Pavement Surface Texture and Highway Traffic Noise.”
NCHRP19 Synthesis of Highway Practice 268, Transportation Research Board of the
National Academies,20 Washington D.C., 1998.
Yamaguchi, M., Nakagawa, H., Mizuno, T., (1999).“Sound Absorption Mechanism of porous
Asphalt Pavement.” Journal Acoustic Society of Japan, Vol. 20, pp 29-43.
Zhu, H., Liu, C., Tom, K., Norasit, T., (2008). “Crumb Rubber Blends in Noise Absorption
Study.” Materials and Structures, Vol. 41, No. 2,383-390.
Zhu, H., Carlson, D.D., (1999). “A Spray Based Crumb Rubber Technology in Highway Noise
Reduction Application.” Rubber Pavement Association.

93