Asphalt Material Characterization, Accelerated Testing, and Highway Management

GEOTECHNICAL SPECIAL PUBLICATION NO.
190
ASPHALT MATERIAL CHARACTERIZATION,

ACCELERATED TESTING, AND
CONSTRUCTION MANAGEMENT
SELECTED PAPERS FROM THE 2009 GEOHUNAN INTERNATIONAL CONFERENCE
August 3–6, 2009

Changsha, Hunan, China
HOSTED BY
Changsha University of Science and Technology, China
CO-SPONSORED BY
ASCE Geo-Institute, USA
Asphalt Institute, USA
Central South University, China
Chinese Society of Pavement Engineering, Taiwan
Chongqing Jiaotong University, China
Deep Foundation Institute, USA
Federal Highway Administration, USA
Hunan University, China
International Society for Asphalt Pavements, USA
Jiangsu Transportation Research Institute, China
Korea Institute of Construction Technology, Korea
Korean Society of Road Engineers, Korea
Texas Department of Transportation, USA
Texas Transportation Institute, USA
Transportation Research Board (TRB), USA
EDITED BY
Lubinda F. Walubita, Ph.D.
Louw du Plessis, Ph.D.
Shin-Che Huang, Ph.D.
Geoffrey S. Simate
Zhaohui Liu, Ph.D.
Published by the American Society of Civil Engineers

Library of Congress Cataloging-in-Publication Data
Asphalt material characterization, accelerated testing, and construction management :

selected papers from the 2009 GeoHunan International Conference, August 3-6, 2009,
Changsha, Hunan, China / hosted by Changsha University of Science and Technology,
China ; co-sponsored by ASCE Geo-Institute, USA … [et al.] ; edited by Lubinda F.
Walubita … [et al.].
p. cm. -- (Geotechnical special publication ; no. 190)
Includes bibliographical references and indexes.
ISBN 978-0-7844-1042-4
1. Pavements, Asphalt--Testing--Congresses. 2. Pavements, Asphalt concrete--Testing--
Congresses. 3. Pavements, Asphalt--Design and construction--Congresses. 4. Pavements,
Asphalt concrete--Design and construction--Congresses. I. Walubita, Lubinda F. II.
Changsha li gong da xue. III. American Society of Civil Engineers. Geo-Institute. IV.
GeoHunan International Conference on Challenges and Recent Advances in Pavement
Technologies and Transportation Geotechnics (2009 : Changsha, Hunan Sheng, China)
TE270.A765 2009
625.8'5--dc22 2009022740
American Society of Civil Engineers

1801 Alexander Bell Drive
Reston, Virginia, 20191-4400
www.pubs.asce.org
Any statements expressed in these materials are those of the individual authors and do not necessarily
represent the views of ASCE, which takes no responsibility for any statement made herein. No
reference made in this publication to any specific method, product, process, or service constitutes or
implies an endorsement, recommendation, or warranty thereof by ASCE. The materials are for
general information only and do not represent a standard of ASCE, nor are they intended as a
reference in purchase specifications, contracts, regulations, statutes, or any other legal document.
ASCE makes no representation or warranty of any kind, whether express or implied, concerning the
accuracy, completeness, suitability, or utility of any information, apparatus, product, or process
discussed in this publication, and assumes no liability therefore. This information should not be used
without first securing competent advice with respect to its suitability for any general or specific
application. Anyone utilizing this information assumes all liability arising from such use, including
but not limited to infringement of any patent or patents.
ASCE and American Society of Civil Engineers—Registered in U.S. Patent and Trademark
Office.
Photocopies and reprints.

You can obtain instant permission to photocopy ASCE publications by using ASCE’s
online permission service (http://pubs.asce.org/permissions/requests/). Requests for 100
copies or more should be submitted to the Reprints Department, Publications Division,
ASCE, (address above); email: [email protected]. A reprint order form can be found at
http://pubs.asce.org/support/reprints/.
Copyright © 2009 by the American Society of Civil Engineers. All Rights Reserved.
ISBN 978-0-7844-1042-4
Manufactured in the United States of America.
Geotechnical Special Publications
1 Terzaghi Lectures 48 Soil Suction Applications in Geotechnical
2 Geotechnical Aspects of Stiff and Hard Engineering
Clays 49 Soil Improvement for Earthquake Hazard
3 Landslide Dams: Processes, Risk, and Mitigation
Mitigation 50 Foundation Upgrading and Repair for
7 Timber Bulkheads Infrastructure Improvement
9 Foundations & Excavations in Decomposed 51 Performance of Deep Foundations Under
Rock of the Piedmont Province Seismic Loading
11 Dynamic Response of Pile Foundations- 52 Landslides Under Static and Dynamic
Experiment, Analysis and Observation Conditions-Analysis, Monitoring, and
14 Geotechnical Aspects of Karst Terrains Mitigation
15 Measured Performance Shallow Foundations 53 Landfill Closures-Environmental
16 Special Topics in Foundations Protection and Land Recovery
17 Soil Properties Evaluation from Centrifugal 54 Earthquake Design and Performance of
Models Solid Waste Landfills
18 Geosynthetics for Soil Improvement 55 Earthquake-Induced Movements and
19 Mine Induced Subsidence: Effects on Seismic Remediation of Existing
Engineered Structures Foundations and Abutments
21 Hydraulic Fill Structures 56 Static and Dynamic Properties of Gravelly
22 Foundation Engineering Soils
23 Predicted and Observed Axial Behavior of 57 Verification of Geotechnical Grouting
Piles 58 Uncertainty in the Geologic Environment
24 Resilient Moduli of Soils: Laboratory 59 Engineered Contaminated Soils and
Conditions Interaction of Soil Geomembranes
25 Design and Performance of Earth Retaining 60 Analysis and Design of Retaining
Structures Structures Against Earthquakes
27 Geotechnical Engineering Congress 61 Measuring and Modeling Time Dependent
28 Detection of and Construction at the Soil Behavior
Soil/Rock Interface 62 Case Histories of Geophysics Applied to
29 Recent Advances in Instrumentation, Data Civil Engineering and Public Policy
Acquisition and Testing in Soil Dynamics 63 Design with Residual Materials:
32 Embankment of Dams-James L. Sherard Geotechnical and Construction
Contributions Considerations
33 Excavation and Support for the Urban 64 Observation and Modeling in Numerical
Infrastructure Analysis and Model Tests in Dynamic Soil-
34 Piles Under Dynamic Loads Structure Interaction Problems
35 Geotechnical Practice in Dam Rehabilitation 65 Dredging and Management of Dredged
37 Advances in Site Characterization: Data Material
Acquisition, Data Management and Data 66 Grouting: Compaction, Remediation and
Interpretation Testing
39 Unsaturated Soils 67 Spatial Analysis in Soil Dynamics and
40 Vertical and Horizontal Deformations of Earthquake Engineering
Foundations and Embankments 68 Unsaturated Soil Engineering Practice
41 Predicted and Measured Behavior of Five 69 Ground Improvement, Ground
Spread Footings on Sand Reinforcement, Ground Treatment:
42 Serviceability of Earth Retaining Structures Developments 1987-1997
43 Fracture Mechanics Applied to Geotechnical 70 Seismic Analysis and Design for Soil-Pile-
Engineering Structure Interactions
44 Ground Failures Under Seismic Conditions 71 In Situ Remediation of the Geoenvironment
45 In Situ Deep Soil Improvement 72 Degradation of Natural Building Stone
46 Geoenvironment 2000 73 Innovative Design and Construction for
47 Geo-Environmental Issues Facing the Foundations and Substructures Subject to
Americas Freezing and Frost
iii
74 Guidelines of Engineering Practice for 107 Soil Dynamics and Liquefaction 2000
Braced and Tied-Back Excavations 108 Use of Geophysical Methods in
75 Geotechnical Earthquake Engineering and Construction
Soil Dynamics III 109 Educational Issues in Geotechnical
76 Geosynthetics in Foundation Reinforcement Engineering
and Erosion Control Systems 110 Computer Simulation of Earthquake
77 Stability of Natural Slopes in the Coastal Effects
Plain 111 Judgment and Innovation: The Heritage
78 Filtration and Drainage in and Future of the Geotechnical
Geotechnical/Geoenvironmental Engineering Profession
Engineering 112 Soft Ground Technology
79 Recycled Materials in Geotechnical 113 Foundations and Ground Improvement
Applications 114 Soils Magic
80 Grouts and Grouting: A Potpourri of 115 Expansive Clay Soils and Vegetative
Projects Influence on Shallow Foundations
81 Soil Improvement for Big Digs 116 Deep Foundations 2002: An International
82 Risk-Based Corrective Action and Perspective on Theory, Design,
Brownfields Restorations Construction, and Performance
83 Design and Construction of Earth 117 Discrete Element Methods: Numerical
Retaining Systems Modeling of Discontinua
84 Effects of Construction on Structures 118 A History of Progress: Selected U.S. Papers
85 Application of Geotechnical Principles in in Geotechnical Engineering
Pavement Engineering 119 Soil Behavior and Soft Ground
86 Big Digs Around the World Construction
87 Jacked Tunnel Design and Construction 120 Grouting and Ground Treatment
88 Analysis, Design, Construction, and Testing 121 Probabilistic Site Characterization at the
of Deep Foundations National Geotechnical Experimentation
89 Recent Advances in the Characterization of Sites
Transportation Geo-Materials 122 Sinkholes and the Engineering and
90 Geo-Engineering for Underground Environmental Impacts of Karst
Facilities 123 Recent Advances in Materials
91 Special Geotechnical Testing: Central Characterization and Modeling of
Artery/Tunnel Project in Boston, Pavement Systems
Massachusetts 124 GeoSupport 2004: Drilled Shafts,
94 Performance Confirmation of Constructed Micropiling, Deep Mixing, Remedial and
Geotechnical Facilities Specialty Foundation Systems
95 Soil-Cement and Other Construction 125 Current Practices and Future Trends in
Practices in Geotechnical Engineering Deep Foundations
96 Numerical Methods in Geotechnical 126 Geotechnical Engineering for
Engineering: Recent Developments Transportation Projects
97 Innovations and Applications in 127 Recycled Materials in Geotechnics
Geotechnical Site Characterization 128 Soil Constitutive Models: Evaluation,
98 Pavement Subgrade, Unbound Materials, Selection, and Calibration
and Nondestructive Testing 129 Advances in Designing and Testing Deep
99 Advances in Unsaturated Geotechnics Foundations
100 New Technological and Design 130 Advances in Pavement Engineering
Developments in Deep Foundations 131 Contemporary Issues in Foundation
101 Slope Stability 2000 Engineering
102 Trends in Rock Mechanics 132 Advances in Deep Foundations: In
103 Advances in Transportation and Memory of Michael W. O’Neill
Geoenvironmental Systems Using 133 Earthquake Engineering and Soil
Geosynthetics Dynamics
104 Advances in Grouting and Ground 134 Soil Dynamics Symposium in Honor of
Modification Professor Richard D. Woods
105 Environmental Geotechnics 135 Erosion of Soils and Scour of Foundations
106 Geotechnical Measurements: Lab & Field
iv
136 Innovations in Grouting and Soil 165 Geosynthetics in Reinforcement and
Improvement Hydraulic Applications
137 Legal and Liability Issues in Geotechnical 166 Educational Activities in Geotechnical
Engineering Engineering
138 Site Characterization and Modeling 167 Geotechnics of Soil Erosion
139 Calibration of Constitutive Models 168 Grouting for Ground Improvement:
140 Slopes and Retaining Structures under Innovative Concepts and Applications
Seismic and Static Conditions 169 Soil and Material Inputs for
141 International Perspectives on Soil Mechanistic-Empirical Pavement Design
Reinforcement Applications 170 Probabilistic Applications in Geotechnical
142 Waste Containment and Remediation Engineering
143 Geomechanics: Testing, Modeling, and 171 Advances in Shallow Foundations
Simulation 172 Soil Improvement
144 Sinkholes and the Engineering and 173 Advances in Measurement and Modeling
Environmental Impacts of Karst of Soil Behavior
145 Seismic Performance and Simulation of 174 Designing Our Underground Space
Pile Foundations in Liquefied and 175 Field Measurements in Geomechanics
Laterally Spreading Ground 2007
146 Asphalt Concrete: Simulation, Modeling 176 Analysis of Asphalt Pavement Materials
and Experimental Characterization and Systems: Emerging Methods
147 Unsaturated Soils 2006 177 GeoCongress 2008: Geotechnics of Waste
148 Advances in Unsaturated Soil, Seepage, Management and Remediation
and Environmental Geotechnics 178 GeoCongress 2008: Geosustainability and
149 Site and Geomaterial Characterization Geohazard Mitigation
150 Soil and Rock Behavior and Modeling 179 GeoCongress 2008: Characterization,
151 Advances in Earth Structures: Research to Monitoring, and Modeling of GeoSystems
Practice 180 From Research to Practice in Geotechnical
152 Ground Modification and Seismic Engineering
Mitigation 181 Geotechnical Earthquake Engineering and
153 Foundation Analysis and Design: Soil Dynamics IV
Innovative Methods 182 Pavements and Materials:
154 Pavement Mechanics and Performance Characterization, Modeling, and
155 Underground Construction and Ground Simulation
Movement 183 Sinkholes and the Engineering and
156 Geomechanics II: Testing, Modeling, and Environmental Impacts of Karst
Simulation 184 Pavements and Materials: Modeling,
157 Computer Applications in Geotechnical Testing, and Performance
Engineering 185 Contemporary Topics in Deep Foundations
158 Contemporary Issues in Deep Foundations 186 Contemporary Topics in In-Situ Testing,
159 Case Studies in Earth Retaining Structures Analysis, and Reliability of Foundations
160 Dynamic Response and Soil Properties 187 Contemporary Topics in Ground
161 Embankments, Dams, and Slopes: Lessons Modification, Problem Soils, and Geo-
from the New Orleans Levee Failures and Support
Other Issues 188 Advances in Ground Improvement:
162 Problematic Soils and Rocks and In Situ Research to Practice in USA and China
Characterization 189 Characterization, Modeling, and
163 Geoenvironmental Engineering Performance of Geomaterials
164 Innovative Applications of Geophysics in
Civil Engineering
v
This page intentionally left blank
Preface
This Geotechnical Special Publication (GSP) of the ASCE is a compilation of 25

rigorously peer-reviewed papers that were presented at the GeoHunan International
Conference on Challenges and Recent Advances in Pavement Technologies and
Transportation Geotechnics; held from August 3rd to 6th, 2009 in Hunan, China. The
book addresses a variety of issues in pavement engineering, along with the latest
research advances in various areas including materials, design, testing, and, highway
construction. Topics covered in the book proceedings include:
(1) Asphalt Binder and Mixture Characterization,

(2) Recycled Asphalt Materials and Pavements,
(3) Accelerated Testing of Pavement Structures and Materials,
(4) Economic and Management of Highway Construction, and
(5) Embankment Stabilization and Tunnelling.
In particular, the data/information contained in this GSP book will be found valuable
to all those involved in the fields of Pavement and Geotechnical Engineering.
Acknowledgments
Bhaven Naik, University of Nebraska at Lincoln, NE, USA.

Joeng Oh, PhD, Texas Transportation Institute , TX, USA.
Joseph S. Mayunga, PhD, ARDHI University, Dar-Es-Salaam, Tanzania
Sheng Hu, PhD, Texas Transportation Institute, TX, USA
Sihle Ndlovu, SiRaL Consulting Engineers Ltd, JHB, South Africa.
Tao-ming Cheng, PhD, Chaoyang University of Technology, Taichung, Taiwan.
Xiaodi Hu, PhD, Texas Transportation Institute, TX, USA.
vii
Contents
Asphalt Binder and Mixture Characterization

Effects of Magnesium Hydroxide on Asphalt Pyrolytic Characteristics
and Kinetic Analysis.................................................................................................................1
Tao Xu and Xiaoming Huang
Research on Rheological Performance of SBS Modified Asphalt
at Low Temperature.................................................................................................................9
Liping Cao, Zejiao Dong, and Yiqiu Tan
Laboratory Performance Evaluation on Polymer Modified Porous
Asphalt Concrete....................................................................................................................15
Xin Qiu, Winggun Wong, and Changbin Hu
Laboratory Study of Porous Asphalt Mixture Made with Rubber Bitumen....................22
Jun Yang and Chaoen Yin
Research on Permanent Deformation Property of Asphalt Mixture
with Varying Temperature.....................................................................................................32
Liping Cao, Lijun Sun, and Zejiao Dong
Rutting in Asphalt Pavement under Heavy Load and High Temperature........................39
Qi-sen Zhang, Yu-liang Chen, and Xue-lian Li
Recycled Asphalt Materials and Pavements

Development of Asphalt Emulsion Cold In-Place Recycling Specifications .....................49
Jinhai Yan, Fujian Ni, Zhuohui Tao, and Jonathan Jia
Accelerated Testing of Pavement Structures and Materials

Mechanical Properties of Concrete Pavement with Different Isolation Layers
on Lean Concrete Base ..........................................................................................................56
Yu-liang Chen, Xue-lian Li, Qi-sen Zhang, and Zhi-gang Zhou
Evaluation of the Compaction Characteristics of Unbound Material
Using the Superpave Gyratory Compactor .........................................................................65
Nicholas Lambert, Kyle Denny, Beena Sukumaran, and Yusuf Mehta
Preliminary Design of Testing Segment for Accelerated Loading Facility
Based on Finite Element Simulation Analysis .....................................................................72
Zejiao Dong, Yiqiu Tan, and Fengchen Chen
Backcalculation of Anisotropic Pavement Properties Using Time History
of Embedded Gauge Readings ..............................................................................................79
Eyal Levenberg
Provisional Results from Accelerated Testing of Ultra Thin-Layer
Reinforced Concrete in South Africa....................................................................................86
L. du Plessis, P. J. Strauss, and A. Kilian
ix
Evaluation of the Effect of Tire Loads with Different Contact Stress Patterns
on Asphalt Rutting .................................................................................................................98
W. J. vdM. Steyn
Test Study on Abrasion of Skid-Resistant Textures on Concrete Pavements .................107
Yu Chen and Qisen Zhang
Economy and Management of Highway Engineering

Theory and Application of Total Project Management.....................................................116
Jianbo Yuan and Qisen Zhang
Decision-Making Model of Highway Route Plan Based on Entropy
and Entropy Weight Theory................................................................................................127
Yingxue Zhang, Chunhua Han, and Qisen Zhang
Application of GIS to Freeway Management System .......................................................133
Xiaoge Tian
Highway Bridge Construction Process Simulation Base on 4D Visualization ................138
ChengHan Zhou and WeiDong Wang
On the Construction Organization Design for the Widening of Expressways................146
Qingbo Li, Guanqun Lu, and Xiaohong Fu
Application of the Agent Construction System Management Mode
of Government Investment Highway Project ....................................................................153
Yingxue Zhang and Qisen Zhang
Economical Risk Analysis Model and Intelligent Optimization Methods Study
of the Soil Slope Based on the Reliability...........................................................................159
Shi-jiao Yang, Sheng Zeng, and Hai-fei Yang
A Study of the Cost Analysis of the Porous Pavement on a Freeway ..............................166
Chien-Ta Chen, Ching-Tsung Hung, Ming-Chen Yeh, and Jyh-Dong Lin
A Correlation Study of the Existing Bridges for Failure Analysis—Case Study
of Taichung County..............................................................................................................173
Tung-Tsan Chen, Yao T. Hsu, and Chun-Yuan Wang
Embankment Stabilization and Tunnelling

Application of Air Foam Stabilized Soil for Bridge-Embankment Transition
Zone in Thailand ..................................................................................................................181
Pichit Jamnongpipatkul, Montri Dechasakulsom, and Jiraroth Sukolrat
Simulation of NATM Tunneling Construction in Gravel Formation—Lessons
Learned from Pakuashan Highway Tunnel Project in Taiwan ........................................194
Chih-tsang Lin, Wen-ta Hsiao, Tao-ming Cheng, and Hsien-tang Wu
Indexes
Author Index.........................................................................................................................203
Subject Index ........................................................................................................................205
x
Effects of Magnesium Hydroxide on Asphalt Pyrolytic Characteristics and Kinetic
Analysis
Xu Tao1 and Huang Xiaoming2

1
Graduate Research Assistant, Southeast University, 2 Sipailou, Nanjing 210096, China; [email protected]
2
Professor, PhD, Southeast University, 2 Sipailou, Nanjing 210096, China; [email protected]
Abstract: Thermogravimetry experiments and thermal analysis kinetics are used to

analyze the effects of magnesium hydroxide on asphalt pyrolysis characteristics. The
pyrolysis models are developed based on experiments to evaluate the flame-retarded
mechanism of magnesium hydroxide to asphalt. The results indicate that asphalt
pyrolysis process in N2 is a single stage, and presents unimodal character in DTG
curves. Owing to endothermic decomposition and crystal water release of magnesium
hydroxide, the temperature rise of asphalt is inhibited. Pyrolysis of SBS modified
asphalt follows a one-dimensional pyrolysis mechanism model, but flame-retarded SBS
modified asphalt follows a three-dimensional pyrolysis mechanism model. After adding
magnesium hydroxide, asphalt pyrolysis kinetics parameters are raised and become
available to decrease reactive activity and rate. Based on this study, it is obvious that the
thermal stability of asphalt is improved by magnesium hydroxide under high
temperatures. Equally, the flame-retarded performance of asphalt is also enhanced by
using magnesium hydroxide.
INTRODUCTION
With the rapid development of highway construction in China, more and more road
tunnels are being built. Compared with concrete pavements, asphalt pavement has some
advantages. The asphalt pavement is getting applied increasingly in large scale road
tunnels. However, due to asphalt pyrolysis and combustion under tunnel fires, massive
poisons and smoke are released that result in serious secondary disasters. All these bring
great difficulties for personnel escape and fire rescue. So, studies on flame-retarded
asphalt pavements in tunnels is increasingly becoming an issue of great concern (1).
Because of the asphalt flammability, the usage of flame retardants in asphalt
pavement has potentials to decrease fire accidents in tunnels (2). Currently, the method
is to add various flame retardants into asphalt, such as organic bromides, antimony
trioxide, zinc borate, and so on. However, this leads to the formation of poisons and
smoke during tunnel fires (3). In addition, different flame retardants have generally
different flame-retardation mechanisms (4). The above flame-retarded asphalt research
mainly focused on asphalt flammability (5). However, the actual asphalt
flame-retardation mechanism is seldom regarded.
1
2 GEOTECHNICAL SPECIAL PUBLICATION NO. 190
Magnesium hydroxide (abbr.MH), as a kind of environmental protection type

flame retardant, has some advantages of flame retardancy, smoke suppression, and no
pollution. In this work, thermogravimetry (abbr.TG) experiments and thermal analysis
kinetics were used to discuss the influence of MH on asphalt pyrolysis characteristics.
But also, pyrolysis kinetics parameters are calculated to quantitatively evaluate the
flame-retardation effects of MH. Based on the above research, pyrolysis reaction
mechanism of flame-retarded asphalt is developed, and this is helpful to fully
understand the flame-retardation mechanism of metal hydroxide in asphalt.
EXPERIMENTAL DESIGN
Raw materials
SBS modified asphalt is produced by Shell Corporation. The flame retardant, namely
MH, is made by Guangzhou Yaxin Chemical Limited Company in China. The flame
retardant has some physical properties such as white powder, average particle size of
2.2 to 3.0ȝm, density of around 2.36g/cm3, non-toxic, competitive price, crystal water
content of 31.0%, and flame retardancy characteristics.
Sample preparation
SBS modified asphalt is put in a stainless steel cup to heat on an electric furnace. When
the asphalt is heated up to (170 ± 5) oC, flame retardant powder is added into the hot
asphalt in weight percent amounts of 15%, 20% and 25%, respectively. First, stir for 15
minutes at that temperature, then the turbine stirrer (as shown in FIG. 2) is used to stir at
the high speed of 5000(rpm) for 30 minutes, and then to stir at a low speed of 500(rpm)
for 15 minutes to expel air bubbles from the asphalt. Finally, flame-retarded asphalt is
placed in clean vessels. During the asphalt cooling process, hand stirring is done
to prevent segregation.
Experimental instrumentations
The experiment was on a ThermoGravimetry-Differential Thermal Analyzer (TG-DTA)
made by SETARAM Company of France. The instrumentation has some technology
parameters such as a thermo-balance precision of 0.1ȝg, sample weight range of 0 to
around 200mg, experimental temperature range of room temperature to around 1600 oC,
heating rate of 0.01 to 99.99 oC /min and experimental temperature precision of ± 2 oC.
Temperature control and data acquisition are accomplished automatically by computer
on line.
Thermogravimetry experiments
The TG experiments are on the TGA-92 type thermal analyzer system. The samples are
heated from room temperature to around 700°C at 5°C/min under a nitrogen atmosphere.
Non-isothermal kinetic experiments are carried out using approximately 10mg of
GEOTECHNICAL SPECIAL PUBLICATION NO. 190 3
flame-retarded SBS modified asphalt (6).
RESULTS AND DISCUSSION
Influence of MH on pyrolysis characteristics of asphalt

The depolymerization reaction, the end chain scission, and the random chain breaking,
are the main manifestations of asphalt pyrolysis in N2. TG and DTG curves under
different flame retardant (MH) doses are shown respectively in FIG.1.

0+
0+
0+
0+
7*

R
7HPSHUDWXUH &

'7*PJg PLQ

0+
0+

0+

0+

7HPSHUDWXUHR&
FIG. 1. TG-DTG curves of asphalt pyrolysis under different MH doses.
FIG.1 shows the thermal weight loss of asphalt begins at about 280 oC, and over
90% of all the weight loss occurs in the temperature rang of 300oC ~ 500oC. Due to a lot
of asphalt molecular chain breaking during the asphalt pyrolysis process, numerous
different of complicated components are released including carbon monoxide(CO),
carbon dioxide (CO2), methane (CH4), saturated hydrocarbons, unsaturated hydrocarbon,
aromatic mixture, etc. All these are reflected obviously by the concave unimodal DTG
curves in FIG.1. When asphalt samples are heated up to 430 oC ~ 480 oC, the weight loss
reaches the maximum. After the temperature is over 500 oC, the weight loss of samples
is lesser, and the remainder of the constituent composition is mainly un-decomposed
solid coke and ash.
Kinetics analysis of asphalt pyrolysis

Based on the above experiments, kinetics parameters and reaction velocity constant can
be calculated. Thermal analysis kinetics parameters including activation energy and
frequency factor (or pre exponential factor), are used to reflect quantitatively the law of
asphalt thermal reaction during temperature fluctuations. It is helpful to study asphalt
flame-retarded mechanisms and pyrolysis behaviors.
Pyrolysis mechanism models

Pyrolysis reaction velocity is a function of the heating rate, temperature, and pyrolysis
products mass. When non-isothermal reaction is assumed to divide into an infinite
number of short time segments, the reaction in every single segment is taken as an
isothermal reaction. So, it is reasonable that the pyrolysis reaction can be described by
the Arrhenius method (7).
According to the Arrhenius equation, the pyrolysis reaction rate equation of asphalt may
be written as:
dα / dt = k ⋅ f (α ) = Ae( − E / RT ) f (α ) (1)
where: A –frequency factor(or pre exponential factor), s–1; E – activation energy,

kJ·mol–1; R – universal gas constant, kJ·mol–1·K–1; t – time of pyrolysis process, s; T
–pyrolysis temperature at time t, K; Į – residue matter mass, kg; f(Į) – function related
to reaction rate and Į.
When constant heating rate ȕ is assumed as ȕ=dT/dt, equation (1) becomes:
dα A
= e( − E / RT ) dT (2)
f (α ) β
α 1
Defining g (α ) = ∫
0 f (α )
d α , then equation (2) converts to:
A T
g (α ) =
β ∫
T0
e ( − E / RT ) dT (3)
Where: g (Į) – integral function; T0 – initial temperature.

Coats-Redfern method was used to solve the kinetic equations in this study (6);
which is based on an approximate integral model type described in equation (4):
⎡ g (α ) ⎤ ⎡ AR 2 RT ⎤ E
ln ⎢ 2 ⎥ = ln ⎢ (1 − ) − (4)
⎣ T ⎦ ⎣ β E E ⎥⎦ RT
Common integral functions g (Į) used to study the kinetic analysis are shown in Table.1
(8). As is known, when a kinetic model is applied to describe the pyrolysis reaction of
material, it must be verified by various methods. Kinetic analysis is one of methods to
verify the rationality of built models. The linear association degree of ln[g(Į)/T2] versus
1/T is an evaluation of built models.
From equation (4), owing to 2RT/E<<1, 2RT/E can be neglected. This makes
ln[AR(1-2RT/E)/(ȕE)] approximately equal to ln(AR/ȕE). So, when the selected
mechanism function g(Į) is appropriate, ln[g(Į) /T2] versus 1/T curve should
approximately be a line of least square fit. Moreover, the g(Į) with maximum
correlation coefficient R2 is the mechanism function of the pyrolysis reaction. Then, the
slope of the regressive line is -E/R, and the intercept includes a frequency factor
A. Therefore, kinetics parameters(E and A) can be calculated based on the slope and
intercept of the regression line. Fitting results of correlation coefficient of the asphalt
pyrolysis mechanism functions under different MH doses are shown in Table.2.
Table 1. Common Pyrolysis Mechanism and Integral Mechanism Functions.

Mechanism function
Pyrolysis mechanism Symbol
g(Į)
One-dimensional diffusion (Parabolic law) D1 Į2
Two-dimensional diffusion (Valensi equ.) D2 (1-Į)ln(1-Į)+Į
Three-dimensional diffusion (Jander equ.) D3 [1-(1-Į)1/3]2
Three-dimensional diffusion (G-B equ.) D4 (1-2Į/3)-(1-Į)2/3
One-dimensional phase boundary reaction
R1 Į
(Maple powder law)
Two-dimensional phase boundary reaction
R2 1-(1-Į)1/2
(Shrinking column shape symmetry)
Three-dimensional phase boundary reaction
R3 1-(1-Į)1/3
(Shrinking sphere shape symmetry)
Two order chemical reaction F2 (1-Į)-1
Three order chemical reaction F3 (1-Į)-2
Nucleation and nuclei growth
A1 -ln(1-Į)
(Maple single law, one order)
A2 [-ln(1-Į)]2/3
(Avrami-Erofeev equ.)
A3 [-ln(1-Į)]1/2
A4 [-ln(1-Į)]1/3
A5 [-ln(1-Į)]1/4
Table 2. Correlation Coefficient of Asphalt Pyrolysis Mechanism Functions Fitting

Under Different MH Doses.
g(Į) MH Doses 0% 15% 20% 25%
D1 0.9923 0.9894 0.9894 0.9859
D2 0.9889 0.9924 0.9920 0.9909
D3 0.9828 0.9927 0.9911 0.9890
D4 0.9871 0.9942 0.9934 0.9921
R1 0.9906 0.9878 0.9881 0.9842
R2 0.9847 0.9928 0.9916 0.9905
R3 0.9812 0.9919 0.9906 0.9883
F2 0.4167 0.6262 0.5506 0.4573
F3 0.5177 0.6625 0.5840 0.4774
A1 0.9721 0.9860 0.9785 0.9626
A2 0.9694 0.9849 0.9769 0.9600
A3 0.9662 0.9836 0.9751 0.9572
A4 0.9580 0.9805 0.9707 0.9506
A5 0.9458 0.9761 0.9647 0.9422
By optimal identification of the mechanism function (as shown in Table 2), the
asphalt pyrolysis reaction follows a D1 model in fluid N2, namely, one dimensionnal
diffusion mechanism model. It shows that modified asphalt pyrolysis belongs to a
diffusion process control mechanism and represent a deceleration shape Į-t curve,
which follows the parabolic rule (6). Accordingly, the reaction mechanism function of
modified asphalt is as follow: g(Į)=Į2, and reaction rate function is as follow: f
(Į)=Į-1/2.
However, pyrolysis reaction of flame-retarded modified asphalt follows a D4 model,
namely, a three dimensional diffusion mechanism model. It shows that flame-retarded
modified asphalt pyrolysis belongs also to a diffusion process control mechanism, but
represents a spherical symmetry and a deceleration shape Į-t curve that follows the
Ginstling-Broushtein equation(6). In addition, the reaction mechanism function of
flame-retarded modified asphalt is as follow: g(Į)=(1-2Į/3)-(1-Į)2/3, and the reaction
rate function is as follow: f (Į)=3/2·[(1-Į)-1/3-1]-1.
Pyrolysis kinetics parameters analysis
When pyrolysis mechanism function of asphalt is determined, kinetics parameters are

calculated based on the slope and intercept of the fitting line. Calculated results are
shown in Table.3.
Table 3. Calculation Results of Asphalt Pyrolysis Kinetics Parameters in N2

Kinetics MH doses
parameters 0% 15% 20% 25%
E ( kJ·mol–1) 169.36 214.17 232.86 250.97
-1
A(s ) 9.79E+09 4.83E+11 1.08E+13 4.37E+14
From Table.3, when flame retardant (MH) is mixed into asphalt, frequency factor
(A) and activation energy (E) increase as an obvious function of the MH amount added.
This implies that flame-retarded asphalt needs more energy for activating the thermal
decomposition reaction and consequently, inhibits asphalt pyrolysis. This leads to a
decrease in the reaction activity. So, compared with modified asphalt, flame-retarded
modified asphalt is more difficult to decompose. That is to say, thermal stability and
flame-retarded effect are better under high temperature environments. Thus, the kinetic
parameters in Table.4 can evaluate quantitatively the flame retardation effect of MH.
The main reason is that MH releases crystal water and absorbs heat during MH
decomposing. This minimizes the temperature rise in the asphalt matrix and decreases
the reaction rate, consequently retarding asphalt pyrolysis. On the other hand, the char
yield ratio increases due to active magnesia, and the charring layer hinders further
pyrolysis of the asphalt.
CONCLUSIONS
(1) Due to massive asphalt molecular chain breaking, asphalt decomposition process is
a single stage pyrolysis in N2. The DTG curves present an obvious unimodal
character. Temperature rise of asphalt matrix is inhibited by the endothermic
decomposition and crystal water release of MH. This improves thermal stability and
delay the release of flammable volatiles under high temperatures.
(2) The pyrolysis reaction of modified asphalt follows a one dimensional diffusion
mechanism model (D1). The model belongs to a diffusion process control
mechanism that follows a parabolic rule, with its reaction mechanism function given
as g(Į)=Į2. However, pyrolysis reaction of flame-retarded modified asphalt follows
a three dimensional diffusion mechanism model (D4). The model also belongs to a
diffusion process control mechanism, but follows Ginstling-Broushtein equation and
its reaction mechanism function is given as g(Į)=(1-2Į/3)-(1-Į)2/3.
(3) MH has an obvious influence on the kinetic parameters of asphalt pyrolysis.
Analytical results indicate that flame retardant (MH) inhibits asphalt pyrolysis, and
decreases both the reaction activity and rate. All these factors cause the thermal
stability of flame- retarded asphalt to be stably better under high temperatures.
ACKNOWLEDGEMENTS
This study was funded by the Research Program of Guangdong Provincial

Communications Department (No. GCKY-1-24), China.
REFERENCES
(1) Gao P Z, Liu S L, Chow W K, et al. (2004). Large eddy simulations for studying
tunnel smoke ventilation [J].Tunnel and Underground Technology, 19 (6):577-586;
(2) Shao peng Wu, Liantong Mo, Peiliang Cong, et al. (2008) Flammability and
rheological behavior of mixed flame retardant modified asphalt binders [J].Fuel,
87:120-124;
(3) Hu Shuguang, Zhang Houji Wang Jiaolan. (2006). Research on Alkaline Filler
Flame-Retarded Asphalt Pavement [J]. Journal of Wuhan University of Technology
(Mater. Sci. Ed.), 21(3):146-148;
(4) Demir H, Arkis E, Balkose D, Ulku S. (2005). Synergistic effect of natural zeolites
on flame retardant additives[J]. Polym Degrad Stab, 89(3):478–83;
(5) Lonnermark A, Ingason H. (2005). Gas temperatures in heavy goods vehicle fires in
tunnels [J]. Fire Safe, 40(6):506–527;
(6) Hu Rongzu, Shi Qizhen. (2001). Thermal Analysis Kinetics [M]. Beijing: Science
Press, 127-131;
(7) X.X. Han, X.M. Jiang, Z. G. Cui. (2006). Thermal Analysis Studies on Combustion
Mechanism of Oil Shale [J]. Journal of Thermal Analysis and Calorimetry, 84:
3631–3636.
(8) N.A. Liu, W.C. Fan. (2002). Kinetic modeling of thermal decomposition of natural
calluses materials in air atmosphere[J].Journal of Analytical and Applied Pyrolysis,
63:303-325.
Research on Rheological Performance of SBS Modified Asphalt at Low
Temperature
Cao Liping1, Dong Zejiao2, Tan Yiqiu3

1
PH.D. of Civil Engineering; School of Transportation Science & Engineering, Harbin Institute of
Technology; No.202, Haihe Road, Harbin 150090, China; PH(086)0451-86282120; [email protected]
2
Technology; No.202, Haihe Road, Harbin 150090, China; PH(086)0451-86282120; [email protected]
3
Professor of Civil Engineering; School of Transportation Science & Engineering, Harbin Institute of
Technology; No.202, Haihe Road, Harbin 150090, China; [email protected]
ABSTRACT: In this paper, phase angle( δ ) and viscous modulus( G′′ ) were used to
study the low temperature rheological performance(LTRP) of nine kinds of Styrene-
butadiene-styrene (SBS) modified asphalts, with three penetration grades of base
asphalts (90, 110, and 130) and three contents of weight (3%, 4.5% and 6%) of linear
SBS modifiers. The LTRP of modified asphalts were tested using the Gemini 150
dynamic shear Rheometer produced by Bohlin Company at the temperatures ranging
from -10 °C to 10 °C at intervals of 5 °C. Then, the influence of the modifier contents
on the LTRP of SBS modified asphalt was analyzed. The result indicated that the
modifier contents have a great influence, which is apparently confined by base
asphalt’s penetration grade. The higher the grade is, the greater the influence. Lastly,
the influence of the base asphalt’s grade on the LTRP of SBS modified asphalt was
investigated. The analysis showed that it plays an important role in the LTRP. What’s
more, the impact degree is dependent on the modifier’s contents. The more the content
is, the greater the impact degree is.
INTRODUCTION
Low-temperature cracking can be classified as one of the main distresses for

asphalt pavements in the Northern China. Therefore, asphalt technologists have tried
many methods to solve this problem. SBS modification has been the most common
method used in China, in recent years. As we all know, SBS modified asphalt is a
complex compound, and the LTRP of SBS modified asphalt is influenced by many
factors (e.g., SBS modifier’s structure and contents, base asphalt’s penetration grade
and oil sources, processing methods, processing temperature, etc). As a result, to what
extent these factors affect the LTRP of SBS modified asphalt can be a significant
problem.
9
In recent years, there are several ways to evaluate the LTRP of asphalt. In the
Superpave specification and others’ researches, different techniques were used to
characterize the LTRP of modified asphalts (1-8). However, the effect of polymer
modification on the LTRP of modified asphalts has been more undecided or
controversial (4, 5-8), and remained so even after the introduction of the original
Superpave low temperature parameters, creep stiffness, S, and the m-value from the
bending beam Rheometer (BBR) and direct tension test (DTT) failure stress values
and the critical cracking temperature (Tcritical) from the DTT. In this paper, two
rheological basic parameters (į and viscous modulus) were chosen to assess the LTRP
of nine kinds of SBS modified asphalts.
MATERIALS AND EXPERIMENT
In this research, storage-stable SBS polymer modified asphalts were obtained from
Beijing Asphalt Research and Development Center, KOCH Material. The penetration
grades of base asphalts with the same oil source were receptively, 90, 110, and 130.
Three contents (3%, 4.5% and 6%) of the SBS linear modifiers were used for each
base asphalt with different penetration grades. The details of 9 types of SBS modified
asphalts prepared by high-speed shear machine at 180 °C are shown in Table 1,
including the names of SBS modified asphalts, modifier’s contents, and base asphalt’s
grade, penetration, ductility, and softening point. As shown in Table 1, for the same
base asphalt’s grade, the ductility increases with SBS content: however, the softening
point decreases with SBS content.
Table 1ˊ Properties of 9 kinds of SBS modified asphalts
Base asphalt’s SBS content, Penetration, 0.1mm Softening Ductility, cm

Name
grade wt% 5°C 15°C 25°C point, °C 5°C
A1 90 3.0 7.8 22.3 77.8 63.5 67.4
A2 90 4.5 7.7 21.3 55.0 75.7 57.5
A3 90 6.0 9.0 22.0 51.4 82.9 49.1
B1 110 3.0 9.3 22.8 64.2 60.6 73.7
B2 110 4.5 9.0 25.0 63.3 69.2 74.4
B3 110 6.0 11.3 28.0 62.4 81.6 72.1
C1 130 3.0 10.8 31.7 81.8 57.9 107.7
C2 130 4.5 13.3 36.0 87.7 77.8 72.3
C3 130 6.0 14.7 35.0 85.6 78.3 59.2
The samples were tested using the Gemini 150 dynamic shear Rheometer, which
can test the performance at low temperatures up to -20 °C with the same resolution,
with parallel plate geometry (25 mm in diameter) produced by Bohlin Company.
About 1.0g sample was put onto the lower plate. After the sample was heated to flow,
the upper parallel plate was lowered to contact tightly with the sample and the sample
trimmed. The final gap was adjusted to 1.0 mm. The temperatures ranged form -10
to 10 °C at intervals of 5 °C and the frequency was 0.1Hz.
Several rheological parameters can be received from the experiment. In this paper,
δ and G′′ were used to appraise the LTRP of modified asphalts for the reason that
they can reflect the LTRP of SBS modified asphalts. The better the LTRP is, the
higher δ and G′′ are. Therefore, the influence of modifier’s contents and base
asphalt’s grade on the LTRP will be researched.
INFLUENCE OF MODIFIER’S CONTENTS ON THE LTRP
Polymer modifier’s contents not only relate to the cost of asphalt mixtures, but
also involve in whether SBS modifiers form net-structures or not. As a result, linear
SBS polymer was chosen to analyze the effect of SBS polymer modifier’s contents.
Figure 1 is a summary of the graphs of δ and G′′ versus temperature for a set of
binders of three modifier contents (3%, 4.5%, and 6%) and the same base asphalt
grade (90). As shown in Figure 1, δ increases with temperature and modifier’s
contents. However, there is a peak value for curves of G′′ with temperature. The peak
temperature, which can be defined as glass transition temperature, indicates the LTRP.
The lower the peak temperature is, the better the LTRP is. It can be seen from Figure1
that the higher the contents are, the lower the peak temperature is. That means the
LTRP will be better with higher modifier’s contents.
60.0 6.20
A1
50.0
A2 6.00
40.0 A3
į(degree)
5.80
logGą
30.0
5.60
20.0 A1
5.40 A2
10.0 A3
0.0 5.20
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
Temperature(ć) Temperature(ć)
FIG. 1. Graphs of δ and G′′ versus temperature for asphalts A1, A2 and A3.
Figures 2 and 3 show the graphs of δ and G′′ versus temperature for three
modifier contents (3%, 4.5% and 6%) and two base asphalt grades (110 and 130).
Graphs of δ and G′′ versus temperature have the same trends as Figure 1, but the
distance of graphs between three contents becomes larger. In addition, the distance
increases with base asphalt grade. What’s more, there is a more clear decrease in the
peak temperatures for base asphalt grade of 110 and 130.
As discussed above, the contents of SBS modifiers have a great influence on the
LTRP of SBS modified asphalts, and that increasing the contents will be an effective
way to improve the LTRP. At the same time, it is very interesting to find that the
influence of modifier’s contents is confined by the base asphalt’s grade, and that the
influence degree of contents increases with the grades increasing.
60.0 6.20
B1
50.0 B2 6.00
40.0 B3
į(degree)
5.80
logGą
30.0
5.60
20.0 B1
5.40 B2
10.0 B3
0.0 5.20
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
FIG. 2. Graphs of δ and G′′ versus temperature for asphalts B1, B2 and B3.
60.0 6.20
C1
50.0 C2 6.00
C3
40.0
į(degree)
5.80
logGą
30.0
5.60
20.0 C1
5.40 C2
10.0 C3
0.0 5.20
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
FIG. 3. Graphs of δ and G′′ versus temperature for asphalts C1, C2 and C3.
INFLUENCE OF BASE ASPHALT’S GRADE ON THE LTRP
Another important influencing factor of the LTRP of SBS modified asphalts is

base asphalt’s grade. From some point of view, base asphalt’s grade reflects the LTRP
of base asphalt binders. Therefore, linear SBS modifier was chosen to analyze the
effect of base asphalt’s grade. For three contents of modifiers (3%, 4.5%, and 6%),
influence of grades were discussed separately as followings.
Figure 4 depicts the graph of G′′ versus temperature for a set of binders of
different base asphalt’s grades (90, 110, and 130) with a similar content of 3%. δ
increases with temperature and base asphalt’s grade. However, there is a peak value
for the graphs of G′′ . The higher the base asphalt’s grade is, the lower the peak
temperature is. That means the LTRP will be better with higher base asphalt’s grade.
60.0 6.40
A1
50.0 6.20
B1
40.0 C1 6.00
į(degree)
logGą
30.0 5.80
A1
20.0 5.60 B1
10.0 5.40 C1
0.0 5.20
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
FIG. 4. Graphs of δ and G′′ versus temperature for asphalts A1, B1 and C1.
Figures 5 and 6 show the graphs of δ and G′′ versus temperature for a set of
binders of different base asphalt grades (90, 110, and 130) and the same contents (i.e.,
4.5% and 6%). Graphs of δ and G′′ versus temperature have the similar trends as
Figure 4, but the distance between the three graphs of grade becomes larger.
Furthermore, the distance increases with modifier’s contents.
As analyzed above, the base asphalt’s grade has a significant impact on the LTRP
of SBS modified asphalt, and that choosing high penetration grades will be a better
way to improve the LTRP. Moreover, it is obviously that the influence of the base
asphalt’s grade is confined by modifier’s contents, and that the degree of influence of
the contents increases as the grades increases.
60.0 6.40
A2
50.0 6.20
B2
40.0 C2 6.00
į(degree)
30.0 logGą 5.80

20.0 5.60 A2
B2
10.0 5.40 C2
0.0 5.20
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
60.0 6.40
A3
50.0 6.20
B3
40.0 C3 6.00
logGą
į(degree)
30.0 5.80
20.0 5.60 A3
B3
10.0 5.40 C3
0.0 5.20
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
To sum up, the LTRP of SBS modified asphalts is effected by both the base
asphalt’s grade and the SBS modifier’s contents. The base asphalt binders supply the
foundation for SBS polymer modifiers and the modifiers impacts the degree of
influence. The base asphalt binders with higher penetration grades will get better
results of modification; on the contrary, those with lower penetration grades will get
worse results of modification. Therefore, the modifier will have greater impact on the
LTRP only if the base asphalt’s grade is higher. In contrast, modifiers won’t have any
obvious impact with lower grades of the base asphalt binders even if the modifier’s
contents are considerably high. In other words, the higher base asphalt’s penetration
grades will be a better choice when aiming at improving the LTRP of the SBS
modified asphalts.
SUMMARY AND CONCLUSIONS
Based on the testing and analysis presented, the following conclusions may be
made:
1. SBS modifier’s contents have a great influence on the LTRP, and that increasing
the contents will be an effective way to improve the LTRP. Moreover, the influence of
modifier’s contents is confined by base asphalt’s grades, and that the influence degree
of contents increases with the grades increasing.
2. The base asphalt’s grades have a significant impact on the LTRP, and that
choosing high penetration grades will be a better way to improve the LTRP.
Additionally, the influence of base asphalt’s grades is confined by SBS modifier’s
contents, and that the influence degree of contents increases as the grades increasing.
3ˊBase asphalt binders supply the foundation for SBS polymer modifiers, and the
modifiers provide the influence degree. Therefore, modifier will have greater impact
on the LTRP only if base asphalt’s grade is higher. In a word, higher base asphalt’s
penetration grades will be better choice when aiming at improving the LTRP of SBS
modified asphalts.
ACKNOWLEDGEMENTS
This project was supported by the Development Program for Outstanding Young
Teachers of the Harbin Institute of Technology (HITQNJS.2008.047) and the
Research Fund for the Doctoral Program of Higher Education (200802131011).
REFERENCES
(1) AI. (1994). Superpave Performance Graded Asphalt Binder Specification and
Testing. Superpave Series No. 1 (SP-1), Asphalt Institute, Lexington/KY.
(2) Brule, B. and M. Maze. (1995). Application of SHRP Binder Tests to the
Characterization of Polymer Modified Bitumens. AAPT, Vol. 64: 367-392.
(3) Stock, A.F. and W. Arand.(1993). Low-temperature Cracking in Polymer
Modified Binders. AAPT, Vol. 62: 23-53.
(4) Fortier, R. and T.S. Vinson. (1998). Low-temperature Cracking and Aging
Performance of Modified Asphalt Concrete Specimens. Transportation Research
Record 1630, TRB, Washington, D.C.: 77-86.
(5) Hoare, T.R. and S. Hesp. (2000). Low-temperature Fracture Testing of Asphalt
Binders, Regular and Modified Systems. Transportation Research Record 1728,
TRB, National Research Council, Washington, D.C.:36-42.
(6) Hesp, S., T. Terlouw and W.C. Vonk. (2000). Low-temperature Performance of
SBS-Modified Asphalt Mixes. AAPT, Vol. 69: 540-567.
(7) Button, J.W., and C.P. Hastings. (1998). How Well Can New Binder Tests Predict
Cracking? The Canadian Technical Asphalt Association Annual Meeting,
Vancouver: 48-72.
(8) Anderson, R.M., D.E. Walker and P.A. Turner. (1999). Low-Temperature
Evaluation of Kentucky Performance-Graded 70-22 Asphalt Binders.
Transportation Research Record 1661, TRB, Washington, D.C.: 69-74.
Laboratory Performance Evaluation on Polymer Modified Porous Asphalt
Concrete
Xin Qiu1, Winggun Wong2,Changbin Hu3
1
PhD, College of Civil Engineering, Fuzhou University, Fuzhou, P.R.China, 350108.
Email: [email protected]
2
Associate Professor, Civil and Structural Engineering Department, Hong Kong Polytechnic University, Hong
Kong, P.R.China. Email: [email protected]
3
Associate Professor, College of Civil Engineering, Fuzhou University, Fuzhou, P.R.China, 350108.
ABSTRACT: This paper presents the details of the laboratory performance evaluation
of the three types of polymer modified binders and their respective porous asphalt
concrete. The same gradation was applied to all the mixtures and the base bitumen was
Shell 60/70. The binder test results indicate that the SBR and SBS modified binders
have the smaller temperature susceptibility and exhibit better performance in resisting
ageing, rutting, and fatigue than the conventional penetration 60/70 binder. When
evaluated in a series of laboratory tests designed to assess the fundamental mechanical
properties and durability performance of porous asphalt concrete, the results show that
the SBS modified porous asphalt concrete exhibit the most overall satisfactory
performance with respect to stiffness, dynamic creep deformation, rutting, and
moisture resistance damage among the three types of porous asphalt concretes studied.
Based on the lab results and the cost-effectiveness analysis, the SBS modified porous
asphalt concrete is recommended as the most favorable material to be used as a friction
course in Hong Kong.
KEY WORDS: Polymer modified binder, Porous asphalt concrete, Performance
evaluation, Friction course
INTRODUCTION
Porous asphalt concrete(PAC) is one of the safest pavement surfacing materials used as
a friction course by improving friction resistance, minimizing hydroplaning, reducing
splash and spray, improving night visibility and lowering pavement noise levels
(Golebiewski ,2003). Although it has been a long time that extensive study of PAC had
been conducted in the UK, the current use of PAC is not satisfactory because of the
less durability and the earlier replacement of the surfacing than the conventional
asphalt mixtures (Nicholls, 1997).
15
PAC has been used in the United States of America since at least 1944, but their use
across the country was limited for several reasons. These mixtures require a different
mix design method and special construction and maintenance considerations. However,
they will tend to fail suddenly at the end of their design life (Huber, 2000). PAC is
comparatively new in South America, particularly in Argentina where it started in 1997
with the first trial on Provincial Road No.2 in the Buenos Aires Province, east central
of Argentina (Bolzan, 2001).The overseas experiences show that the search for the
superior durable PAC has become a major goal for the asphalt industries all over the
world and every effort towards its enhancement is worthwhile.
In order to improve service performance and reduce the initial damages to PAC
pavements, the ultimate purpose of this study is to conduct performance tests on the
three types of binders and their respective PAC, to compare their mechanical
properties and durability performances, and then, to recommend the best suitable PAC
materials to be used widely in the vital highways in Hong Kong.
BINDER TESTING AND EVALUATION

Binders Used in the Investigation
The three types of binders were studied including conventional penetration 60/70
binder, conventional polymer (SBR) modified binder, and pre-blended polymer (SBS)
modified binder.
Test Programme
Four tests (penetration, softening point, viscosity, and DSR) were performed at three
conditions of each of the three types of binders. The test plan is accordingly illustrated
in Fig.1.
FIG.1. Test Procedure on Binders.

RESULTS AND ANALYSIS

Penetration Test
The test results of penetration for the three studied binders are shown in Table1. It
indicates that the two polymer modified binders can provide better resistance against
deformation than the conventional binder in the high temperature domain.
Table 1. Penetration Testing Results.
Test Condition at 25ć
Binder Type
Un-aged (0.1mm) RTFO(0.1mm) PAV(0.1mm)
60/70 67 36 22
SBR 46 34 28
SBS 50 31 25
Softening Point Test

The test results of the softening point for the three studied binders are given in the
Table 2. The results shows that the temperature susceptibility of the two modified
binders is less than that of 60/70 binder, i.e., their softening point temperatures are
comparatively higher.
Table 2 Softening Point Testing Results.
Test Condition
Binder Type
Un-aged (ć) RTFO (ć) PAV (ć)
60/70 49.3 54.1 61.1
SBR 58.2 60.3 67.2
SBS 57.3 59 64.1
Viscosity Test
The test results of viscosity for the three studied binders are shown in the Table3.
Based on their comparatively higher viscosity values in magnitude at 135 °C, i.e. stiffer,
the results in Table 3 indicate that the two polymer modified binders have lower
temperature susceptibility than the 60/70 binder.
Table 3. Viscosity Testing Results.
Test Condition at 135ć
Binder Type
Un-aged (10-3Pa s) RTFO (10-3Pa s) PAV (10-3Pa s)
60/70 475 614.2 445
SBR 1053 1255 532.5
SBS 762 1427 231.7
Dynamic Shear Rheometer (DSR) Test

The DSR test results of the three studied binders are shown in the Fig.2. The critical
high temperature of the binder is defined as the temperature, at which the stiffness
value, G*/sinį, of the binder just exceeds 1.0kPa and 2.2kPa on the original condition
and RTFO condition, respectively. The critical high temperatures of the three studied
binders are 70.7°C for 60/70 binder, 81.4°C for SBR, and 81.8°C for SBS, respectively.
The values of SBR and SBS are comparable and are much higher in magnitude than
that of the 60/70 binder. It means that two modified binders have higher resistance to
permanent deformation under the condition of high pavement temperature when
compared to the conventional 60/70 binder.
25
60
60/70
60/70
SBR SBR
20 50
SBS SBS
G*/sin(į)(kPa)
G*/sin(į)(kPa)
40
15
30
10
20
5 10
0 0
50 55 60 65 70 75 80 85 90 50 55 60 65 70 75 80 85 90
Temperature(deg.c) Temperature(deg.c)
(a) (b)
FIG. 2. G*/sin(į) Versus Temperature at Unaged e and RTFO Aged State.
POROUS ASPHALT CONCRETE TESTING AND EVALUATION
Stiffness Characteristics
The indirect tensile stiffness modulus (ITSM) test can be conducted to evaluate the
load spreading ability of mixtures in pavement. After compaction and cooling, the
same three specimens (100mm×50mm) for each mixture were tested at 30°C with
Nottingham Asphalt Tester (NU-10). The test results of indirect tensile stiffness
modulus are shown in the Table 6.
Table 6. Testing Results of ITSM.
Mixture Type PA1 PA2 (SBR) PA3 (SBS)
Average Stiffness (MPa) 591.8 509.8 462.5
Binder Content (%) 4.5 5.7 5.7
Notes: PA1denotes conventional PAC
PA2 denotes conventional polymer (SBR) modified PAC
PA3 denotes pre-blended polymer (SBS) modified PAC
The test results indicate that the ITSM results of the three types of PAC are 591.8MPa
for PA1, 509.8MPa for PA2, and 462.5MPa for PA3. On comparative basis, the
stiffness modulus value of PA1 is about 16 percent higher than PA2 and about 28
percent higher than PA3. Stiffness modulus of a mix is considered to be related to its
rutting resistance. Low stiffness modulus of a material normally would result in larger
rutting. PA2 and PA3 are shown to have lower stiffness modulus values than PA1.
However, from the wheel tracking testing results, both PA2 and PA3 exhibited higher
resistance against rutting. It implies that the lower stiffness modulus of the two
mixtures does not affect their load bearing capacity. The drop of the stiffness modulus
of the PA2 and PA3, under this circumstance, is mainly caused by their higher binder
content compared with the conventional one.
Permanent Deformation Characteristics
Two test types including the dynamic creep and wheel tracking test, were conducted to
simulate the load associated permanent deformation of the three studied mixtures. The
testing temperature of the wheel tracking and dynamic creep test were 60ć and 30ć,
respectively. The average results of the wheel tracking test of the three different kinds
of PAC mixtures are presented in Table 7. The test results indicate that PA3 has the
best satisfactory performance with respect to rutting resistance due to its lower rutting
potential at the end of test compared to the other studied mixtures. The rutting
performance ranking is followed by PA2 and PA1.
Table 7. Wheel Tracking Results for Mixtures at 60ć.
Mixture Type Rutting Depth(mm) Rutting Rate(mm/h) Ranking
PA1 7.46 7.79 3
PA2 (SBR) 5.80 4.00 2
PA3 (SBS) 2.43 1.60 1
The average results of dynamic creep tests of the three different kinds of PAC
mixtures are presented in Fig.6. It can be seen that after 1800 load repetitions, the
values of axial permanent strains of the three PAC mixtures are 0.69% for PA1, 0.67%
for PA2, and 0.50% for PA3, respectively. PA3 exhibits the best resistance to
permanent deformation among the three types of PACs studied. A similar ranking in
resistance to permanent deformation as the wheel tracking test was obtained with the
dynamic creep test. Each of the two kinds of tests gives a similar qualitative evaluation
on the effect of the different binders on permanent deformation of the same PAC
mixtures.
0.80
0.60
Axial Strain(%) 0.40

PA1
PA2
PA3
0.20
0.00
5 305 605 905 1205 1505 1805
Number of Load Repetition

FIG. 6. Axial Permanent Strain versus Load Repetitions.
Moisture Damage Test

The test results of moisture damage of the three types of PAC are given in the Table 8.
The results indicate that the SBS polymer modified PAC (PA3) has the best resistance
to moisture damage as indicated by its comparatively higher ITS ratio in magnitude. It
should be noted that the SBR polymer modified PAC (PA2) exhibits the lowest values
among the three types of PAC. The main reason for this is that the dry mixing method,
which is defined as adding SBR directly into the PAC mix process, is adapted by the
local asphalt industry.
Table 8. Testing Results of ITSR.

Average Tensile Strength(kPa)
Mixture Type ITS Ratio (%)
Unconditioned Conditioned
PA1 500.02 370.34 74.1%
PA2 (SBR) 446.02 312.86 70.1%
PA3 (SBS) 486.00 369.50 76.0%
CONCLUSIONS
The results of the penetration and soften point test indicate two polymer modified
binder, SBS and SBR, exhibit less temperature susceptibility and better ageing
resistance during construction phase and after years in services than the conventional
60/70 binder. The two modified binders, SBS and SBR, have less temperature
susceptibility and exhibit better resistance to pavement deformation than the 60/70
binder. The testing results of Permanent Deformation Characteristics show that two
modified PAC (SBR and SBS) have less rutting depth and rutting rate than the
conventional one. The results of moisture damage indicate that the polymer (SBS)
modified PAC has the best ability of moisture damage resistance shown by the higher
values in ITSR before and after the water treatment. Based on the overall satisfactory
performance with respect to stiffness ability and good resistance to ageing, moisture
damage and permanent deformation, the polymer (SBS) modified PAC is
recommended as the safest paving material available to use for friction courses used in
Hong Kong.
ACKNOWLEDGMENTS
The writers would like to acknowledge the Road Research Group of the Hong Kong
road research laboratory and the R&D division of highways department of HKSAR.
REFERENCES
R. Golebiewski, R. Makarewicz, M. Nowak, A. Preis. (2003). “Traffic noise reduction

due to the porous road surface.” Applied Acoustics, 64, 481-49.
Nicholls, J C.(1997). “Review of UK Porous Asphalt Trials.” TRL Report 264.
Transport Research Laboratory, Crowthorne, UK.
Huber G. (2000).”Performance Survey on Open-Graded Friction Course Mixes.” Rep.
No.Synthesis 284, Strategic Highway Research Program, National Research
Council, Washington, D.C.
P.E.Bolzan, J.C.Nicholls and G.A.Huber. (2001).” Searching for Superior Performance
Porous Asphalt Wearing Courses.” Rep.No.Synthesis 852, Strategic Highway
Research Program, National Research Council, Washington, D.C.
Laboratory Study of Porous Asphalt Mixture Made With Rubber Bitumen
YANG Jun and YIN Chaoen

School of Transportation, Southeast University, Nanjing 210096, China; [email protected]
ABSTRACT: A porous asphalt mix was designed with high viscosity modified
bitumen and rubber bitumen at an aimed air void content of 20%. A Hamburg wheel
tracking device was employed to test rutting potentials in both 60 °C air bath and 40
°C water bath conditions. Triaxial compression tests and triaxial cyclic compression
tests were performed to determine the resistance to shear stress and permanent
deformation at 60 °C. The study results shows that despite the similar rheological
behavior between rubber bitumen and high viscosity modified bitumen, rubber
bitumen is far less applicable than high viscosity modified bitumen in the aspect of
high temperature performance and shear resistance. However, the rubber asphalt
mixture shows a longer creeping stage and a stiffer resilient modulus in cyclic
loading and lower level. Under the conditions defined herein, porous asphalt made
with rubber bitumen can only be used for lighter volume of traffic.
KEYWORDS: rubber bitumen, porous asphalt, high viscosity modified bitumen
INTRODUCTION
Porous asphalt is of great research interest in China in recent years. Some test roads
have been built on freeways, which mostly recommend high viscosity modified
bitumen as the binding material (Ni 2003). However, the cost of this material
supersedes what is saved when fewer aggregates are used (Chen 2006). Meanwhile,
amounts of rubber waste are increasing with the boom in the rubber industry. Reuse
of waste rubber in bitumen is still on a small scale; current applications include tack
coating and use as a stress absorbing interlayer, but rarely used for asphalt mix.
Large scale studies on rubber bitumen proved the advantage of crumb rubber used as
a modifier. Studies on viscoelastic characteristics of rubber bitumen with 9 wt.%
crumb rubber revealed enhanced resistance to permanent deformation and similar
properties as SBS-modified bitumen with 3 wt.% SBS at -10 °C and 7 wt.% SBS at
75 °C (Navarro 2002). These enhanced properties make rubber bitumen a possible
solution for porous asphalt mixes.
22
The main objective of work in this paper is to study the applicability of rubber
bitumen in porous asphalt on a laboratory scale. The properties of rubber bitumen
from 30-mesh size reclaimed rubber were compared with those of high viscosity
modified bitumen. The various performances of the mixtures designed with the two
bituminous binders are investigated. The porous asphalt mixture with high viscosity
modifier (hereinafter abbreviated as HVM) is taken as a comparison, since this has
already proved successful in test road practices (Yang 2008).
MATERIAL DESIGN AND PERFORMANCE TEST
Asphalt binders Characterization

The base bitumen, abbreviated as AH-70, has a penetration of 6.55 mm at 25 °C and
a ring & ball softening point of 53.1 °C. The passing rate of the crumb rubber under
each sieve is shown in Table 1.
Table 1. Crumb rubber sieve gradation.

Mesh size No. 16 30 50 100 200
Pass rate% 100 56.2 17.7 5.8 1.0
The density of crumb rubber is 1.16g/cm3. It is uncontaminated from steel or other

particles, with fiber less than 0.5% by weight. To obtain the optimum properties of
asphalt binders, these two additives were added to the base bitumen in different
compositions; high viscosity modifier at percentages of 12%, 14%, and 16%, and
crumb rubber at 15%, 18%, 21%, and 24% respectively.
The following tests were performed according to the SHRP test method
procedure (Superpave 2003): (1) ductility at 5 °C with 2mm/min elongating rate, (2)
viscosity at 60 °C according to the vacuum tube method, (3) G*/sin į at 64 °C using
the dynamic shear rheometer (DSR) test and (4) stiffness S(t)|t=60s at -12 °C using the
bending beam rheometer (BBR) test. All the test samples were subjected to thin film
oven conditioning at 163.5 °C for 5 hours. The temperature at which the rheometers
were operating was deduced from the climatic information of Nanjing weather
station in southern China, according to the method promoted by Solaimanian (1993).
The shear frequency for the DSR test was 1.60 Hz, and the shear stress was 110 Pa
without normal force. All the results are shown in Table 2.
Table 2. Properties of different bitumen.

High viscosity
Rubber bitumen with
bitumen with HVM
Test items Condition AH-70 rubber in weight percent
in weight percent
15% 18% 21% 24% 12% 14% 16%
Softening unaged 53.1 64.2 62.5 67.7 67 80.1 91 92.7
point °C aged 55.8 62.5 73.6 74.4 68.7 79.6 87.3 94.8
Ductility unaged 16 11 15 17 14 44 50 31
cm aged 4 15 12 15 16 29 29 24
Viscosity unaged 0.21 1.7 2.4 1.7 1.9 13.6 76.6 63.9
kPa.s aged 0.26 1.2 2.2 1.6 1.7 11.4 68.6 51.1
G*/sin į unaged 1.6 12.3 13.3 10.0 7.2 8.2 10.5 14.5
kPa aged 2.7 14.7 14.5 10.7 7.7 7.6 9.8 12.6
S(t)|t=60s unaged 115 68.7 71.5 59.3 36.5 70.2 50.6 113
MPa aged 119 88.2 98.1 94.6 58.2 88.5 64 119
The results in Table 2 demonstrate that the softening point of both rubber
bitumen and HVM modified bitumen increases with the increase of modifier content.
G*/sin į of HVM modified bitumen also rises when more HVM is added, while that
of rubber bitumen reaches a maximum at 18%. Ductility of rubber bitumen is
relatively low, but that of HVM reaches maximum at 14%. The viscosity of HVM
modified bitumen also reaches maximum at 14%, and all three are in excess of
10,000 Pa.s. The viscosity of rubber bitumen at 60 °C is about 2,000 Pa.s. The
stiffness of rubber bitumen of different concentrations is almost the same, but that of
HVM modified bitumen is lowest at 14% HVM content, which means it has better
low temperature properties. With regard to the high temperature and low temperature
properties, the content of crumb rubber and HVM is selected to be 18% and 14% in
weight, respectively.
Volumetric Design
The main objective of porous asphalt design is to achieve an aimed air void of 20%
in this study, which meets the drainage needs in peak rainfall within an hour in most
parts of southern China. Mix aggregates are with the maximum nominal grain size of
13mm. Marshall and Cantabro tests after water bath conditioning and freeze-thaw
indirect tensile tests were performed to investigate the water susceptibility, which is
taken as the major factor for proportion selection. Two sets of optimal gradation of
the aggregates and asphalt content were selected. One is identical to the way that
porous asphalt is normally designed with HVM bitumen, namely AR-PAC gradation
(RIOH 2004).
The other optimal designs takes into account the American experiences in
open-graded friction courses with rubber bitumen (abbreviated as AR-O), which is
characterized with high asphalt content and open gradation (Caltrans 2003). The
envelope and the actual gradation of the tested formations are listed in Table 3.
However, in the experiment, gradations that were close to the center of the AR-O
envelope showed weak stability and scattered when moulding. From those within the
PAC gradation envelope, the best formation was selected for rubber bitumen. The
result is the AR-PAC with optimum asphalt content (OAC) of 6.1%, and for HVM
bitumen is the HVM-PAC with OAC of 5.0%.
Table 3. Gradations of the tested mixtures.

Pass rate under certain Sieve size (mm) %
Gradation
16.0 13.2 9.5 4.75 2.36 1.18 0.6 0.3 0.15 0.075
min 100 90 40 10 10 7 6 5 4 3
PAC Limits
max 100 100 71 30 20 17 14 12 9 7
AR-PAC 100 96.6 49.9 16.5 13.8 11.2 9.0 7.5 6.5 6.0
HVM-PAC 100 97.1 56.0 17.8 14.6 11.5 8.8 7.0 5.8 5.3
AR OGFC min 100 85 45 5 3 0
- - - -
Limits max 100 100 80 30 15 4
AR-O 100 97.8 66.2 17.1 8.5 6.3 5.1 4.3 3.8 3.5
Performances of the Mixtures

Hamburg Wheel Tracking Test
Due to the open-graded proportion of the mix, rutting potentials related to scattering
and instability will be the major problem. The Hamburg wheel tracking device was
used to test the rutting trend for a possible longest length of 20,000 wheel repetitions.
The rubber wheel, of which the outer diameter is 200 mm and the width is 50 mm,
runs a distance of 230 mm at a rate of 26 times to and fro per minute. Under a
prescribed load, the rubber tire keeps a pressure of 700 kPa on the asphalt mixture
slab specimen. In this test, the specimens are seperated into two groups, respectively
in a 60 °C air bath and a 40 °C water bath. The pre-heat time is 4 hrs for the 60 °C air
bath and 2 hrs for the 40 °C water bath. Rut depth development was recorded and
graphed as shown in Figures 1 and 2.
Rut development for AR-PAC (OAC=6.1%)

1600
60 °C air bath
1400 40 °C water bath
rut depth in 0.01 mm
1200
60 °C air bath
1000
y = 0.2834x + 97.432
800
R² = 0.992 40 °C water bath
600
400
y = 0.0835x + 147.36
200
R² = 0.9854
0
0 1000 2000 3000 4000 5000
tracking number
Figure 1. Rut development for AR-PAC (OAC=6.1%)
Rut development for HVM-PAC (OAC=5.0%)

700
60 °C air bath
600 40 °C water bath 60 °C air bath
rut depth in 0.01 mm
500
y = 0.0259x + 132.56
400
R² = 0.9957
300
40°C water bath
200
y = 0.0076x + 153.95
100
R² = 0.9928
0
0 5000 10000 15000 20000
tracking number
Figure 2. Rut development for HVM-PAC (OAC=5.0%)
According to the British standard, the rut rate between 5,000 and 10,000 times is
taken as an index for evaluating the rutting potential. Those developing excessive
rutting before 10,000 times will be evaluated with a rate in any steady stage for more
than 2,000 times (BSI 2003).
The rut developing rate is obtained from a linear regression of certain stages of
the curves. It can be seen that, in both the 60 °C air bath and the 40 °C water bath
conditions, AR-PAC mixtures developed rut more than ten times faster than
HVM-PAC mixtures. In the 40 °C water bath, the rut depth was much smaller than in
the 60 °C air bath. AR-PAC developed excessive rutting after 4,000 times application
of the wheel load. By contrast, the overall rut depth of HVM-PAC was much smaller.
In summary, AR-PAC showed greater susceptibility for rutting.
The Triaxial Compression Test

The triaxial compression test can reflect to a large extent the on-road stress
conditions of the asphalt mixture. According to the South African experience,
unconfined uniaxial tests may underestimate the strength of porous asphalt
(Verhaeghe 1995). The shear strength indexes can be obtained from this test and used
for road design based on a shear strength criterion (Zhu 2007). The specimen was
made by a gyratory process and was cored and sawn into a cylinder specimen with a
diameter of 100 mm and a height of 135 mm. After pre-heating in a 60 °C air bath
for 4 hrs, it was tightened with a watertight rubber sleeve and then, attached to the
load platen for loading. The loading chamber is shown in Figure 3, where the
ambient pressure is transferred by water pressure.
Load cell
Ambient pressure inlet

Chamber
ı1
ı3
Water inlet
Figure 3. Test facilities
Ambient pressures were set as 0 kPa, 138 kPa, and 276 kPa, that is uncombined
condition with loading pressures of 20 psi and 40 psi, respectively. The loading
platen descends at a rate of 1.27 mm/min until the specimen is destroyed. The
loading cell is driven by a pneumatic servo system with a loading capacity of 14 kN.
During the test, the temperature of the test chamber and specimen is maintained at 60
°C. The maximum compressive stress will be recorded. Based on the Mohr-Coulomb
theory, the shear strength indexes can be obtained by constructing the common
tangent of the Mohr-circles, as shown in Figure 4. The results of the AR-PAC,
HVM-PAC, and another SMA mixture made with SBS PmB are also listed in Table
4.
Figure 4. The common tangent of Mohr-circles
Table 4. Triaxial test settings and results.

Average Ambient Maximum Failure Inner
Cohesive
Formation air void pressure norm stress strain friction
c kPa
VV % ı3 kPa ı1 kPa İ% angle ĳ °
0 212.7 3.250
AR-PAC
18.5 138 452.7 4.757 73.22 24.36
(OAC=6.1%)
276 596.9 8.001
0 383.1 3.766
HVM-PAC
19.9 138 815.4 5.401 108.45 33.99
(OAC=5.0%)
276 1075.1 9.624
SMA mixture 0 836.4 4.661
with SBS PmB 3.9 138 1322.0 4.554 197.68 39.47
(OAC=5.9%) 276 1799.6 4.071
Despite similar gradations, the results reveal that the shear indexes of AR-PAC are all
about 30% smaller than those of HVM-PAC, only because of the different bitumen
binders. As a result, the AR-PAC has much less strength than HVM-PAC. Related
research shows that the cohesive for hot mix asphalt (HMA) is a round 35 - 240 kPa
(5 -35 psi), and the inner friction angle is about 35 - 48 ° (Witczak 2002). For porous
asphalt, it is relatively lower. This is due to the open-graded formation, which is
primarily due to the relatively lower content of fine aggregates and fillers, which
form a binding bituminous mortar. Compared to the PAC gradation, the SMA
formation only adjusts the filler content from 5% to 10% or so. With SBS PmB,
which is about the same grade as rubber bitumen, the SMA mixture shows much
higher strength. Taking into account the difference between AR-PAC and HVM-PAC,
the results shows that the two shear strength indices c and ĳ relate to both the
gradation of aggregates and binder properties.
Triaxial Cyclic Compression Test

Cyclic compression is deemed to be a good way to test the behavior and
development of permanent deformation. The test facilities and the specimen
treatment are the same as those in the triaxial compression test. The test temperature
is also 60 °C. The ambient pressure was set as 138 kPa (20 psi), which ties-in well
with the actual pavement conditions. The contact pressure between the load platen
and the upper face of the specimen is 30 kPa, and the cyclic load is in a haversine
form with crests of respectively, 200 kPa, 300 kPa and 400 kPa, which last for
100 ms and then pause for 900 ms and repeats the cycles in this way. The test stops
when the cyclic loading reaches to 1,000 repetitions or the accumulative strain
reaches 10%. Figure 5 shows the graph of accumulative strain and the strain rate.
5 250
Accumulative strain %
4 200
Creep strain rate 10e-6

Accumulative strain III
3 150
II
Turning point
2 100
I
Creep strain rate
1 50
0 0
0 400
800 1200 1600
Cyclic repetitions
Figure 5. Strain and strain rate under cyclic loading
The creep strain development can usually be divided to in three stages. The
strain rate firstly descends to a certain value. Then, in the second stage, the strain rate
remains at its lowest and is almost constant, while the accumulative strain increases
stably and slowly. Lastly, the creep strain increases quickly until the specimen is
damaged. The cyclic repetition, at which the rate is zero, is taken as the flow number.
American researchers pointed out that the flow number, the creep strain rate in the
second stage, and the accumulative strain at 1,000 repetitions are well correlated with
the rutting potential of the mixture on the road (Witczak 2002). The results are shown
in Table 5.
Table 5. Triaxial cyclic compression test results.

Partial Strain
Average Flow Strain at 1000 Resilient
stress rate
Formation air void number repetitions Modulus
ı1-ı3 dİ/dt
VV % n İ1000 % Er MPa
kPa 10-6
200 791 19.2 3.61 115 - 130
AR-PAC 18.3 300 355 18.4 5.86 120 - 130
400 227 13.4 3.85 154 - 185
200 429 31.5 5.40 95 - 110
HVM-PAC 20.0 300 287 33.6 4.18 120 - 135
400 171 29.4 4.55 145 - 165
In Table 5, HVM-PAC has a much higher flow number and strain rate in the
stable creep stage than AR-PAC. This shows that the AR-PAC exhibits better
permanent deformation resistance. However, the flow numbers for both mixes are
relatively small and below 1,000. So, the accumulative strain at 1,000 repetitions
makes less sense. In the test, AR-PAC shows a bigger resilient modulus. This is not
contradictory with the relatively weaker performance from the wheel tracking test
and triaxial compression test, because the loading scale is relatively minor in this test.
Porous asphalt with rubber bitumen shows higher stiffness at a lower loading scale
here. The low flow number at such a light loading level shows that the open-graded
PAC mixture has limited strength in resisting cyclic loading.
CONCLUSIONS
Through design and laboratory testing of porous asphalt mixture with rubber bitumen
and high viscosity bitumen, the performance at the normally in-service high
temperature of 60 °C was investigated. Based on the findings, the following
conclusions can be drawn:
1. The ultra open-graded formation with fillers less than 4.0% in total aggregates
may be inappropriate for porous asphalt made with rubber bitumen.
2. Greater G*/sin į at the testing temperature from the DSR test doesn’t
necessarily mean better high temperature performance for rubber bitumen mixes.
Viscosity may be a decisive factor for these bitumen binders used in porous asphalt.
3. Wheel tracking and triaxial compression tests proved that porous asphalt mixes
made with rubber bitumen have a higher rutting potential and thus, are more suited
for use in lighter traffic loading conditions.
REFERENCES
British Standards Institution, BSI (2003), "Bituminous mixtures – Test methods for
hot mix asphalt, Part 22: Wheel tracking." BS EN 12697-22.
Chen, W. Z., Wang, K. P. (2006). "Cost evaluation on porous asphalt pavement on
Yan-Tong Freeway." Journal of Highway and Transportation Research and
Development, Application Technology Edition, Vol. 2006(10): 31-34.
F. J. Navarro, P. Partal, F. Martínez-Boza, C. Valencia, C. Gallegos (2002).
"Rheological Characteristics of Ground Tire Rubber-Modified Bitumen."
Chemical Engineering Journal, Vol. 89, Issues 1-3: 53-61.
Ni, F. J., Xu H., and Leng, Z. (2003). "Influence of asphalt properties on porous
asphalt mixture performance." Journal of Traffic and Transportation Engineering,
Vol. 3 (4): 1-4.
Solaimanian, M., Bolzan, P. (1993). "Integrated Model of Climatic Effects on
Pavements." SHRP-A-637, Washington, D C: National Academic Press.
Superpave (2003), "Performance Graded Asphalt, Binder Specification and Testing."
Superpave Series No. 1 (SP-1), Third Edition, Asphalt Institute.
Research Institute of Highway Ministry of Communications China, RIOH (2004),
"Technical Specification for Construction of Highway Asphalt Pavements", China
Communications Express. 94-95.
Verhaeghe, B., CSIR Division for Roads and Transport Technology (1995). "SABITA
Manual 17 Porous asphalt mixes: design and use." SABITA.
Witczak, M. W., Kaloush, K., Pellinen, T., et al (2002). "Simple Performance Test for
Superpave Mix Design." NCHRP Report 465. Washington, D C: National
Academic Press.
Yang, J., Yin, C. E., Cao, D. W., Ge, S. M., Liu, Q. Q., Qian, G. C. (2008). "Practice
of Porous Asphalt Pavement on Yan-Tong Freeway in China." Proceedings of
International ISAP Symposium on Asphalt Pavements and Environment, ISAP,
Zürich, Switzerland, 330-339.
Zhu, H. R. (2007). "Evaluation on the shear properties of asphalt mixtures through
triaxial shear test." MSc Thesis, Southeast Univ., Nanjing, China.
Research on Permanent Deformation Property of Asphalt Mixture with
Varying Temperature
Cao Liping1, Sun Lijun2, Dong Zejiao3

1
Technology; No.202, Haihe Road, Harbin 150090, China; PH(086)0451 86282120; [email protected];
2
Professor of Civil Engineering; School of Transportation Engineering, Tongji University; No.4800,
Caoan Road, Shanghai 201804, China; e-mail: [email protected];
3
Technology; No.202, Haihe Road, Harbin 150090, China; PH(086)0451-86282120, [email protected].
ABSTRACT: Partial hourly pavement temperature variation measured in Shanghai

was analyzed. Cumulative frequency of high temperature (above 50ć defined in
this paper) was more than 90%. However, cumulative frequency of moderate
temperature (20 ć to 50 ć defined in this paper) was less than 10%. Then
permanent deformation (PD) of two asphalt mixtures at the temperatures ranging
from 20ć to 60ć was measured using the APA (asphalt pavement analyzer). The
results show that PD at 40ć can be as high as 50% of that at 60ć. Therefore, the
PD property should be researched systemically. Firstly, the cumulative PD curve of a
typical asphalt mixture was measured and defined by two zones. In the primary zone,
the PD accumulates rapidly. The incremental PD tends to decrease, reaching a
constant value in the secondary zone. Then, the PD property of asphalt mixture under
varying temperatures was researched in laboratory. The result shows that the trend of
the PD curves at the second stage was the same as but less than that at the first stage.
Based on this study, it was concluded that temperature, number of load cycles, stress
history, and intermission time are some of the major influencing factors on the PD of
asphalt mixtures.
INTRODUCTION
Rutting of asphalt pavement is a serious concern experienced for a long time due
to the increased traffic and wheel loads in China. Thus, it is important to estimate the
permanent deformation (PD) response of a mixture before construction. Several test
methods are in practice to assess the PD (1). However, almost all the tests evaluate
the PD at constant high temperature such as 60ć or PG (performance grade) high
temperature, while the PD at moderate temperature is often neglected.
32
In fact, rutting occurs at both high temperature and moderate temperature. At the
same time, pavement temperature changes every hour, every day, and every year. For
a certain asphalt mixture, it is often not sure whether it is enough to evaluate high
temperature performance at one temperature or not. Consequently, the PD property
can be researched considering the distribution of pavement temperature at different
depths of the asphalt layers and taking pavement temperature variation into account.
So, the primary objective of this paper was to research the PD property of mixtures
with varying temperatures and with distribution of pavement temperature considered.
DISTRIBUTION OF PAVEMENT TEMPERATURE IN SHANGHAI
The temperature profile in an asphalt pavement structure has a significant effect

on performance of asphalt pavement because of the susceptibility of asphalt mixture.
Numerous researches have been carried out to measure and predict the temperature
distribution in asphalt pavements. Qin jian (2) measured pavement temperature data
of asphalt pavement sections in Shanghai. In the test section, there were 12
thermal-probes installed at different depths in the asphalt layers. The frequency
histograms of hourly average pavement temperature for different averaging intervals
are given in Figure 1 based on collected data at four depths in the asphalt layers from
August 20 to September 14, 2002. The time region represented typical high
temperature season in Shanghai, and the depths were 5mm, 17mm, 39mm, and
61mm, respectively, below the surface.
50.0
5mm
40.0 17mm
Frequency(%)
39mm
30.0 61mm
20.0
10.0
0.0
30 35 40 45 50 >50
Temperature(ć)
FIG. 1. Hourly Average Pavement Temperature Distribution for Different
Averaging Intervals in Shanghai.
As shown in Figure 1, the frequency of pavement temperature interval from 30ć

to 50ć is about 90%, while the frequency of pavement temperature above 50ć is
less than 10% at the four depths. As a result, even if in high temperature season,
moderate temperature accounts for a large proportion of the pavement temperature
distribution in the asphalt layers. Therefore, is it proper that only the PD property of
asphalt mixtures at constant high temperature is evaluated? How about the PD at
moderate temperatures?
HIGH TEMPERATURE PERFORMANCE OF ASPHALT MIXTURES AT

DIFFERENT TEMPERATURES
PD of Asphalt Mixtures Tested by APA

PD can be evaluated using the APA, which is not a fundamental test but a
simulative test simulating the traffic loading and temperature effects. The APA
allows for an accelerated evaluation of rutting potential after volumetric design of
mixes. Cylindrical specimens compacted using the Superpave gyratory compactor
(SGC) of two gradations with the same asphalt and aggregates were used. The testing
time was 135 minutes (8000 cycles). A load of 100 lb and a hose pressure of 100 psi
were used for all the tests (3). The control temperatures were respectively 20, 30, 40,
50, and 60ć, respectively.
In order to simplify the explanation, two surface course mixtures (13.2 mm
mixtures) with the same aggregate sources (database for coarse aggregates and
granite for fine aggregates) and asphalt binder were used. The penetration of the base
asphalt was 68 at 25ć. These two gradations satisfied the requirements of Technical
Specifications for Construction of Highway Asphalt Pavements in China (4). The
aggregate gradations for the two mixtures are given in Table 1.
Table 1. Aggregate Gradations for Two Mixtures

Percent Passing (%) Actual air
Sieve Size(mm)
16 13.2 9.5 4.75 2.36 1.18 0.6 0.3 0.15 0.075 void
Upper limit of gradation 100 100 85 68 50 38 28 20 15 8 content
Lower limit of gradation 100 90 68 38 24 15 10 7 5 4 (%)
Design gradation 1 100 95 70 41.5 30 22.5 16.5 12.5 8.5 6 4.03
Design gradation 2 100 96 84 60 45 30 19.5 13.5 8 5 4.11
The PD of the two mixtures at five temperatures is shown in Figure 2. Evidently,

PD increases with temperature for the two mixtures. What’s more, the PD of design
gradation 1 is higher than that of design gradation 2 at all the temperatures except for
40ć. Moreover, PD at 40ć is almost 50% of that at 60ć for these two mixtures,
which indicates that the PD at moderate temperature can be compared to that at high
temperature even if without consideration of the pavement temperature distribution.
5
4.5 Design gradation 1
Permanent deformation(mm)
4 Design gradation 2
3.5
3
2.5
2
1.5
1
0.5
0
20 30 40 50 60
Temperature(ć)
FIG. 2. PD at Five Temperatures for Two Mixtures Tested by APA.
DS of Four Asphalt Mixtures Tested by Wheel Tracking Device

The wheel tracking test measures rut depth and the number of passes of tire loads
at high temperatures. Test results are presented by a curve relating deformation with
loading time, and the rate of rut depth monitored during 15 minutes of testing (from
45 to 60 minutes during the test) is then recorded as an indicator to estimate dynamic
stability (DS) of the mixture. The wheel tracking test is at presently widely used in
China to assess high temperature performance of asphalt mixtures. Therefore, test
results of four mixtures (HXL, LHE, SLI and CHA) typically used in China were
cited and analyzed in this paper (5). Results of these four mixtures were Figure 3.
6000
Dynamic stability(numbers/mm)
HXL
5000 LHE
SLI
4000 CHA
3000
2000
1000
0
35 40 45 50 60
Temperature(ć)
FIG. 3. DS of Four Mixtures Tested by Wheel Tracking Device.
As shown in Figure 3, the DS of four mixtures decreases with temperature. The

general trend is that HXL presents higher DS, followed by LHE, CHA, and SLI at 60
ć. The trend changes as HXL presents higher DS followed by CHA, LHE, and SLI
at 50ć. Then, for 45ć, DS of SLI ranks first, followed by HXL, CHA, and LHE.
For 40ć, SLI still ranks first, however followed by HXL, LHE, and CHA. HXL
ranks first again at 35ć, followed by LHE and SLI. The change of ranking at
different temperatures indicates that these four mixtures have different temperature
sensitivity. In other words, the trends of DS variation with temperature are different.
The mixture with higher DS at 60ć may have a lower DS at 40ć. Therefore, the
PD property at moderate temperature is of the equivalent importance as that at high
temperatures. Furthermore, research on the PD property with varying temperature is
of benefit to better understand the PD property of asphalt mixture and to more
rationally predict the rutting in asphalt pavement.
PERMANENT DEFORMATION PROPERTY
The PD property at constant temperature and varying temperatures were

researched , respectively, using the APA simulation test.
Permanent Deformation Property at Constant Temperature

The PD of the asphalt mixture (design gradation 2) at the temperatures ranging
from 20ć to 60ć for 8,000 cycles was experimented using the APA simulation test.
As shown in Figure 4, results of five curves indicate that the cumulative PD

curve can be divided into two zones. Taking the cumulative PD curve at 60ć as an
example, in the primary zone, the PD accumulates rapidly. The incremental PD tends
to decrease, reaching a constant value in the secondary zone. However, the critical
point of these two zones changes with temperatures.
4.0
3.5 Critical point

3.0
2.5
2.0
1.5
1.0 20ć 30ć
0.5 40ć 50ć
60ć
0.0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Numbers of loading repetitions
FIG. 4. PD of design gradation 2 at five temperatures tested by APA.
Permanent Deformation Property with Varying Temperature

The PD property with varying temperatures should be researched for the reason
that pavement temperature changes hourly and that the PD property of asphalt
mixtures is affected by temperature. In order to simplify the research, the PD
property at two pavement temperatures was simulated by the APA test under four
test modes which are listed in the text below:
Test mode 1: The first PD was tested for 4000 loading cycles at 30ć at the
first stage, then the test was stopped and the temperature was raised to 40ć
and maintained at 40ć for 6 hours. The second PD was tested for 4000
loading cycles at 40ć in the second stage.
Test mode 2: 5000 loading cycles at 40ć at the first stage and 5000 lb
cycles at 50ć in the second stage, were used.
Test mode 3: 5000 loading cycles at 40ć at the first stage and 5000 loading
cycles at 30ć at the second stage were used.
Test mode 4: 2000 loading cycles at 50ć at the first stage and 2000 loading
cycles at 40ć at the second stage were used.
A load of 100 lb and a hose pressure of 100 psi were used for all tests (3). Graphs
of PD with temperature rising and cooling are shown in Figures 5 and 6, respectively.
Figures 5(a), 5(b), 6(a), and 6(b) correspond to test modes 1 through 4, respectively.
2.4 2.8
2.0 2.4
1.6 2.0
1.6
1.2
1.2
0.8
0.8
0.4 0.4
0.0 0.0
0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000 12000
Numbers of loading repetitions Numbers of loading repetitions
(a) Test Mode 1. (b) Test Mode 2.
FIG. 5. Graphs of PD with Temperature Rising.
\
2.1 2.0
1.8
1.6
1.5
1.2 1.2
0.9 0.8
0.6
0.3 0.4
0.0 0.0
0 3000 6000 9000 12000 0 1000 2000 3000 4000 5000
Numbers of loading receptions Numbers of loading repetitions
(a) Test Mode 3. (b) Test Mode 4.
FIG. 6. Graphs of PD with Temperature Cooling.
As shown in Figures 5 and 6, the curves don’t go into the secondary zone directly
after the first stage. On the contrary, the trends of PD curves at the second stage are
the same as those at the first stage. The curves can be divided into two zones for the
viscoelastic property of asphalt mixtures. However, the difference in the curves
between the first stage and the second stage is that the PD of both zones at the first
temperature is larger than that at the second temperature. At the same time, the PD
under temperature rising mode is higher than that under temperature cooling mode,
which indicates that the mode of varying temperature plays an important role in the
PD property of mixtures.
For further studying, Figure 5(b) is compared with Figure 6(a). In these two
figures, both the parameters at the first stage and numbers of loading cycles at the
second stage are the same, however, the temperature at the second stage in Figure
5(b) is higher than that in Figure 6(a), which leads to the PD at the second stage in
Figure 5(b) to be higher than that in Figure 6(a). Therefore, the temperature at the
second stage is the first important factor influencing the PD property with varying
temperature.
In addition, this research also finds that intermission time is another factor.
Deformation will recover gradually with intermission stop time. The longer the
intermission time is, the larger the recovery deformation is. Accordingly, the PD at
the second stage will be affected. Furthermore, stress history also affects the PD
property due to the asphalt mixture’s viscoelastic nature; i.e., loading time dependent.
The PD property will be different when the conditions are the same except the stress
history. In other words, temperature, numbers of loading, stress history, and
intermission time are PD influencing factors.
CONCLUSIONS
Based on the testing and analysis presented in this paper, the following
conclusions can be drawn:
1) The frequency of high pavement temperature is less than 10%, while that of
moderate temperature is more than 90%; based on the measured pavement
temperature data in Shanghai.
2) Permanent deformation or other evaluation indexes at moderate temperature is
close to or compared with that at the high temperature regime.
3) The permanent deformation property at moderate temperatures is of the
equivalent importance as that at the high temperature regime.
4) Based on this paper, some of the major influencing factors on the permanent
deformation of asphalt mixtures can be concluded to be the temperature, the
number of loading cycles, stress history, and the intermission time. However,
further research on the fundamental relationship between these factors and
permanent deformation will be discussed in more details in later publications.
ACKNOWLEDGEMENTS
This project was supported by Development Program for Outstanding Young

Teachers in Harbin Institute of Technology (HITQNJS.2008.047) and the Research
Fund for the Doctoral Program of Higher Education (200802131011).
REFERENCES
(1) Jorge B. Sousa, Joseph Craus, and Carl Monismith. (1991). Summary Report on
Permanent Deformation in Asphalt Concrete. SHRP-A-318. Strategic Highway
Research Program, National Research Council, Washington, D.C..
(2) Qin Jian. (2005). Research on prediction of temperature field in asphalt
pavements. Master’s thesis. Tongji University, Shanghai, China.
(3) Pavement Technology INC. (2003). Asphalt pavement analyzer user’s guide
[2003-03-28]. http://www.pavementtechnology.com/ manuals/APAManual.pdf.
(4) Highway Scientific Academe of Ministry of Communications. (2004). JTG
F40-2004 Technical specifications for construction of highway asphalt
pavements. Beijing: China Communications Press.
(5) Liu wei. (2005). Research on Causation and Prevention of Rutting in Asphalt
Pavements. Master’s thesis. South China University of Technology, Guangzhou,
China: 48-49.
Rutting in Asphalt Pavement under Heavy Load and High Temperature
Zhang Qi-sen1, Chen Yu-liang2, Li Xue-lian3
1
Professor and PhD supervisor. School of Communication and Transportation Engineering, Changsha
University of Science and Technology, Chiling Road 45#, Changsha, Hunan, China, 410076.
E-mail: [email protected]
2
Lecturer, School of Engineering & Architecture, East China Jiaotong University, Nanchang, Jiangxi,
China, 330013. E-mail: [email protected]
3
Lecturer and PhD, School of Communication and Transportation Engineering, Changsha University
of Science and Technology, Chiling Road 45#, Changsha, Hunan, China, 410076.
E-mail: [email protected]
ABSTRACT: Rutting is one of the main distresses in the asphalt pavements,

especially in higher summer temperatures and/or under heavy loads. Efforts were
made in this study to investigate rutting failures on Jing-Zhu expressway in China to
determine the cause and the source of the rutting. The characteristic of the rutting on
each asphalt pavement layer were analyzed by surveying the rutting depth,
excavating transverse profile sections, and boring samples in the typical sections of
expressways where different degrees of rutting damage were observed. It was found
that 60% of the rutting was generated by the asphalt middle course layer and the
remaining 40% came from the other surface layers, which was consistent with the
mechanical analysis of the asphalt pavement that showed that the shear stress in
middle asphalt course layer was the highest in magnitude, under load or temperature.
With the combination of the indoor rutting tests on the samples from a corresponding
location of the road shoulders, the influence of the high temperature stability of each
asphalt pavement layer on rutting was analyzed using the method of ANOVA
(Analysis Of Variance). The results indicated that the stability of the middle course
layer impacted the rutting development the most. Finally, the relationship between
the RD (rut depth) and the DS (dynamic stability) of each asphalt pavement layer
was established using a nonlinear regression method. The governing criterion for the
DS at the bottom course asphalt mixture was studied and recommendation made
based on regression analysis.
KEYWORDS: road engineering; asphalt pavement; variance analysis method;

correlation analysis method; rutting; mechanical properties
39
INTRODUCTION
Rutting is the load-induced permanent deformation of asphalt pavements and may

occur in any layer of a pavement structure (Witezak et al. 1997). It is one of the main
distresses occurring in asphalt pavements and badly affects the comfort-ability,
ride-ability, motorist safety, and general performance (Haas et al. 1994). The United
States and other countries have taken rutting as one of the design criteria for asphalt
pavements (AASHTO2002). In China, a large number of highway asphalt pavements
have been constructed in recent years, while deflection is the only design criterion
(Zu-kang Yao 2003). With the increase in both traffic volume and heavy overloading,
the asphalt pavement rutting problem has become more and more serious. Under the
function of such factors as heavy loading and bad weather that happened to some
expressway asphalt pavements, serious damages in terms of rutting appeared shortly
after opening to traffic (Qing-lin Sha 2001). With the combination of surveying the
rutting on the southern section of Jing-Zhu Expressway, testing, and the statistical
analysis, the impact of each asphalt pavement layer on the rutting formation were
studied, and the relevant recommendations for improvement have been suggested in
this paper.
RUTTING SURVEY AND TRANSVERSE PROFILE ANALYSIS
Rutting Investigation and Typical Section Selection
Due to the joint function of heavy traffic loading and continuous peculiar high
temperature, serious rutting failure appeared after the south section of Jing-Zhu
Expressway was opened to traffic for two years. The total average rut depth (RD)
reached 10mm and the maximum RD was about 30mm. The RD distribution is
shown in Fig.1, 90% is larger than 5mm and 5% is more than 20mm. So, rutting is a
very serious problem in this expressway and a cause for concern.

3HUFHQWDJH˄˅

5' 5' 5' 5' 5'
5'˄PP˅
FIG..1. Distribution of Rutting Depth.

Shape of Rutting Transverse Profile Analysis
In order to visually reflect the rutting and deformation in each asphalt layer, the
cross section excavation or core boring in a few main lanes where there were deeper
ruts in the typical section of the expressway were carried out, and the thickness and
the deformation of each layer was measured. The section shape of a typical rutting is
shown in Fig. 2.

'HSWK˄FP˅

7RSRIVXUIDFH 7RSRIPLGGOHOD\HU
7RSRIERWWRPOD\HU %RWWRPRIERWWRPOD\HU

'LVWDQFHEHWZHHQPHDVXULQJSRLQWVPP
FIG..2. Typical Cross-Section Shape of Rutting.
From the typical rutting section in Fig. 2, it was known that: (1) there was no
subsidence or cracks on the top of the semi-rigid base. All the rutting deformation
came from the asphalt layers, and the deformation was different in the different
asphalt layers. (2) Analyzing the magnitude of deformation in each asphalt layer, it
was known that the largest deformation occurred in the middle course layer, about
60% of the total rutting, while 25% come from the surface layer, and the other 15%
come from the bottom course layer. These results were inadvertently consistent with
the results of elastic layer theory that the largest compression and shear stress
appeared in the middle course layer of the three-layered asphalt pavement system
(Qing-lin Sha 2001). In addition, with the samples from the rutting of the wheel
marks in a few sections and from the lane, they reflected mostly the same as the
findings mentioned above.
MECHANICAL ANALYSIS OF THE ASPHALT PAVEMENT
Based on the typical road design concepts, FEM (Finite Element Method)
common software, ANSYS, was used to analyze the road structure shown in Fig.3.,
where loads of 100kN, 120kN, 140kN, 160kN, 180kN, and 200kN, respectively,
were applied vertically and the load evenly distributed over the effected area.
4cm AK-16A E1 =0.25

6cm AC-20I E2 =0.25
6cm AC-25I E3 =0.25
1cm Seal layer
20cm 6% Cement stabilized macadam E4=1300MPa =0.25
36cm 4% Cement stabilized macadam E5=1000MPa =0.25
Soil subgrade E0=40MPa =0.25
FIG..3. Analysis Structure of Pavement.
To simplify the analysis, the tire ground contact area was turned from a circle (JTJ
D50-2006) into square, 189mmh189mm, while keeping the contact area invariant.
The distance between the wheels is 320mm. The horizontal load is produced as the
vehicle brakes or starts to accelerate. To simplify the calculation further, the
horizontal load was assumed proportional to the vertical load, and the proportional
coefficient, f, of 0.2 was utilized. The dynamic modulus, E*, of the asphalt mixture
was analytically calculated using the Asphalt Institute (AI) model, considering
temperature, loading frequency, asphalt, and the aggregate. The results are shown in
Fig.4 and Fig.5.
From Fig.4 and Fig.5, the following conclusions can be drawn: (1) the maximum
shear stress of the pavement structure increases linearly as the load increases. For
example, at 70 ć, the maximum shear stress of the middle course layer is 0.275MPa
under 100kN, while under 200kN it is 0.550MPa, increasing by about almost 2 times.
(2) The maximum shear stress of the pavement structure also increases with the
temperature increase. For example, under 100kN, the maximum shear stress of the
middle course layer is 0.220MPa at 40ć, while at 70ć it is 0.275MPa, increasing
more about 25%. (3) Considering the impact of the axle or temperature, the shear
stress of middle course layer is larger in magnitude than that of the bottom course
layer or surface. It means that the anti-shear capacity of the middle course layer is
very important to solving the rutting problem in this pavement structure.

6XUIDFHć 0LGGOHOD\HUć %RWWRPOD\HUć
6XUIDFHć 0LGGOHOD\HUć %RWWRPOD\HUć

6KHDUVWUHVV˄03D˅

/RDG˄N1˅
FIG..4. Maximum Shear Stress in the Pavement under Varying Load.
Ϟ䴶ሖ
Surface Ё䴶ሖlayer
Middle ϟ 䴶ሖ layer
Bottom

࠾ᑨ࡯˄03D˅

Shear stress (MPa)

⏽ᑺ˄ć˅
Temperature (ć)
FIG..5. Maximum Shear Stress in the Pavement under BZZ-100.
PREPARATIONS AND TESTING OF RUTTING SPECIMENS
To scientifically evaluate the influence of the asphalt pavement layers on rutting,

samples from the corresponding shoulders at 12 typical sections were taken for
indoor rutting test. Firstly, based on the requirements of "Rutting Test for Asphalt
Mixtures” in the “Standard Test Methods of Asphalt and Asphalt Mixtures of
Highway Engineering” (JTJ052-2000), every sample was cut into a plane size of
300mm×150mm dimensions; the thickness was the same as that of the actual layer.
There were at least 3 trials in a group for each respective layer. The specimens were
then placed in the middle of the standard rutting mould of 300mm×300mm
dimensions, two sides of the sample were fixed with splints, and the other sides were
filled with newly asphalt mixture, as shown in Fig.6.
˄a˅Specimen Preparation. ˄b˅Specimen after Rutting Test.

FIG.6. Rutting Test Specimens.
The tests for 115 trials from 12 locations were carried out altogether. The
evaluation index is DS. The test results of the DS from the samples are listed in
Table 1.
Table 1. Result of Rutting Test.
DS˄times/mm˅ Pavement condition

Index üüRD in roadway
No.
Surface Middle course Bottom course ˄mm˅
1 5237 4442 1264 6
2 2754 1409 586 10.3
3 3691 3065 743 8.2
4 987 1348 1091 14
5 918 1014 1564 8.3
6 1256 652 406 24.5
7 3500 3700 980 23.8
8 650 791 405 11.9
9 3536 2820 1089 13.3
10 1219 413 548 21.8
11 1593 506 626 18.8
12 2988 2581 1872 8.3
STATISTICS AND ANALYSIS ON THE TEST RESULTS
Analysis of Variance (ANOVA)
To determine the impact of control variables (Jing-xiang Ou 2003), ANOVA was

used to analyze and study the contribution of the different variables. It was applied to
analyze the significant factors of DS in each layer to RD so as to determine the
influence of each factor.
For the convenience of ANOVA, there was a need to classify the data measured in
accordance with the required parameters, but there should not be too many levels.
Otherwise it was very difficult to find adequate samples to analyze, which would
affect the accuracy of the analysis. So, the classification of the relevant factors is
presented as follows:
DSsurface: Level 1İ1000 times /mm, 1000<Level 2<3000 times/mm,

Level 3ı3000 times /mm;
DSmiddle: Level 1İ800 times /mm, 800<Level 2<1500 times/mm,
Level 3ı1500 times /mm;
DSbottom: Level 1İ600 times /mm, 600<Level 2<1200 times/mm,
Level 3ı1200 times /mm.
Based on the above classification, the data was classified into different groups. If
the significance was pre-set at 5% and dependent variable was RD, then ANOVA
was carried out, and the results were shown in Table 2. We can see that the
significance of F in the analysis model is 0.003, that is, the model has a statistical
significance. So, the model can be used to determine the significance of various
factors on the RD. As the statistical significance level is 5%, Table 2 shows that the
DS of all the layers has a significant effect on the RD (Sig. of F0.05), especially that
of the middle course layer at 0.002. Based on the significance level between RD and
DS, the impact ranking is: (1) the middle course, (2) the bottom course, and (3) the
surface layer.
Talbe 2. Result of Variance Analysis under Factor Effect.

Sum of Mean Testing
Sources of ANOVA Sig. of F sigificance
Squares square Value F
Corrected Model 7.659(a) 1.276 18.690 0.003 **
Intercept 21.589 21.589 316.128 0.000
DSsurface 0.762 0.381 5.579 0.053
DSmiddle 3.750 1.875 27.454 0.002 **
DSbottom 0.859 0.429 6.286 0.043 *
Errors 0.341 0.068
Total 56.000
Corrected Total 8.000
Correlation Analysis
The correlation analysis is another statistical method to study the intimate degree
of variables. The correlation coefficient between the RD and DS of each layer is

listed in Table 3.
Table 3. Relation Analysis of Rutting Depth.

DS
Factors
DSsurface DSmiddle DSbottom
Related coefficient -0.511˄*˅ -0.883˄**˅ -0.375
From Table 3, the following conclusions can be drawn: (1) RD is negatively

correlated with the DS of all the layers. Considering the high absolute magnitude
value (|0.883|), the DS of the middle course layer related most significantly with the
RD, which is consistent with the field investigation results. So, these results imply
that the middle course layer (4-10cm below the surface) should be designed, selected,
and constructed carefully. Modified asphalt is suggested to be used in the middle
course layer, especially when expressways are under high temperatures and heavy
loads. (2) The surface and bottom course layers also have some impact on rutting
failure, i.e., on RD.
Regression Model for Rutting Depth
According to the characteristics of the data and the results discussed above, it was
found that Eq. (1) was better fitted to the data when the multiple nonlinear regression
models were applied.
RD = k1 + k 2 ln( DS surface ) + k3 ln( DS middle ) + k 4 ln( DSbottom ) (1)
Where k1, k2, k3, and k4 are regression coefficients, showed in Table 4; DSsurface,
DSmiddle and DSbottom are the DS of surface, middle course and bottom course,
respectively.
Table 4. Regression Coefficients.
Regression Testing Value Sig. of Related
k1 k2 k3 k4
coefficients F F coefficient R
Results 94.42 -2.34 -2.02 -7.13 10.92 0.003 0.897
From Table 4, the value of the related coefficient is 0.897, which means that the
equation fits well with the data. Moreover, the significance of F at 0.003 is smaller
than the level of significance of 0.05. Therefore, the regression model can be used to
evaluate the influence of the DS on RD of the various asphalt mixtures in all the
pavement layers.
In the “Technical Specifications for Construction of Highway Asphalt Pavement”
(JTG F40-2004), the rutting test is required for an asphalt mixture with a nominal
particle size equal to or less than 19mm only if it is used on expressway and first
class highways. While for nominal particle size greater than 19mm (generally used
for the bottom course layers), there are no specification indices or rutting test
requirement demands.
However, from the investigation and statistical analysis in this study, we can see
that the effect of the bottom course layer on rutting also can not be ignored. Under
heavy loading and high temperatures, the DS of the middle course and surface layers,
using modified asphalt, is 3,000 times /mm, and the tolerance limit of RD is 15mm.
The DS of the bottom course layer, computed from Eq. (1), is 515 times/mm.
Considering the impact of the test conditions and if the guarantee coefficient can be
taken as 1.5, then the DS of the bottom course layer will be 770 times/mm. So, it is
fairly reasonable to take 800 times/mm as the DS governing criterion of the bottom
course asphalt mixtures under normal traffic loading conditions. Under special
conditions such as steep long longitudinal, big super elevation, and heavy loads, the
DS of the bottom course asphalt mixture should be increased to 1000 times/mm.
CONCLUSIONS
Through the field rutting investigation, mechanical analysis, and indoor testing,
the major conclusions that can be drawn from this study are: (1) the survey on typical
sections shows that most of rutting is generated by asphalt middle course layer. (2)
The maximum shear stress of the pavement structure not only increases linearly with
as the load increases but also increases with temperature. Considering the impact of
the axle or temperature, the shear stress of the middle course layer is bigger in
magnitude than that of the bottom course or surface layer. This means that the
anti-shear capacity of the middle course layer is very important to solving rutting
problems. (3) With the method of ANOVA and the correlation analyses, it is found
that the DS of middle course layer has the greatest effect on RD, which is consistent
with the field surveys and the mechanical results. So, the middle course layer
(4-10cm below the surface) should be designed, selected, and constructed carefully.
Consequently, modified asphalt is suggested to be used in the middle course layer. (4)
The relationship between the RD (rut depth) and the DS (dynamic stability) of each
asphalt pavement layer was established using the nonlinear regression method. The
governing criterion for the DS of the bottom course asphalt mixture was studied and
recommendations made based on regression analysis.
REFERENCE
AASHTO. (2002). Pavement Design Guide.

Haas, R. W R Hudson, J Zaniewski. (1994). Modern Pavement Management. Florida:
Krieger Publishing Company.
Ou Jingxiang. (2003). Experimental Statistics. Guangzhou : High Education Press of
Guangdong.
SHA Qinglin. (2001). Premature Damage Phenomena and Preventive Techniques in
Expressway Asphalt Pavement. Beijing: People's Transportation Press.
Witczak, M. W., Von Quintas, H., and Schwartz, C. W. (1997). “Superpave support
and performance models management: Evaluation of the SHRP
performancemodels system.” Proc., Eighth Int. Conf. on Asphalt Pavements, Vol.
III, 175–195.
YAO Zukang. (2003). “A Review on Design Criteria of Asphalt Pavements”
Highway, 43-49.
JTJ D50. (2006). Design Specifications of Highway Asphalt Pavement. Beijing:
People's Transportation Press.
JTG F40-2004 ˈ Construction Specifications of Highway Asphalt Pavement[S].
Beijing: People's Transportation Press.
Development of Asphalt Emulsion Cold In-Place Recycling Specifications
Jinhai Yan1, Fujian Ni2, Zhuohui Tao3, Jonathan Jia4
1
PhD. Candidate, School of Transportation, Southeast of University, No.2 Sipailou Road, Nanjing, Jiangsu
province, P.R. China, 210096; [email protected]
2
PhD, P.E., Prof, School of Transportation, Southeast of University, No.2 Sipailou Road, Nanjing,
Jiangsu province, P.R. China, 210096; [email protected]
3
Technical Engineer, Shanghai application center, MeadWestvaco Corporation, 5FL,Nami Building, No.18
Lane1305, Huang Jing Road , P.R. China, 200231; [email protected]
4
Emulsion Field Support Manager, Shell (China) Limited, No.43 Liangmaqiao Road, Chao Yang
District, Beijing, P.R. China, 100016; [email protected]
ABSTRACT: In the past few years, asphalt emulsion Cold In-Place Recycling (CIR)
has been used successfully with asphalt emulsions in asphalt concrete pavement
rehabilitation in China. However, there are currently no existing performance
specifications using the Superave Gyratory Compactor (SGC) in China for the
development and verification of asphaltic CIR mix designs. The objective of this study
was to develop performance based specifications through laboratory testing and field
verification of three CIR projects. Laboratory testing was completed at different
emulsion contents through the Marshall Stability and indirect tensile strength tests. In
addition, the indirect tensile test, using a freeze/thaw cycle, and a rutting test were
performed at the predicted optimal emulsion content. The test results were compared
with field data recorded before and after the pavement rehabilitation process and
incorporated into a set of performance specifications. The results confirmed that
asphaltic CIR is an effective treatment for asphalt pavement preservation and
maintenance.
INTRODUCTION
Results of recent research studies and field projects have shown that asphalt
emulsion Cold In-Place Recycling (CIR) is a cost-effective solution to the challenges
of asphalt pavement maintenance and rehabilitation (Alkins et al., 2008). There are
several benefits of using CIR, including re-establishing crowns, maintaining
49
clearances, improving the pavement grade, reducing maintenance costs, reusing

existing materials, conserving material resources, and protecting the environment. In
addition, using CIR can eliminate pavement distresses, including rutting, shoving,
cracking, and raveling (Epps, 1990). The current asphalt emulsion specification for
CIR in Jiangsu province is based on the Marshall Design method. However, three
projects in 2007 used modified specifications from the United States (Thomas et al.,
2003) in order to develop local specifications. The objective of this study was to
develop performance based specifications using the Superpave Gyratory Compactor
(SGC) through laboratory testing and field verification through three CIR projects in
Jiangsu province, China.
EXPERIMENT
Three roads within the Jiangsu province highway network, S229, S239, and S336,
were selected for this study. Before the CIR treatment, these sections were evaluated
using the falling weight deflectometer (FWD) test and visual investigation. Current
traffic volumes on these three roads are between 8,000 and 16,000 AADT. The
inspection indicated that the roads had good structural bearing capacity, but distresses
of the pavements included rutting, moisture damage, cracking, and flushing. For
highway S229, the subgrade and semi-rigid base both have good modulus values as
shown in Table 3. Although the pavements had severe distresses, the strong structural
bearing capacity made these roads ideal candidates for CIR rehabilitation.
TEST PROGRAM
Materials and Characteristics

There were four components to the CIR pavement rehabilitation: recycled asphalt
pavement (RAP), Portland cement, coarse aggregate, and asphalt emulsion. The RAP
used in these three projects was obtained from surface and binder courses of the
project roads. The surface courses were approximately 4 cm and the binder courses
were 6 cm. A typical CIR project in Jiangsu province, the top 7.5 cm is recycled.
Ordinary Portland cement (P.O.42.5 and/or 32.5) were used in these projects. Cement
content was 2.0% by weight for the three projects. Coarse virgin aggregate was added
with the aim of adjusting the RAP gradation to ensure that recycling mixtures had
enough coarse aggregate to form a load-bearing aggregate skeleton and provide higher
inner frictional resistance. The content of added coarse virgin aggregate was between
10% and 15% by weight based on RAP gradation.
A cationic slow setting asphalt emulsion was formulated with combined emulsifiers
and chemical additives. A residue of greater than 63% was used in order to guarantee a
thick asphalt emulsion film coating on the RAP and new aggregate.
Mixing water has multiple functions in these emulsion related processes. Chief
among these functions is to improve coating of the RAP and virgin aggregate, to
lubricate mixture during compaction, and to facilitate and accelerate the cement
hydration reaction.
Gradation Design
Table 1 lists the aggregate gradations of the blended gradations. Due to cement used
as an additive, 2% cement by weight was not in included the gradation.
Table 1. Recycling Blended Aggregate Gradations.
Passing Percent (%)

Sieve Size (mm)
S229 S239 S336
37.5 100 100 100
26.5 100 100 100
19 88 91 93
16 77 83 84
13.2 67 74 73
9.5 56 61 61
4.75 36 38 39
2.36 21 21 23
1.18 15 12 17
0.6 9 5 10
0.3 6 2 4
0.15 3 1 2
0.075 1 0.5 0.7
Compaction and Curing

A laboratory mixer was used for mixing the formulation ingredients. Cement was
first added in RAP aggregate and mixed for 30 seconds. Next, water was added and
mixed for another 30 seconds. Finally, emulsion was added and the material was
mixed for an additional 60 seconds. When the mix process was completed at room
temperatures, the Superpave Gyratory Compactor (SGC) was used to compact loose
mix into specimens having 100 mm diameter and 63.5 ± 2.0mm in height. All
specimens were prepared at 30 gyrations (Cross, 2002). The specimens were initially
cured at room temperature for 12 hours and then, they were cured at 60°C for 48
hours. Finally, the specimens were cooled to room temperature and used for the
performance tests.
TEST RESULTS AND DISCUSSION
The experimental items used to analyze the performance properties of the SGC
compacted specimens were based on domestic Chinese recycling technical
specification (Huang, et al., 2008). Laboratory testing was completed at four emulsion
contents through the Marshall Stability and indirect tensile strength tests. In addition,
the indirect tensile test, using a freeze/thaw cycle, and rutting test was performed at
the predicted optimal emulsion content. For both Marshall and indirect tensile testing,
six replicate samples were prepared and separated into two groups for conditioning
(wetting) and non-conditioning, respectively. The rutting test had three replicate
samples.
Results of Marshall and Indirect Tensile Testing

Table 2 shows the results for both unconditioned and conditioned Marshall Tests at
40 °C and indirect tensile tests at 15 °C with different emulsion contents. As shown in
Table 2, Marshall Stability (MS) has no obvious trend with increasing emulsion levels.
The air voids decrease with increasing percent emulsion. Retained Stability (RS)
values are all greater than the 75% threshold. IDT strength increases with increasing
percent emulsion for S229 and S239, and IDT strength first increases and then
decreases with increased emulsion dosage for S336. Dry-Wet Tensile Strength Ratio
(DWTSR) values all surpass the 80% threshold.
Table 2. Marshall Testing and Indirect Tensile Testing.
Unconditioned Conditioned
Air
Emulsion IDT IDT RS DWTSR
Projects Voids MS MS
(%) strength strength (%) (%)
(%) (kN) (kN)
(Mpa)) (Mpa)
3.0 13.1 13.44 0.37 11.91 0.51 88.6 137.8
3.5 12.3 11.78 0.54 10.07 0.55 85.5 101.9
S229
4.0 12.1 12.94 0.56 9.93 0.78 76.7 139.3
4.5 12.0 10.71 0.61 10.42 0.63 97.3 103.3
3.0 14.7 6.98 0.53 6.31 0.57 90.4 107.5
S239 3.5 14.4 6.67 0.60 5.95 0.67 89.2 111.7
4.0 13.5 7.21 0.63 6.10 0.51 84.6 81.0
3.0 11.0 6.61 0.61 7.53 0.59 113.9 96.7
S336 4.0 9.7 5.79 0.75 5.75 0.62 99.3 82.7
5.0 8.8 5.56 0.70 6.40 0.66 115.1 94.3
Optimal Emulsion Contents

As shown in Figure 1, for S229, when the emulsion content is 3%, MS has the
maximum value, while its corresponding value of IDT strength is the lowest and less
than 0.4MPa. When the emulsion content is 4.5%, the highest IDT strength is acquired
and MS has the lowest value. However, considering limited pavement maintenance
and preservation capital, the percent emulsion 4.5% is not economical. Therefore,
when emulsion content is between 3.5% and 4%, the values of MS and IDT strength
are both reasonable. Furthermore, the air voids are all less than 14%. Combined with
the results of voids, Marshall Stability, IDT strength, and consideration of project
economics, 4% was selected as the optimal emulsion content and the mixing water
content was 2.4% for S229. The same method was used for S239 and S336, the
optimal emulsion contents and moisture contents were: emulsion content of 3.8% and
3.8% with mixing water contents of 2.4% and 2.6%.
14 0.7
IDT (MPa)
13 0.6
MS (kN)
12 0.5
11 0.4
10 0.3
2.5 3 3.5 4 4.5 5 2.5 3 3.5 4 4.5 5
Emulsion Content (%) Emulsion Content (%)
FIG. 1. MS and IDT Strength Trend (Unconditioned).
The freeze-thaw IDT testing (one cycle) and rutting test were conducted on the
mixes formulated with optimal emulsion contents and mixing water. For freeze-thaw
indirect tensile tests, one group of three replicate specimens were vacuum saturated to
55-75%, placed in sealed plastic bags for 16 hours at -18°C, and then (after removal
from the plastic bag) placed in a water bath at 60°C for 24 hours. Two groups were
soaked in 25°C water bath for at least 2 hours before testing at 25°C. The results show
that tensile strength ratio is all more than 70%. Rutting test was performed using a 0.7
MPa rubber wheel load at a test temperature of 60±1°C for three replicate specimens,
using rutting specimens having the dimensions of 300 mm × 300 mm × 60 mm. The
results indicate that dynamic stability is all not less than 2500 times/mm for three
projects.
Post-Project Monitoring
For each project, a 3 to 5 cm surface course (based on traffic volume) was applied
on the CIR binder course. FWD inspection of three recycled roads showed good
structural bearing capacity after more than one year of traffic opening with no
cracking. Rutting was also minimal with about 3 mm on the three projects. For S229,
the cracking numbers, rutting depth, and modulus of pre- and post- recycling are listed
in Table 3. The surface course for the recycled road includes 10cm CIR course and
4cm SMA.
Table 3. Field Data of Pre- and Post- Recycling on S229 Road.
Cracks Rut Modulus (MPa)

S229 (number depth Semi-rigid Surface
/km/lane) (mm) Subgrade
base course
Old road 109 20 112 7878 2098
Recycling
0 3 119 8093 2584
road
Discussion
When the emulsion content is within a reasonable range, MS and IDT strength
showed that recycled mix has good strength performance: MS is greater than 6kN and
IDT strength is more than 0.5MPa. The values of the retained stability, tensile strength
ratio and dry-wet tensile strength ratio indicated that the recycled mix exhibited good
moisture stability. Rutting test showed that recycled mix has good high temperature
stability which was vital for the Jiangsu provincial highway network, because rutting
is the major distress in summer. Post-project monitoring indicated that the CIR had
good cracking resistance, moreover, rut depth and higher modulus showed that CIR
road had a higher structural bearing capacity than the previous asphalt pavements.
Overall, the results of performance evaluations and field investigations indicated that
cement asphalt emulsion recycling mixtures has good performance.
PERFORMANCE SPECIFICATIONS
Table 5 gives a summary of the performance specifications for CIR that was used as
a binder course using the SGC in Jiangsu province. Future performance specifications
will include the raveling test and low temperature cracking test (Thomas et al., 2003).
Table 5. The Performance Specifications Summary of CIR.
Property Criteria Purpose

1.25 degree angle
Compaction effort,
600 kPa stress Density Indicator
Gyratory Compactor
30 gyrations
Void content 8-14 % Compacted Density
Gradation for RAP Report Mix design
Marshall stability, 40ć 6 kN min Stability Indicator
Retained stability 75 % min Moisture susceptibility
Indirect Tensile test, 15ć 0.5 MPa min Strength Indicator

Retained tensile strength 75 % min Moisture susceptibility
Freeze-thaw IDT, 25ć 70 % min Moisture stability
Rutting test, 60ć 2500 times/mm min High temperature stability
CONCLUSIONS
This research studied the performance characteristics of cement asphalt emulsion

Cold In-Place Recycled (CIR) mixes using the Superpave Gyratory Ccompactor
(SGC). A mix design procedure using performance evaluation tests coupled with the
recommended performance specifications has been used on three projects. Recycled
mixes exhibited satisfactory cracking, rutting, and modulus properties. The results of
lab tests and field verification indicated that asphalt emulsion CIR is a suitable
technique for the rehabilitation of deteriorated asphalt pavements in Jiangsu province,
China.
ACKNOWLEDGEMENTS
This study was supported by the Jiangsu Provincial Transportation Science and
Technology Project. We would like to thank Li Xiaoming, Dr. Ma Xiang, Xu Biao, Li
Jian, and Dr. Everett Crews who assisted in the lab testing and field construction.
Thanks are also extended to Dr. Andrew Franz Braham for his revisions, edits, and
comments on the paper.
REFERENCES
Alkins A., Lane B., Kazmierowski T. (2008). “Sustainable pavement-environmental,

economical and social benefits of in-situ pavement recycling.” 87th TRB Annual
meeting, Washington, D.C..
Cross, S.A. (2002). “Determination of Ndesign for CIR mixtures using the Superpave
gyratory compactor.” RMRC research project No.15 final report, Kansas: 52-57.
Epps J.A. (1990). “Cold-recycled bituminous concrete using bituminous materials,
NCHRP synthesis of highway practice 160.” TRB 69th Annual meeting,
Washington, D.C..
Huang S.C., Xu J., Dong, P.R., Zhou, G.L., et al., (2008). “Technical specifications for
asphalt pavement recycling.” JTJ F41-2008, Research Institute of Highway
Ministry of Communications, Beijing.
Thomas T. and Kadrmas A. (2003). “Performance-related tests and specifications for
cold in-place recycling: lab and field experience.” TRB 82nd Annual meeting,
Washington, D.C..
Mechanical Properties of Concrete Pavement with Different Isolation Layers on
Lean Concrete Base
Chen Yu-liang1, Li Xue-lian2, Zhang Qi-sen3, Zhou Zhi-gang4

1
Doctoral candidate, School of Communication and Transportation Engineering, Changsha University
of Science and Technology, Chiling Road 45#, Changsha, Hunan, China, 410076; Lecturer, School of
Engineering & Architecture, East China Jiaotong University, Nanchang, Jiangxi, China, 330013.
Phone: (86731) 230-9023; E-mail: [email protected]
2
Lecturer and PhD, School of Communication and Transportation Engineering, Changsha University
of Science and Technology, Chiling Road 45#, Changsha, Hunan, China, 410076. E-mail:
[email protected]
3
Professor and PhD supervisor, School of Communication and Transportation Engineering, Changsha
University of Science and Technology, Chiling Road 45#, Changsha, Hunan, China, 410076. E-mail:
[email protected]
4
Professor and PhD supervisor, School of Communication and Transportation Engineering, Changsha
University of Science and Technology, Chiling Road 45#, Changsha, Hunan, China, 410076. E-mail:
[email protected]
ABSTRACT: With increasing traffic volumes and higher wheel loads, the structures
of Cement Concrete Pavement (PCC) with lean concrete base (LCB) have been used
widely in China. However, reflection cracking is one of the primary distresses in
PCC on LCB. Although bond breaker mediums have been utilized with success to
reduce and minimize undesirable cracks reflecting from lean concrete bases, the
mechanisms of delaying or/and reducing reflection cracking propagation have yet to
be fundamentally understood. In this paper, with the basic parameters obtained from
Qing-Lian highway, three-dimensional FEM considering contacts between layers
was utilized to analyze the pavement systems with different bond breaker media. It
was found that pavement performance is less favorable under wheel loading when
there is a good bond breaking. However, the pavement performance is more
favorable under thermal loading when there is a good bond breaking. A good bond
breaking is essential for PCC on LCB as the undesirable cracks are the results of the
thermal loading. Based on this study, it was found that wax can be used effectively to
reduce the tensile stresses at the interface between PCC slab and LCB such that the
erratic cracks due to the temperature variations can be minimized.
KEYWORDS: road engineering, isolation layer, lean concrete, waxy isolation

layer, slurry seal, load stress, thermal stress
56
INTRODUCTION
To improve the Cement Concrete Pavement (PCC) performance, cement-treated or

lean concrete base (LCB) has been utilized worldwide to reduce PCC pavement joint
deflections and stresses. Because of the increasing traffic volumes and higher wheel
loads, the structures of PCC on LCB have been used widely in China in the last 10
years. Literature reviews indicate that many other countries have been utilizing LCB,
including Britain, Germany, Brazil, Australia, and Belgium. The engineering
practices of Britain showed that using LCB can provide good base support and load
transfers between concrete slabs (Zhao et al. 2003). Recent researches demonstrate
that there are high frictions between LCB and concrete slabs (Hu et al. 2007). The
high frictions have been found to be the culprit of the reflection cracking. Scott M
Tarr et al. (1999) analyzed interlayer bonding conditions between different bond
breaker media and concrete slabs and found bond breaker media have significant
impacts on pavement stresses. To support and provide a better guidance for an
on-going highway (referred as Qing-Lian highway) construction, a research project
was funded to investigate the influences of the different bond breaker media for LCB
and concrete slabs. Theoretical analyses were performed to quantify the effects of
different bond breaker mediums on reducing pavement stresses.
Three-Dimensional Finite Element Model with Consideration of Interlayer

Effects
The accuracy of the mechanical behavior modeling depends greatly on the method
of handling the interlayer system. The traditional models such as friction model,
sandwich model, or equivalent rigidity method using combination factor, were not
able to fully simulate the actual stress states (Tan et al. 2004). To overcome those
shortfalls, this study uses interlayer contact model combining both the friction and
the spring effects. Spring model describes plane deformation of the weak interface,
using normal stiffness and tangential stiffness to describe the relationships between
the top of the spring and the bottom of the mechanics. It simulates both compression
and shear stresses. Tangential stiffness and normal stiffness are interrelated with the
modulus of the top and the bottom material and thickness. On other hand, friction
model described the interface force state, the state of the friction between concrete
surface slab and LCB. It normally changes with time, from a rough surface with a
gradual transition to a smoother surface. To consider the attenuation phenomenon of
interface friction effects, different friction coefficients depending on the
circumstances, were utilized in the interlayer analyses. Three-dimensional model of
contact elements provided by the ANSYS finite element package was employed in
this study to simulate the interlayer effects.
Figure 1 illustrates the 3-D finite element model with 8 nodes isoparametric
element utilized in this study. It demonstrates the free slabs with three-layer system.
The following boundary conditions were used: (1) complete restraint for the bottom
of the soil (2) the surface layer is free with no any restraint (3) the displacements at
the side level direction are zero.
y
x
o
FIG.1. Model Simulation for Analysis.
Fundamental Material Parameters
The accuracy of the model prediction depends on the simulation of the interlayer
contact condition. It is very critical to determine appropriate parameters such as
normal stiffness, tangent stiffness, cohesion and friction coefficient. In this study,
cohesion and friction coefficient values of bond breaker media were determined
directly from the field shear and pull-out tests (Figure 2), and the values are shown in
Table 1. Three different bond breaker media were utilized on the Qin-Lian highway:
(1) slurry seal, (2) wax, and (3) control section with no bond breaker media. Other
pertinent pavement material parameters are presented in Table 2.
˄a˅Field Shear Test. ˄b˅Field Pull-Out Test.
FIG.2. Field Tests on Qin-Lian Highway.

Table 1. Cohesion and friction coefficient of different bond breaker mediums.

The type of isolation layer Cohesion˄MPa˅ Friction Coefficient
Control 1.00 0.14
Wax 0.78 0.47
Slurry seal 0.54 0.58
Table 2. Basic material parameters of pavement structure layers.

Thickness Modulus of resilience Poisson
Material
˄cm˅ ˄MPa˅ Ratio
Cement concrete slab 30 31000 0.15
LCB 20 20000 0.15
4%cement stabilized subbasel 18 1000 0.2
Subgrade - 40 0.35
Loading Conditions
Based on the Chinese Asphalt Pavement Design Criterion (JTG D50-2006), the
standard axle load was used in the analysis. To simplify the analyses, the wheel tire
contact area was modeled as a square, which is 189mm×189mm. The distance
between the wheel tires is 320mm. In addition, the wheel loads were assumed to be
uniformly distributed by taking the vertical and horizontal loads into account in the
calculation. The value of the vertical load was 0.7MPa. The horizontal load was
assumed to be proportional to the vertical load with the relationship given in
Equation 1. As given in Equation 1, the proportional coefficient (f ) has a constant of
0.5.
q=f×p=0.5p (1)
where: q is the horizontal load, and p is the vertical load.
MECHANICAL PROPERTIES OF PAVEMENT STRUCTURES WITH

DIFFERENT BOND BREAKER MEDIUMS
Pavement Structure Analysis under Vehicle Load
Base on the material parameters presented in Tables 1 and 2, the results from 3D
FEM analyses for legal load are presented in Table 3. As aforementioned, the
material parameters presented in Tables 1 and 2 were obtained from the in-situ tests
on Qin-Lian highway. To consider the overloading condition, the vertical load was
raised up to 50% of the legal load. The vertical load for the 50% overloading is 1.05
MPa. The comparisons of the surface deflections, tensile stresses, and shear stresses
between the standard legal load and 50% overload are presented in Figure 3.
Table 3. The calculation result of Qin-Lian road under standard vertical load.
maximum tensile stress Maximum shear
surface
Index The bottom stress between
deflection The bottom of
Types of slab slab and base
˄0.01mm˅ base˄MPa˅
˄MPa˅ (MPa˅
Control 20.3 0.1646 0.3679 0.1144
Wax 20.9 0.2752 0.3759 0.1109
Slurry seal 20.3 0.1704 0.3676 0.1143
VWDQGDUGORDG RYHUORDG
VXUIDFHGHIOHFWLRQ˄PP˅

QRLVRODWLRQ ZD[\LVRODWLRQ VOXUU\VHDO
OD\HU OD\HU
(a) Surface Deflection.

WKHERWWRPRIVODE03D
PD[WHQVLOHVWUHVV

OD\HU OD\HU
(b) Max. Tensile Stress at the Bottom of Slab.


WKHERWWRPRIVODE03D

PD[WHQVLOHVWUHVV

OD\HU OD\HU
(c) Max. Tensile Stress at the Bottom of the Base.

EHWZHHQVODEDQGEDVH03D
0D[LPXPVKHDUVWUHVV

OD\HU OD\HU
(d) Max. Shear Stress between Slab and Base

FIG.3. Comparisons of the Surface Deflections, Tensile Stresses, and Shear
Stresses between the Standard Legal Load And 50% Overload.
In view of the results presented in Table 3 and Figure 3, the observations are given
as follows:
(1) It is interesting to see that the surface deflections are about the same
regardless of the use of bond breaker mediums. As expected, the surface deflections
under 50% overloads were much higher than those under the standard legal loads.
(2) Tensile stress at the bottom of the concrete slab was the greatest when wax was
used as the bond breaker medium. In comparison to the control section with no bond
breaker medium, tensile stresses at the bottom of slab were 67.2% and 3.5% higher
when wax and slurry seal were used, respectively.
(3) Tensile stresses at the bottom of the base were approximately the same,
regardless of the types of bond breaker mediums. Same observation holds true for the
maximum shear stress between the slab and the base.
(4) Tensile and shear stresses were significantly higher (e.g., 20%) than those
under the standard legal loads as they were compared to the 50% overload. As it is
believed, overloading would no doubt cause substantial damages to the highway
pavements.
Temperature Stress Analysis of Pavement Structure with Different Isolation

Layers
The bonding conditions between PCC slab and LCB are very complex, as they are
impacted by, among others, the use of material and construction techniques. In
addition, temperature plays a critical role on the performance of the PCC pavements.
Unfortunately, the existing PCC design specifications did not address the bonding
conditions between PCC slab and base under the impacts of temperature loading.
Thus, it is necessary to conduct thermal stress analyses for different bond breaker
mediums. With the aid of 3D FEM, it is expected that computational results would
further improve pavement design theory in China.
Temperature Gradients
The thickness of PCC slab in Qin-lian highway is 300mm. Based on the

literature reviews, the two critical parameters were adapted in this study: (1)
maximum temperature gradient correction coefficient of 0.79, and (2) maximum
temperature gradient of slab of 21.33ć (Zhang et al. 2005). As reported by Zhao et
al. (2007), the variation of temperature would decrease significantly when it is below
a certain depth. The temperature gradients shown in Table 4 were utilized in this
study. Note that the reference temperature was 30ć.
Table 4. Pavement structure temperature distribution of Qin-lian road.
Location Temperature(ć)
Pavement surface 55.33
The bottom of slab 39
The bottom of base 36
The bottom of subbase 35
The bottom of subgrade 30
Results and Analysis
Table 5 presents stress analysis results for the three different bond breaker
mediums: (1) slurry seal, (2) wax, and (3) control section with no bond breaker
medium.
Table 5. Temperature stress analysis results of Qin-Lian Highway.

maximum tensile stress Maximum shear
Index
The bottom of The bottom of stress between slab
Types
slab˄MPa˅ base˄MPa˅ and base (MPa˅
Control 1.276 0.042 0.396
Wax 1.020 0.033 0.496
Slurry seal 1.148 0.037 0.446
In view of Table 5, it is observed that:

(1) Tensile stress at the bottom of the slab with wax as the bond breaker medium is
the least, as it has the least friction at the interface between PCC slab and LCB. At
the same location, the tensile stresses in the control and slurry seal are 25.1% and
12.5% higher than those with wax as the bond breaker medium, respectively.
(2) Tensile stresses at the bottom of base are approximately the same, regardless of
the types of the bond breaker mediums.
(3) The maximum shear stress between the slab and the base is the least for the
control with no bond breaker medium. However, the maximum shear stress between
the slab and base for wax and slurry seal are 25.3% and 12.6% higher than those with
the control with no bond breaker medium, respectively.
CONCLUSION
To improve the PCC performance, LCB has been utilized worldwide to reduce
PCC pavement joint deflections and stresses. However, there is a high potential for
bonding of the PCC slab to a stabilized or LCB that would induce uncontrolled
cracking because of the high friction and/or bonding between the base and the PCC
slab as the friction or bond restrains the concrete's volume change (shrinkage or
contraction), inducing higher stresses that result in cracking. Bond breaker mediums
have been utilized with success to reduce and minimize undesirable cracks reflecting
from lean concrete bases. In this paper, with the basic parameters obtained from
Qing-Lian highway, three-dimensional FEM considering contacts between the layers
was utilized to analyze the pavement systems with three different bond breaker
mediums. The observations and conclusions are given as follows:
(1) Tensile and shear stresses were significantly higher (e.g. 20%) than those
under standard legal loads as they were compared to the 50% overload. It is believed
that the overloading would cause substantial damages to the highway pavements.
(2) Wax is the best bond breaker medium as it has the least friction at the interface
between PCC slab and LCB. Consequently, it yielded the least maximum tensile
stress at the bottom of slab under thermal loading.
(3) It was interesting to find that the tensile stresses at the bottom of base were
approximately the same, regardless of the type of the bond breaker mediums.
(4) Based on this study, it was found that wax can be used effectively to reduce the
tensile stress at the interface between PCC slab and LCB such that cracking due to
temperature variations can be minimized.
REFERENCES
Scott M Tarr, Paul A Okamoto O, Mathew J, Sheehan, and Robert G Packard. (1999).
“Bond interaction between concrete pavement and lean concrete base”
Transportation Research Record 1668: 9~7.
Hu Chang-bin and Zeng Hui-zhen. (2007). “A Study on Monitor Measurement of
Temperature Field of Cement Concrete Pavements Structure in Fujian Province”
Journal of Highway, Vol. 8:69-76.
JTG D40-2002ˈCement Concrete Pavement Design Specification.
Tan Zhi-ming, Yao Zu-kang, and Liu Bo-ying. (2004). “Analysis of Thermal Stresses
in Cement Concrete Pavements” Journal of Highway, Vol. 11:63-67.
Zhao Jun, Tan Zhi-ming, and Liu Bo-ying. (2007). “Effect and Correction on Slab
Interior Deflection of Cement Concrete Pavement Due to Temperature Gradient”.
Journal of Tongji University (Natural Science), Vol. 35(5):617-620.
Zhao Wei-cheng, Xu Zhi-hong, Huang Wen. (2003). “Influence of load stress and
deflection on interfacial interaction between Portland concrete pavement and lean
concrete base” China Journal of Highway and Transport, Vol. 16(4):9-15.
Zhang Hong-bo, Zhou Zhi-gang, Chen Xiang, Zhang Qing-ping. (2005). “Research
on the Interfacial Characteristics between PCC Surface layer and Poor Cement
Concrete Base Course”. Journal of Changsha Communications University, Vol.
4:17-20.
Evaluation of the Compaction Characteristics of Unbound Material using the
Superpave Gyratory Compactor
Nicholas Lambert1, Kyle Denny2, Beena Sukumaran 3, A.M. ASCE and Yusuf
Mehta4, A.M. ASCE
1
Graduate Assistant, Civil Engineering, Rowan University, Glassboro, NJ-08028, USA; PH: (856)
256-5324; FAX: (856)256-5242; e-mail: [email protected]
2
Undergraduate Assistant, Civil Engineering, Rowan University, Glassboro, NJ-08028, USA; PH:
(856) 256-5324; FAX: (856)256-5242; e-mail: [email protected]
3
Associate Professor, Civil Engineering, Rowan University, Glassboro, NJ-08028, USA; PH: (856)
4
Associate Professor, Civil Engineering, Rowan University, Glassboro, NJ-08028, USA; PH: (856)
ABSTRACT: During previous full scale tests at the National Airport Pavement
Facility (NAPTF) at Atlantic City, - USA, significant consolidation of the subbase
layer occurred during aircraft trafficking accompanied by shear flow failure in the
underlying low strength subgrade material. In order to understand the compaction and
shear flow characteristics of the subbase layer during construction and trafficking, the
material was compacted in the Superpave Gyratory Compactor (SGC) at various
stress levels and at different moisture contents. Since the SGC is primarily used for
compaction of asphalt concrete, the compaction characteristics in the Superpave
Gyratory Test were calibrated to field compaction of the subbase material during
construction at similar moisture contents. The field data collected included the change
in density of the subbase layer during compaction with number of passes of the roller.
The calibration was done by comparing the shape of the curve. The calibration
provided the benchmark values for laboratory evaluation of unbound material in
SGC, such as stress level and the gyratory angle of testing. The shape of the
compaction curve and the density at various gyrations levels provided an insight into
the understanding of the compaction characteristics of the subbase material under
roller compaction. This paper provides the airport pavement designers the efficacy of
using the SGC as a means of evaluating the performance of unbound material during
compaction.
65
INTRODUCTION
With the introduction of larger planes, such as the Airbus A380, which weigh
approximately 0.6 million kg (1.3 million pounds), the Federal Aviation
Administration (FAA) is testing airport pavements under heavy aircraft loading at the
National Airport Pavement Test Facility (NAPTF). As a result of the heavy aircraft
loading, the FAA has found that the subbase layer compacted past the maximum
Modified Proctor Density (ASTM D1557) during the simulated trafficking (Hayhoe
and Garg, 2007) with very little crushing of the material (Santino and White, 2006).
The subbase layer, which the FAA has labeled as P-154, is a non-cohesive granular
material with properties as indicated in Table 1. To improve the understanding of
compaction characteristics of unbound materials such as the P-154, it was decided to
utilize the Superpave Gyratory Compactor (SGC) because of the larger energy
capabilities it possesses. The gyratory compactor is currently used in the Superpave
mix design method, (AASHTO M323), which only applies to asphalt concretes and
not unbound materials. The gyratory compactor applies a constant pressure as well as
a shear force which simulates traffic more closely.
Table 1. FAA P-154 Material Properties and Classification.

Properties Results
Maximum Modified Proctor (kN/m3) 20.4
Optimum Moisture Content (%) 6.9
Coefficient of Uniformity, Cu 26
Coefficient of Curvature, Cc 3.25
Mean Particle Size, D50 1.5 mm
AASHTO A-1-b
The evaluation of the relationship between laboratory compaction methods and field
compaction is important because each construction cycle and analysis of trafficking is
expensive and time consuming. The FAA spends millions of dollars for each
construction cycle, therefore the material properties must be established in the
laboratory prior to construction. This approach has limitations because the material in
the field does not compact similar to material in a laboratory test because in the field
the unbound aggregate has more freedom for particle rearrangement whereas in the
laboratory, the Proctor and SGC mold are confined apparatus.
The gyratory compactor has many variables including pressure, angle, number of
gyrations, and gyration rate. When trying to match the compaction curve from the
field, any of these variables can be changed. In most of the tests that were conducted,
the maximum pressure of 1000 kPa (145 psi) was used with an angle of 1.25 degrees
and up to 500 gyrations. The gyratory compactor outputs the height versus the
number of gyrations for every gyration conducted in the test. By knowing the cross
sectional area of the compaction mold, the volume of the soil can be calculated from
the sample height shown in Fig. 3.
The field compaction methods used in this study will be based upon the methods
chosen by the FAA National Airport Pavement Testing Facility (NAPTF). The SGC
testing was not performed at or above the Optimum Moisture Content (OMC)
because the FAA cannot compact in the field at high levels of moisture. The primary
compactor used for the subbase material is a Pneumatic tire roller which was
configured to weigh 20,956 kg (46,200 lb) giving it a ground contact pressure per
wheel ranging from 345 kPa (50 psi) to 965 kPa (140 psi) depending on the tire
pressure. A vibratory drum roller was also used on the subbase during construction
which uses a 2,134 mm (84-in.) wide, single-drum vibratory compactor with an
operating weight of 11,990 kg (26,433 lb.) and centrifugal weight of 26,127 kg
(57,600 lb.) Each roller makes a predetermined number of passes or until the
maximum achievable density is reached. As with the gyratory compactor, the
material would eventually reach a point where additional passes would no longer
increase the density.
The density was measured in the field using a nuclear density gauge which
produces fast and accurate results for soil density and moisture measurements. The
measurement of soil density has been a problem for years mostly because of
uncertainty in the density at different positions in the soil. A single probe gauge uses
one probe lowered into a hole using a push rod with a receiver that remains at the
surface. The nuclear density gauge can work at depths ranging between 51 to 305 mm
(2 to 12 in.) below the surface at 25.4 mm (1 in.) increments. Nuclear gages can
measure a density range of 1,121 to 2,723 kg/m3 (70 to 170 lb/ft3) and within 4 kg/m3
(0.25 lb/ft3) for a 1 minute interval (Ayers et al. 1985).
SUPERPAVE GYRATORY COMPACTOR RESULTS
The SGC was used to test the P-154 at many different moisture contents in order to
make comparisons to the field compaction at similar values. The SGC results are
consistent with the Modified Proctor results which indicate that as you reduce the
moisture content below a particular level, the maximum Modified Proctor density is
unattainable. This is evident as you go below three percent moisture and are unable to
reach the maximum Modified Proctor density. The results from the Superpave
gyratory compactor were split into separate fields from percent modified proctor in
order to evaluate the compaction characteristics at different levels. The average
number of gyrations to compact the aggregate from one level to another is shown in
Table 2 for different moisture contents and pressure applied from the Superpave
gyratory compactor. In Table 2, 85% Modified Proctor density was considered the
baseline or loose compaction from the field compaction results and allows for a
starting point.
Table 2. SGC Results for P-154 Field Compaction Comparisons.

Pressure
1000 1000 1000 1000 1000 1000 600 600
(kPa)
Moisture 6.0- 5.0- 4.0- 3.0- 2.0- 1.0- 2.0-
1.0-2.0
(%) 7.0 6.0 5.0 4.0 3.0 2.0 3.0
No. of Test 3 4 6 5 3 7 2 4
% Mod.
Average Number of Gyrations
Proctor
85-90% 3.00 3.25 3.83 6.20 5.67 6.57 10.00 14.25
90-95% 6.67 7.00 10.00 23.40 23.00 33.00 53.50 81.25
95-100% 22.33 19.25 34.33 50.75 - - - -
100-105% 95.67 94.00 216.17 264.00 - - - -
It is evident that the pressure applied by the gyratory compactor and moisture content
both affect the compaction rate because the tests performed below three percent
moisture were unable to exceed 100% maximum Modified Proctor density after more
than 200 gyrations.
Additional testing with the Superpave gyratory compactor was done with different
samples sizes to simulate lift height. This testing was done in order to give the FAA a
recommendation on achieving the highest density without over saturating the
material. In Fig. 1, the Superpave gyratory results are shown as a function of moisture
content to help illustrate the effect of moisture content and lift height on the P-154
material. The results from the Superpave gyratory compactor indicated that at smaller
lift heights 100% maximum Modified Proctor density was attainable above 4%
moisture. The recommended lift height was approximately 4 inches, which was used
for the final two lifts of the P-154.
1 45
50 0 g yrat io ns & 50 00 g
1 40
30 0 g yrat io ns & 50 00 g
f)
c
(p M o di fie d Pr o cto r
y
it
s1 35
n 20 0 g yrat io ns & 30 00 g
e
D
y
r 30 0 g yrat io ns & 30 00 g
D
.
x1 30
a 20 0 g yrat io ns & 30 00 g
M (a d di tio nal test)
30 0 g yrat io ns & 30 00 g
(a d di tio nal test)
1 25
3 4 5 6 7 8 9 10
M o ist u re C o n te n t (% )
FIG. 1. Evaluation of SGC results Below OMC3 with Simulated Lift Height
(1 pcf = 16.02 kg/m ).
FIELD COMPACTION CHARACTERISTICS
During construction cycle five (CC5) at the NAPTF, the field compaction was
closely monitored and resulted in density measurements below the maximum
Modified Proctor density value. This lower Modified Proctor value could be due to
the low amount of moisture being added prior to compaction to prevent over
saturating of the weak subgrade as stated previously. The field construction
compaction was monitored for all five lifts of the P-154 as the layers were
compacted. The results from lift one, two, four, and five are shown in Fig. 2.
Different rollers were used for compaction for the various lifts. Lift three was not
included in Fig. 2 because it was from a different screening which resulted in a higher
maximum Modified Proctor density. The first data point for all the lifts are the
baseline results which were considered loose compaction from placing the material.
This value varied for all the lifts but did not seem to effect the overall compaction.
The first lift of the P-154 consisted of only six passes with the roller to prevent
penetrating the clay subgrade layer with the P-154 and resulted in the lowest density
recorded. After a specific number of passes with the roller, the change in height and
density were measured at six separate locations at the NAPTF. The placement and
testing of each point was not the exact same point because the nuclear density gauge
uses a probe and can potentially disturb the surrounding soil. Therefore, it is more
accurate to look at the average for a series of points to find any trend in compaction.
Lift four was done on two separate days with a different nuclear density gauge based
on equipment availability, which resulted in a gap in the data. For the analysis of the
P-154, lift four day two was not used in finding the compaction characteristics but
shown in the figure for completeness due to the discontinuity in the data obtained by
using different gauges and performing the compaction over two days. The
compaction of the P-154 in Fig. 2 approaches 95% Modified Proctor density in lifts
two and four and exceeds 95% in lift five. Lift five resulted in a lower average
density reading after two additional passes with the rubber tire roller.
100
95
90
% Modified
Proctor
Field Data Lift# 1

85
Field Data Lift# 2
80 Field Data Lift# 4
Field Data Lift# 5

75
70
0 5 10 15 20
No. of Passes
FIG. 2. Field Construction Compaction Results for P-154 during CC5 with the
Following Moisture Contents: Lift 1 – 2.9%, Lift 2 – 2.6%, Lift 4 – 2.5% and
3.4%, Lift 5 – 3.0%
After analysis of the field construction compaction and the Superpave gyratory
compactor testing there was a close relationship with the compaction characteristics.
This can be seen very well in Fig. 3. In this figure, the number of passes of the roller
and number of gyrations in the SGC are plotted on the x-axis and the dry density
appears on the y-axis. Because the scale of the figure can misrepresent the actual
data, the field compaction and gyratory compaction can be evaluated using the
number of gyrations and passes to achieve a specific density. Another factor that must
go into this evaluation is the energy input with the different rollers used during
construction. This can be accounted for by increasing or decreasing the pressure
applied with the Superpave gyratory compactor. Further testing must be completed in
order to verify the number of passes versus the number of gyrations for the P-154 for
consistency.
No. of Passes
0 2 4 6 8 10 12
135
130
125
Dry Density
120
(pcf)
115
110
Gyratory Compactor 03.17.08 Test#5
105 Lift # 3
100
0 10 20 30 40 50
No. of Gyration
FIG. 3. Comparison of P-154 Gyratory Compactor Results and Field

Construction Compaction from Lift # 3 during CC5 (1 pcf = 16.02 kg/m3).
CONCLUSIONS
Based on the study comparing the SGC results with field compaction results, it can
be seen that the SGC is capable of replicating field compaction results even with
different field compaction equipment being used. It is also capable of achieving much
higher densities than the Modified Proctor test, where a standard energy input is used.
Further studies are currently underway to determine if the SGC can be used to
replicate field trafficking under heavily loaded aircraft wheel loads.
ACKNOWLEDGEMENTS/DISCLAIMER
The work described in this paper was supported by the FAA Airport Technology
Research and Development Branch. The contents of the paper reflect the views of the
authors who are responsible for the facts and accuracy of the data presented within.
The contents do not necessarily reflect the official views and policies of the FAA.
REFERENCES
Ayers, P. D. and Bowden H. D. (1985), “Laboratory Investigation of Nuclear Density

Gauge Operation,” ASAE Tech. Paper No. 85-1541.
Hayhoe, G. F., Garg, N. (2007), “Consolidation of aggregate base and subbase layers
under heavy aircraft loading at the National Airport Pavement Test Facility,”
Proceedings of the Advanced Characterisation of Pavements and Soil
Engineering Materials, Athens, June 2007.
White, B., and Santino, D. (2007), “Particle crushing in base and subbase during
aircraft trafficking,” Rowan University Junior/Senior Clinic Report, 2007.
Preliminary Design of Testing Segment for Accelerated Loading Facility based on
Finite Element Simulation Analysis
Dong Zejiao1, Tan Yiqiu2 and Chen Fengchen3

1
Doctor, School of Transportation Science & Engineering, Harbin Institute of Technology, No.202
Haihe Road, Nangang District, Harbin 150090, China; [email protected]
2
Professor, School of Transportation Science & Engineering, Harbin Institute of Technology, No.202
Haihe Road, Nangang District, Harbin 150090, China; [email protected]
3
Candidate of PhD, School of Transportation Science & Engineering, Harbin Institute of Technology,
No.202 Haihe Road, Nangang District, Harbin 150090, China; [email protected]
ABSTRACT: It has been recognized that full-scale Accelerated Loading Test (ALT) is
the most important and direct way to obtain structural behavior of pavements subjected
to traffic loading, as it avoids many assumption compared to that in the laboratory. As
one part of ”the Eleventh- Five-Year National Grand Technology Infrastructure
Program”, a full-scale Accelerated Loading Facility (ALF) with full-controlled
environment conditions has been designed in China. At first, a brief introduction of
ALTs all over the world was given followed by a description of track line types of two
proposed schemes. Then, a comparison between stress state of the circular track and that
of the oval one was analyzed through a three dimensional finite element simulation to
identify a proper track line type. Finally, the testing pavement dimensions, i.e. planar
sizes and depth, were determined based on the variation rules of stress components. The
proposed design parameters according to finite element simulation will be helpful to the
ultimate design of testing track.
INTRODUCTION
Accelerated Loading Test (ALT) is one of the most important means to acquire
structural behavior of pavements. It is crucial to the structural design, performance
prediction, and maintenance decision of pavements. It is urgently needed in China due
to some engineering problems which have arisen in the recent years, such as premature
failure of asphalt pavements. As a result, promoted by the National Development and
Reform Commission, an establishment item aimed at service safety evaluation for
important engineering materials, was launched in March 2007 and includes
establishment of the ALF especially for testing under special regional environmental
conditions. The proper size of the test segment, i.e., preliminary design of testing
pavement geometry, is analyzed by finite element (FE) simulation here. The first job is
to determine an appropriate scheme by comparing two proposed testing line-types, i.e.,
72
the circular track and the oval one. Then, the proper dimensions for the testing pavement
section, the depth and width, are obtained through the analysis of decay rules of
pavement response. A preliminary design of pavement segment is performed
considering a series of influential factors in the final part. The work presented here gives
a base and reference for the establishment of similar facilities.
SUMMARY OF EXISTING ACCELERATED LOADING FACILITIES
There are more than forty ALFs being used in more than twenty countries in the world
today. Generally, these ALFs can be classified into four types based on their
geometries: large full-scale real track, linear testing track, circular testing track, and
oval testing track.
Fig. 1. Typical ALFs. a) Large-scale real track, American MnROAD (top left);
b)Circular testing track, French LCPC(top center); c)Oval testing track, Spanish
CEDEX (top right); d)Linear testing track, American FHWA ALF.
Fig.1a shows a typical large-scale real track, taking MnROAD(Minnesota test road) as
an example. It has the best correlation with the actual traffic and pavement conditions,
and can provide the most convincing testing data about pavement performance. It is
time-consuming and its cost for construction and operation is rather higher; this makes
it difficulty for world-wide application. Fig.1b shows a typical circular testing track;
giving the French LCPC (Laboratoire Central des Ponts and Chaussees) near Nantes as
an example. The circular track applies loading through the wheel rotating the
underpinning at the center, which can reach higher vehicle speeds and thus able to
simulate high-speed traffic action. Because the transverse shear force is difficult to
eliminate completely, this often results in a different stress distribution within pavement
structure. Fig.1c shows a typical oval testing track, which is the Spanish CEDEX
(Center for Public Works Studies and Experimentations). This kind of a test road
contains some advantages of the linear and circular tracks. It can eliminate transverse
shear forces in the linear segment compared with the circular track and can reach
relatively higher vehicle speeds compared with the linear track. Fig.1d shows a typical
linear testing track, American FHWA ALF for instance. For this kind of ALF, the area
occupied is small, and its construction and operation cost is comparatively lower.
However, its speed is usually restricted by the length, which may not reflect the effect of
high-speed traffic.
COMPARISON BETWEEN CIRCULAR AND OVAL TRACK
In the preliminary design, there are two schemes proposed: circular track subjected to
half-axle vehicular loading by a simulator, and an oval track subjected to the whole-axle
loading by a real truck, as shown in Fig. 2, which are both to be built under the ground in
order to easily control temperature.
Test sections
Fig. 2. Two proposed schemes. a) Circular track with Simulator (left); b) Oval
track with truck (right) (not to scale).
As the tire scrub and transverse shear force are difficult to eliminate unconditionally in
the circular track, there is no proof showing that the response is comparable between
these two cases, especially for the given boundary and loading conditions. As a result,
the static responses of the circular and linear track subjected to its corresponding
loading are analyzed here by FE simulation.
A typical rectangular loading, with an area of 0.28m long and 0.2m wide with a contact
pressure of 700kPa, was used to represent the half-axle vehicular loading in the circular
track model. Two of them were used in the straight track model (a typical 1.4m is
chosen for the distance of two tires). Typical elastic properties and layer thickness of
pavements in China were used, shown in Table 1. The bottom of both models was
completely restrained in the FE model. Horizontal movement perpendicular to the
perimeters was also restrained whereas the remaining two directions were free.
Table 1. Material and structure properties for finite element modeling.
Upper Middle Lower Upper Lower

Properties Subgrade
asphalt layer asphalt layer asphalt layer base layer base layer
Thickness (cm) 5 6 8 20 20 140*
Elastic modulus
1400 1400 1400 1600 600 50
(MPa)
Poisson’s Ratio 0.30 0.35 0.35 0.20 0.30 0.40

*Note: The thickness is used in the comparison between the circular and oval track.
Fig. 3 shows the surface deflections of the two cases. It can be seen that the
unsymmetrical boundary between the inboard and outboard sides has an observable
influence on the deflection distribution in the circular track. It indicates that the
pavement response is more susceptible to the boundary conditions compared with that
in the straight track, which will result in more difficulties in in-situ pavement response
monitoring. Further more, it can also be seen that there is an interaction between the two
tire loadings in the straight track case, namely the superimposing effect of vehicular
loading, which results in a different response distribution and peak value compared to
that in circular track case. These phenomena can also be- observed when comparing the
vertical strain, transverse strain, and longitudinal strain between them. This also
indicates that the incapability of applying the whole axle vehicular loading in circular
track makes the simulation of the full-scale pavement response restricted.
Fig. 3. Surface deflection results. a) Circular track; b) Straight track.
Table 1 shows the comparison of maximum response value between the two cases, in
which the relative percentage is defined as the percentage ratio of response in a circular
track to that in straight track. As can be seen from Table 2, the surface deflection and
the transverse strain for the circular track are lower than that of straight track by75.4 and
65.6% of the latter, respectively; while the vertical and longitudinal strain are 24.0 and
9.4% higher than the latter, respectively.
Table 2. Comparison of maximum response values.
Relative
Item Circular Straight
percentage
Maximum surface deflection (m) 0.0002362 0.0003134 75.4
Maximum vertical strain İzz (ȝİ) 478.3 385.7 124.0
Maximum transverse strain İyy (ȝİ) 113.8 166.0 65.6
Maximum longitudinal strain İxx (ȝİ) 150.8 137.8 109.4
SIZES OF TESTING SEGMENT SECTION
Since the testing pavement structure will be built in a U-shaped concrete box under the
ground to provide a full-controlled environmental condition, the rigid boundaries will
be used (concrete wall) on both sides, and a thin concrete slab with low strength will be
incorporated in the soil subgrade. An evaluation of the effects of rigid boundaries on the
response of the testing pavement was made using FE simulation. The same vehicular
loading as proposed in the previous text was used to conduct the analysis. Although the
structure and loading are both symmetric, no symmetry was considered in the three
dimensional FE model to simulate the whole-axle tire loading action.
The surface deflection, transverse stress at the bottom of hot-mix asphalt (HMA) layer,
and vertical strain on the top of subgrade are chosen as critical response parameters for
the size- analysis, as shown in Fig. 4. The width of the model analyzed is 10.0 m, while
the length and depth are 8.0 m and 4.0 m, respectively, after a series of mesh sensitivity
analyses. The relative deflection is defined as surface deflection divided by the
maximum deflection under loading expressed as a percentage. Fig. 5 illustrates the
decay of the selected response with the distance to the loading center. It can be seen that
at a distance of 2.0 m, the deflection is about 15.0% of the maximum deflection, the
strain is about 10.0% of its maximum, and the transverse stress is about 5.0% of its
maximum.
120 Displacement
Pecent of Maximum
100 Vertical strain

80 Transverse stress
60
40
20
0
-20 0.0 1.0 2.0 3.0 4.0 5.0
Distance to the loading center(m)
Fig. 4. Decay rules of critical responses with the distance to the loading center.
Table 3. Critical responses vs. Subgrade thickness.
Thickness(m) 0.9 1.4 1.9 2.4 2.9 3.4

deflection(mm) 0.2677 0.3156 0.3478 0.3702 0.3864 0.3987
Percent (%) 0 117.9 129.9 138.3 144.3 148.9
Vertical stress (kPa) 45.4 47.8 49.4 50.6 51.4 52.1
Percent (%) 0 105.3 108.9 111.5 113.4 114.7
Subgrade strain (ȝİ) 175.1 172.0 170.1 168.6 167.2 166.0
Percent (%) 0 99.2 97.1 96.3 95.5 94.8
Table 3 shows the decay of critical responses (the same responses defined as in Fig. 5)
with the variation of the subgrade thickness. The results indicate that the transverse
stress of the HMA layer converges quickly with the thickness, while the surface
deflection converges very slowly. The difference in strain on top of subgrade between
the thickness of 1.4 m and 3.4 m is 3.4%, which is negligible.
PLANAR SIZE OF TEST SEGMENT GEOMETRY
Width: The width of the testing track will be affected by the following factors: Wheel
spacing A , Wandering W , tire width T , and boundary effect length B , thus the final
segment width L can be expressed by:
L = 2( B + W + T ) + A (1)
The boundary effect length is used to prevent the rigid wall from influencing the
parameters to be measured. This distance can be determined by the decay of critical
response parameters with distance. W is the transverse wandering width, which reflects
realistic traffic operations. Taking the previous vehicular loading as an example, the
final minimum width should be more than (2.0×2+0.4×2+0.2×2+1.2=6.4m), in which
0.4m is used for wandering space (Kokkins, 1992; Owusu, 1992).
Length: A typical test section consists of transition, boundary effect length, and loading
vehicle, shown in Fig. 5.
1/2 1/2
Boundary Loading vehicle Boundary
Transition Transition
Fig. 5. Typical testing segment length.
The transition segment is used to be a transition between different testing sections and
smooth the vehicle in case the adjacent testing section fails. Loading vehicle segment
represents the length of truck used in the experiment. Boundary segment is
incorporated to consider the influence of longitudinal boundary on the quality of the test
data, which can be gotten by analyzing the decay of critical responses along the
traveling direction. Fig. 6 presents the results of FE simulation, whose parameters are
the same as previous part.
100 Displacement
Longitudinal stress
Percent of Maximum
80
Vertical strain
60
40
20
0
0.0 1.0 2.0 3.0 4.0
-20 Distance to the loading center(m)
Fig. 6. Critical responses at the loading centerline along longitudinal direction.

From the results, it can be determined that at a distance of 2.0 m, the deflection reduces
by about 18.0% of the maximum surface deflection, and the strain decreases by nearly
2.0% of the maximum vertical strain. Taking the proposed vehicular loading as an
example, the final minimum length should be more than 2.0×2+3.5+3.5 (=11.0m), in
which 3.5m is used for transition part, the same as the length of the testing vehicle.
CONCLUSION
A preliminary design for a full-scale ALF with full-controlled environment was

presented in this paper, which includes the comparison between the circular track and
the straight track and the preliminary design of the testing pavement dimensions. The
proposed design parameters according to finite element simulation will be helpful in the
final design of the testing track dimensions.
REFERENCES
Kokkins, S.J. (1992). "FAA Unified Pavement Analysis 3-D Finite Element Method,"
Proceedings of Unified Airport Pavement Design and Analysis Concepts Workshop,
Report No. FAA-RD-92/17, Federal Aviation Administration, Washington DC.
Owusu-Antwi, E. (1992). "Development of a Unified Airport Pavement Analysis and
Design System," Proceedings of Unified Airport Pavement Design and Analysis
Concepts Workshop, Report No. FAA-RD-92/17, Federal Aviation Administration,
Washington DC.
Backcalculation of Anisotropic Pavement Properties using Time History of
Embedded Gauge Readings
Eyal Levenberg
Purdue University, North Central Superpave Center, West Lafayette, Indiana 47906;
[email protected]
ABSTRACT: An instrumented asphalt pavement was built at the INDOT/Purdue

accelerated pavement testing facility. The construction processes induced preferred
aggregate orientation and lock-in of horizontal stresses, resulting in mechanical
anisotropy. This anisotropy is investigated by means of inverse analysis. The
pavement system is modeled as a linear elastic transversely-isotropic layered half-
space; anisotropic properties are back-calculated using the time history of all
embedded gauge readings, collected during one load pass. The experiment and
analysis method are presented; results are compared with the isotropic case and
discussed.
INTRODUCTION
The resilient response to load of asphalt pavement systems is typically modeled
using isotropic material properties (e.g., Yoder and Witczak, 1975; NCHRP 1-37A
2004). However, compaction operations produce preferred aggregate orientation (Oda
et al., 1985; Saadeh et al., 2002) and lock-in of horizontal stresses within the different
layers (Uzan, 1985; Duncan et al., 1991), resulting in inherent material anisotropy.
Under traffic loads further symmetry changes may take place due to the (recoverable)
tendency of the materials to dilate in shear (Uzan, 1992). The symmetry of the
resilient properties may also be influenced by permanent (irrecoverable) deformations.
The study of such anisotropy in the laboratory (e.g., Adu-Osei et al., 2001; Tutumluer
et al., 2003) is problematic because of the uncertainty associated with recreating the
conditions prevailing within the pavement system. In the present work resilient
anisotropic material behavior is investigated by means of inverse analysis. Use was
made of an instrumented asphalt pavement loaded in the INDOT/Purdue accelerated
pavement testing (APT) facility; the time history of all embedded gauge readings,
collected during a single APT pass, was used for the back-calculation.
THE APT EXPERIMENT

The INDOT/Purdue APT facility was fabricated in early 1990’s; it is housed in a
186 sq. m (~2,000 sq. ft) hangar located near INDOT’s Research Division in West
Lafayette. The testing area includes a pit embedded in a concrete floor for
accommodating prototype scale pavements. The pit is 1.83 m (6 ft) deep, shaped as a
79
square with 6.1 m (20 ft) long sides. The loading system is capable of producing a
downward force of up to 9,080 kg (20,000 lb) applied to the pavement surface through
a half-axle dual-wheel assembly that is traveling at a constant speed of 8 km/h (5
mph). In the present study the temperature in the testing area was set and maintained
throughout the experiment at 15.5ºC (60ºF). The wheel assembly was loaded to 6,810
kg (15,000 lb) with tires inflated to 0.70 MPa (100 psi). Passes were applied without
wander in unidirectional mode. Note that due to internal tire reinforcement, circular
imprints resulted on the pavement surface, each 203 mm (8 in.) in diameter.
The pavement structure was relatively thin, comprised of 127 mm (5 in.) of hot mix
asphalt (HMA) overlaying 152.4 mm (6 in.) of crushed granite aggregate base course;
placed on top of an A-4(0) soil serving as subgrade. The HMA course was made of
dense-graded Superpave mixtures constructed in three lifts. The aggregate base course
was compacted in a single lift and the subgrade soil was compacted in several lifts,
each up to 152 mm (6 in.) thick. During construction the pavement was instrumented
with four pressure gauges and eight strain gauges. The instrumentation layout plan is
shown in Fig. 1. The Y-axis in the figure denotes the travel path of the center of the
APT carriage (loading from left to right); this line itself was not directly loaded as it
passed between the dual wheels. Pressure gauges #1178 and #1185 measured vertical
stresses on top of the base course; these were installed at a depth of 127 mm (5 in.)
from the surface. Pressure gauges #1179 and #1184 measured vertical stresses on top
of the subgrade; these were installed at a depth of 279.4 mm (11 in.) from the surface.
All eight strain gauges were attached to the bottom of the HMA course, i.e., at a depth
of 127 mm (5 in.) from the surface. Gauges G-2, G-4, G-5 and G-7 measured
horizontal strains in the loading direction (i.e., strain in Y) while gauges G-1, G-3, G-6
and G-8 measured horizontal strains in the transverse direction (i.e., strain in X). More
experimental details are contained in the following sources: Llenín & Pellinen (2004),
Llenín et al. (2006), and Levenberg (2008).
FIG. 1. Plan of embedded instrumentation in the APT experiment.
MODELING AND ANALYSIS

The APT pavement system was modeled as a four layered half-space (disregarding
the test pit walls) with linear elastic and transversely-isotropic material properties. The
three HMA lifts were treated as one (top) layer, 127 mm (5 in.) thick. The second
layer from the top represented the aggregate base course, with a thickness of 152.4
mm (6 in.). The subgrade was modeled as a single layer (third from the top) having a
total thickness of 1.55 m (61 in.). The fourth and final layer, with semi-infinite
thickness, represented the concrete floor of the test pit. The dual-wheel loading was
represented by two circular areas, spaced 343 mm (13.5 in.) apart (center to center),
transferring uniform vertical stresses of 1.03 MPa (150 psi) in magnitude to the
pavement surface. The moving APT carriage was simulated using the quasi-static
approach, by applying the loading at different locations along the Y-axis in Fig. 1
(with the tire imprints on each side of this line). Computation of stresses and strains in
the transversely isotropic layered half-space was done using a new program named
ELLEA2 (Levenberg, 2008), for which the theoretical derivation was based upon the
works of Lekhnitskii (1963), Singh (1986), and Huang (2004). The underlying
equations were programmed into an Excel workbook without the use of VBA code -
resulting in fast real-time calculations.
In terms of a Cartesian coordinate system ( x , y , z ), with x − y as the plane of
material isotropy, the constitutive law of a transversely-isotropic material is:
⎡ ε x ⎤ ⎡ a11 a12 a13 ⎤ ⎡σ x ⎤

⎢ε ⎥ ⎢ ⎥ ⎢σ ⎥
⎢ y ⎥ ⎢a12 a11 a13 ⎥ ⎢ y⎥
⎢ ε z ⎥ ⎢a13 a13 a33 ⎥ ⎢σ z ⎥
⎢ ⎥=⎢ ⎥⋅⎢ ⎥ (1)
⎢ε yz ⎥ ⎢ a44 / 2 ⎥ ⎢τ yz ⎥
⎢ε xz ⎥ ⎢ a44 / 2 ⎥ ⎢τ xz ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢⎣ε xy ⎥⎦ ⎣⎢ (a11 − a12 ) / 2⎦⎥ ⎢⎣τ xy ⎥⎦
in which a11 = 1 / E x , a33 = 1 / E z , a12 = −ν xy / E x , a13 = −ν zx / Ez and a44 = 1 / Gxz (zero

elements have been omitted for clarity). Hence, five elastic constants are included,
namely: E x (= E y ) , E z , ν xy (= ν yx ) , ν zx (= ν xz ⋅ E z / E x ) and Gxz (= G yz ) . The
thermodynamic requirement for positive strain energy imposes the restrictions (Poulos
and Davis, 1974): E x , E z , Gxz > 0 ; 1 −ν xy − 2 ⋅ν xz ⋅ν zx > 0 ; and 1 − ν xy > 0 .
Analysis of the APT experiment was performed using data collected in the initial
phase of the experiment, during pass #5,000 (out of a total of 90,000 passes). The
analysis consisted of obtaining anisotropic material properties through back-
calculation by requiring that model generated stress and strain histories match the
gauge measurements as best as possible. To accomplish this, a scalar error term
defined to represent the overall goodness of fit (see Levenberg et al., 2008) was
minimized using a nonlinear optimization algorithm by manipulating the numerical
values of the free elastic constants (without violating the property restrictions). It
should be noted that matching was only performed for data collected in the
approaching branch of the response traces; starting from an offset distance of 1.8 m
(70 in.) and ending when the wheels were in line with a gauge along the Y-axis in Fig.
1. The receding branch was not considered because the measured time histories were
nonsymmetrical with respect to the load location; a characteristic that cannot be
simulated with the time-independent model.
Given that each layer of interest (i.e., HMA, base, and subgrade) is characterized by
five independent elastic constants (Eq. 1), there are 15 unknowns to be back-
calculated in the most general case (note that the concrete properties were prefixed). In
order to simplify matters and reduce the number of free parameters to a more realistic
level, it was decided to focus only on the vertical and horizontal moduli, namely E z
and E x . Subsequently, assumptions were made regarding the remaining elastic
constants. The shear modulus G xz in each layer was related to the other elastic
properties of the same layer according to the expression (e.g., Wolf, 1935; Barden,
1963):
Ex ⋅ Ez
Gxz = (2)
Ez + Ex ⋅ (1 + 2 ⋅ν zx )
where it can be seen that when E z = E x = E and ν xy = ν zx = ν the isotropic shear
modulus is attained. Also, the two distinct Poisson’s ratios ν zx and ν xy were assumed
identical and their values were predetermined (for each layer separately). Different
assumptions are known to exist, most noteworthy of which (although not applied
herein) were proposed by Graham and Houlsby (1983).
RESULTS AND DISCUSSION

Table 1 presents the back-calculated anisotropic elastic moduli; also listed are the
Poisson’s ratios and the resulting shear moduli calculated according to Eq. 2. It can be
seen that both the subgrade and aggregate base were found to be stiffer in the vertical
direction compared to the horizontal direction; this outcome is expected, at least
conceptually. For the HMA, however, the trend is reversed with greater stiffness in the
horizontal direction. Quantitatively, the ratio of E z / E x is 1.3, 5.6 and 0.4 for the
subgrade, base and HMA,respectively. In the vertical ( z ) direction, the HMA is only
2.5 times stiffer than the underlying aggregate base; this ratio seems relatively low.
The base itself is about 5.0 times stiffer than the subgrade; a ratio that is relatively
high considering the thinness of the layer and structure. In the horizontal ( x − y )
direction, the HMA was found to be 33 times stiffer than the aggregate base; the value
of E x seems too low for the base (only 1.2 stiffer than the subgrade), but reasonable
for the HMA. For comparison, the corresponding isotropic case given in Levenberg et
al. (2008) yielded the following moduli (in MPa): E HMA = 2,412 , Ebase = 165 , and
E subgrade = 83 . Hence, it can be seen that the isotropic analysis gives moduli values that
more or less range between Ez and Ex (separately for each layer); the stiffness ratios,
however, do not match. In both cases the moduli sets may be considered valid for
forward analysis given that they were obtained through matching a large array of
responses.
Table 1. Back-calculated anisotropic layer moduli.

Poisson's ISackcalculaK'tl Moduli, Shear Modulus
Thickness, Ratio MPa (psi) Gvz Eq. (2)
# Layer
mm (in.)
"„ = ".-,, Ez Ex MPa (psi)
1 IIMA 127(5) 0.30 1,050(152.000) 2.470 (358,500) 520 (75,100)
2 Base 152(6) 0.35 422(61.300) 75(10,850) 58 (8,340)
3 Subgradc 1,549(61) 0.40 82(11,950) 64 (9,250) 27(3,865)
4 Concrete Scmi-inf. 0.20 4.000,000 (27,560) 11,490(1,666,670)
Measured and computed responses are graphically contrasted in Fig. 2 which

contains six charts, each corresponding to a different response; the abscissa in the
charts denotes the offset distance along the Y-axis (see Fig. 1) between the specific
gauge location and the center of the APT loading carriage. Gauge measurements are
shown using solid markers and computed responses are shown using solid lines; the
isotropic case (Levenberg et al., 2008) is also included for comparison using dashed
lines. As can be seen, differences were recorded between pairs of gauges that should
have provided identical readings. These differences are large enough to overshadow
the improvement of using more sophisticated modeling techniques. Partially for this
reason the computational complexity needed for introducing the test pit walls was
avoided. The anisotropic model is seen to offer only a slight advantage in capturing
the measured responses compared to the isotropic case. Both models underestimate the
peak vertical stresses on top of the base (V) and the peak horizontal strains in X (I and
III); the other three responses (II, IV and VI) are relatively well reproduced. In a future
study, these response histories will be reanalyzed using layered viscoelastic theory
(both isotropic and anisotropic) to assess which of the present findings and results may
have transpired from imposing a time-independent model on the test data.
FIG. 2. Measured vs. computed responses (refer to Fig. 1).
REFERENCES
Adu-Osei, A., Little, D. N., and Lytton, R. L. (2001). “Cross-anisotropic
characterization of unbound granular materials.” Transp. Res. Rec. 1757, 82-91.
Barden, L. (1963). “Stresses and displacements in a cross-anisotropic soils.”
Géotechnique, 13 (3), 198-210.
Duncan, J. M., Williams, G. W., Sehn, A. L., and Seed, R. B. (1991). “Estimation of
earth pressures due to compaction.” J. Geotech. Eng., 117(12), 1833-1847.
Graham, J., and Houlsby, G. T. (1983). “Anisotropic elasticity of a natural clay.”
Géotechnique, 33(2), 165-180.
Huang, Y. H. (2004). Pavement analysis and design, 2nd Ed., Prentice Hall, New
Jersey.
Lekhnitskii, S. G. (1963). Theory of elasticity of an anisotropic elastic body, Holden-
Day, San Francisco.
Levenberg, E. (2008). “Validation of NCAT structural test track experiment using
INDOT APT facility: final report.” Joint Transp. Res. Prog., SPR-2813 project.
North Central Superpave Center, Purdue University.
Levenberg, E. McDaniel, R. S., and Pellinen, T. K. (2008). “Backcalculation of layer
moduli using time history of embedded gauge readings.” 3rd Inter. Conf. Accel.
Pave. Test., Madrid, Spain.
Llenín, J. A. and Pellinen, T. K. (2004), “Validation of NCAT structural test track
experiment using INDOT APT facility: interim draft final report.” Joint Transp.
Res. Prog., SPR 2813 Project, Purdue University.
Llenín, J. A., Pellinen, T. K., and Abraham, D. M. (2006). “Construction management
of a small-scale accelerated pavement testing facility.” J. Perform. Const. Facil.,
20(3), 229-236.
NCHRP 1-37A (2004). “Guide for mechanistic-empirical design of new and
rehabilitated pavement structures.” National Cooperative Highway Research
Program, Project 1-37A Final Report (PDF version).
Oda, M., Nemat-Nasser, S. and Konish, J. (1985). “Stress-induced anisotropy in
granular masses.” Soils Found., 25(3), 85-97.
Poulus, H. G., and Davis, E. H. (1974). “Elastic solutions for soil and rock
mechanics.” Center for Geotech. Res., University of Sydney (1991 reprint).
Saadeh, S., Tashman, L. Masad, E. and Mogawer, W. (2002). “Spatial and directional
distribution of aggregates in asphalt mixes.” J. Test. and Eval., 30(6).
Singh, S. J. (1986). “Static deformation of a transversely isotropic multilayered half-
space by surface loads.” Phys. Earth and Planetary Inter., 42, 263-273.
Tutumluer, E., Little, D. N. and Kim, S. (2003). “Validated model for predicting field
performance of aggregate base courses.” Transp. Res. Rec. 1837, 41-49.
Uzan, J., (1985). “Characterization of granular material.” Transp. Res. Rec., 1022, 52-
59.
Uzan, J. (1992). “Resilient characterization of pavement materials.” Int. J. Num. Ana.
Meth. in Geomech., 16, 453-459.
Wolf, K. (1935). “Distribution of stress in a half-plane and a half-space of anisotropic
material.” ZAMM J. App. Math. Mech., 15(5), 249-254.
Yoder, E. J., and Witczak, M. W. (1975). Principles of pavement design, 2nd Ed.,
Wiley, New York.
Provisional Results from Accelerated Testing of Ultra Thin-Layer
Reinforced Concrete in South Africa
L du Plessis, P.J Strauss* and A Kilian**
CSIR-Built Environment, PO Box 395 Pretoria 0001, South Africa

* Consultant, PO Box 395 Pretoria 0001, South Africa
**Gauteng Department of Public Transport, Roads and Works, South Africa
ABSTRACT
The CSIR Built-Environment, in conjunction with the University of Pretoria and the
Cement and Concrete Institute of South Africa, developed a low cost option for the
upgrading of unsurfaced (gravel) roads. The proposed solution is the placing of a thin
layer of normal concrete reinforced with 5.6mm diameter steel with a mesh grid size
of 200mm. This thin layer is placed on top of the existing unsurfaced road with
minimal preparation to the existing road surface using labour-intensive construction
methods.
Through full-scale trials this type of upgrading proved to be adequate for low-volume
traffic applications (e.g. residential streets) as well as for higher-volume applications
(e.g. bus routes). During the trials test sections were subjected to a total of over
700,000 ESALs over a period of 5 years without showing any deterioration.
In order to determine the structural capacity of this type of overlay full-scale Heavy
Vehicle Simulator tests were conducted. This paper summarizes the initial results from
the accelerated pavement testing (APT) tests and is aimed at building confidence in
the use of thin-layer CRCP, with cognizance being taken of the pavement structure,
support conditions, construction, climate and traffic.
INTRODUCTION
The upgrading of unsurfaced residential roads has become a priority for many
metropolitan areas in South Africa. Coupled with this is the need to construct roads
using labour-intensive construction techniques. One solution is the use of an ultra thin
reinforced concrete pavement (UTRCP). UTRCP has been shown to offer good
performance during a previous trial study (1,2) but a number of questions still remain
to be answered, particularly with regard to the limits of application of this technology.
86
The primary objective of this project is to assess the performance of 50-mm thick
CRCP placed on various support conditions through accelerated pavement testing
(APT), coupled with a laboratory-testing programme. The aim is to determine the
limits for safe application of the technology in South Africa. The outcomes of the
project will also be used to update the South African design method for rigid
pavements (cncPAVE) (3).
This study is aimed at building confidence in the use of UTRCP, taking cognizance of
the pavement structure, construction, climate and traffic, by providing practitioners
with reliable design information.
Expected Benefits
The use of UTRCP on roads offers multiple benefits by comparison with more
conventional approaches. The benefits include:
• UTRCP is labour-friendly and is highly suitable for labour-based construction;
• Only simple, inexpensive construction equipment is required;
• Existing subgrade and alignment can be used;
• It can be used as an overlay on existing roads;
• It requires less lighting energy at night by comparison with bituminous surfaces
because of the reflectivity of the surface; and
• Significant reduction in construction energy can be realised as the mix is hand-
placed at ambient temperatures.
THE CONSTRUCTION OF THE UTRCP SECTIONS
The Ultra Thin Reinforced Concrete Pavement sections were constructed on a 130m
long test section using a mix of 13mm and 9mm quartzite aggregate to achieve a dense
aggregate packing. A high-strength cement (CEM I - 42.5N) was used and the mix
was reinforced with a steel mesh consisting of a 200 x 200mm grid size using 5.6mm
diameter steel wire. The total thickness of concrete was 50mm and the steel was
placed on the neutral axis, 25mm from the surface using plastic chairs. The mix was
hand placed using shutters to control the width and thickness of the concrete. A double
vibrating roller screed was used for compaction and the pavement was cured under
plastic for 7 days to minimize differential shrinkage. The sections were constructed on
the R80 Highway north-east of Pretoria, South Africa and Figure 1 shows some
aspects of the construction process.
Figure 1: The UTRCP construction process
The average 28-day cube crushing strength was 37.5 MPa (without the steel).
The UTRCP was placed on a prepared road bed consisting of various support types,
which included a granular base section, a cemented base section, and sections with a
50mm of emulsion-treated base. One section was constructed using bottom dump ash
(from a coal-fired power station) as a replacement for the quartzite aggregate. The aim
of this was to determine the structural bearing capacity of the UTRCP under various
support conditions with the aid of the HVS.
ENVIRONMENTAL INFLUENCE ON SLAB CURL AND WARPING
Prior to the start of APT loading (during the curing period), a set of measurements was
taken to measure the slab curl and warping movements as a result of daily temperature
fluctuations and differential shrinkage under the influence of the environment. Joint
Deflection Measuring Devices (JDMDs) were installed immediately after placement
of the concrete to record the vertical movements of the edges of the concrete slab.
Thermocouples (TCs), which measured both top and bottom temperatures were also
installed in the concrete at each JDMD location. The data captured during the first 23
days are graphically represented in Figure 3. This graph shows the maximum
movements as captured by instrument JDMD 10. The instruments were installed as
soon as the concrete had developed enough strength to carry the weight of the slugs of
the various LVDTs.
Two distinct edge movements were identified:

a) The daily (elastic) up/down movement due to daily temperature variations.
As the temperature increased during daytime, the concrete expanded elastically

but, because the surface was hotter at the top (being exposed to the sun) than
the bottom, the top expanded at a faster rate than the bottom, causing the
longitudinal edge of the slab to curl downwards. This process was reversed at
night when the top was cooler than the bottom, causing upwards slab
movement along the edge of the concrete slab.
b) Slab warping.
During the initial curing period in particular, a significant amount of shrinkage
in concrete took place but, because the top was exposed to the environment,
the degree of hydration was greater towards the top than the bottom of the slab.
This differential shrinkage caused the longitudinal edge of the slab to curl
upwards. This permanent non-recoverable movement is termed ‘slab warping’.
The degree of slab warping depends on environmental factors such as
humidity, temperature and wind, as well as on the type of concrete and of the
reinforcement.
Both these movements can be identified in Figure 3. Analysis of the first 23 days of
curing showed that the maximum separation of the concrete from the base was 1.6mm
during the day and increased to 3.1mm during the night. The data are summarized in
Table 1.
4 45
3.5 40
Slab Edge Movement (mm)
3 35
Temperature (°C)
2.5 30
2 25
1.5 20
1 15
0.5 10
0 5
-0.5 0
25-00-00
26-00-00
27-00-00
28-00-00
29-00-00
30-00-00
01-00-00
02-00-00
03-00-00
04-00-00
05-00-00
06-00-00
07-00-00
08-00-00
09-00-00
10-00-00
11-00-00
12-00-00
13-00-00
14-00-00
15-00-00
16-00-00
17-00-00
18-00-00
19-00-00
Time (Day - hour - minutes)

JDMD 10 Surface temp
Figure 3: Maximum slab edge movements and associated surface temperatures

Table 1: Summary of slab edge movements due to environmental influences
Permanent Warp Daily Curl Total daily elastic

(upward) curling movement
Instrument (mm) up (mm) down (mm) (mm)
JDMD 10 2.30 0.80 0.70 1.50
Maximum surface temperature (Deg C) 42.40
Minimum surface temperature (Deg C) 10.50
Maximum positive temperature differential (top - bottom) (Deg C) 3.60
Maximum negative temperature differential (top - bottom) (Deg C) -3.40
3.1
2.3
1.6
Unfortunately data collection was stopped after 23 days to accommodate the HVS
tests which followed directly afterwards.
APT TESTING PLAN
In order to address the aims of this study a series of Heavy Vehicle Simulator tests
(HVS) were planned as shown in Figure 4. The rectangular boxes indicate possible
HVS testing areas
Plan View:
Section H Section F
Tied To adjacent lane
458A4 457A5
0m 20m 40m 60m 70m 90m 110m 130m
Section G Section E Section D Section A Section B Section C
Pavement Structures:
50mm UTRCP ash aggregate
50mm ETB No ETB No ETB 50mm ETB
150mm imported aggregate base 150mm imported aggregate base
150mm 3% Lime stabilized SUBBASE In-Situ rip & compacted SUBBASE
Figure 4: Planned HVS testing areas
The main objectives of these tests can be summarized as:
• To assess the effects of various input parameters (i.e. type of aggregate, support
conditions, longitudinal joints, traffic loading) on the structural integrity and
performance of the UTRCP layer;
• To determine the structural strength across joints and transverse cracks as these
are believed to be the weak areas in the UTRCP system; and
• To establish the success of patchwork and rehabilitation options in the event of
the early UTRCP layer failure.
In order to address the last mentioned point, certain HVS testing will be conducted
until complete failure has been observed. Those failed areas will then be patched
using the standard fixing techniques for continuously reinforced pavements. The same
areas will then be subjected to a second round of accelerated loading.
LATEST HVS RESULTS
To date only two tests have been conducted. These tests (457A5 and 458A4) are
according to the layout in Figure 4. A total of over 1.3 million standard axle loads
(80kN) were applied to each of the sections without any signs of structural failure.
Only two small transverse cracks had appeared after 5 days of trafficking but no signs
of pumping or any other form of deterioration could be detected, despite the fact that
over 540 mm rain fell on the test pavement during the test. The pavement has been
subjected to accelerated trafficking for 3 months and loading will continue until
structural failure. Figure 5 shows the peak deflection with repetitions.
1.4 40
1.2 35
Elastic Surface Deflection (mm)
30
1
Rainfall (mm)
25
0.8
20
0.6
15
0.4
10
0.2 5
0 0
0 200000 400000 600000 800000 1000000 1200000
Repetitions
Deflection Rain (mm)
Figure 5: 40kN deflections and rainfall data
The variations in the data are due to daily temperature and rainfall effects. From the
graph it is evident that no significant degree of deterioration in terms of deflections
can yet be detected. After 1.2 million load applications the deflection values are very
similar to those recorded at the beginning of the test.
From this it is concluded that very little damage has been done to the pavement under
loading. This conclusion is supported by the results of visual inspections.
MODELING THE BEHAVIOUR OF THE PAVEMENT
The purpose of modelling was to develop a model which could be used for the design
of a UTRCP and to update the design procedures in cncPAVE (3).
cncPAVE is a mechanistically based design method developed in South Africa to

facilitate interaction with the mechanistic flexible pavement design methods currently
in use and to promote the use of concrete to rehabilitate flexible pavements. cncPAVE,
a user friendly computer program, is based on models developed from finite element
and multi-layer evaluations. The design method has been calibrated against actual
performance of different concrete pavement sections (roads and streets) under normal
traffic loading, as well as under the HVS.
The UTRCP pavement that was modelled and tested consisted of a thin reinforced
concrete pavement on top of a layered system. Properties of the model include:
• Generation of a generic finite element model of a representative piece of road
provided with a reinforced concrete slab on top and various layers underneath;
Application of wheel loads as circular patches of constant pressure moving at a
given speed; and
Non-linear dynamic FE analyses using explicit time integration.
The results from the analyses included principal stress values at the top and bottom of
the concrete, as well as vertical deflections of the concrete. The results were presented
as contour fringe plots as well as in tabular form.
The mesh of the model was selected per layer to be finer at the top and coarser at the
bottom in order to reduce computational effort. The finest mesh was directly in the
wheel path. At both ends of the wheel path, for a distance of about 1m, the mesh was
coarser, allowing basically for the wheel speed to increase to a constant value in these
areas.
The boundary conditions were such that all displacements at the bottom (x, y and z
directions) were fixed. On the plane of symmetry at y=0, the displacement in the y
direction was constrained to zero. At the ends, as well as along the sides, non-
reflective boundaries were prescribed, meaning that pressure waves that propagated
radially outwards from the wheel positions were not reflected back into the modelled
part.
Two circular load patches, 200mm in diameter, simulated the dual-wheel loading of a
typical truck. A surface pressure of 700 kPa was applied to the top concrete surface
along a path that ran at a distance of 900mm from the plane of symmetry.
In the dynamic analysis the wheel load together with the gravity load of the concrete
was first applied in 0.1 seconds from 0 to 700 kPa at an axle position close to one end
of the model (145 mm). Thereafter it was accelerated for a distance of 855 mm along
the path up to its ultimate speed. Thereafter the wheel load moved at a constant speed
along the wheel path.
A transverse crack in the concrete was modelled at the center of the wheel path over
the full width of the model. The crack was through the total thickness of the concrete
and was modelled in such a way that compression forces, but not tension forces, could
develop on the two adjacent surfaces of the crack. Vertical shear could also be carried
in the crack, which meant that there was no vertical slip between the surfaces of the
crack. The reinforcement was modelled as a continuous sheet of steel in both the x and
y directions.
In order to create a void below the slab, the interlayer stiffness was reduced to
practically zero and, for the transition areas at both ends of the void area, reduced
stiffness were introduced by linear interpolation.
The explicit time integration of the dynamic process was quite time-consuming, and
parallel processing with multi processors and about 400 MB of memory was required.
An eight processor server with 3.66 GHz 64bit Xeon processors was used. The time
required for a run using four processors was typically was three to nine hours,
depending on the size of the smallest element in the model and the stiffness of the
material. The analysis was done stepwise in time, with extremely small time steps,
typically 2 to 3 microseconds. With the duration of the wheel passing process from
end to end being about 2 seconds, about 800 000 solution cycles were required.
A non-linear dynamic analysis was required since the wheel patches are moving
objects in the wheel path and sliding contact is defined between the wheels and the
concrete surface. At the simulated crack in the concrete non-linear contact was also
defined between the surfaces of the crack.
The results of the analyses were available after a run as time history displacements,
strains and stresses. These could be presented as time history plots or as fringe or
contour plots at selected time intervals. Deformed plots scaled to enlarge the actual
deformation were also generated.
In view of the time and cost of using sophisticated Finite Element Analysis (FEA), a
limited number of cases were evaluated. The data generated by the FEA was used to
re-calibrate the equations used in cncPAVE and cncPAVE was subsequently used as a
tool to illustrate the relative effect of different parameters on the performance of a
UTCRCP. Figure 6 shows the reliability of the predicted maximum tensile stress on
the surface of the pavement using cncPAVE compared to the values generated by
FEA. The correlation coefficient R2 for the data in Figure 6 is 0.91.
3.00
Calculated stress (cncPave)

2.50
2.00
1.50
1.00
0.50
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00
S t ress calcu lat ed f ro m FEA
Figure 6: Comparison between FEA and cncPAVE calculated tensile stress.
MODE OF FAILURE
Combinations of FEA results and cncPAVE equations can now be used to explain the
behaviour of UTCRCP as observed on the test sections under HVS testing and to
establish the consequences if parameters that may vary during the design, construction
and loading of a UTCRCP.
The modelling and the output of program cncPAVE can be used to establish the
possible mode of failure and to demonstrate the sensitivity of certain characteristics of
UTCRCP:
1. Transverse cracks will develop as a result of shrinkage. It is also inevitable that

transverse construction joints will be introduced into the pavement. Since steel bar
reinforcement will be going through this joint, the joint or crack can be regarded as
a hinge that allows shear but no moment to be transferred across.
2. High stresses develop when a wheel load crosses from one side of the crack or
joint to the other side. According to FEA analyses the critical stresses are tensile
stresses at the bottom of the pavement about 450mm from the crack, those
between the wheel loads at the top of the pavement and the compression stress at
the top of the crack when the wheel is crossing the crack. A plot of these stresses is
shown in Figure 7 where the stress has been calculated for a 60mm slab with the
longitudinal steel bar reinforcement placed at different levels in the slab. It was
found that the tensile stress is at its greatest about 450mm away from the crack and
that a second crack will rapidly develop due to the high tensile stress at that point.
The crack itself may initially not be visible but the stiffness of the slab is reduced,
resulting in an increase in deflection and in a greater vertical stress at the top of the
supporting layer. At the same time high compression stresses develop at the top of
the slab in the crack, resulting in spalling and a risk of loss of shear resistance, as
well as in a risk of water entering the slab (see Figure 7). The crack between the
wheels will later extend from the surface into the pavement itself.
7
Top compr
Tensile stress (MPa)

6
5 Bott. Tens
2
Top tens
1
0
0 10 20 30 40 50
Position steel from top (mm)
Figure 7: Stress in a 60mm slab as a function of the position of the

reinforcement
3. Water that enters the supporting layer through the spalled crack results in a loss of
bond and, with an increase in deflection, a void will develop between the slab and
the supporting layer. The effect of a loss of bond is shown in Figure 8 below where
the maximum stress at the surface of the slab, calculated with cncPAVE, is plotted
as a function of bond and crack width. Figure 9 indicates the effect of the void size
once bond between the slab and the supporting layer is lost. The crack widths in
both Figures 8 and 9 are used to indicate the loss of shear and thus load transfer at
the crack.
1.80
1.60
1.40
1.20 Void= 0. 5m
1.00 Void= 0. 25m
0.80 V oid=0.0m
0.60
0.40
0.20
0.00
0.00 0.50 1.00 1.50 2.00
C rack w id t h (m m )
Figure 8: Maximum tensile stress at the surface as a function of crack width

and bond between slab and support layer.
The FEM analysis shows that the estimated pavement life for this type of structure is
approximately 2 million 80kN load applications.
1.80
1.60
1.40
1.20 Void= 0. 5m
1.00 Void= 0. 25m
0.80 V oid=0.0m
0.60
0.40
0.20
0.00
0.00 0.50 1.00 1.50 2.00
Crack w id t h (mm)
Figure 9: Maximum tensile stress at the surface as a function of crack width and
the size of the void between the slab and the supporting layer.
CONCLUSIONS
This aim of this study was to characterize the structural performance of a 50mm thick
Ultra Thin Reinforced Concrete Pavement under various types of realistic support
conditions with the aid of the Heavy vehicle Simulator.
Early indications are that this type of structure is prone to curling and warping.
Separation of over 3 mm between the bottom of the slab and the base along the
longitudinal edge of the test section was recorded. Special attention will be required to
mitigate the harmful effects of the loss of bond between the concrete and the support
structure.
Although APT testing has not been completed, the visual and structural evaluation of
the current state of the pavement ties in with the predictions from the dynamic Finite
Element Analysis.
ACKNOWLEDGEMENTS
This HVS research is supported by the Gauteng Department of Public Transport,
Roads and Works (GDPTRW) and their contribution towards this study is
acknowledged.
REFERENCES
1. Bergh, A.O., McKay, A. Semmelink, C.J. and Steyn, W.J.vdm. (2005).
“Roodekrans thin concrete experiment sections 4, 5 and 6: Continuously reinforced
thin concrete pavements.” Paper presented at the South African Transport Conference,
July 2005, Pretoria, South Africa.
2. W.J.vdm. Steyn, P J Strauss, B D Perrie, and L du Plessis. “Roodekrans trail

sections: The role of structural support under very thin jointed CRCP pavements
subjected to heavy traffic.” 9th International Conference on Concrete Pavements,
August 2005, Colorado Springs, USA
3. Strauss P.J., Slavik, M. and Perrie, B.D., 2001. A mechanistically and Risk
Based Design Method for Concrete Pavements in South Africa, Proceedings of the 7th
International Conference on Concrete Pavements, September 2001, Orlando, Florida,
USA.
Evaluation of the Effect of Tire Loads with Different Contact Stress
Patterns on Asphalt Rutting
W.J.vdM. Steyn
University of Pretoria, Civil Engineering, Lynnwood road, Hatfield, 0002, South Africa
PH (+27) 12 420 2171, FAX (+27) 12 362 5218, EMAIL [email protected]
ABSTRACT: Road pavements are constructed to carry traffic which applies its load-
to the pavement through the contact area between the tire and the pavement surface.
Previously, the tire-pavement contact area and stresses were idealized, as appropriate
instruments for quantification of these contact stresses were not available. The Stress-
In-Motion (SIM) technology has made it possible to characterize these contact stresses
at low speeds. In a recent Heavy Vehicle Simulator (HVS) test on various layers of
Hot Mix Asphalt (HMA), the effects of non-uniform tire-pavement contact stresses
were directly measured through application of two distinct types of tire-pavement
contact stresses onto the HMA pavement. The rutting response of the pavement
specifically showed the direct effects of these non-uniform contact stresses. In this
paper the background to the tire-pavement contact stresses is discussed briefly,
followed by details regarding the specific rut responses of five HVS tests where the
pavement performance reflected the effects of the non-uniform tire loading conditions.
Analysis of this data is presented together with discussions on the potential effects of
this information on roads carrying real traffic and their rut development.
INTRODUCTION
Road pavements are constructed to carry traffic which applies its load- to the
pavement through the contact area between the tire and the pavement surface.
Investigations around the effects of different types of loading on the response of
pavements have been conducted for many years. Previously, the tire-pavement contact
area and stresses were idealized, as appropriate instruments for quantification of these
contact stresses were not available. The Stress-In-Motion (SIM) technology has made
it possible for the three dimensional characterization of these contact stresses at low
speeds (De Beer et al, 2006; 2008). However, the actual effects of different non-
uniform tire-pavement contact stress patterns have not often been observed
specifically in a controlled comparative study of pavement performance, although
many analyses have shown that the non-uniform contact stresses should affect the
response of the pavement structure, especially, near the surface.
In a recent Heavy Vehicle Simulator (HVS) test on various layers of Hot Mix Asphalt
(HMA), the effects of non-uniform tire-pavement contact stresses were directly
measured through application of two distinct types of tire-pavement contact stresses
onto the asphalt pavement. The focus of this paper is to present and discuss the effects
98
of specific applied tire-pavement contact stresses on a standard HMA to evaluate the

extent of the effects during actual trafficking. Theory around HMA rutting is not
discussed explicitly, as this is well covered in literature and not the focus of this paper.
BACKGROUND
The contact stresses developed between the tire of a vehicle and the surface of the
pavement are complicated and dependent on various factors. These include the vehicle
load, vehicle configuration, suspension type, tire material type, tire inflation pressure,
tire construction, tread pattern, speed and moving action (i.e. acceleration, free-rolling
or braking, cornering or straight). De Beer et al (2006, 2008) developed the Stress-In-
Motion (SIM) technology that can be used to measure the tire-pavement contact
stresses orthogonally while the tire rolls over the device. Various other devices can
also be used to measure these stresses, but mostly only in the vertical direction
(Morgan et al, 2008).
The tire-pavement contact stress typically show patterns similar to those shown in
Figures 1a and 1b. De Beer et al (2006) used the terminology of m-shaped and n-
shaped tire-pavement contact stresses, denoting tires that are typically overinflated (n-
shaped – Figure 1a) and overloaded / underinflated (m-shaped – Figure 1b). These
stress shapes have been observed under numerous truck tires (De Beer et al, 2008).
Although various tests (mostly Accelerated Pavement Testing (APT)) have been
conducted with different tire-pavement contact stresses (mainly due to changes in tire
loading) (Hugo and Epps, 2004), the specific comparative response of the pavement
surfacing to differing contact stresses has not yet been studied in depth.
a – n-shape b – m-shape
FIG.1 Typical n-shape (a) and m-shape (b) vertical maximum tire-pavement
contact stresses.
It is well-known that rut development in a pavement is directly linked to and

influenced by parameters such as the mix design, the supporting pavement structure,
the loading conditions and the environmental conditions. Typical HMA material
response relationships indicate a direct relationship between the stresses imposed on

the pavement (as an input for the evaluation of rut development (NCHRP 1-37, 2004))
and the observed rut development. In this paper the mix design and environmental
conditions are constants and the focus is on the effect that the maximum vertical
contact stress has on the development of downward surface rut in the HMA surfacing.
EXPERIMENTATION
The HVS tests were conducted on road P159/1 west of Pretoria. The tests were part of
the Gauteng Department of Public Roads, Transport and Works investigation into the
rut development of HMA mixes (Steyn and Verhaeghe, 2006). Construction of the
standard 40 mm thick HMA mix took place during October 2006. The original
supporting pavement structure was constructed more than 20 years ago and consisted
of 4 layers. The structural strength of the pavement was relatively high with an
average elastic surface deflection (FWD-measured) of between 134 and 149 micron
(Steyn and Fisher, 2008). A standard production HMA mix was used as an overlay for
the tests. Selected properties of the HMA mix are summarized in Table 1.
Table 1. Selected properties of STD HMA mix.

Property Value
Binder content [%] 5.0
Air voids in mix (constructed) [%] 6.4
Voids in mineral aggregate [%] 15
Voids filled with binder [%] 72
Indirect tensile strength [kPa] 1 187
Gyratory refusal voids (N=300) [%] 2.4
The HVS is an APT device that has been used extensively nationally and
internationally for the controlled evaluation of pavement behavior under real loads
(Du Plessis et al, 2006). The HVS is typically used to apply a controlled wheel load to
a test section while the behavior of the pavement is monitored using dedicated
instruments. The conditions under which the loads are applied are well controlled, and
through the use of an environmental chamber the temperature of the pavement can be
controlled (Steyn and Denneman, 2008). The test conditions used for the five tests
discussed in this paper are summarized in Table 2. All tests were conducted using a
dual set of radial 11R22.5 tires and a surface temperature of 60°C.
The main parameters monitored on these HVS sections were the surface rutting and
surface profiles. It was measured using a laser profilometer (250 mm intervals across
the 8 m test section). The data were summarized in terms of the average downward
surface rut measured on the section at specific load application intervals. A number of
test slots cut after the HVS tests confirmed that all the rut developed inside the HMA
layers. All further analyses focus on the calculated average downward rut developed
inside the HMA layer.
DATA ANALYSIS
The data analysis is focused on establishing the effects of the non-uniform vertical
tire-pavement contact stresses on the rut profile development inside the HMA layers.
In Figure 2 the general rut development against the number of load applications for the
five sections are shown. The data focus on the average rut for the two wheel tracks of
the channelised sections, and the average rut for the wandering sections. Thedata
indicate the expected increasing trend of surface rut development with increasing
number of load applications.
The similar response observed under the uni- and bi-directional trafficking
(specifically those with similar load levels and both channelized - Sections 446A4 and
447A4) was confirmed by further HVS tests conducted on the same HMA not reported
in this paper. It is perceived that this phenomenon is related to the relatively thin
nature of the HMA (40 mm) and the relative fresh nature of the HMA when subjected
to HVS loading (between 12 and 16 months after construction).
Table 2. Nominal HVS test methodology for five HVS tests discussed in this
paper.
Test Loading Conditions Channelised/ Uni- / bi- Total
Number (Dual tyres) wandering directional Repetitions
446A4 N-shape (60 kN, 800 kPa) Channelised Bi 80 300

(no lateral
447A4 N-shape (60 kN, 800 kPa) Uni 75 337
wander of the
448A4 M-shape (60 kN, 420 kPa) test tires) Bi 212 171
449A4 N-shape (60 kN, 800 kPa) Wandering Bi 306 888

(50 mm x 10
step lateral
450A4 M-shape (60 kN, 420 kPa) Uni 303 000
wandering)
In Figure 3 the cross sections of the three channelized test sections towards the end of
each of the HVS tests are shown. The- data show the effect of the vertical tire-
pavement contact stress pattern on the rut development in the HMA surfacing.
Although the rut development for Sections 446A4 and 448A4 (N- and M-shape loads)
shows similar rut rates towards the secondary stage of the test (constant rut increase)
the difference in rut profile remained until the end of the specific tests, indicating that
the effects of the contact stress patterns still influence the cross-sectional profile of the
rut development. This is specifically evident on the edges of the rut cross profiles
where the higher edge stresses (448A4) caused higher localized rut.
In Figure 4 the cross section profile of the two wandering test sections are shown. In
these cases the tires wandered over the width of the section, distributing the location of
the peak vertical contact stresses over the section. The different effects of the two
contact stress conditions (n-shape and m-shape) are again visible in the surface rut
development.
10
8
CS downward rut [mm]
0
0 50000 100000 150000 200000 250000 300000 350000
Load applications
Section 446A4 Section 447A4 Section 448A4 Section 449A4 Section 450A4
FIG. 2. Average downward rut development against number of load applications

for five test sections.
8
m-shape load
6
n-shape load
Permanent Deformation (mm)
-2
-4
-6
-8
900 1000 1100 1200 1300 1400 1500 1600 1700 1800
Cross Sectional Distance (mm)
446A4 447A4 448A4
FIG. 3 Measured cross section rut profiles for three channelized test sections.
DISCUSSION
The data analyzed in this paper supported the hypothesis that the maximum vertical
tire-pavement contact stresses directly influence the surface rut profile of a HMA
surfacing. Figures 3 and 4 indicate that the rut profile under the various contact stress
patterns differs. This leads to the observation that a similar effect would be visible on
real pavements. With real traffic the tire loads are less channelized and therefore the
ultimate effect may be less than that shown in this study. However, field observations
performed during these tests has shown that trucks tend to follow a channelized
pattern when traveling on straight sections of road, specifically when a level of surface
rutting has started to develop. Further, the data indicate that the phenomenon is not
only visible for channelized trafficking, but also for wandering traffic, although to a
lesser degree.
In Figure 5 the ratio between the average downward rut and the maximum vertical
contact stress for the various tests are shown (centre and edge of tire). The data
indicate that all the ratios are within a band of between 3.3 and 5.3 mm/MPa. It does
not appear as if the direction of trafficking affected these ratios (this is probably
related to the relatively thin HMA being tested). The differences in the ratios are
attributed to changes in the thickness of the thin asphalt layer over the width of the test
section (a coefficient of variation of 9 per cent was measured on the test sections) as
well as a possible difference in density of the thin asphalt layer over the width of the
test section. Further evaluation of this is required, although it is logical that similar
stresses on a weaker layer would cause higher rut development.
The question may be posed as to the distribution of specific contact stress patterns in a
real traffic stream, as the contact stresses in the case of the specific tests discussed in
this paper were all of a similar shape. Surveys conducted on highways in South Africa
have indicated that the majority of steering axle tires in a typical traffic stream in
South Africa show n-shaped contact stresses (De Beer, 2008). Previous studies by
Steyn and Visser (2002) indicated that the road profile can give rise to different
contact stress patterns along the length of the road, while spatial repeatability would
tend to concentrate specific contact stress patterns in specific areas – thereby
amplifying the effect in those specific areas.
3.00
2.50 n-shape loading
2.00
Permanent deformation [mm]
1.50
1.00
0.50
0.00
-0.50
m-shape loading
-1.00
-1.50
500 700 900 1100 1300 1500 1700 1900 2100 2300
Cross section
449A4 450A4
FIG 4. Observed cross section rut profiles for two wandering test sections (n-
shaped contact stress (b) (uni-directional), m-shaped contact stress (b) (uni-
directional)).
6.0
Rut / contact stress ratio [mm/MPa]
5.0
4.0
3.0
2.0
1.0
0.0
446A4 centre 446A4 edge 447A4 centre 447A4 edge 448A4 centre 448A4 edge
Test number and location
RATIO [mm/MPa]
FIG 5. Summary of ratio between surface rut and maximum vertical contact
stress.
CONCLUSIONS
Based on the information provided in this paper the following conclusions are drawn:
• The non-uniform vertical tire-pavement contact stresses have a direct and

measurable influence on the downward surface rut development in the HMA mix
evaluated;
• The ratios between maximum vertical tire-pavement contact stress and average
downward surface rut in the HMA ranged between 3.3 and 5.3 mm/MPA, and is
probably influenced by contact stress value, asphalt layer thickness and asphalt
density;
• The uni- and bi-directional trafficking options provided similar responses for the
test conducted on the specific HMA material, and
• The output of this study should be applicable to normal traffic where similar tire-
pavement loading conditions were observed and may be used to explain the
development of specific rut-related deterioration of thin HMA layers.
ACKNOWLEGEMENTS
The funding from the GDPTRW for the research described in this paper is
appreciated, as well as the permission of the acting Director for CSIR Built
Environment to publish this paper.
REFERENCES
De Beer, M. (2006). Reconsideration of tyre-pavement input parameters for the

structural design of flexible pavements. Proceedings of the 10th International
Conference on Asphalt Pavements, Quebec, Canada.
De Beer, M (2008). Stress-in-motion (SIM) – a new tool for road infrastructure
protection? In Proceedings of the 5th International Conference on Heavy Vehicles,
HVParis2008, iSTE.
Du Plessis, L., Coetzee, N.F., Hoover, T.P., Harvey, J.T. and Monismith, C.L. (2006).
Three decades of development and achievements: The Heavy Vehicle Simulator in
accelerated pavement testing. Geotechnical Special Publication No. 154, ASCE,
Geo Institute, Virginia, USA
Hugo, Frederick and Amy Epps Martin. (2004). NCHRP Synthesis 325: Significant
Findings from Full-Scale Accelerated Pavement Testing, Transportation Research
Board, Washington, DC.
Morgan, G., Poulikakos, L., Arraigada, M., Muff, R. and Partl, M. (2008). Stress-in-
motion measurements of heavy vehicles from the Swiss footprint monitoring site.
In Proceedings of the 5th International Conference on Heavy Vehicles,
HVParis2008, iSTE.
NCHRP 1-37, (2004). Guide for Mechanistic-Empirical design of new and
rehabilitated pavement structures. National Cooperative Highway Research
Program, NCHRP 1-37, Washington DC, Transportation Research Board.
Steyn, W.J.vdM & Visser, A.T. (2002). Effects of moving dynamic tyre loads on tyre-
pavement contact stresses. Paper presented at the 6th Conference on Bearing
Capacity of Roads, Railways and Airfields (BCRA), Lisbon, Portugal. ISBN 90-
5809-394-4
Steyn, WJvdM., Verhaeghe BMJA. (2006). HMA Research Plan: Assessment of the
permanent deformation characteristics of a standard HMA Mix: project proposal
number: PP/2005/17, 8, CSIR, Pretoria, February 2006.
Steyn, W.J.vdM. and Fisher, C. (2008). Technical Memorandum: Phase 1 of HVS
testing: Road P159/1 – 441A4, 442A4, 443A4, 444A4, 445A4, 445A4A, 446A4,
447A4, 448A4, 449A4 and 450A4. CSIR BE, Contract report
CSIR/BE/IE/ER/2007/0033/B.
Steyn, W.J.vdM. and Denneman, E. (2008). Simulation of temperature conditions on
APT of HMA mixes. Paper accepted for the 3rd International conference on
Accelerated Pavement testing, Madrid, Spain, 1 to 3 October 2008.
Test Study on Abrasion of Skid-resistant Textures on Concrete Pavements
CHEN Yu1 and ZHANG Qisen2

1
School of Communication and Transportation Engineering, Changsha University of Science and
Technology, 960 South Wanjiali Road, Changsha 410004, China; [email protected]
2
School of Communication and Transportation Engineering, Changsha University of Science and
Technology, 960 South Wanjiali Road, Changsha 410004, China; [email protected]
ABSTRACT: An accelerated analog abrasion instrument, previously developed and

manufactured by the authors, is used to study the abrasion and attenuation of
skid-resistant textures on different cement concrete pavements. The abrasion
processes of groove concrete pavement, exposed aggregate concrete pavement and
porous concrete pavement are investigated in detail. The attenuation of their
skid-resistant textures is also numerically analyzed individually based on test results.
It can be concluded that the abrasion and attenuation processes of skid-resistant
textures on concrete pavements are controlled by their own structural formation and
the actual abrasion resistance. As a preliminary research, this is of much significance
for future research. It may also lay the foundation for predicting the theoretic service
duration of the skid-resistant textures on these three types of concrete pavements.

INTRODUCTION
In recent years the problem of the rapid abrasion of skid-resistant textures on

ordinary groove concrete pavement by moving tires, especially on certain roads such
as highway tunnels and long precipitous slope sections, has been the focus of many
studies (Xueliang et al., 2003; Panagouli and Kokkalis, 1998; Hongzhu et al., 2007).
Studies of exposed aggregate concrete with the random exposed hard aggregates
embedded in the concrete base (Sen et al., 2004) and porous cement concrete, a
skeleton-pore structure with the stochastic uneven appearance (Fujiwara, et al. 2001;
Park et al., 2005; Yu and Xueyi, 2007), have reported strong skid-resistance on the
surfaces. Due to the development of rigid pavement, these two types of concrete are
currently used in highway pavement, mainly to resolve the aforementioned problem.
Abrasion on concrete pavement can be explained as a tire-pavement system and
107
is affected by the surrounding environment, such as wear particles, temperature,

humidity cycles, water, and other factors, a gradual loss of the materials on the
pavement surface, which directly contacts with moving tires. In this paper, an
accelerated analog abrasion instrument, through which the abrasion and attenuation
of skid-resistant textures on different concrete pavement slabs are tested, was
successfully developed and manufactured.
TEST METHOD
In China, concrete abrasion tests are referred to as standards: Test Methods of

Cement and Concrete for Highway Engineering (JTG E30-2005) or Test Method for
Abrasion Resistance of Concrete and Its Products (GB/T16925-1997). Further
abrasion tests have been developed by researchers (Horszczaruk, 2004; Xibin et al.,
2000). Unfortunately, none of these tests meets all the requirements for a study of the
abrasion of concrete pavement for the following reasons:
i. No close link with the practical service circumstances of highway pavement,
ii. No reflection of the abrasion process changing with time, and
iii. No good relevance of 2 indexes: the abrasion speed and the corresponding
attenuation degree of skid-resistant textures on concrete pavements.
Zong et al.(2002) used a rotating ring specimen of asphalt mixture to sustain
abrasion by the above heavy tire in order to avoid repetitive abrasion back and forth
upon a straight line specimen. As far as cement concrete is concerned, the difference
between the ring specimen and line specimen is not significant. However, for the ring
specimen, there exists a linear velocity difference along the tire’s edge, resulting in
apparent distinction of the abrasion on the ring specimen surface.
Fig.1. Accelerated analog abrasion instrument for concrete pavement slab
In an accelerated analog abrasion instrument shown in Fig.1, a concrete specimen

(300h300h50mm) is firmly placed on a sliding track, installed on a test table and
driven back and forth at 50~100 cycles/min by a power system, which includes an
electromotor, a gearshift, an eccentric gear group with concave groove and a crank.
The system will exert pressure on the wear wheel axle and keep it rolling freely in
the vertical plane to the concrete surface. If there is enough friction between the
wheel and the concrete surface, the wheel rolls; if there isn’t, it transforms to slide.
To save test time, a hard emery wheel, with a width 200 mm and a vertical pressure
of 0.3f0.02MPa, was adopted to replace conventional rubber wheel.
Mass loss percentage, Δm , macro-texture depth, TD , as well as pendulum-like
friction coefficient, FB , on concrete surface were all measured. For the reason that
the 200×260mm wheel track region on concrete surface was not in line with the
requirements of handwork sand-paving method stipulated by On-the-spot Test
Methods for Highway Road Bases and Pavements (JTJ059-95) in China, modified
TD was calculated from the sand area paved and amount of sand actually used. Two
parameters defined as: mass loss percentage within wheel track region, Δmi , and
abrasion speed index, κ i , reflecting the abrasion degree and its corresponding speed
are shown in equations 1 and 2, respectively.
m0 − mi 300 × 300 m − mi
Δmi = × × 100% = 1.73 × 0 × 100% (1)
m0 200 × 260 m0
Δmi − Δmi −1
κi = ×1000 (2)
n
where, m0 original mass of concrete specimen; mi the ith measured mass of
concrete specimen; 1.73 size conversion coefficient; n abrasion frequency between
two neighboring measures.
RESULTS AND ANALYSIS
On Groove Concrete Pavement Slabs

Indoor manufacture C35 concrete specimens (9.5~16mm limestone as coarse
aggregate). After being roughly dragged by rag, the surface of the concrete slab was
hand pressed to form horizontal grooves with widths of 4mm, depths of 3mm and
intervals of 20mm. Δmi was measured after sustaining 200 cycles of abrasion. The
test data is shown in Fig.2. It illustrates that the abrasion process of a groove concrete
pavement slab can be divided into 3 phases:
i.Groove macro-texture on the surface of concrete slab, in direct contact with the
heavy wheel, rapidly wore away with Δmi straightly rising (<800cycles). Bad
abrasion resistance is undoubtedly due to only the weak convex part of a groove
bearing the abrasion.
ii.During 1000~1800 cycles of abrasion frequency, with grooves on concrete
surface being grinded away, both hardened mortar and coarse aggregates collectively
undertake abrasion. As a result of the strong coarse aggregates, κ i is by far lower
than that at the first phase.

iii.Abraded to certain degree (>1800cycles), there is increasing κ i again on
account of accumulated internal damages within concrete.
12.0 10.0
Mass loss percentage (%)
10.0 8.0
Abrasion speed index

8.0 6.0
6.0 4.0
4.0 2.0
mass loss percentage (%)
2.0 0.0
abrasion speed index
0.0 -2.0
1 2 3 4 5 6 7 8 9 10
Test occasion (once every 200 cycles)
Fig.2. Abrasion process on groove concrete slab
70 1.0
0.9
Macro-texture depth
60
Pendulum-like friction
0.8
coefficient (BPN)
50 0.7
40 0.6
(mm)
0.5
30 0.4
20 0.3
pendulum-like friction coefficient 0.2
10
macro-texture depth 0.1
0 0.0
1 2 3 4 5 6 7 8

Fig.3. Attenuation of skid-resistant textures on groove concrete slab
FB and TD are shown in Fig.3. In the early stages, when a heavy wheel directly
contacts with groove skid-resistant textures, both FB and TD straightly wane down
for the reason of weak convex hardened mortar, corresponding to the first phase with
quite high abrasion speed index shown in Fig.2. After sustaining about 1000 cycles
of abrasion, the attenuation of skid-resistant textures tends to stabilize, and the final
values of FB and TD are dominated by both microscopic textures of hardened
cement stone and tiny veins of aggregates.
On Exposed Aggregate Concrete Pavement Slabs

Adopting 6~8mm basalt crushed stones embedded in concrete as exposed
aggregates, and 9.5~16mm limestone as coarse aggregate for concrete base to
produce exposed aggregate concrete pavement slabs. Concrete specimens are
measured FB and TD on their surfaces after sustaining every 200 cycles of
abrasion.
80 1.4
Pendulum-like friction
Macro-texture depth
70 1.2
pendulum-like friction
coefficient (BPN)
60 coefficient 1.0
macro-texture depth
50
(mm)
0.8
40
0.6
30
20 0.4
10 0.2
0 0.0
1 2 3 4 5 6 7 8 9 10
Fig.4. Attenuation of skid-resistant textures on exposed aggregate concrete slab
Fig.4 shows the optimal function curves of FB and TD on exposed aggregate
concrete slab, both of which satisfy the nonlinear exponent function y = Ae Bx + C ,
accounting for rapid attenuation in early stages, gradually retarding later, and at last
the tendency of stabilization. ( A + C ) stands for the initial FB or TD , indicating the
original skid resistance of exposed aggregate texture; while C means the final stable
value after attenuating. A , attenuation range, is on behalf of the gap of skid resistance
between the surface and basis of exposed aggregate concrete. Higher absolute
B means more rapid attenuation of skid-resistant textures.
On Porous Concrete Pavement Slabs

Two typical highway porous concrete specimens (5~10mm basalt crushed stones
as coarse aggregates) are manufactured from diverse raw materials and different
mixtures. As shown in Fig.5, though there are different Δmi and κ i for 2 groups of
concrete, their abrasion processes are unanimous in essence. There are 2 main
features during the abrasion of porous concrete slabs:
i. low Δmi and κ i in correspondence with the same abrasion cycles, and
ii. steady abrasion process with no intense change of abrasion speed index.
10.0 4.0
mass loss percentage 3.5
8.0 abrasion speed index
percentage (%)
Abrasion speed
3.0
Mass loss
6.0 2.5
index
2.0
4.0 1.5
1.0
2.0
0.5
0.0 0.0
1 2 3 4 5 6 7
a. No.1
7.0 2.0
6.0
percentage (%)
1.5 Abrasion speed

5.0
Mass loss
4.0
index
1.0
3.0
mass loss
2.0 0.5
percentage
1.0 abrasion speed
index
0.0 0.0
1 2 3 4 5 6 7
b. No.2
Fig.5. Abrasion process on porous concrete slab

FB and TD are marked in Fig.6. Compared to groove concrete and exposed
aggregate concrete slabs shown in Fig.3 and Fig.4, there are no remarkable and
stable descending tendency of both FB and TD with the increase of accumulated
abrasion cycles. Scattered test data in Fig.5 do not show certain definitive law, not
only because of transforming of skid-resistant texture on the abraded surface along
the concrete depth, but also because of possible low precision measures. Yu and
Xueyi (2007) have highlighted that the attenuation of skid-resistant textures on

porous concrete is no more than 5% through a more reasonable parameter; the fractal
dimension of transect contour line of porous concrete specimen.
90 1.4
1.3
coefficient (BPN)
Pendulum-like
80 78 1.2
Macro-texture
1.1 1.0
depth (mm)
77
friction
67
70 1.0
0.9 68
64 58
60 0.8
0.8
pendulum-like 0.7
50 friction coefficient 0.6
macro-texture depth
40 0.4
1 2 3 4 5 6
a. No.1
90 1.4
1.3
coefficient(BPN)
1.1
Pendulum-like
80 1.2 1.2
Macro-texture
75 69 depth (mm)
74 66
friction
70 72 63 1.0
0.9
60 0.8 0.8
pendulum-like friction 0.7
50 coefficient 0.6
macro-texture depth
40 0.4
1 2 3 4 5 6
b. No.2
Fig.6. Attenuation of skid-resistant textures on porous concrete slab

The theoretical abrasion processes of different concrete pavement slabs are
described in Fig.7. For groove concrete, when only weak groove mortar is untaken
abrasion at the early period, mass loss is serious and abrasion is quite rapid. With
groove mortar grinded away step by step, strong coarse aggregates restrain the
abrasion course resulting in a decrease of abrasion speed. Finally, there is accelerated
abrasion again. In comparison, exposed aggregate concrete displays excellent
abrasion resistance in the early stages because it has strong exposed basalt stones that
sustain abrasion, and it exhibits the similar law with that of porous concrete at the
same stage. With the exposed aggregates entirely worn and the internal structure
exposed, the abrasion curve comes together with that of groove concrete at the later
stage.

Fig.7. Theoretical abrasion processes on different concrete pavement slabs

Different from groove concrete and exposed aggregate concrete, porous concrete
has an identical constitution materials and a homogeneous structure along its depth.
During abrasion, every abraded layer, directly contacted with the heavy wheel, is
stochastic but statistically identical. No weak layer within porous concrete leads to
stable abrasion process without stages. A slight rise in the abrasion speed should be
taken into consideration in practice, attributed to the accumulated internal damages
within concrete caused by moving tires for a long period of time.
CONCLUSIONS
Based on tests using the accelerated analog abrasion instrument developed by the
authors, the abrasion and attenuation of skid-resistant textures on different concrete
pavement slabs were investigated in detail. It can be concluded that as far as the
abrasion resistance and theoretical service life of skid-resistant textures are
concerned, porous concrete pavement is the most durable.
It is worth noting that the above-mentioned analysis is qualitative, not
quantitative. This is becausefor every factor of consistent materials such as cement
brand and added amount, coarse aggregate type and gradation, concrete admixture as
well as its dose, may affect test data even on the same test grounds. However, as a
preliminary test research, it is of much significance for future research. It may lay the
foundation for predicting the theoretical service duration of skid-resistant textures on
these three types of concrete pavements.
ACKNOWLEDGMENTS
The authors appreciate the support of Hunan Transportation Science foundation

provided by Hunan Provincial Communications Department.
REFERENCES
Yu, C., Xueyi, W. (2007). “Fractal Features and Surface Texture Parameters of
Porous Cement Concrete Pavement.” Journal of Changsha University of Science
& Technology (Natural Science Edition), 3, 13-17.
Horszczaruk, E. (2004). “The Model of Abrasive Wear of Concrete in Hydraulic
Structures.” Wear, 256, 787-796.
Sen, H., Zhiling, L., Dongsheng Z.(2004)a. “Performance of Exposed-aggregate
Cement Concrete Pavement.” Journal of Chang'an University (Natural Science
Edition), 4, 6-9.
Sen, H., Zhiling, L.,Dongsheng, Z.(2004)b. “Key Technology of Exposed-aggregate
Cement Concrete Pavement.” China Journal of Highway and Transport, 4, 20-23.
Fujiwara, H., Tomita, R.,Okamota, T. (2001). “Properties of High-strength Porous
Concrete. Recent Advances in Concrete Technology.” Singapore: Fifth
CAN-MET/ACI international conference on recent advances in concrete
technology, 173-187.
Xibin, H., Yinzhong, Y.,Shixia, W. (2000). “Analysis of Abrasion of Wall Material by
High Velocity Sediment Laden and Aerated Flow.” Journal of Hohai University, 3,
27-31.
Hongzhu, K., Zhigao, L., Ben-Ming, L. (2007). “A Study on Evaluation Standard of
Skid-Resistance Performance for Expressway Tunnel Pavement.” Highway, 9,
90-93.
Panagouli, O.K., Kokkalis, A.G. (1998). “Skid Resistance and Fractal Structure of
Pavement surface.” Chaos.Solitons & Fractal, 9(3), 493-505.
Park, S. B., Seo, D. S., Lee, J. (2005). “Studies on the sound absorption
characteristics of porous concrete based on the content of recycled aggregate and
target void ratio.” Cement and Concrete Research, 35, 1846-1854.
Xueliang, Y., Liang, Y.,Qun, Y.(2003). “Field Survey and Analysis About Skid
Resistance of Tunnel Pavements.” Highway, 12, 136-140.
Zong, Y., Zhongyin, G., Yun, H. (2002). “Study on Accelerated Test Method for Skid
Resistance of Asphalt Mixture Wear Layer.” East China Highway, 2, 50-54.
Theory and Application of Total Project Management
Yuan Jianbo, Zhang Qisen
Highway Engineering School, Changsha University of Science & Technology, Changsha 410076,
Hunan, China.
ABSTRACT: Based on the fundamental principles of total quality management and

system engineering, total project management is defined and analyzed, which is a
system management for all the fields of a project using the theories and methods of
total quality management. After analyzing the characteristics of total time management,
total information management and total risk management, and integrating these with
the theories of total quality management, total cost management and total contract
management, the scientific project management method is total project management is
systematically demonstrated. The theories of total project management are
corresponded with a new project management software system. This software system
is designed and developed to meet the requirements of total project management.
Using this software system in the project process can promote the development and
application of total project management.
Key words: total project management, total time management, total information
management, total risk management, system design method.
INTRODUCTION
Total quality management was firstly used to control product quality, then, its
application was extended to service fields. Juran brought up the concepts of little Q
and big Q [1]. Quality refers to the characteristics of the products that meet the
customer's needs as defined elsewhere. In the little Q, the customer only includes
outside people who purchases goods, and the product only includes goods. In big Q,
the customer includes inside and outside people, and the product include goods and
service. Therefore, total quality management is adapted to manage goods quality as
well as service quality.
PMI analyze quality control problem in the processes of projects management[2]
who believe that quality includes the quality of the project management and the
quality of the project product, and the quality of project management is the foundation
116
and guarantee for the quality of the project product, and present the theories of project
life cycle management, integration management and plan-do-check-act
cycle(PDCA)management, and validate that total quality control can not only improve
the quality of the project product, but also the quality of the project management.
Westney put forward the idea of total cost management [3, 4]. Then, many specialists
study on it, and put forward total cost management system. Integrated with the
theories and technologies used in the process of cost management, total cost
management is the method and procedures used to manage the project life cycle cost.
In which, “total cost” covers the cost of all sorts of devoted resource, and “total
management" includes the management of project life cycle, all sidedness, all
resources and overall processes. By using PDCA cycle management to manage the
total life cycle cost and asset in total cost management, enterprises asset are be
optimized and their profitability are promoted[5]. Fang brings up the idea and means of
total contract management[6]. Those above theories are the theoretical foundation of
total project management.
INTENSION OF TOTAL PROJECT MANAGEMENT
Total project management applies the theories of methods of total quality

management to manage all the project fields (such as quality management, time
management and cost management, etc) which contains: Firstly, project management
is total quality management. To ensure the quality of the project management, total
quality management should be used to manage the project quality, time, cost and
information, etc., which include setting a high value on project stakeholders’
ascertaining the participation of all members, strengthening the project team, attaching
importance to the impact of requirements
Man-Material-Machine-Method-Environment, namely "4M1E", on the project quality,
strengthening the project management plan, developing the PDCA cycle,
strengthening the project life cycle management. The second, project management is a
overall and system management to all the fields of the project such as quality
management, time management and cost management and so on.
Total Quality Management Characteristic of Project Management
Modern project management theories presents that the project management

quality include the quality of the product and the quality of the project management [2].
In the processes of construction project management, the quality of the construction
management is the same importance as the quality of the construction products. The
quality of construction management belongs to the service quality.
Total quality management is effective methods for improving product and service
quality [1].According to the theory of total quality management, to improve project
management quality effectively and achieve the project goal, total quality management
must be used to manage the project management activities. It is total project
management with total management, the whole team management, overall processes
management, all areas management, all factors management and PDCA cycle
management. In the process of the total management, managers should the same
importance to quality management, schedule management, cost management and
others management. In the process of the whole team management, the success project
management requires the participation of all the members of the team. In the process
of the overall processes management, managers should pay attention to project life
cycle management, including plan phase, construction phase , operation phase, and so
on, by strengthening the project planning, decreasing the project life cost, increasing
the project benefit. In the process of all areas management, the customer requirements
are meted., and all the partners are organized and managed effectively by using the
project contracts. In the process of all factors management, all the factors which have
impact on the project quality must be considered. In addition, in the process of the
project management, according to the theory that the quality is oriented from planning,
managers should strengthen the planning and improving the quality of project
management by using the PDCA cycle.
System Characteristic of Project Management

Project integration management theories presents that project management is an
integrated task, and the success in one area can usually have influence on the others [2].
In the construction project management, the schedule and cost of the project is varied
as the varied quality of the project. Normally, the higher is the quality, the longer is the
duration, the more is the cost. Whereas, poor quality will bring rework and delay the
progress. The project schedule has influence on the project quality and cost. Delaying
or crashing or accelerating construction progress will increase the project cost.
Unpractical crashing has bad influence on the construction quality. Therefore, each
elements of the project management must be properly coordinated by using system
methods.
The above discuss that construction project management is a kind of total project
management using the deductive method. Many specialists at home and abroad
present that construction project management include "three control and two
management ", which are quality control, schedule control, cost control, contract
management and information management, and some specialists suggest which should
include risk management too[8]. They put forward the theories of total quality
management, total cost management [5] and total contract management [6]. Based on
those theories, the paper makes further study on the characteristics of total time
management, total information management and total risk management. And then, it is
presented that construction project management is total project management.
CHARACTERISTICS OF TOTAL TIME MANAGEMENT
From the following analysis, project time management is total time management
with the following characteristics, such as whole team management, overall processes
management, all areas management, all factors management and PDCA cycle
management.
Whole Team Characteristic And All Areas Characteristic of Time Management

Time management is a management activity that requires the participation of all
members of the team. Both technical and construction individuals work together in the
processes of the project. The managers at the middle level are responsible for
developing the project schedules, gathering construction information, and responding
to the execution of engineering schedules. The managers at the top level are
responsible for establishing and controlling the ultimate project time goal. Therefore,
time management has the characteristic of whole team management, which involve
effort from all the individuals.
Project time management cannot be separated from each participant’s efforts and
cooperation. In the processes of construction project management, project time
management will success only with positive participation and cooperation from every
party involved, in particular the contractors, engineers and employer. The construction
schedules can affect contractors' construction cost and liability for breaking the
contract, which can affect supervising cost and performance, which can affect the loan
interest of the construction phase and investment recovery period, which can affect
national economy benefit of the project. Consequently, there are of both contradiction
and unitarily in time management. Including contractors, engineers and owner, they
must be in harmony and fit effectively. In addition, the project progress will be
affected by the other involvers, such as designers. Therefore, time management is a
management activity with all areas characteristic.
Overall Processes Characteristic of Time Management

The progress of the construction project not only depends on the construction
speed of the project, but also depends on the design progress and the progress in the
other project stages. We can accelerate the project progress using the optimize
technique, such as fast-track, etc.. On the contrary, the delay of the design activities
may cause the delay of the project progress, and deflect project construction into
badness working season (as monsoon), which had a very bad impact on the project
schedule, In addition, the project duration and progress are influenced not only by the
design, but also by the construction methods and the construction schemes. Therefore,
to accelerate the project progress, it should be strengthened to manage construction
stage, as well as design stage and the other stages. Time management is project life
cycle management. It should be utilized to guide every project phases, including
project initial phase, project design phase, project construction phase and project final
phase. All factors management in the processes of project time management, there are
many factors which have effects on time management and are named “4M1E”.
Man - Man is a complicated factor which influences the project progress constantly.
From the above first characteristic, we can know that all the individuals of the project
team have influence on the project progress. The quality of time management and the
project progresses not only depend on the management level of the contractors and the
technical level of the constructors, but also depend on their working enthusiasm and
cooperation spirit (team work). In order to manage the process of undertaking project
management more effectively, it is necessary to consider all members’ ideas, promote
cooperation, and coordinate all members’ efforts.
Material - The supplying material and the quality of the material have influence on the
project progress. In the process of undertaking project management, whether materials
are supplied in time or not and whether the qualities of materials meet the needs of the
project can lead to rework and affect the project schedule.
Machine - Whether into the arena on time, working condition, the performance and
quantity of equipment and machine have a direct effect on the project schedule.
Method - The project schedule can be affected by construction techniques and

organization methods. Using advanced construction methods and optimizing
construction organization can accelerate construction progress.
Environment - The surround conditions such as natural environment and social

environment (weather, geology, soil of construction site, etc) can also affect the project
management.
Therefore, project time management is a kind of all factors management. To
manage project schedule effectively, it must be strengthened to manage every involved
factors.
PDCA Cycle Management

Because construction project is of complexity, it is needed to improve time
management using PDCA cycle. Usually, planning department develops and optimizes
schedules. Then, the schedules are putted into execution. In the processes of execution,
the higher level and supervision departments inspect implement of the schedules and
turn inspect feedback to planning department for dynamic adjustment, and then
adjusted schedules are done. By PDCA cycle, the quality of time management is
improved. Therefore, according to the changes of internal and external environments

of the project, the time management process is dynamic and evolves following either a
circulation type or a spiral type, as shown in Figure 1. With PDCA cycle method, the
time management can be upgraded continuously and the management goal can be
achieved eventually. Figure 2 shows the basic theories of PDCA cycle management.
For example, in highway project management, the larger circle is the PDCA cycle of
the employer in the whole project; the smaller circle is the PDCA cycle of the engineer
in the supervision section; and the smallest circle is the PDCA cycle of the contractor
in the construction section.
Improvement
A P
C D
Maintain
Improvement
A P
C D
Maintain
Figure 1. Schedule control based PDCA circulation
AP
A C DP
A
C D
P
C D
Figure 2. PDCA circulation in project schedule management
CHARACTERISTICS OF TOTAL INFORMATION MANAGEMENT
From the following analysis, project information management is total information

management with the same characteristics as total time management, such as whole
team management, overall processes management, all areas management, all factors
management and PDCA cycle management.
Whole Team Characteristic and all Areas Characteristic of Information

Management
In the processes of project management, all individuals produce and accept
information. To communicate effectively, everyone should be ought to be project
information acquisition, process trimmer, information promulgator within oneself
relevance area. Although different individual requires different information,

information management requires the participation of all members of the team. It is an
obpyramidal structure on the information range and occupancy duration among
bottom level, middle level and top level individuals. The top level individuals need
more and more time to manage more and more information. The top level information
management must depend on middle and bottom level information management and
serves them. The information management in the bottom level accept instructions and
collect activities’ information. The middle level managers collect, gather, process and
handle their department’s information, and put forward execution reports to the top
level, and accept instructions from the top level and feedback to involved individuals
in their department. The top level information management almost include various
content, such as accepting and releasing information.
Usually the project managers use 80% of work time to manage information
(communication)[9], which is decided by the quality of the information management
and the efficiency of the information management of the middle and bottom level. The
better are the information management of the middle and bottom level, the higher are
the efficiency and quality of information of the high level. Therefore, project
information management is a kind of whole team management, which involve effort
form all the individuals.
The project information management must depend on various participants. In the
processes of project management, the supervision take contractor' information as
foundation, and the owner rely on information of contractors, supervision, and other
concerned partners (such as design, material supply, etc.). Therefore, information
management is a management activity with all areas characteristic.
Overall Processes Characteristic and all Factors Characteristic of Information

Management
In the process of project management, information of every phase in project life
cycle are interdependent and are constituted to a correlative information chain. The
information from the feasibility study phase is the basis for initial design. The
information from initial design phase is the basis for construction diagram design. The
information from construction diagram design phase is the basis for construction bid
and construction management. The information form construction phase is the basis
for checking, accepting and maintaining the completed project. Therefore, project
information management is a kind of overall processes management.
In the processes of project information management, 4M1E have different effect
on information management. For man factor, the quality of information management
is decided by the accuracy and the feedback speed of information which is provided
from all the individuals of the project. For equipments and method factors, the speed
of information processing, handling, delivering, storing and feedback can be speeded
by using the computer techniques, and the quality of information management can be
improved. The speed and quality of collecting information can be improved by using
GIS, GPS, photograph measure, remote sensing technique and data excavating
technique. Therefore, project information management is a kind of all factors
management.
PDCA Circulation Characteristic of Information Management

The quality of information management can be improved by using PDCA cycle.
Firstly, information management section draws up the information management plans.
Then, the plans are putted into practice by involved departments according to their
respective duty or division. During the implement of the plans, the top level managers
and engineers make dynamic check, and the information management section make
dynamic adjustment in view of checking result. Finally the adjusted plans are carried
out. Therefore, the project information management is a PDCA cycle job. Along with
the changing of the inside and the outside environment, information is dynamically
managed. By using PDCA cycle, project information management improves
continuously to ultimately achieve the target.
CHARACTERISTICS OF TOTAL RISK MANAGEMENT
In the processes of project management, there is risk everywhere. When risk

occurs, their affects will bring serious effect on project cost, schedule and quality. It is
certain that risk management should be strengthened. It is demonstrable that project
risk management is total risk management.
Whole Team Characteristic and all Areas Characteristic of Risk Management

Because there is risk everywhere, the risk consciousness of all the members of the
team need be set up to manage risk effectively. Risk management requires the
participation of every involved individual. They should be acquainted with the
methods to distinguish and deal with risk. The whole member prevention mechanism
against risk should be built up. By reinforcing the quality of activities, the technology
risk can be decreased effectively. Therefore, the project risk control is a kind of whole
team management. Usually project risk is beard by all participants together. There are
owner’s risk, contractors’ risk, engineers’ risk and designers’ risk. Some risks are
undertaken by the owner and contractors together. Identifying risk, preventing risk and
controlling risk are the responsibility for all participants. The project risk management
is a king of all areas, which require the participation and effort of every participant.
Overall Processes Characteristic and all Factors Characteristic of Risk

Management
Risk exists in the whole processes of project construction, and the earlier is the
time, the more is the risk[2]. Accordingly, risk management should be undertaken in the
whole project proceeding. In the earlier period, risk is paid more attention to,
including identifying risk, analyzing risk and dealing with risk.
Moreover, the possibility of risk occurrence exists in every management process.
There is the quality trouble and product safe risk in the quality management process.
Lots of factors incur to delay the progress in time management. In process of cost
management, various reasons cause project cost increasing etc. Therefore, risk
management should be strengthened in every phase.
The all factors characteristic of risk management means 4M1E can initiate risk in
the implement processes. Man may have the work accident. Material may have a
certain blemish. Machine appear breakdown probably. Inappropriate technique or
management methods affect quality seriously. Environment may cause the natural
disaster etc. Therefore, the project risk management is a management activity with all
factors characteristic.
PDCA Circulation Characteristic of Risk Management

PDCA cycle can keep risk management on the continuous improvement and
exaltation. First risk management plans are be established. By risk management plans
risk management activities are undertaken including dealing with risk and controlling
risk in time to escape risk loss as far as possible. The implementation of risk
management plans are checked timely, and based on the checking result, risk
management plans are adjusted and improved. By using PDCA cycle, risk
management improves continuously to achieve target ultimately. Therefore, the project
risk management has PDCA cycle characteristic.
APPLICATION OF TOTAL PROJECT MANAGEMENT
Applying Total Project Management Theories in System Design of Project

Computer-Aid Management System
Design methods of computer-aid management system are systematic original
shape and object oriented developing technique etc. According to the characteristic
and request of total project management, the computer-aid management system is
designed to meet requires of total project management with project technical economic
characteristics. The project computer-aid management system is analyzed and
designed guided by using the theories of total project management, which can be able
to ensure function integrity of computer-aid management system, and satisfy different
manager's need to the system, and overcome existing defects in general computer-aid
management system, including absent function, meeting nothing on whole team
management, all areas management, overall processes management, all factors
management and PDCA cycle management. In addition, computer-aid management
system can promote the development of total project management to raise the quality
of project management. The author has ever undertaken the development of highway
engineering management system. At the beginning stage, the system need analysis was
insufficient because of be short of the total project management guidance, and only the
system 'present needs were investigated, which have affected development progress
and system quality. Thanks to the instruction of total project management, the project
development quality has an obvious improvement, its function gets better, and its
application promotes the highway engineering progress along with improving
management quality.
Total Project Management Applied on Linchang-Expressway Construction

Linchang-expressway is a section of Beijing-zhuhai expressway who lays in
Linxiang to Changsha of Hunan province, which is invested by the World Bank loan.
the total invest is up to 5,200,000,000 Yuan, and it is planed to be constructed for 3
years. In the processes of project management, the owner adopted the theories of total
project management and applied project computer-aid management system (in 5.1) to
manage the project. The theories of total quality management, total time management
and total cost management were adopted to manage quality, time and cost[10].
In total time management, a new network technique for highway considering

weather and other various random factors was built up[11]. As a result, the quality of
project management got an obvious exaltation. The project was open to traffic 6
months in advance, and decreased 200,000,000 Yuan of construct loan interests. The
construction load was returned in advance, and strained traffic situation was alleviated.
After opening to traffic for 5 years, the project is judged as a high grade project by
ministry of communications, and receives the honor of ZHAN TIAN YOU civil
engineering prize, and obtains a good economic and social performance.
CONCLUSIONS
All above adopt deduction and induction together to approve that project
management is a kind of total project management, and introduce its application. From
the above analysis and study, total quality management should include product quality
and service quality, which is adopted to manage both of them well. While carrying out
total quality management to product quality, those theories need to be used to manage
time, cost, contract, information and risk etc., including total time management, total
cost management, total contract management, total information management and total
risk management. In addition, the quality, time and cost need to be managed as a
system to improve the quality of project. Moreover, the theories of total project
management are the important basis for the computer-aid management system, using
the system in the project process can promote the development and application of total
project management.
REFERENCES
1. Juan, J. M. Juran on Quality by Design The New Steps for Planning Quality Into
Goods and Services {M}. Juran Institute Inc. 1992: pp. 33-38.
2. PMI. A Guide to The Project Management Body of Knowledge (3rd Edition) {M}.
Project Management Institute, Inc. 2004, pp. 8-20.
3. Westney, R.E. Project Management Trends For The 90’s {J}. Cost Engineering
1991(10), pp. 533-539.
4. Westney, R.E. Total Cost Management. AACE – Vision of Growth [J]. Cost
Engineering, 1992(10), pp. 541-548.
5. Qin Anban. Engineering Project Total Cost Management [M]. Tianjing, Nankai
University Press. 2000, pp. 56-145 (in Chinese)
6. Fang Ying. Theory System of Total Contract Management [J]. China Economist.
2003 (2), pp. 154-155 (in Chinese).
7. Yuan Jianbo. Highway Economics [M]. Beijing, China Communications Press.
2002, pp. 91-92 (in Chinese)
8. Qin Wanhua. Project Management [M]. Beijing Science Press. 2001, pp. 269-293.
(in Chinese).
9. Kathy Schwalbe. IT Project Management [M]. Beijing, China Machine Press.
2003, pp. 348-364 (in Chinese).
10. Yuan Jianbo and Liu Weijun, Total Time Management and Highway Engineering
Time Management System – 1st International Conference on Construction
Engineering and Management, Seoul, Korea. 2005.
11. Yuan Jibo, Liu Weijun, Zhang Jianren. Network Planning Technique for Highway
Engineering Considering Influence of Random Factors [J]. China Journal of
Highway and Transport. 2006, 19(1), pp. 104-107. (in Chinese)
Decision-Making Model of Highway Route Plan Based on Entropy and Entropy
Weight Theory
Zhang Yingxue1, Chunhua Han2 and Zhang Qisen1.

1
School of Traffic and Transportation Engineering,Changsha University of Science and Technology,
Changsha, Hunan, 410004, China, e-mail: [email protected]
2
Braun Intertec Corporation 11001 Hampshire Avenue S Minneapolis, MN 55438, U.S.A,
e-mail:[email protected]
ABSTRACT: It requires many factors to choose highway route plan. Different

decision making aims will decide to select different decision making variables, namely
indexes. There are also many different evaluation methods. This paper considers the
general correlation of multiple objectives and multiple attributes based on Context
Sensitive Design, and it presents the index system of decision-making route by
investigating, analyzing, consulting experts and so on. In the decision-making problem
of route plan which including indexes of M and plans of N , the paper gets the
evaluation matrix of single object about multiple index by combining with the
qualitative and quantitative principles, then gathers single index of different plans and
forms a synthesis index based on entropy and entropy weight theory. At last, the best
plan can be chosen by comparing the synthesis index value of different plans. It shows
that the index system and decision-making model have reference value for engineering
practice.
INTRODUCTION
It is a multiple objective and multi-attribute decision making problem for selecting

highway route design plans. How to use reasonable making decision method to
analysis indexes impersonality to reduce the synthesis impaction of highway on its
context as natural environment, community, landscape, scenic resources, historic sites
and so on are very important and practical. The paper introduces entropy and entropy
weight into decision making route design plans of highway, and uses entropy to
measure indexes and sets up a decision making model to select optimal design plan.
SET UP THE INDEXES SYSTEM OF HIGHWAY ROUTE DESIGN PLAN
In order to protect the natural environment, scenery and landscape resources, historic
sites and so on along highway route, coordinate highway with its surroundings, we
have made a lot of investigation about highway context and analyzed the existing
127
route design decision making indexes of China, then put forward a new index system
to evaluate highway route design based on context sensitive design concept. The
indexes are shown in Table 1 and their criteria values are obtained by questionnaire
investigation and are graded by experts. For the eight qualitative indexes as
constructions aesthetic treatment, culture heritage integrality, effect of exploitation and
using culture heritage, impaction on history and culture, impaction on community,
development of tour trade, scenic resource preservation, sight status of the highway
route in its area, it uses ten point system to get index value. When selecting and
analyzing the indexes, we put highway project into its context, and regard it as one
part of its context. Traditionally, we mainly consider highway construction from
technology and cost and ignore other factors easily. So this can not only consider
highway existing criterions and cost but also protect physical environment, aesthetic,
scenic, historic, and cultural resources and the physical characteristics of the area that
highway will travel along.
Table 1. Evaluation indexes system of highway route design
Indexes Character Indexes Character

culture heritage
Maximum height of fill slope(m) quantitative qualitative
integrality
Maximum depth of cutting slope Exploitation & using
quantitative qualitative
(m) culture heritage
impact on history and
Average fill height(m) quantitative qualitative
culture
Average cutting depth (m) quantitative Impact on community qualitative
Average amount of each kilometer development of tour
soil and rock square (104m3/km) trade
Right of way acquisition area(the scenic resource
sum of all sorts of soil area (hm2)) preservation
sight status of the
constructions aesthetic treatment qualitative highway route in its qualitative
area
ENTROPY AND ENTROPY WEIGHT DECISION MAKING MODEL
Entropy concept comes from thermodynamics originally which describes the no-
reversible phenomenon of movement process. It was introduced for the first time into
information area by C.E.Shannon. But now it is applied widely in many areas as
engineering technology, society, economy, and so on.
The process of the entropy weight decision making model has eight steps as below:
Step 1: There have n schemes and m evaluation indexes in the decision making
system. If the eigenvector of the m evaluation indexes values of number j scheme are
showed as the vector X j = ( x1 j , x2 j
,⋅ ⋅ ⋅ , x )
T
mj
, then the eigenvectors of the m evaluation
indexes values of n schemes can show as the matrix X which is
m by n : X = ( xij )m×n xij

.Where is the eigenvalue of number i index for
number j objective, i = 1 ,2,⋅ ⋅ ⋅ ,m ˗ j = 1,2,⋅ ⋅ ⋅ ,n . There are two types of index; one is
the cost type which is better if it is less, and the other is the benefit which is better if it
is larger. After normalizing the matrix X according to the following equations, (1) and
(2), the matrix R = ( r ij )m×n can be deduced (Zuo1991, Song 2004), where ij ( 0 ≤ rij ≤ 1 )
r
j
is the normalized eigenvalue of number i index for number objective.
The cost type˖ r = max xij − xij (1)

ij
max xij − min xij
The benefit type˖ xij − min x ij (2)
rij =
max x ij − min xij
Step 2: for the entropy of evaluation indexes, there have m evaluation indexes and
n evaluation objectives in the evaluation problem (( m, n ) for short as following),
then the entropy H i of number i index can be expressed as
n
. H i = −k ∑ f ij ln fij
j =1
n
fij ln fij = 0
Where i = 1,2,<<<, m ; j = 1,2,<<<, n ; fij = rij / ∑ rij , k = 1/ ln n .If fij = 0 , then ,
j =1
the entropy value is smaller. This shows that the index is useful for decision makers. It
is necessary for us to select k to ensure 0 ≤ H i ≤ 1 when the value of H i is more then
1.
Step 3: for the entropy weight of index, the entropy weight
wi of number i index
in the ( m, n ) decision making problem can be defined as:
1 − Hi
ωi = m (3)
m − ∑ Hi
i =1
w
Where, the value range of i is from zero to 1, and the sum of i equals to 1.
w
According to the entropy theory, when the difference of the same index for different
schemes is larger and the entropy weight value is smaller, it illustrates that the index
carries useful information for decision makers and should be considered as an
important factor for each scheme in that problem. If the index entropy is larger while
the entropy weight is smaller, then the index is more unimportant. In order to evaluate
more accurately and reliably for the evaluation objective, the indexes can be adjusted,
added or reduced according to the values of entropy weight.
Step 4: Construct the weighted normalized matrix C
The weight value of m index firstly can be obtained by expert method, then we
combine it with formula (3), the synthesis weight
λi about index i can be deduced
from the equation (4):
χω
λi = m i i
(4)
∑ χiωi i =1
After having introduced entropy and entropy weight, we add a synthesis entropy
weight for each index, so the matrix C with synthesis entropy weight is as follows:
⎛ λ1 r11 λ1r12 λ1r1n ⎞
⎜ ⎟
λr λ2 r22 λ2 r2 n ⎟
C = ⎜ 2 21
⎜ ⎟ (5)
⎜⎜ ⎟⎟
⎝ λm rm 1 λm rm 2 λm rmn ⎠
Then gathering each index of each scheme and forming one synthesis index, the
optimal scheme can be obtained by comparing the synthesis index value (ZHANG
2005).
Step 5: Definite ideal point and negative ideal point. Negative ideal point means
that it is a muster of all the worst points of each index. The muster is a zero matrix. It
is maybe negative value if the indexes have not been normalized. The optimal ideal
point of multi-objective is a muster matrix of the maximum point value of each index.
By calculating the distance of each scheme from the ideal point and negative ideal
point and the fidelity that describes in step 7, the scheme that is best close to the
positive ideal point can be selected.
The ideal point: P = ( P , P ,⋅ ⋅ ⋅ , P )
* * * *
T
1 2 m
In which
{
P1* = max cij j = 1,2 ,⋅ ⋅ ⋅ ,n;i = 1,2 ,⋅ ⋅ ⋅ ,m
j
}
The negative ideal point: P0 = ( 0 ,0 ,⋅ ⋅ ⋅ ,0 ) T
Step 6: Distance calculation. The distance of the evaluation object point to the ideal
point is calculated by using Euclidean distance.
m
(6)
d *j = c j − p* = ∑( c
i =1
ij − p*i )
Step 7: The fidelity of the evaluation object to ideal point.

m
∑( c ij pi ) (7)
Tj = 1 − i =1
m
∑( p
i =1
*
i )2
The fidelity indicates the ratio value that the distance of the evaluation object point
to the ideal point divided by the distance of ideal point to negative point. If the ratio is
smaller, the scheme is better.
Step 8: Scheme selection optimization priority
Prioritize and select each scheme according to the fidelity Tj value after it is
*
calculated. If the value is small, the scheme is optimal. If Tj is equal, the value of d j
d *j
must be calculated and prioritized. The distance value of is small, its scheme is
optimal from the synthesis index.
APPLICATION EXAMPLE
There are three route design plans named A1 ˈ A2 ˈ A3 to be selected of some

freeway with design speed of 80km/h in mountainous area. The data and fraction
value of each index and their expert weights show in Table 2. The decision making
process of plan is as follows:
Table 2 The form of analysis and evaluation for route plans
Expert Plan Plan Plan

Index weight χ i A1 A2 A3
Maximum height of fill slope (m) 16 15 0.04 19

Maximum depth of cutting slope (m) 30 70 0.04 60
Average fill height(m) 11 8 0.12 10
Average cutting depth (m) 8 23 0.20 16
Average amount of each kilometer soil and
0.02 10 16 22
rock square (104m3/km)
2
Right of way acquisition area (hm ) 0.22 5 8 7
constructions aesthetic treatment 0.08 8 6 7
culture heritage integrality 0.07 9 10 8
exploitation and using culture heritage 0.01 6 8 7
impact on history and culture 0.02 5 4 6
impact on community 0.05 4 6 5
development of tour trade 0.02 5 7 6
scenic resource preservation 0.04 9 7 8
sight status of the highway route in its area 0.06 4 3 5
According equations (1) and (2), the normalized matrix R is obtained.
T
⎛ 0.751011100.510.50100.5 ⎞
⎜ ⎟
R = ⎜ 10100.50100 01010 ⎟
⎜ 00.250.330.47 00.330.5 10.510.50.50.51 ⎟⎠
⎝
After calculating the entropy Hi , entropy weight

ωi and synthesis weight λ , the i
results are shown in Table 3.
Table 3 the entropy H i , entropy weight ωi and synthesis weight λi of each index
H1 H 2 H3 H 4 H5 H6 H7 H8 H9 H 10 H11 H12 H 13 H14

ω1 ω2 ω3 ω4 ω5 ω6 ω7 ω8 ω9 ω10 ω11 ω12 ω13 ω14

λ1 λ2 λ3 λ4 λ5 λ6 λ7 λ8 λ9 λ10 λ11 λ12 λ13 λ14

The ideal point and negative ideal point can be deduced when considering the
synthesis entropy weight by formula (5) in the normalized matrix. The weighted
matrix is as follows: T
⎛ 0.0200.07000.1510.0150.19000.0620.0150.00800.01500.022 ⎞
⎜ ⎟
C = ⎜ 0.02600.02100.00800.0590000.037 00.0300 ⎟
⎜ 00.0180.007 0.07100.0610.0300.1230.0080.0150.0080.0150.0020.044 ⎟
⎝ ⎠
The ideal point:
P * = ( 0.026 ,0.070 ,0.021,0.151,0.015 ,0.190 ,0.059 ,0.123 ,0.015 ,0.015 ,0.037 ,0.015 ,0.030 ,0.044 )
T
The negative ideal point: P0 = ( 0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 )T

From equations (6) and (7), the distance d *j and fidelity Tj of each scheme can be
calculated. The results are shown in Table 4.
Table 4 The distance, fidelity to ideal point and priority for plans
plans A1 A2 A3
distance d *j
priority
fidelity Tj
priority
Table 4 shows that the priority of distance is consistent with the fidelity. So the route
plan A1 is optimal.
CONCLUSIONS
1) Comparing with other methods, the weight of index by using the entropy weight
method is related to objective data not subjective factors when select optimal route
scheme. This makes the calculated result is more objective and realistic.
2) When making decision route scheme, it always consider the weight of each index
separately and ignore the relationship of each other. The paper considers the
relationship by introducing entropy weight and expert weight and calculates the
synthesis weight, presents decision making model based on entropy weight theory to
select optimal scheme. This can not only overcome the common methods shortage but
also ensure that the selected route is more benefit to its context.
3) The index system for route design selection in this paper needs to be farther
improved and perfected in engineering projects for the future.
ACKNOWLEDGMENTS
The authors appreciate the support of the West Transportation Construction Science
and Technology of China (2003 318 798 05) and Hunan Provincial Natural Science
Foundation of China (07JJ3111).
REFERENCES
Zuo Jun.(1991). Multiple Objective Decision Making Analysis. Zhejiang University

Press.27-91.
Song Ximin, Zhou Shujing.( 2004). Based on Entropy Weight Method on Investment
Plan’ Appraisal of Building Project. Building Technique Development. Vol.31,
No.7, p113-114.
ZHANG Yang, WU Liu-yuan.(2005) Study on the design ideology and model of
highway landscape environment. Journal of Xi'an University of Architecture
&technology. Vol. 37, No.3, p366-369.
Application of GIS to Freeway Management System
Tian Xiaoge 1
Changsha University of Science and Technology, Changsha, Hunan, P.R.China, P.O.Box 410076;
PH(+86-731-2309375); FAX(+86-731-5219375); Email: [email protected]
ABSTRACT: An application scheme of Geographic Information System (GIS) in

Freeway Management System (FMS) is proposed. The main contents of the system and
the dynamic segmental technology adopted in the FMS were introduced. This FMS
realized the integration management of spatial data and attribute data, offered explicit
and efficient proofs for assistant decision-making by combining the characteristic and
predominance of GIS with original FMS function. It fully utilizes the advantage of
computer and GIS technology to modernize the management of freeway.
INTRODUCTION
Corresponding with the rapid development of freeway construction in China, it is

the time to build standard and highly efficient information management systems.
Traditional pavement management systems (PMS) store data in a database, and present
them using datasheets and text modes, so the user has to face large number of complex
and bald data. As an important tool, except directly reflecting the conditions of freeway,
retrieve the correlative information and make beautiful thematic maps with figures;
furthermore, GIS can also offer evidence of assistant decision-making in routine
management business. Therefore, applying GIS to strengthen and expands efficiency of
management has been a trend in the FMS.
SYSTEM DATABASE DESIGN
Based on GIS, the database of FMS is mainly composed of spatial database,

attribute database and system model base. A brief description of each is given below.
Spatial Database: - It mostly includes fundamental geography data layer, route data
layer and maintenance geography data layer, and so on.
133
Freeway Attribute Database: - This includes;

i. Freeway basic database. It mainly stores freeway attribute data, such as routes,
pavement structure, traffic volume and subsidiary facilities, and so on.
ii. Freeway service performance database. It mostly stores functional
performance data of pavement in FMS, including deflection, distress,
smoothness, skid resistance, and so on.
iii. Bridge (Tunnel) management database. It mainly stores attributes of bridge
(tunnel), such as structures, diseases, and so on.
iv. Maintenance database. It mostly keeps attribute database, such as investment
cost of important maintain engineering, development organization,
construction units, supervision units, types and technical class, and so on.
System Model Base: - It mainly includes evaluation model of pavement functional

performance, prediction model, maintenance costs model and decision-making model,
and so on.
LAYERS OF SPATIAL DATABASE
Freeway engineering contains numerous structures, such as tunnels, bridges and

interchanges. Spatial object has many types and large quantities. So when we design the
database, it is necessary to consider the requirements of freeway management; and
reasonably organize different types of spatial objects into different layers, in order to
apply management system conveniently.
Spatial layer data has point, line and surface. One layer only keeps the same type of
spatial data, thus, not only physical types are unified but also figure formats are unified
in the same layer when layers are being used.
According to the requirements of the freeway management, the spatial data of
freeway are divided into four species and eighteen layers in this paper, as shown in
Table 1.
THE DESIGN OF SYSTEM FUNCTIONAL MODELS
This system uses the architecture of client/server. The functional modules

configuration is shown as fig1.
(1) Map Display

Traditional PMS only displays attribute data in text mode, but all kinds of
information can be displayed in map mode when applying GIS. Different line types,
signs and colors can be used to denote different geographic entities. And different colors
express different pavement function statuses. It also can be carried out many
manipulations on embedded maps in pages, such as spatial orientation, magnifying,
de-magnifying and ramble, and so on.
(2) Retrieve in Figure and Character Mode

The FMS can retrieve information in both figure and characters mode, which means that
it can retrieve character information from figure and figure information from character.
Figure inquiring characters means that choosing one geography body from maps to
inquire correlated attribute information. Characters inquiring figure, that is to say, some
attribute conditions defined by consumers, particular geography bodies are thrown out
on the map.
(3) Information Output

Information output involves the output of text, diagram and thematic maps. The
output of text and diagram is mainly using original PMS diagram output modules.
While the output of thematic map is a particular function of GIS, it means that a
thematic map can be made and printed according to the selection of special topic, like
pavement status graphs, traffic volume graphs, and so on.
Table 1 – Layers of the spatial data.
Type Figure Layer Name Spatial Data Format

Reference Datum mark Point object
Systems Central line Line object
Sub-grade Surface object
Pavement Surface object
Bridge1 (Big Bridge) Surface object
Bridge2 (Small Bridge) Surface object
Interchange1(Intercommunicated
Surface object
Management Unit Interchange)
Interchange2 (Separated Interchange) Surface object
Tunnel Surface object
Culvert Surface object
High slope Surface object
Service station Surface object
Terrain line Line object
Background Layer House building Line object
River Line object
Structure name Text
Mark Layer District name Text
Stake mark Text
(4) Assistant Decision Support

All kinds of models such as evaluation, prediction, expenditure and
decision-making were built. The mid database and end-all results of decision-making
can be displayed graphically, retrieved and exported.
Freeway Management System

Management and
Chart and Thematic
Decision-making
Searches Inquiry
Making statistic
Map Display
Maintain
Assistant
Data’s
Map
Spatial Database, Attribute Database, System Model
Base
Figure 1. System Function Module Graph.
DYNAMIC SEGMENTATION TECHNOLOGIES AND ITS REALIZATION
Linearity and multilayer are the distributing character for Freeway geography
information. Linearity indicates that information is distributed linear along the
geography routes, and is in connection with mileage. And this is a notable difference
from other geography information system (such as mine and land management
systems). Multilayer indicates that manifold information is distributed along the same
route; segment is represented by different mileage in every information layer. For
example, one road has many kinds of information, such as technical grade, pavement
type, traffic volume and environment along the route, but the length and position of
different kinds of information segment are different. Therefore, segmentation must be
well settled to fulfill inquiring and analyzing of various properties using GIS in freeway
database. Segment types of linearity objects are mostly have three methods: fastness
segment method, alterable segment method and dynamic segment method. According
to analyze excellent and shortcoming of these three segment methods, Dynamic
segmental method is chosen in this paper.
Dynamic segmental method adopt topology data structure, based on topology graph,
route is built. Facilities data are regard as event happened on the route and are stored in
the corresponding event tables, and the relation to route is established according to
consumer identification code. GIS software carries out real-time inquiring to event and
divides and displays the route based on user’s requirement. Route system is adept by
dynamic segmental method; route is a logic combination which is built on topology
graph, but not a real “Entity”. Segment of route for different attribute is also a logic
concept. GIS software stores them in the corresponding event table, and creates the
corresponding segments when it is needed.
This has much flexibility because it does not really segment figure. Dynamic
segment method can express all of facilities’ attribute information on the same route,
and attribute information can be saved in the traditional relation database, and
maintained by relation database to decrease the redundancy data. At the same time,
spatial data and attribute data can be maintained separately, which is convenient to
management.
To build the route system and event attribute table of the road is the main works to
realize the dynamic segment in freeway management system. The implementing step of
dynamic segment mainly includes:
(1) Establishing Route System. Starting point of the road is determined according to
the road tendency, the distance (or mileage stake) of each node to starting point is the
measurement of the node.
(2) According to the spatial relations of route to build the attribute table of freeway
sections corresponding to spatial database. Road coding, starting mileage and terminal
mileage are added into the attribute table. Thus attribute tables of freeway sections is
founded based on different properties
(3) After building the attribute database and spatial database based on dynamic
segmental method, road can be segmented according to one or more different selected
properties.
CONCLUSIONS
The advanced GIS technology was applied into a freeway management system. The
paper introduced the main contents of the system and studied the realizing method of
route segment technology in the progress of system developing. This FMS realized the
integration management of spatial data and attribute data, offered explicit and efficient
proofs for assistant decision-making by combining the characteristic and predominance
of GIS with original PMS function.
REFERENCES
1. Maguire, D. J., and GoodChild M.F. (1991). Geographic Information System:

Principle and Application, Longman, London.
2. Mapinfo V5.0 User Guide. Mapinfo Corportion
3. MapX Reference, User guide, Mapinfo Corportion
4. Poiker T.K., and Chrisman N. (1998). Proceeding of 8th International Symposium
on Spatial Data Handling, International Geographical Union Geographic
Information Science Study Group
5. Tang Guoan, and Zhao Mudan. (2000). Geographic Information System, Science
Press, Beijing
6. Yao Na (2000). Tutorial of GIS, Mapinfo and MapBasic, Beijing University Press,
Beijing.
Highway Bridge Construction Process Simulation Base on 4D Visualization
ChengHan Zhou 1 and WeiDong Wang 2 ,*
ABSTRACT: It is very important that 4D visualization technology is applied in

bridge construction management since traditional tools of construction management
cannot present intuitive construction progress. This study established a 4D simulation
model for bridge construction management by integrating the construction schedule
and resource consumption progress with the bridge 3D model. Based on this model, a
4D simulation system of bridge construction was developed. With this system, a
dynamic simulation of bridge construction schedule and resource requirements were
realized, which make it more convenient for the users to forecast and inquire
construction progress and resource consumption intensively in a visual environment.
In brief, it realized visualization management and control in bridge construction. The
Bridge Construction 4D Simulation System was applied in the construction
management of DaWu River Bridge, a 340m pre-stressed concrete continuous rigid
frame bridge.
Keywords: 4D simulation model; Visual construction management; Construction

schedule; Construction resource consumption; Bridge
INTRODUCTION
Compared with any general construction project, the large-scale bridge, as a

control project in the highway engineering, has the characteristics of large invest, long
construction periods, making a great impact and a lot of data and information been
processed in construction management.
As an important part of construction schedule, construction planning has a
limiting characteristic that is the ability of superintends, project managers, and
1
School of Civil Engineering and Architecture, Central South University, 22 Shaoshan South Rd.,
Changsha 410075, China
2
Virginia Tech Transportation Institute, Virginia Polytechnic Institute and State University, 3500
Transportation Research Plaza (0536), Blacksburg 24061, VA, U.S.
*
Corresponding Author. Tel: +1-540-231-1058; Fax:+1-540-231-1555;
E-mail:[email protected]
138
foreman to visualize and communicate construction plans. Especially, a detailed and

intuitional construction planning will identification of work sequence conflicts,
identification of potential safety hazards and improved communication of plan to
crews in a large-scale bridge construction process (Mckinney and Fisher 1998).
Obviously, traditional method, such as bar chart and logical network chart, are very
difficult to work effectively. These 2D performances have the following flaws:
1. It seldom allows the planner to detect omissions, conflicts, or logical errors,
which will result in reduced productivity, delays in project completion, and claims
between involved parties (Riley and Sanvido 1995; Varghese and O'Connor 1995; Guo
2002), because a large-scale bridge construction procedure is so complex that the
static 2D charts are difficult to intuitionally present the relationships among the
sequences of work activities. The general people, such as investors of the facilities,
may have little experience in construction project. As a result, they may never be able
to participate in the process of developing an error-free construction plan unless they
have an easy way of reviewing the construction schedule (Kang et al. 2007).
2. Increasing the frequency and time of communication among the investors,
superintends, contraction team members and foremen and reducing the efficiency of
construction. In order to explain the necessity of 4D model introduced to contraction
management, Kang et al. (2007) has done a interesting empirical study in which some
tasks of construction schedules and planning with some conflicts and logical errors
were given to the different owner teams and contractor teams. Then, recorded and
analyzed the frequency and time of communication for identifying the conflicts or
logical errors in construction schedules between the owner teams and constructor
teams using traditional 2D form and 4D model tools, respectively.
Results of the research presented that teams using 4D model detected more
logical errors, more accurately, and faster, with less need for intrateam communication
than teams using 2D form. These findings show that the 4D construction visualization
model can enhance team performance and may also improve team collaboration in
construction planning and scheduling.
This paper presents a 4D simulation model for large-scale complex bridge
construction management. Based on this model, the study developed the Bridge
Construction 4D Simulation System. As a engineering practice, the system was
applied in the construction management of DaWu River Bridge, a 340m prestressed
concrete continuous rigid frame bridge.
4D Simulation Model of Bridge Construction Progress and Resource

Management
Definition
Collier and Fischer (1996) proposed one kind of 4D model, and applied this
model in the construction of the San Mateo County Health Center. 4D model,
integrating the 3D CAD representation of physical components with the time

dimension from a schedule, dynamically displays the three-dimensional model of this
structure with time changed during the period of construction.
The 4D-CAD technology has provided the possibility to realize a bridge virtual
construction process with dynamic simulation.
It must be pointed out that a lot of 3D simulation models, at present, are usually
used to present a virtual bridge in a static view or in a 3D dynamic roaming, which has
not unified the construction progress or the raw material management.
4D Simulation Model
A bridge is composed of several components, such as beams, piers, abutments,
foundations and so on. In practical, each component of bridge, as a part of bridge
construction process, is associated with 3 dimension data, construction progress and
raw material consumption. According to principle of Object Oriented Programming,
the researches define each component as an abstract object which is an unbreakable
mixture of attributes and behaviors. The object, as shown in Fig. 1 and Table 1, is
composed of 3 sub-objects, which named dimension, progress and raw materials
respectively. In each sub-object, there are some attributes and behaviors (Table 1).
Therefore, the object and its attributes and behaviors compose a 4D simulation model
shown as Fig. 1 and Table 1. The 4D simulation model, having a high reusability, can
simulate each component of bridge when it been input relative data of the component.
Dimension Sub-object
This sub-object is the simplest one in the 4D simulation. It responses to input and
output the component dimension data and type. In order to draw a 3D component
model, the system provides a serials interface to input and inquire all dimension data
(Fig. 2). Also, there is a standard component database which is continence for user to
select normal components of bridge.
Progress Sub-object
As a foundation of simulation of bridge construction process, the sub-object of
progress is key technology of the 4D simulation model. Before the progress control
and simulation, a detail schedule should be identified by the 4D simulation model. The
critical path method (CPM) is an efficient way to schedule and analyze the
implementation and performance of projects in general. Popescu and Charoenngam
(1995) stated that CPM is widely recognized as the most effective method for keeping
large and complex construction projects on schedule, under budget, and up to
professional standards. The construction schedule was built into the 4D simulation
model based on CPM techniques in this study. The attributes in the progress
sub-object recorded the names and durations of work sequences. The system can
present bar chart and CPM network chart based on construction schedule.
In this study, given a certain time that after the bridge construction start time, the
simulation system makes the judgment which components of bridge had been finished
and which components of bridge are being constructed at that time according to the
construction schedule. Then, the system calculates a percentage that the components
of bridge have completed, presents the all or part of components that have been
finished in 3D view (Fig. 3). Project participants can simulate the construction
progress at any time to find any conflicts and logical errors which may be exist in the
construction schedule.
Raw Materials Sub-object

In this sub-object, the raw materials consuming model is more important than
others (Table 1). This study thinks that the quantity of raw materials consuming is
definition and always comes in line with the raw materials planning.
Table 1. Attributes and behaviors of the 4D simulation model.

Dimension sub-object:
Attributes Component type, appearance dimension et al.
3D model of components creating model
Behaviors
Component dimension inquiring method
Progress sub-object:
Attributes Names and durations of work sequences et al.
Bar chart and CPM network chart creating model
Behaviors Progress simulation model
Progress inquiring method
Raw materials sub-object:
Attributes Names and quantity of row materials et al.
Raw materials consuming model
Behaviors
Resource requirement inquiring method
Dynamic presentation Bridge three-dimensional animation according to
model bridge construction progress
Figure 1. 4D simulation model.
Figure 2. Input and inquire dimension data.

Figure 3. Bridge presentation in 3D view according to construction progress

simulation.
A general method to assess the raw materials consuming of construction activities

is calculating the percentage of completion in terms of the raw materials necessary to
perform these activities. In this method, the sum of raw materials used to perform all
activities represents 100% project completion. The raw materials consuming achieved
at a certain date can be measured by dividing the total raw materials expended on all
the activities completed until that date by the total raw materials necessary for the
entire project. The raw materials consuming situation in construction process was
shown as Fig. 4.
Figure 4. Raw materials consuming list.
Dynamic Presentation Model

This model is used to presents the bridge 3D model in a time sequence
corresponding to construction schedule. Also, the user assesses actual progress by
observing and recording the physical accomplishments on the construction site. The
system allows the user to input actual progress information. Then, a comparison
between the schedule progress and actual progress is made in the system. It provides a
tool to engineer to clearly and quickly identify the different between the schedule and
the actual in 3D view. Therefore, the engineer will make a decision to adjust schedule
to satisfy the actual progress, or vice versa.
CONCLUSIONS
The study developed the Bridge Construction 4D Simulation System that will
assist construction managers to reduce conflicts and logical errors in construction
schedule, promote construction progress, avoid the resources waste, enhance the
efficiency of construction management. Some important results shows as following:
(1) Integrated the bridge 3D model, construction progress and the raw materials
requirement planning, 4D simulation model was created for the bridge construction
progress and the resource management.
(2) Proposed a dynamic simulation technology based on the 4D simulation model,

which can simulate the large-scale bridge construction process and realize the
construction progress and raw materials visual inquiry.
(3) Applying the Bridge Construction 4D Simulation System in actual engineering
projects has proven the system validity and usability.
ACKNOWLEDGEMENTS
The research work was supported by Guizhou Communication Planning Survey

and Design Academy, China. The authors appreciate Dr. Hesham Rakha, the professor
of Virginia Polytechnic Institute and State University, for providing good research
condition. Also, we are grateful to anonymous referees for their useful comments and
careful manuscript reviewing.
REFERENCES
1. McKinney, K., and Fischer M. (1998). "Generating, evaluation and visualizing

construction schedules with 4D-CAD tools." Automation in Construction, 7(6),
433-447.
2. Riley, D. R., and Sanvido, V. E. (1995). "Patterns of construction-space use in
multistory buildings." J. Constr. Eng. and Mgmt., 121(4), 464-473.
3. Varghese, K., and O'Connor, J. T. (1995). "Routing large vehicles on industrial
construction sites." J. Constr. Eng. and Mgmt., 121(1), 1-12.
4. Guo, S. J. (2002). "Identification and resolution of work space conflicts in building
construction." J. Constr. Eng. and Mgmt., 128(4), 287-295.
5. Kang, J. H., Anderson, S. D., Clayton, M. J. (2007). “Empirical study on the merit
of web-based 4D visualization in collaborative construction planning and
scheduling.” J. Constr. Eng. and Mgmt., 133(6), 447-461.
6. Collier, E. and Fischer, M. (1996). “Visual-based scheduling: 4D modeling on the
San Mateo County Health Center.” Proc., 3rd Congress on Computing in Civil
Engineering, ASCE, New York, 800-805.
7. Popescu, C. M., and Charoenngam, C. (1995). Project planning, scheduling and
control in construction, Wiley, New York.
On the Construction Organization Design for the Widening of Expressways
Li Qingbo 1 ; Lu Guanqun 2 and Fu Xiaohong 3
ABSTRACT: Expressways in China are busy and unable to meet needs of traffic and
transportation capacity, which hinders the economic development in many areas.
These busy roads need to be widened, namely the four-lane roads need to be widened
to be eight-lane roads. But a suitable design plan for the widening of expressways is of
vital importance. This paper puts forward some construction organization design
principles and methods of the widening projects and focuses on three designing plans
for how to construct sub-bases, pavements and bridges, which are valuable and useful
for other similar widening projects.
INTRODUCTION
The widening of expressways in China has a history of about seven years. In

recent years, there has been some research on the construction organization plans for
expressway widening. For example, Chai Xiaolong(2006) compared some pavement
construction plans , Han Yi and Li Jie(2007) studied some basic theories and concepts,
and Zhang Fengyan(2006) discussed some designs and methods for traffic
organization. However, some studies have weaknesses such as introduction of pure
theories and simple design plans, and lack of specific and practical implementation
processes. The construction organization for the widening expressways is different
from that of a new road. Widening an expressway involves construction and traffic
organization with traffic open (Han Yi and Li Jie 2007). Besides we should take into
account the smooth and safe traffic operation, the safety of construction workers and
equipment and construction quality. Widening expressways is very difficult and
complex. It is necessary to study how to design suitable plans. Based on the widening
project of the Xu-Luo expressway, a part of Jing-Zhu freeway, this paper discusses the
designs of construction organization for each sub-project.
1
Professor, School of Communications and Transportation Engineering(SCTE), Changsha University
of Science and Technology(CSUST), P.O. Box 410076, Changsha city, China, 45,Chiling road; PH (86)
013974802320; FAX(86)0731-5219375;e-mail: [email protected]
2
SCTE, CSUST, P. O. Box: ditto, PH (86) 013875882904; e-mail: [email protected]
3
SCTE, CSUST, P. O. Bo: ditto, PH (86) 013755025499; e-mail: [email protected]
146
DESIGN OF THE CONSTRUCTION ORGANIZATION
Design Principles
Expressways are the traffic and transportation backbone, widening expressways is
closely related to the economic development. A suitable plan is of vital importance for
a smooth implementation of the new widening project, which usually consists of two
parts: the construction plan and the traffic organizations plan. The principles of the
former should be technically reliable, convenient, applicable, rapid and economical.
The principles of the latter should be: avoiding traffic accidents and not affecting
traffic operation.
Construction of the Sub-Base

The construction of the sub-base does not affect the traffic. Two temporary roads
are constructed on both sides of the sub-base for the convenience of other work related
to the sub-base construction .The construction must meet the design requirements (see
Figure 1).
.5 1:1
1:1 .5
The original road embankment The original road embankment

1:1 1:1
.5 .5 .5 .5
1:1 1:1
(a) Step1 (b) Step2
Figure1.The construction of the sub-base
The sub-base construction and the traffic organization are carried out in two steps:
1. As shown in Figure 1, clean and consolidate the basement of the bilateral
mosaic part, make steps on the original slope and install the temporary
drainage facilities, and the traffic of the four-lane road is open;
2. As shown in Figure 1, the construction for the foundation of the bilateral
mosaic part is that to make the steps one by one on the east and the west slopes
and fill earth timely, compact it layer by layer to the top surface of the new
road bed, at the same time bury the geo-grille and the horizontal drain pipes,
and the traffic of the four-lane road is open.
Construction of the Pavement

The construction of the pavement affects the traffic. It must meet the design
requirements. Besides, the traffic organization plan is particularly important. The
following three plans "side and section pavement construction ", " temporary
pavement and side construction ", and "transition pavement and side construction" are
compared and analyzed as follows:
Plan1: side and section pavement construction (see Figure 2)

The pavement is divided into sections according to the structures (bridges,
channels etc.) of the whole line and the distribution of sub-exchanges.
(1)The east side of each construction section is closed first and the traffic is open
to the west side of the corresponding section, the travel speed must be limited and then
the pavement construction of the entire east construction section can be carried out: i)
to remove the old pavement, ii) to construct the sub-base of the new roads, and pave
the base layer and the surface at the same time.
(2) Transfer the traffic of the west to the east side with the four-lane traffic open,
and then proceed to the pavement construction of the west section;
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
T h e o rig in a l e m b a n k m e n t T h e o rig in a l e m b a n k m e n t
× 3. 5
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
× 3. 5 × 3. 5
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
Figure2. Plan1 “side and section pavement construction”
(3)Complete the construction of traffic facilities; and the traffic is open again;
Plan 2: temporary pavement and side construction (see Figure 3)

(1) The traffic of east side is open. The speed is limited on the west side. The west
side of the temporary pavement is first constructed. The traffic of west side is open;
(2) This construction is the same as (1) of Plan1.
(3)First transfer the traffic of the west to the east, the east four-lane traffic is open,
and then proceed to the pavement construction of the west section, and complete the
construction of traffic facilities;
Plan3: transition pavement and side construction (see Figure 4)

(1) The traffic of east side is open; the speed is limited on the west side. The
construction of this part is carried out as follows: to construct the west side of the
new pavement; to pave the road to the middle surface according to the structure
level of new roads (The surface
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
× 3. 5
.5 .5 1: 1 :1
1 .5
1 :1 1 :1 .5
× 3. 5 × 3. 5
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
Figure3. Plan2 “temporary pavement and side construction”
layer is not constructed until the middle layer is done); locate the slurry seal which
acts as the provisional water-proof layer on the surface,. These three steps form the
transitional road, four-lane road on the west side;
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
The original embankment The original embankment
× 3. 5
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
× 3. 5 × 3. 5
.5 .5 1 :1 1 :1
1 :1 1 :1 .5 .5
Figure4. Plan3 “transition pavement and side construction”
(2) This construction is the same as (1) of Plan1.

(3) Transfer the traffic of the west to the east side with four- lane traffic open, and
then proceed to the pavement construction of the west section: to remove the old
pavement of the west side, controlling the removing depth according to the treatment
of old pavement; to construct the middle layer of the new roads, and then pave the
surface layer and the surface at the same time;
Plan2 and Plan3 involve more complex construction organizational work and more
processes. But only a small number of heavy vehicles are needed. Plan1 has simpler
construction organization, less processes, and continuous operations. But the two-lane
traffic can not meet the busy traffic .The heavy-duty vehicles should be appropriately
used to avoid congestion, and the unity management traffic is needed because of the
complexity of the management of transportation organizations. The construction of the
new widened roads of Plan1 and Plan2 is completed once, it is easier to control
construction quality. The construction of the west old road of Plan3 is not completed
once, it is harder to control the quality. Plans 1 and 3 have less waste of temporary
material, and lower costs. Plan2 has more waste of temporary material and higher
costs. Considering the advantages and disadvantages of the three plans, the plan “side
and section pavement construction” is recommended.
Construction of the Bridge

The construction of the bridge greatly affects the traffic of the expressway. The
bridge construction must meet the design requirements. The traffic organization is
particularly important. According to the present traffic of the expressway and the
conclusion of the engineering feasibility study, and combining with the road
construction and the traffic organization, the plan of "side traffic and side construction
"(see Figure 5) is recommended for bridge construction.
Construction and transport organizations are as follows.
(1) Construct the lower part of the bridge when the sub-base is constructed, the
four-lane traffic is open;
800(The extended part) 1200(The original bridge) 1200(The original bridge) 800(The extended part)
50 50
Mid-line
Figure5. The cross-section of the sub-base
(2) Install upper hollow boards on the widened part on the both sides (excluding
the new inner-plate)ˈthe four-lane traffic is open;
(3) Due to the limited construction time and conditions, the two added roads can
be constructed at the same time, and can also be constructed in order, which is
outlined as follows: to set a temporary guardrail on the original east bridge, to limit the
two-lane traffic speed, to remove the guardrail and side rails of the original east bridge,
to cut the flange plate outside the original bridge. The traffic of west side is open, and
the traffic can also flow away;

(4) Fix concrete iron on the out-lateral board of the east original bridge, sling the
in-board of the widening part, lap steel welding, and place concrete, pre-press the
widened bridge;
(5)Close the right side traffic of the bridge, remove the temporary guardrail,
remove the asphalt concrete pavement layer and the leveling layer, place concrete into
the wet joints, unburden the bridge at the same time, transfer all vehicles to the west,
the two-way two-lane traffic on the west is open with limited speed;
(6) Place asphalt concrete into pavement layer, open the traffic of the east two-way
four-lane bridge, transfer the traffic of the west side to the east side, construct the west
side of the bridge (there are no such steps if both sides of the bridge are constructed at
the same time), repeat the above steps, and finally the eight lanes can be put into
service.
CONCLUSIONS
A suitable plan for the widening project can reduce the impact on the existing
traffic. Reducing this impact is very important. Appropriate construction organization
plans can ensure the safety of the construction and the traffic when the widening
project is going on. Through the analysis of the construction organization for the
expressway, implementation plan of each project can be determined. The conclusions
can be drawn as fallows:
(1) The construction organization for the widening project is different from that of
the new road. A suitable plan can improves the service level. Road maintenance and
traffic organization for treatment of traffic accidents should be taken into consideration
for the planning of construction organization;
(2) The construction of the sub-base has little impact on the traffic. Construction of
each project can be carried out by the temporary pavement. But we should especially
address the problems such as the difference of the sedimentation between the new and
the old sub-base, and the ground treatment of the widened sub-base;
(3) For the construction of the pavement, the plan “side and section pavement
construction” can be adopted, the appropriate diverging of the heavy vehicles will
reduce traffic pressure of the widened road, and the waste of temporary material in
construction will be reduced, the costs of construction will be lower. But the lanes
should be divided again prior to the construction to meet the traffic demand of the
direction imbalance;
(4) For the construction of bridges, the plan “connecting the upper structure
without connecting the lower one” can be applied. In other words, the upper part of the
new bridge and that of the original bridge should be connected to form an overall
bridge, which is beneficial for the force of the upper structure, and reducing the
additional internal forces produced by the differential settlement between the old and
the new bridge, and ensures that old and new bridges are structurally safe.
ACKNOWLEDGMENTS
We are grateful to the GeoHunan conference’s appreciation of the abstract of this

paper. Without its guidance and help, we could not have finished the writing of this
paper. Our thanks also go to the host, Changsha University of Science of Technology,
for its encouragement to participate in this conference.
REFERENCES
1. Cai Weidong and Han Changgen (2003). "The Construction Management of

the Reconstructing and Extension Project." Petroleum & Chemical
Construction. (6, 2003).
2. Chai Xiaolong(2006). "Study on Construction Organization Scheme of
Roadbed Widening in Jing-Zhu Expressway." Communications
Standardization. (11, 2006).
3. Du Weiqian and Liu Junfeng (2008). "Subgrade Joint Construction."
Communications Standardization. (4, 2008).
4. Han Yi and Li Jie(2007). "Study about Traffic Operation Design for
Reconstruction and Expansion Engineering of Expressways." Highway
Transportation Science & Technology.(8, 2007).
5. Hu Juxian and He Xinyuan (2006). "Construction Organizing and Quality
Controlling for Rebuilding and Extension Engineering of Shanghai-Hangzhou
Expressway ( Songjiang-Fengjing Section ). " Communications
Standardization. (4, 2006).
6. Huang Xin (2006). "Simple Analysis for Construction on Extension Project of
Highspeed Highway." Subgrade Engineering. (2, 2006).
7. Liu Jingjun and Zheng Yuguo (2004). "Freeway Rebuilding Engineering
Subgrade Construction Points." Northern Communi Cations. (8, 2004).
8. Zhang Fengyan, Zhou Wei, Wang Yuanqing, and Zhang Jia.(2006).
"Discussion about Traffic Operation Design for Reconstruction and Expansion
Engineering of Expressways." Highway.(1, 2006).
Application of the Agent Construction System Management Mode of
Government Investment Highway Project
Zhang Yingxue1 and Zhang Qisen1

1
School of Traffic and Transportation Engineering,Changsha University of Science and Technology,
Changsha, Hunan, 410076, China, e-mail: [email protected]
ABSTRACT: With the process of the investment system reform in China, the
construction management mode of government investment projects which centralizes
the functions of investment, construction, management and use has being translated to
the agent construction system management mode to separate the functions. As a new
mode, it can effectively supervise the action of government, hold back corruption and
stimulate market competition. At present, some provinces have carried out agent
construction system. But the modes are a little different in different areas because of
their different management means and facts. The paper analyzes the agent
construction system management mode and puts it in practice of Hainan province
government investment highway project after investigating and analyzing of the
Hainan existing management modes and local fact. The mode is being carried out step
by step on the emphases highway projects that be managed by transportation
department of Hainan province, such as the extended project of freeway that surrounds
Haikou city, the rebuilding freeway that surrounds Sanya city and so on. This helps to
separate the duty and power among government department that in charge, agent
construction units and agent contracts, makes the best use of resource and advantage
of different specialty. It also helps to supervise and manage the engineering quality
and the practitioner qualification and gains the higher benefit for government
investment.
INTRODUCTION
Agent construction System of construction projects is a management mode adopt

internationally, it is a project management method (commonly known as turn-key
project) that, to the government investment capital construction project, the
government departments in charge set construction enterprises (the construction agent)
down for constructing, and then turn over to the user after completion and completion
acceptance records. Agent construction System is more progressive than project legal
person responsibility system. In ‘Agent construction System’, the government
departments in charge, the construction agents and the users make an agent
153
construction contract to well define their power and responsibility and standardize
their conducts. The project construction is implemented by law of tendering and
bidding, contract management and supervision, demonstrating fully the principle of
‘openness, fairness and justice’.
Along with the in-depth reform of investment in China, the government investment
projects management mode has gradually transformed from the gradual integration
management mode to the separation management functions.
In recent years, it is started to try out the project management mode of ‘Agent
Construction System’ on the government investment projects in the whole country,
and much has been accomplished. Practice proves that, ‘Agent Construction System’
achieves a specialization construction management, and accomplish a control of
investment, gain better investment returns and raise the management level, which is an
innovation in government investment projects management.
Highway investment project is a system engineering, because of its inherent laws,
system engineering needs to through with its suitable management mode, management
program, management method and management technology to achieve (JIA2006).The
agent construction system makes a breakthrough to the old management style, so the
current quaternity management mode which includes investment, construction,
management and using, is changed into an advanced mode of these "four sections
dividing and acting mutually each other".
THE AGENT CONSTRUCTION MANAGEMENT MODE OF HAINAN

HIGHWAY PROJECTS
Analysis the existing management mode of Hainan highway projects

The existing management pattern of Hainan highway project carries out the mode of
"construction, supervision and service", although which has played a certain role for a
long time, but now it exists many disadvantages as follows.
1) The Governments interferes with the corporations each other, causes unknown
responsibility, weak supervision, and lower benefit, even corruption.
2) Some government investment of highway projects is organized and implemented
by a temporary discipline inspection team, and some units lack of engineering
management professional and technicians which are all needed for construction, and
don’t have professional technology and experience about project management, once
project ends, the organization will be dismissed immediately. This mode will causes
loss of resources of persons, wealth and properties.
3) Highway projects of government investment is self-built and self-used by the
construction department, which will lead to the separation of the responsibility, benefit
of the owners and the users. Users will be driven by their self-interest, and take all
kinds of means to compete and gain projects and capitals, thus will be easy to cause
the phenomena of the super-size and above-standard construction to emerge.
4) Government investment in highway projects and their respective organization
construction will be difficult to implement effective supervision, to correct illegal and
irregular problem in the process of construction promptly.
Therefore, it must to reform the existing government investment management
approach in investment process and the institutional implementation, design an
appropriate management flow. This not only can rectify and regulate the market
economic order and fundamental strategy, ensure project quality, improve investment
benefit, but also can prevent the source of corruption, reduce costs and improve
efficiency.
During the "11th Five-Year Plan" period, the Hainan highway construction projects
invest over 110 billion yuan. Some important projects which are taken over directly by
the transportation department of Hainan Province will all be managed with the agent
construction management mode, including the extending construction of Haikou
district freeway project, the freeway through Sanya district, projects of repair and
support facilities rebuilding, the exit roads of old city exploitation zone, the exit road
of Oriental Chemical City, exits road of Changjiang Industrial zone, Wenchang Fuyao
Glass exit road of quartz sand and the sea-yu eastward link road of No.223 national
highway. Therefore, for a very long time in the future, the highway construction of
Hainan province will enter a new golden period. Under the huge investment
macroeconomic context, on the one hand, it needs to speed up engineering project
management system reformation, on the other hand, it needs to explore and research
highway agent construction system of China's highway construction continuously to
ensure the healthy development of highway.
The flow of the agent construction management mode

In the project management of agent construction system, the main departments of
government, the agent construction corporation, the contractor and the consumer
perform their respective duties and supervise each other to achieve their common
goals, which can overcome the disadvantages of traditional mode of “investment,
construction, management, use” (DENG2006). After having investigated and analyzed
the existing management mode and the disadvantage of the Hainan freeway projects as
well as the management modes of the same condition construction projects in other
cities and areas of China, combing with all kinds of existing policies and regulations
formulated by the different local governments of Hainan and adapting the
characteristic of local construction projects, the paper puts forward the flow about
agent construction management mode of Hainan province government investment
highway project which shows in Fig. 1.
The characteristic of the mode is as follows:
1) Highway project owners (Transportation Department of Hainan Province) is
responsible for the bidding affair of the agent construction units, investigation units,
design units, contractors, supervision units and the consultation units, and signs
contracts with them. The agent construction units supervises and inspects the entire
bidding process of the owners units, and design units, contractors, supervision units
and the tender of consultation units.
2) The Transportation Department establishes market access conditions, authorizes
the project management units with the qualified approval to compete the projects and
the agent construction units takes the open manner to bid for them.
3) Set up the special account for agent projects and manage by the transportation
department. The fund must be agree by the agent company and pay directly to the
stakeholders.
4) Establish a strict supervision system, including the discipline supervision,
inspection audit supervision and the public voice.

the examination and the transportation Signed a contract
approvement
department of the department of (the rights,
government project Hainan province obligations, etc.)
The individual projects are
formulated by the government, project
Organizate the consumer
other projects are established the
Project Management
access conditions by the
Companys to invite
government, introducing the
bidding
competition system, and achive
them through competition
among the project management the Project
company that has good Management
performance of construction Company
management and can bear the Organizate
investment risk of project contractors to
management and the project invite bidding
replacement onstruction.
contractor contractor other

1 2 contractor
FIG. 1 The agent construction management mode of Hainan highway projects
The basic functions of each department in Agent Construction System

The project management mode of agent construction system actually is the process
of the three units of government department, project management and project use units
which aim at the common goals, they coordinates with each other. To implement agent
construction system, it should be distinguished each responsibilities, this is a decisive
factor which related to whether the management process is smooth or not.
1) The responsibilities of government departments
(1) Audit approval
Approving the project feasibility study report and decide design and design budget
scale.
(2) Supervise and urge the schemes
Issue investment schemes, arrangement the construction funds, pay funds timely
according the construction plans.
(3) Supervision the invite bidding
Choose agents by inviting public bidding openly, supervise it performance.
(4) Inspection Coordination
Auditing standards and scale of projects in the implementation of major changes and
financial adjustment plan and coordinate construction of major issues.
(5) Supervision according to laws
Supervise project management company by the laws and deal with lawless behave.
2) The responsibilities of project management company, namely the agent company.
Project Management company is the main entity which is commissioned by the
government during the process of construction. It takes the charge of the projects in all
directions and full system and the entire responsibility within its responsibility. The
agent management company is responsible for the government and takes advantage of
their professional, technical strength and strict management experience during the
project planning, design, initial preparations, the project budget, selecting construction
unit bid and buying the materials and equipments period. Control and management the
project investment, quality and construction time to ensure the contract aim can be
achieved smoothly. Responsible for the management of the construction funds to
ensure construction funds can be used in its appropriate way and accept the relation
departments’ examination and supervision.
3) The duties of the organization which uses the project
(1) Put forward the project proposal document according to the actual needs and
development plan of the project.
(2) Put forward the necessary function at the stage of the project planning.
(3) Put forward some advice and suggestions during construction, supervising the
quality, the construction time and the rational use of funds, and as well as the behavior
of the representative.
(4) Participate the examination and take-over of the project after it has been finished,
take charge the completion project and its service and maintenance.
APPLICATION OF THE AGENT CONSTRUCTION MODE
Project overview
This project is an important section of Qionghai-Linshui rebuild road project of east
Haiyu route in Hainan province, and the route starts from the south region of Jiaji
bridge in Qionghai city, comes by the town of Zhongyuan in Qionghai city, the
Longgun town, the Damao town, the Wangcheng town, the Changfeng town, the
Xinglong farm, the Nanlin farm, the Nanqiao town of Wangning city, and the
Guangpo town of Lingshui county, then ends in the intersection of Lingbao road in
the north of Lingshui Town, which total length is 110.583 kilometers. In accordance
with the standard of secondary classification road, the width of road-bed is 8.5~12m
and the lanes width is 7.5-12m, it is cement concrete pavement. There will construct 3
new-build large bridges, 1 medium-scaled bridge and 8 small bridges in the route.
This project is made up of 7 construction contract sections and 2 supervision
contract sections.
Agent construction management effect

The project practices agent construction system. According to the agent construction
management mode in this paper, it sets up a market access conditions, implements the
invite public bidding, and selects the project management company, that is, the China
Branch of the consultative (Beijing) Technology Co., Ltd. which affiliates with the
China Highway Engineering Consulting Group Co, Ltd. transports construction
project management. In counseling, the staff of the company set up by the project
construction headquarters is already in place and has carried out their work at the site.
The key for the effect of construction projects lies in the units’ management level
and coordination capacity. According to the units’ construction experience and
management level, in strict compliance with the relevant provisions of the contract, in
accordance with the work of the current situation, the project will be built and the
results are as follows:
(1) The agent construction system makes the information communication transfer
fast among the project management departments, supervision units and construction
units and makes the construction projects be carried out in an orderly manner.
(2) As the project management company undergoes market competition, its staffs
hold a wealth of experience and excellent technology, this can further strengthen the
management power, make full use of resources, greatly improve work efficiency.
(3) The detailed management work about the project is managed by the specialized
and market-oriented project Management Companies when adopted agent construction
system. This can reduce management cost and enhance the management benefit.
Obviously, it enables to save the highway project management fees of Qionghai to
Lingshui section of Line Yudong.
(4) For both the management and implementation, the cost is reduced, and the
benefit is raised.
(5) Agent construction system achieves “investment, construction, management,
using” separation, forms the advantageous aspect for the departments each other to
control engineering investment strictly.
(6) The effect of engineering construction achieves the anticipated target, the project
quality is good. Project investment is controlled in the budget and the project is
completed on time.
CONCLUSIONS
The agent construction system is an important action and the new attempt about
engineering project management system in China. Agent construction system
management applied in Hainan province government investment highway projects can
effectively regulate the government acts, hold back the corruption, stimulate the
market competition and use the fund in reason. Simultaneously it offers the chance for
the agent construction company to show their management experience, to guarantee
the project quality, to control the engineering schedule and investment and to increase
the engineering construction benefit. For investing the road project management of the
Hainan province government, the agent construction system management mode
suiting the local characteristic still have some faultiness and limitation, it needs further
study and application combining with Hainan province engineering project to
improve.
REFERENCES
Hu Yu, Yan Jingfu. (2003). New Model of Project Management for Government
Investment - Entrust to a Company to Build and Manage System.JOURNAL OF
BEIJING INSTITUTE OF CIVIL ENGINEERING AND ARCHITECTURE. Vol.
19 (3): 87-89.
JIA Hu (2006).The applica tion and research of con struction agen t system in
projectmanagement.JOURNAL OF CHONGQING JIAOTONG UNIVERSITY.
Vol. 25 Supplement. 122-124
DENG Zhong-mei. (2006). Construction-agent system project management pattern
study based on principle-agent theory.JOURNAL OF CHONGQING JIAOTONG
UNIVERSITY. Vol.25 (2):128-131.
Economical Risk Analysis Model and Intelligent Optimization Methods Study of
the Soil Slope based on the Reliability
Yang Shi-jiao, Zeng Sheng, Yang Hai-fei
School of Nuclear Resources and Safety Engineering, University of South China, Hengyang, China,
421001
ABSTRACT: The study of modeling and optimization methods of the highway soil
slope stability has been carried out based on the reliability theory in this paper,
because of the extensive uncertainties of the cost and safety factors in highway slope
design. First, considering the overall construction cost and risk cost, using the
cost-benefit analysis method, taking the minimum total expected cost as the objective
function, using the slope-type and slope-ratio as design variables and the reliability
index of the slope stability as the main constraint conditions, and then the
economical risk analysis model of the soil slope is established. Considering high
nonlinearity of the model, genetic algorithm is used to iteratively solve and the
optimum program in slope engineering design is developed based on the platform of
MATLAB. The program was used to design slope engineering in some highway and
the optimum slope-type and slope-ratios were obtained. The result shows that the
program has good stability, high computational speed, high precision and practical
value.
INTRODUCTION
When the stable analysis and engineering treatment of the slope are performed,
the physical and mechanical parameters and the load of the rock soil mass generally
present a probabilistic change. The slope probability analysis is often used to reflect
the factors’ influential effect on the slope stability. At present, there is no question of
the design specifications to form the basis of the reliability analysis, including how to
use the results of the reliability analysis, and how to evaluate the slope stability using
different treatment methods. The slope analysis tends to focus mainly on the
uncertainty of the soil mass and its’ quantification, so as to give the slope failure
probability or reliability indicators, but the economic risk assessment of the slope are
often not seriously addressed. The economic risk level and failure probability of the
slope are two different concepts. The failure probability is a pure technical indicator
and economic risk level has the meaning of economic loss. Risk can not be separated
from the economy, so it is a technical and economic comprehensive indicator after
the judgment and assumption. Therefore, the author determines the acceptable
reliability or failure probability based on the principle of minimum investment,
studies the economic risk analysis of the slope, and then judges the slope stability. As
a result, the optimal treatment of the project is obtained thereby overcoming the
situation of non-standard-based methods of analyzing the reliability of slope
engineering. The- method can also make the slope safety issues and economic issues
much better.
159
STUDY OF THE ECONOMICAL RISK ANALYSIS OF SOIL SLOPES
MODEL ESABLISHMENT
A design is determined by some variables. Design variables which are adjusted
accordingly include slope pattern, slope height, slope angle, and platform width.
These are the basic parameters in the process of slope optimization, which should be
independent parameters, and have an important impact on the settlement of the issue
or the merits of design options. The purpose of optimization is to find the optimal
combination of these parameters. The objective function is used to depict the
problems and targets to be achieved by mathematical methods, which measures the
virtues or defect degree of a design. It also acts as evaluation function or value
function, which is a function of design variables. Optimal design seeks to find the
maximum objective function, so that the engineering design can achieve the best
economic results within the scheduled function. The objective function depends on
various specific issues. Generally, the investments on slope engineering include the
initial cost, the loss, and maintenance cost due to the slope instability. To make
optimal investment project, the total cost of present value of slope must be minimum.
The loss caused by slope instability is difficult to estimate because of several losses
involved which include loss of tangible and intangible losses, loss of economic
indicators including direct and indirect losses, the loss of public utilities, personal life
and property losses, and the loss of prestige. The economic loss due to slope
instability is the most difficult to estimate. In view of this, the invisible loss
coefficient is identified based on the local economic indicators, expert advice and
specific project conditions.
Assume a steep slope has n steps (Fig.1), and every step has a height h , a slope
angle θ , a last slope angle α , step width L , and the total height of the slope is H . The
intangible loss coefficient is λ , the price of earth excavation is A yuan/m3, the slope
maintenance costs is B yuan/m2, the slope failure probability is Pf , the equation of
slope surface under natural conditions is f (x ) ˈthe total excavated volume of the
slope is V m3, and the total cost of the slope engineering is C yuan, then the
objective function of the slope can be determined as follows:
a a
C = ( ∫ f ( x)dx − V ) × A + Pf × ( ∫ f ( x)dx − V ) × A +
0 0
b a b
˄1˅
λ × Pf × (V + ∫ f ( x)dx − ∫ ydx) × A + τ × Pf × ∫ 1 + (tgα ) 2 dx × B
a 0 0
a
In the equation, (∫ f ( x)dx − V ) × A is the initial cost of the slope- from the product
0
a
of the earth excavation and the price; Pf × (∫ f ( x)dx − V ) × A is the additional cost of
0
instability, including initial construction cost, inspection and management cost, etc;
a a
λ × Pf × (V + ∫ f ( x)dx − ∫ ydx) × A is the cost of slope instability loss;
0 0
b
τ × Pf × ∫ 1 + (tgx) 2 dx × B is the maintenance cost due to the slope instability.
0
x)
f( 1
H
h
Į
Ǉ
0 a b x
FIG. 1. Graphics of the slope calculation.
The equation of the slope sliding surface representing the worst case scenario
can be written as follows:
y = ( x3 − ( x − x1 ) 2 + x2 )
2
˄2˅
x1, x2, x3 is the central abscissa, longitudinal coordinates, radius of the most
dangerous sliding surface.
The expression for V is then deduced as follows:
n2h2
V = ctgα ˄3˅
2
1
−
( x − ( x − x1 ) 2 ) 2 • (−2 x + 2 x1 )
2
tgα = y = 3
'
˄4˅
2
In this model, the slope stability and the value of the total cost are two important
indicators of the optimal design, and the purpose of optimizing is seeking the best
economic benefit under the stable slope. The model seeks the value of the slope
height and the slope angle which can make the total cost of the present value of the
slope minimal.
The failure probability Pf is selected as a measure indicator- of reliability and a
major design variable- when establishing the objective function. The reliability of the
slope engineering is not that the higher the better, because the higher reliability
incurred the more costs. Generally, the cost increases exponentially when the order
of magnitude of the reliability is improved from 99 percent to 99.9 percent. The
acceptable limit of the failure probability has not yet been standardized; engineering
experience is still the general approach that is commonly utilized. So to determine
the design reliability or the acceptable risk level is not an easy task under different
conditions. The design of slope reliability engineering is still at an initial stage and
there is lack of practical experience. This paper gives a general method of optimizing
the design of the slope. The study shows that usually higher failure probability is
acceptable. Based on the reference by Yu (1993), the acceptable limit of failure
probability for excavation slope design is Pf ≤ 0.1 . Therefore, the constraints which
ensure soil slope safety and reliability is,

0 ≤ Pf ≤ 0.1 ˄5˅
According to the results of slope stability of the Northwest Agriculture and
Forestry University of Science and Technology Department of Water Resources in
Taiwan, the maximum height of the single step-by-the-slope is,
C
Hi = ˄6˅
0.11Fs γ
Where, C is the average cohesion (kPa) of Hi slope; Fs is the stability safety
factor; γ is the average natural soil stability ˄KN/m3˅of Hi slope. Determine the
range of the slope angle, single-stage slope angle by experience, e.g.,
30 0 ≤ α ≤ 50 0 ˄7˅
45 0 ≤ θ ≤ 70 0 ˄8˅
Experience has shown that the width of platform is 1.5 or 2 m.
Based on the above variables, the objective function, constraints, the economic
risk analysis model of the optimal slope design can be established as follows,
⎧ min C = ( a f ( x )dx − V ) × A + P × ( a f ( x )dx − V ) × A +
⎪ ∫0 f ∫0
⎪ b a b
⎪λ × P f × (V + ∫a f ( x )dx − ∫0 ydx ) × A + τ × P f × ∫0 1 + (tg α ) dx × B
2
⎪
⎪ 0 ≤ P f ≤ 0 .1
⎪ ˄9˅
⎨ h ≤ 15 m
⎪30 0 ≤ α ≤ 50 0
⎪
⎪ 45 0 ≤ θ ≤ 70 0
⎪
⎪1 .5 m ≤ L ≤ 2 m
⎪
⎩
THE MODEL OPTIMIZATION BASED ON MATLAB

On the basis of the Bishop method (Zu yu. C, 2003), the most dangerous sliding
surface arc is searched by genetic algorithm using the toolbox of MATLAB, i.e., the
reliability index calculation program module of the soil slope and the optimal
investment programme module which ensures that the minimum present value are
compiled by the MATLAB program. The specific procedures are shown in Fig. 2.
EXAMPLE OF APPLICATION

INTRODUCTION OF THE PROJECT
A 360-metre-long slope was formed after the highway was excavated in the
right side. The drilling data of the slope indicates that the slope is made mainly by
high-level liquid clay, quartz sandstone strong weathered layer, sandstone strong
weathered layer and sandstone-weak weathered layer. The slope will only be initially
cut after the slope is excavated. Many times the slope had slide, one slide appeared
because of a holing through crack, the farthest level distance of the crack from the
road axis is about 46 m, and local collapse occur between the second-class and the
first-class step, the slip plane was developed along two group of joints.
Another slide appeared because of a holing through crack and slip wall surface,
the farthest level distance of the crack from the road axis is about 50 m. The sliding
face was in the quartz sandstone strong weathered layer and sandstone strong
weathered layer, the form of the sliding face was circular sliding surface.
Fig. 2. The program flow chart.
SLOPE OPTIMAL DESIGN

The natural slope surface line is fitted by the MATLAB using equation (10)
below,
f ( x) = 1.0209 + 1.7068 × x − 0.0227 × x 2 (10)
Using the optimization process of economic risk analysis, based on experts’
advices and reference (LI Dongsheng, 2006) identify risk factor λ = 10 .Considering
the maintenance and repair costs of sliding surface, T =1.0. The unit price is 8
yuan/m3, and that of the masonry slope is 127 yuan/m3. Application of the
optimization process has been obtained with the results shown in Table 1.
Table 1. The optimized results.

Grade˄°˅ Safety factor Stability Failure probability Toal cost˄Yuan˅
40 1.20 1.91 0.0281 1761.4
42 1.18 1.71 0.0456 1656.6
45 1.15 1.61 0.0537 1553.2
48 1.11 1.52 0.0643 1597.6
50 1.06 1.32 0.0934 1659.7
53 1.02 0.94 0.1736 1682.8
From Table 1, it can be seen that when the slope is 45e, the total cost is
minimal and the slope is safe. Therefore, when the slope is 45e, it is safe and
economical. The slope line before and after handling is shown in Fig. 3.
132.88
129.67
.7 5
25
1.
1 :0
125.61 1:
1.5 1.5
.7 5
1 :0
1
1:
before handling slope shape line
115.61 after handling slope shape line
1
1:
.5
1 :0
105.61
Fig. 3. Slope shape of before and after slope excavation.
CONCLUSIONS
(1) Using the genetic algorithm toolbox of MATLAB, location of the most
dangerous arc sliding surface was found to simple, rapid, and practical; the results
were accurate.
(2)The checking point method was applied to calculate the reliability of the soil
slope in the paper; this method was found to be more convenient than the others.
(3) For slope design, if the slope angle is smaller, the slope is safer, and failure
probability is less, but the earthwork will be more, thus the greater the cost of
construction, but the loss due to slope failure is smaller. By contrast, if the slope
angle is larger, the construction cost is reduced, but the failure probability will
increase with a subsequent larger price for failure. Thus it is necessarily that the best
slope angle is designed so that the total cost of the present value is minimal.
(4) The optimal model was proposed, which takes safety and reliability as the
main constraints and the total cost of present value as the minimum objective
function. This approach has the potential to carry out an optimal design of slope
stability.
REFERENCES
Srbulov, M M.( 1987,). “Limit equilibrium method with local factors of safety for
slope stability. Can.Geotech”J. 124(4):652-656
Li, K S., White W. (1987). “Rapid evaluation of the critical slip surface in slope
stability problems”J. Internationl Journal for Numerical and Methods in
Geomechanics, 11:449~473
Zuyu, C., Changming, S.(1987). “The Use of the Method of Optimization for
Minimizing Safety Factors in Slope Stability Analysis”J. Chinese Jounal of
Geotechnical Engineering. 10(4):1-13
Cuiping, L., Li, W.(2004). “The Method of Project Risk Analysis and Control”J.
Journal of Shijiazhuang of University of Economics, 27(2):179-182

Chaohong, T., Haihong, M . (2000). “The Application of the Pattern Search Method
For Slope Stability Analysis”J. Journal of South China University of Technology
(Natural Science) 28(2):42-46
Yu xue. Z .( 1993). “The reliability analysis of slope” M. Metallurgical Industry
Press.
Zu yu. C. (2003). “The stable analysis of soil slope” M. China WaterPower Press.
LI Dongsheng, ZHU Zhengwei, LIU Dongyan. (2006). “Slope Resistant Engineering
Decision-making Method Based on Risk Theory”J. Journal of Shenyang Jianzhu
University (Natural Science) 20(6):907-910
A Study of the Cost Analysis of the Porous Pavement on a Freeway
Chien-Ta Chen 1, Ching-Tsung Hung 2, Ming-Chin Yeh3, and Jyh-Dong Lin 4
1
Dept. of Civil Engineering, National Central University; [email protected]
2
Dept. of Logistics & Shipping Management, Kainan University; [email protected]
3
4
ABSTRACT: Porous pavement performances are better than traditional pavements,

but for some reasons which include the higher cost and complex construction
techniques, Porous Asphalt is not widely used in Taiwan. Therefore, the concept of
LCCA (Life Cycle Cost Analysis) is introduced in this study in order to analyze the
cost required by different kinds of porous pavements. HDM-4 related parameters
were first used to estimate the service life of porous pavement. As a result, the
different types of porous pavements and asphalt pavement thickness showed different
degrees of service life results. Type III of polymer modified asphalt concrete
(conforming to CNS 14184 requirement) with asphalt pavement thickness 2.5 cm
had the longest service life.
The results of LCCA also indicates- that type III of polymer modified asphalt
concrete 2.5 cm has the lowest cost, but is still more expensive than “Open-Graded
Asphalt Friction Course.” For the sake of cost, therefore, the traditional friction
course is still a better option.
INTRODUCTION
Porous asphalt is a form of gap-graded asphalt containing a high percentage of coarse

aggregate. Porous asphalt also contains a fiber which stabilizes the large amount of
asphalt, mineral filler such as lime or fly-ash with the main function of preventing
striping, and also contains appropriate asphalt cement or modified asphalt. For more
than 30 years since it was developed, porous asphalt’s performance has been
acknowledged by road administrators in Europe and Japan for its durability and
rutting resistance ability that are both superior to those of the traditional, dense-grade
asphalt concrete (DGAC). Considering the characteristics of Taiwan’s freeways, the
researchers of this study conducted trial-paving on the northbound sections of
132k+200 ~132k+400 on National Highway No.3 and made measurements before
and after the paving in order to determine how appropriate porous asphalt is for
Taiwan’s road conditions. This study used the concept of “life cycle” to discuss the
cost of -maintenance of porous asphalt and compare it to that of open-graded asphalt
concrete in order to determine the cost differences between different types of porous
asphalt concrete.
166
Life Cycle Cost Analysis

Life Cycle Cost Analysis (LCCA) is a concept developed by the U.S. government in
the 1960s for increasing the cost-effectiveness in the purchases they make, and it was
gradually applied in the domains of business, product development, and accounting.
Even though the focus of LCCA is to follow-up on maintenance, the International
Electro-technical Commission (IEC) established an LCCA standard (IEC 60300) for
reliability management and LCCA procedures.
In the “Last Policy Phase” that was published by Federal Highway
Administration (FHWA) in 1996, LCCA was given a definition. In 1998, FHWA
applied the concept of LCCA on pavement design, and their personnel even held
seminars in over 40 states to explain and provide consultations regarding pavement
design. The FHWA policy points out that LCCA is a decision-making tool, thus
FHWA encourages the application of LCCA when making any investment-related
decision. Khaled (2000) proposed that when conducting optimized pavement design,
one must deal with the performance and cost of its life cycle. One life cycle is known
as the “life cycle durability,” which is the cost-to-performance ratio. An optimized
plan is to minimize the numerical value of life cycle durability, and the evaluation
conducted based on such model converts the potential “maintenance” and
“improvement” activities into two choices of policies. The first policy requires a
fixed analysis period, and the second one requires a flexible analysis period; both
periods of time are the multiplied value of improvement cycles. The cost of
pavement life cycle includes initial construction, main improvement cycle, routine
maintenance, and additional user cost. Pavement life cycle performance is defined as
the area under the pavement cycle performance curve.
Papagiannakis and Delwar (2001) proposed a pavement LCCA model that
simultaneously analyzes user cost and maintenance cost. Maintenance costs include-
the cost of improving, resurfacing, or rethinking different pavement methods and
pavement structural damage. User costs include- the cost regarding vehicle operation
or non-vehicle operation such as roughness and user delay cost. User benefit is
calculated based on the user cost caused by the different levels of roughness between
existing and new pavements. The user costs for different improvement plans in the
network level all need to be calculated and included in the project level. Developed
by Papagiannakis et al. (2001), the PID (Pavement Investment Decision) program
allows customization of pavement management databases based on actual traffic
conditions in each state.
Analysis of Porous Pavement Service Life with HDM-4

In 1993, International Study of Highway Development & Management (ISOHDM)
started expanding the scope of the original HDM-III and established a set of road
management system, known as the Highway Development and Management Tool
(HDM-4) through a variety of easy-to-use software. HDM-4’s capacity goes beyond
the traditional evaluation methods and is a powerful system for road management
and investment decision analysis. The key tasks are to organize and apply existing
knowledge rather than to conduct new studies, and to create new development
methods that help apply the latest knowledge to solve different countries’ technical
and managerial issues. Furthermore, HDM-4 has more comprehensive models and
obviously has better methods for improved road maintenances budgets, as well as a
“maximum economic interest analysis model” that was developed based on the
experience of road investment in more than 100 countries. HDM-4 allows road
investors to conduct more comprehensive analysis during the different weathers and
conditions (World Bank, 2000).
RESEARCH HYPOTHESIS
Relevant information must be established before conducting an analysis through

HDM-4, and the vast amount of complex data calls for data input via the computer or
an external database. The database established prior to HDM-4 analysis includes (1)
Network data, (2) traffic data, (3) vehicle data, and (4) climate data.
The study conducted in HDM-4 that was applied on national highway
maintenance management was based on the data provided by the rigid pavement
section of the Northern Region Engineering Office. The porous pavement examined
in this study is located in this area, thus data such as traffic, vehicles, and climate
came from this region. Through these data, relevant parameters and measurements
need to be provided for network data; therefore, the measurements made in June of
2006 are used to indicate the conditions of each road section. The input parameters
for each pavement are shown in Table 1.
Table 1. Input of Pavement Data.

Type of AC Deflection Structural Roughness Striping Potholes Rutting
Value of Strength (IRI m/km) Area (quantity/km) (mm)
Benkelman (SPN) (и)
Beam
(mm)
Factory-mixed
2.5cm
5.6 1.08 2.33 20 0 5.55
Pre-mixed
2.5cm
5.6 1.08 2.46 0 10 5.56
Mineral fiber
5.7 1.07 2.43 0 10 5.71
3cm
Lignin fiber
3cm
4.7 1.21 2.72 0 10 4.71
Lignin fiber
2.5cm
7.4 1.03 3.58 0 0 7.36
Pre-mixed 4cm 6.6 1.03 2.87 15 5 6.56
Type ҉
modified asphalt 3.0 1.60 2.98 0 0 3.20
2.5cm
Factory-mixed
4cm
9.9 1.03 3.74 15 0 9.89
Service Life of Porous Pavement

The analysis time in HDM-4 is 10 years, during which no maintenance efforts are
taken in order to analyze pavement damage. The results are shown in Table 2. The
results of the analysis indicate- that “Roughness” is the main standard for pavement
maintenance. The cracking analysis show that the “cracking model” developed has
accuracy issues when making long-term estimations. This is because the coefficient
of adjustment was -based on the Roughness within two years after the construction
was completed, and the forecast accuracy was, therefore, compromised when the
Roughness was greater. The LCCA in this study is based on the average service years
as the service life of porous pavement.
Table 2. Service Years of Porous Pavement - Cracking & the Maintenance

Threshold.
Roughness Stripping Potholes Rutting Average Minimum
Area (Quantity/km) (mm) Service Service
(IRI m/km) Life Life
(и)
Maintenance
5 30 30 10
Standard
Factory-mixed
2008(4) 2007(3) 2008(4) 2010(6) 4.25 3
2.5cm
Pre-mixed
2.5cm
2007(3) 2014(10) 2007(3) 2010(6) 5.5 3
Mineral fiber
3cm
2008(4) 2014(10) 2007(3) 2010(6) 5.75 3
Lignin fiber
3cm
2007(3) 2013(9) 2007(3) 2007(3) 4.5 3
Lignin fiber
2.5cm
2007(3) 2014(10) 2015(11) 2008(4) 7 3
Pre-mixed 4cm 2007(3) 2007(3) 2008(4) 2007(3) 3.25 3
Type ҉
modified 2007(3) 2014(10) 2015(11) 2014(10) 8.5 3
asphalt 2.5cm
Factory-mixed
4cm
2007(3) 2007(3) 2008(4) 2009(5) 3.75 3
LCCA of Porous Pavement

Since FHWA developed LCCA for pavement design, many types of LCCA software
have been introduced. For example, Highway Economic Requirement System
(HERS) is a cost-effective analysis system developed by EHWA, and it is used to
conduct economic analysis of highway renewal, modification, and expansion.
HDM-4 can also be used for conducting LCCA, but the relevant parameters in the
work areas are needed, and each section needs to come up with two or more types of
maintenance strategies before LCCA could take place. The purpose of the analysis is
to maximize the effectiveness of the maintenance of the entire network during the
entire life cycle, and this is different from other types of LCCA software that only
compare the life cycle cost of different plans.
In this study, LCCA 2.2 (FHWA, 2004) is utilized, which is developed by

FHWA for reports made in 2002 or earlier. Its main function is to analyze the cost
differences of lifecycles between different plans. Created with Visual Basic
Application that is based on Microsoft Excel, this program is user-friendly for
different management levels.
Parameter Input and Hypothetical Conditions

1. Analysis Cycle and Discounted Rate
LCCA researchers from other countries believe that the analyzed lifecycle
should go beyond the original pavement service year, and usually a length between
35 and 40 years is recommended. Since Real Cost only allows 6 improvement tasks
within an analysis cycle, a length of 20 years is chosen in this study. The
recommended value of discounted rate is between 3% and 15%, and most researchers
choose 5%.
2. Traffic Flow and Vehicle-per-hour Distribution
The data of traffic flow requires the input of AADT, percentage of large trucks,
percentage of semi trucks, traffic flow growth rate, and speed limits.
3. User Time Value
Since there are few studies in Taiwan on users’ time effect, the findings of user
time value by the Department of Transportation of Washington are used in this study.
The time value for a passenger car is US $13/hr, US $21/hr for a truck, and
US $26/hr for a semi truck.
4. Additional Time and Cost due to Constructions and Traffic Congestions
Since the traffic flow changes caused by road constructions are not considered
in this study and only changes of lifecycle cost of different types of porous
pavements are analyzed, the default value was chosen.
LCCA requires the data on the initial construction cost and follow-up
maintenance strategies. This study is a simplified analysis, thus it is assumed that
when the service life of a porous pavement increases, the improvement is done based
on the same thickness and composite. It is also assumed that there are no serious
damages to the road within the service year, thus preventative maintenance is not
required. Furthermore, in order to have the same basis for analysis, the road sections
for analysis is all 1 km long without constructions.
RESULTS OF POROUS PAVEMENT LIFECYCLE COST ANALYSIS
Table 3 shows that without considering user cost, maintenance cost is determined
by the pavement’s service year. A longer service year means less number of
improvements required, resulting in lower maintenance cost. Porous pavement of
type III modified asphalt mastic asphalt costs less and performs well in many aspects,
thus its maintenance cost during the analysis period was the lowest. However, its cost
was US $13,569/km, which is three times as much as that of the traditional
open-graded friction course, and this is because the construction cost for the latter is
lower. Looking at cost-effectiveness, however, the average life of friction course is
only 2.3 years, which is much worse than that of porous pavement. This finding
matches with that of a study on pavements by the U.S. Department of Energy (Ting
et al., 2001), which also indicates that the lifecycle cost of porous pavement is higher
than that of traditional asphalt concrete, but the former offers a longer service life
and less maintenance efforts. The sensitivity analysis of discounted rate shows that
the higher the discounted rate, the lower the lifecycle cost. The cost drops more if the
cost of the porous pavement is higher. At a discounted rate of 15%, from the
construction cost view point, there is no significantly difference between type III of
polymer modified asphalt 2.5cm and lignin fiber 2.5 cm.
Table 3. Construction Cost of Porous Pavement.

Construction Cost Number of Net Present Value
(USD/km) Improvements of Maintenance
within the Cost within the
Analysis Cycle Analysis Cycle
(USD/km)
Factory-mixed 2.5cm 63,212 4.7 222,158
Pre-mixed
2.5cm
45,550 3.6 113,831
Mineral fiber 3cm 36,300 3.5 89,684
Lignin fiber 3cm 32,505 4.4 114,238
Lignin fiber 2.5cm 27,088 2.9 58,561
Pre-mixed 4cm 72,880 6.2 333,965
Type ҉ modified asphalt 2.5cm 21,650 2.4 421,41
Factory-mixed 4cm 101,140 5.3 355,455
SUMMARY AND RECOMMENDATIONS
In summary, analyzing the cost of different types of porous pavements through

LCCA requires taking many aspects into consideration which include the initial cost
as well as the follow-up maintenance. LCCA indicates that the life cycle cost of type
III of polymer modified asphalt concrete 2.5cm is the lowest, but it is still three times
as much as that of the traditional open-graded asphalt friction course. This study has
not considered the user cost influenced by the construction areas. Since open-graded
friction course has a shorter service life, the cost incurred in re-constructions should
be higher than that of the porous pavement, bringing the life cycle cost closer to each
other.
REFERENCES
Khaled, A. A. (2002). “Optimum Flexible Pavement Life-Cycle Analysis Model,”

Journal of Transportation Engineering ,ASCE.
Papagiannakis, T. and Delwar, M. (2001), “Computer model for life-cycle cost
analysis of roadway pavement,” Journal of Computing in Civil
Engineering ,ASCE.
The World Bank (2000),The Highway Development and Management SeriesΔVolume
one: Overview of HDM-4.
FHWA (2004). Life-Cycle Cost Analysis RealCost User Manual
Ting, M., Koomey, J., and Pomerantz, M. (2001). “Preliminary evaluation of the
lifecycle costs and market barriers of reflective pavements,” LBNL-45864, U. S.
Department of Energy.
A Correlation Study of the Existing Bridges for Failure Analysis-
Case Study of Taichung County
Tung-Tsan Chen1, Yao T. Hsu2, Chun-Yuan Wang3

1
Department of Construction Engineering, National Kinmen Institute of Technology, No.1 University
Rd., Kinmen County, 892, Taiwan, Email: [email protected]
2
Department of Transportation Technology and Management, Fung Chia University, Taiwan, R.O.C.,
3
Department of Business Administration, Nanya Institute of Technology, Taiwan, R.O.C., Email:
[email protected]
ABSTRACT: In this study, a total of 935 data was used to analyze the correlation
factors and positions of the common bridge damage through logistic regression.
There are three aspects of the factor of bridge damage which include essential bridge
character, external bridge condition, and internal bridge structure. Based on the
whole bridge condition index, in terms of different damage positions, we performed
the cross analysis which considered the different materials and bridge types. We
constructed the logistic regression models and found all the factors which influence
bridge damages in Taichung County, Taiwan.
INTRODUCTION
Effective utilization of the bridge structure and its durable service life are
closely linked with quality of designing, construction, and materials used. Bridges
may have some defects because of design, traffic, material, and structure. Various
factors of damaging effects on the bridge are closely related to the environmental
exposure of the bridge.
To find the damage of the bridge structure on time, one must first asses and
evaluate the bridge condition. The bridge structure is measured in order to guarantee
that the bridge is being used normally, and for maintaining, or mending the bridge
such that we can grasp the situation of the bridge structure and security state at any
time, and we can also assess the real bearing capacity in the existing bridge correctly.
This procedure is needed so as to lengthen the life cycle of the bridge. We probe into
the damaging situation of the structure in the life of the bridge if some level of bridge
damage has already occurred. Thus, we could provide proper maintenance decisions,
173
and set up the perfect management criterion of the bridge to lengthen its life-span.
DAMAGING REASONS IN THE STRUCTURE OF THE BRIDGE AND

CONDITION INDICES
Damaging reason
Two kinds of damaging factors are natural and human related. The natural
factors include earthquakes, typhoons, floods, corrosion due to the natural
environment, etc., and the human factors include design mistake, bad construction,
vehicle overloads, etc.
The above-mentioned factors can further be divided into three kinds: design
factors, construction factors, and external factorsΰLin, 2004; Miyamoto, et al.,
2001α.
Condition index
We divide the degradation situation of the bridge into degradation severity
(Degree), degradation range (Extent), the security of the bridge structure (Relevancy),
and the urgency of component maintenance (Urgency). We denote them as D.E.R.U.
They have some characteristics such as: 1) they can be simplified and measured, 2)
especially consider the impact on the importance of the bridge degradation, 3)
reduce and fill the materials, 4) put forward change, 5) maintenance urgency, propose
the law for maintenance and estimate the budgetary funds, and 6) determine the
optimum maintenance strategyΰLi et al., 2002; Zhan, 1999α.
Every component condition of the bridge
According to D.E.R.U., we assess them with bridge component state indicator,
Icij, of which the calculation formula is as followsΰLin and Li, 2000:
Icij = 100 − 100

(D + E )* R a …………………………….……………. ( 1)
(4 + 4)* 4 a
,where a represents the relevant importance parameter.
Whole condition indicator of the bridge

To understand the state of the whole structure of the bridge, we calculate the
whole state indicator CI of the bridge, called Condition Index ΰLin and Li, 2000α
as follows:
20
∑ Ic
i =1
i * wi
CI = 20
……….………………………………….………………. ( 2)
∑ wi
i =1
,where Ici is the state value of component i ( item condition) , wi is a weight of

component i to the bridge, n is total number of components of the bridge. ICi can be
calculated via the following equation:
n
∑ Ic
j =1
ij
IC i = ………….………………………………………….…. ( 3)
n
The higher the whole state indicator value, CI, of the bridge is, the better the
whole state of the bridge is. The lower the value is, the worse the whole state of the
bridge is. CI is a demarcation line value, representing the standard threshold setup by
the user based on rationality, historical materials, structure characteristic, project
practice and other specific purposes. Once comprehensive indicator values of some
of the bridge’s structures are lower than this point, we must measure, mend, and
maintain the bridges prior to total failure.
RESEARCH APPROACH
The key environmental factors that damage the bridge structure are generally
divided into earthquakes, erosion, and corrosion. Bridge materials are divided into
two parts: one part is the basic materials, which include the management of materials,
geometry of materials, structure of materials, and design of materials. The other part
is the measuring of materials, which include the management of materials, abutment
of materials, pier materials, and opening materials.
Our research on the bridge structure damage is divided into 2 stages. Firstly, we
analyzed the damaging factor. Secondly, we analyzed the damaging position.
Damaging factors are divided into three aspects as shown in Fig.1.
Fig. 1. Analysis of the damaging factors of the bridge.
Another classification of the bridge damaging positions which is divided into

upper-structure, under-structure, and the interface is shown in Fig. 2.
Fig. 2. Analysis of the bridge damaging positions.
This research collected 1268 materials altogether, but the effective materials
utilized were only 935. We judge, according to the student's /-value test whether the
parameter of choosing damaging factors is significant or not. A t-value of 1.96 was
used as a threshold in this study. That is an absolute t-value greater than 1.96 means
that the factor has a significant damaging impact on the bridge damage and vice
versa if the absolute t-value is less than 1.96.
Damage model in the whole bridge

Based on our study, the fundamental characteristics of bridges included bridge
age, total length of the bridge, number of lanes, and the longest span. External
environmental materials of the bridge included the situation of crossing the river,
distance from the sea, and the condition of the river under the bridge. The forms of
the bridge construction included structure form, girder form, pier form, abutment
form, pavement material, and aseismatic facilities. Table 1 shows that there are 8
parameters with the absolute t -value larger than 1.96. We find that the bridge’s age
of more than 15 years, the farthest district (or location) in terms of the sea distance,
and aseismatic facilities have positive effect upon the bridge damage. On the other
hand, we find that the bridge’s age less than 10 years, total lane more than 4, and
beamed girder have negative effect upon the bridge damage.
Table 1. Damage assessment of the best model in the whole bridge.

Variable' s win value
Significant variable (95%Confidence interval) t-value
coefficient ratio
The bridge' s age less than 10 years -0.466 -2.39 0.63
The bridge' s age more than 15 years 0.612 2.20 1.84
Total lane more than 4 lanes -0.654 -2.10 0.52
From the farthest district of the sea distance 0.577 3.73 1.78
T roof beam (Girder model) -0.680 -4.55 0.51
I roof beam (Girder model) -1.085 -4.61 0.34
Aseismatic facilities 0.815 2.49 2.26
Coefficient of constant 0.602
Log Likelihood Ratio -591.479
Damage model of the severity degree in the bridge

Table 2 shows that there are 6 parameters with the absolute student’s t-value
larger than 1.96. We find that the farthest location in terms of the sea distance, poor
river function, and the gravity type abutment have significantly positive influence
upon the bridge damage. On the other hand, we find that the plate of the bridge and
the asphalt concrete pavement could significantly reduce the bridge damage.
Table 2. Damage the assessment of the best model of the severity in the bridge.
coefficient ratio
From the farthest district of the sea distance 1.019 3.02 2.77
River function is relatively poor under the bridge 2.469 5.23 11.80
Plate bridge (Structure model) -1.927 -4.18 0.15
Gravity type abutment 1.835 5.56 6.27
Asphalt concrete pavement -0.935 -2.50 0.39
Coefficient of constant -2.389
Damage model in the upper-structure

The upper-structure includes structure form, girder form, pier form, abutment
form, and the pavement material. Other damage positions of the bridge include
damages of under-structure, interface and general function of the whole bridge. Table
3 shows that there are 4 parameters with the absolute t-value larger than 1.96. We
find that the damaging upper-structure and T roof beam have positive effect on the
bridge damage. On the other hand, we find that the half- gravity type abutment could
significantly reduce the bridge damage.
Table 3. Damage the assessment of the best model in the upper-structure.

coefficient ratio
Under-structure is damaged 1.789 5.07 5.98
T roof beam (Girder model) 0.824 3.37 2.28
Half-gravity type abutment -0.567 -2.31 0.57
Damage model in the under-structure

Table 4 shows that there are 8 parameters with the absolute t-value larger than
1.96. We find that the damaged upper-structure, the damaged interface, the damaged
general function, wall type pier, and pillar type pier have positive effect on the bridge
damage. In contrast, we find that the half-gravity type abutment and asphalt concrete
pavement have a negative effect upon the bridge damage.
Table 4. Damage assessment of the best model in the under-structure.

t-value
Significant variable (95%Confidence interval) coefficient ratio
Upper-structure is damaged 1.975 4.90 7.21
The interface is damaged 1.453 2.09 4.28
The general function of the whole bridge is damaged 0.973 2.43 2.65
Half-gravity type abutment -0.733 -2.03 0.48
Wall type pier 2.960 6.63 19.30
Pillar type pier 2.710 5.51 15.01
Asphalt concrete pavement -1.404 -3.35 0.25
CONCLUSIONS
From this study, we have found that the following 7 factors can significantly
reduce damage to an existing bridge based on the t-value tests: 1) I roof beam; 2) T
roof beam; 3) plate beam; 4) asphalt concrete pavement; 5) the bridge’s age less than
10 years; 6) half-gravity type abutment; 7) total lane more than 4.
Conversely, we also find that the following 12 factors could significantly
increase the existing bridge’s damages: 1) wall type pier; 2) gravity type abutment; 3)
pillar type pier; 4) bad river’s function; 5) damaged under-structure; 6) damaged
upper-structure; 7) district of the sea distance; 8) T roof beam; 9) aseismatic facilities;
10) damaged general function; 11) damaged interface; 12) the bridge’s age (more
than 15 years).
Conclusively, these results show that we should pay attention to the interacting
effect amongst the damaging influencing factors. Further research could focus on
temperature, humidity, traffic flow and using frequencies of the bridge.
REFERENCES
Li, Y.F., Xie ,S.X., Wang ,L.C., Chen, Q.Q., Zhan, Q.Z., Zheng, Y. X. and Chang, F.C.
(2002). “ Set up bridge measure system the methods and the research highway
and road bridges of criterion” Institute of Trans., Ministry of Trans. and
Communications.
Lin, S.Z. (2004). “The bridge is checked and maintained” Chinese tech., 21(1): 3-21.
Lin A.Y. and Li Y.F., (2000). "The bridge is measured and assesses and mends
strongly", Chuan Hwa Book Company, Ltd., Taiwan, Taipei.
Miyamoto, A., Kawamura, K. and Nakamura, H., (2001). "Development of a bridge
management system for existing bridges? " Advances in Engineering Software,
Vol.32, pp821-833, 2001.
Zhan, Q.Z. (1999). “A study on the different management levels of bridge Manage.
Systems.” Master thesis of Department of Civil Engineering, National Taipei
University of Technology.
Application of Air Foam Stabilized Soil for
Bridge-Embankment Transition Zone in Thailand
Pichit Jamnongpipatkul 1, Montri Dechasakulsom2, and Jiraroth Sukolrat3

1
Chief Engineer of Research and Development, Department of Highways, Bangkok Thailand
2
Deputy Director, Bureau of Road Research and Development, Department of Highways, Bangkok,
Thailand
2
Engineer, Bureau of Road Research and Development, Department of Highways, Bangkok, Thailand
ABSTRACT: Large settlement is a problem of most highway bridge approaches

constructed in the lower Choa Phraya river basin of the Kingdom of Thailand. The
presence of thick and highly compressible soft clay layer in these areas usually
results in differential settlement between bridge superstructures and highway
embankments. Moreover, it also affects driving comfort of the traveling public.
There are many alternatives that can be used to prevent the vertical movement, and
use of lightweight material is one of the alternatives. Recently, Thailand’s
department of highways (DOH) has initiated a project for use of light weight soil to
reduce settlement under bridge approaches. The test section was located on highway
number 35, at the kilometer post of 72+712.5 to 72+775.0. Geotechnical
instrumentation was also installed at the tested site for material behavior study.
INTRODUCTION
Road and bridge construction on soft clay usually encounters problems of short
term instability and long term excessive settlement. Typically, bridge structures and
pavement are built on different types of foundations. A bridge is usually supported
on either shallow or deep foundation. In contrast, pavement is constructed on
compacted base material overlying the subgrade. Therefore, load generated on
different types of supports usually causes differential settlement especially at a
bridge approach, a joint connection between the bridge and the embankment. The
differential settlement at the bridge approach can reduce driving comfort and perhaps
it may cause injury and deadly accident if the vehicle loses control.
181
Recently, Thailand’s Department of Highways (DOH) has initiated a program to

improve the methods of design, construction and correction of the bridge approach
using air foam stabilized soil to replace conventional embankment material under the
bridge approach. The air foam stabilized material will reduce the weight applied to
soft clay foundation thus reducing differential settlement. The project was a
collaboration of DOH and Public Work Research Institute of Japan (PWRI) who
gave supports in terms of providing the air foam mixing plant and technical
expertise. Results of the study were expected to reduce differential settlement, thus
reduction of the risk of the traveling public and cost in maintenance. The use of air
foam stabilized soil also promotes effective use of local material and saves the
environment degradation during highway construction.
This paper focuses on the air foam stabilized embankment constructed at the tested
site on the bridge approach underlain by soft Bangkok clay. The discussion of
manufacturing and construction of the air foam stabilized embankment is also
presented in details
PROPERTIES OF THE BANGKOK CLAY
Bangkok clay contains a large amount of water, resulting in very low shear strength
and high compressibility. Normally, Bangkok clay is found in the central plain of
Thailand especially in Bangkok city and its vicinity. The stratum varies from 4-5 m
in thickness to more than 20 m in depth. The much thicker Bangkok clay layer is
usually found close to the shoreline and underneath roads locating on the riverside.
Figure 1 shows typical cross-section of marine clay deposited in the Choa Phraya
river basin.
FIG. 1. Typical cross-section of Marine clay deposited in Choa Phraya river

basin in the direction from North to South (Cox, 1968).
Based on previous laboratory test results e.g., Balasubramanium et al. (1985),

Bangkok clay is normally consolidated clay or slightly consolidated clay depending
upon where the clay is found. The presence of Bangkok clay in construction sites
affects project operation cost which include design, planning, construction, and
maintenance costs. Also, the structural behavior in terms of short and long term
performance was considered very carefully.
MIXING AND TESTING OF AIR FOAM STABILIZED SOIL
Air foam stabilized soil is a light weight material having high compressive strength
and low permeability. The strength of the air foam stabilized soil depends upon the
amount of mixed Portland cement. The standard properties of air foam stabilized soil
are shown in Table 1 together with the properties of Bangkok clay taken from the site
without stabilization.
Table 1. Typical properties of airfoam soil and Bangkok clay from the site.
Properties Air Form Bangkok Clay

Stabilized Soil
Density (t/m3) 0.60-1.20 1.56
Water Content (%) - 82
Liquid Limit - 96
Plasticity Index - 55
Unconfined Compressive Strength (ksc) 1-10 -
Flow (cm) 16-20 -
Permeability (cm/s) 1¯10-5-1¯10-6 -
The process of manufacturing air foam stabilized soil can be described as follows:
Firstly, soft clay is mixed with water until it becomes slurry, then it is pumped into a
mixing chamber to mix with Portland cement. The next step is to mix the product
obtained from the first step with the prepared air foam. The air foam can be prepared
by mixing chemical foaming agents (Polyfoam-200A which is non-toxic organic
polymer) with water, and is fed through a nozzle under high pressure. The density of
the air foam stabilized soil depends on the amount of added air foam. For the purpose
of the embankment construction, the prepared material can be pumped to the
construction site. After the on-site air foam stabilized soil installation process is
completed, curing process starts. The shear strength of the air foam stabilized soil
increases with curing time.
The advantages of utilizing air foam stabilized soil to reduce the weight of
embankment are; (i) it is a light weight material, (ii) its flowing ability, which allows
it to be pumped through pipes, and (iii) no compaction is needed with this material.
In the past, an air foam stabilized soil of 7 m in height, and 22 m by 22 m in length
and width, respectively, was constructed at a pilot test site in Japan.
After a year of construction, the vertical settlement was reported to be 29 mm,

which was minimal when compared to the settlement of the conventional
embankment (approximately 125 mm) (Miki et al., 2003). Recently, in Thailand, an
air foam stabilized embankment of 4 m in height and 14 m by 14 m in length and
width, respectively, with 1:1 slope was constructed at the Ladkrabang interchange of
Bangkok-Chonburi motorway. The study found that the settlement of the light
weight embankment was reduced dramatically compared to conventional
embankment. The settlement at the center of the embankment of 0.06 m was
recorded after 49 days of construction compared to that of 0.1 m taken from the
analysis of conventional embankment (Dechasakulsom, 2006).
Optimal mixtures of air foam stabilized soil
The mix design of the air foam stabilized soil was referenced and adapted from a
technology manual of the foam mixing stabilized method by the High Grade Soil
Research Consortium (2005). The unit weight of the air foam stabilized mix was set
at 1.0 t/m3. The flowing ability required was 180 mm. Also, the unconfined
compressive strength of the sample at 28 days was 6 ksc, about 1.5 times of the
required strength of 4 ksc, equivalent to CBR of 10%. According to the preliminary
setup value, the different content of Portland cement in the mixture was tested. The
amounts of Portland cement used in four different mixtures were 150, 175, 200 and
250 kg/m3, respectively.
Preparation of air foam stabilized soil
The air foam stabilized soil consists of three components including soil slurry,
Portland cement, and air foam. Slurry can be prepared by adding water to about three
times the liquid limit in order to increase its flowability to approximately 450 mm.
The prepared slurry was kept in a sealed container for moisture loss protection.
Portland cement then mixed with the slurry so as to gain strength of the air foam
stabilized soil. During the same time, air foam was prepared using a foaming agent
mixed with water. The optimal mixture ratio between foaming agent and water was
1:19 and they were mixed thoroughly in air foam mixing device with careful control
of the water and air pressure in order to get homogeneous air foam mixture. The unit
weight of air foam was controlled at 0.05 g/cm3. The air foam stabilized soil
mixtures are shown in Figure 2.
Mixing and quality control
The mixing process started with thoroughly mixing soil slurry and Portland cement
by hand mixer, and then mixed the prepared solution with air foam. Then, the
material was thoroughly mixed by hand until a homogeneous solution was achieved.
After that, the unit weight of the air foam mixed solution and flowing ability were
checked, and the air foam mixed solution was poured into a prepared mold and
covered with a plastic sheet to prevent moisture loss from the samples. The unit
weight and flowability of different mixtures are shown in Table 2.
(a) (b)
(c) (d)
FIG. 2. Preparation of air foam stabilized soil mixture (a) soft clay, (b) soil
slurry, (c) flowing ability test, and (d) air foam.
Table 2. Unit weight and flowing ability of the samples.
Mix No. 1 2 3 4
Cement Content (kg/m3) 150 175 200 250
Moisture Content (%) 288 312 288 312
Density (t/ m3) 1.026 1.024 1.004 0.967
Flow Value 145 165 160 190
Unconfined compression tests of prepared samples
The unconfined compressive strength was tested on the prepared samples that
cured for 7 and 28 days. The sample was extruded from the mold and the size of the
tested sample was trimmed to 50 mm in diameter and 100 mm in height. Figure 3
shows the results of unconfined compressive strength of the samples of 7 and 28
days, respectively.
Following the design, the unconfined compressive strength of the air foam mixed
sample should achieve 6 ksc for 28 days of curing time. According to the results
plotted in Figure 3, the amount of Portland cement needed to achieve the strength of
6 ksc was 218 kg/m3. This proportion of cement mixing was used to mix the air foam
stabilized soil on the construction site.
Unconfined Compressive Strength

10.0
8.0
UCS (kg/cm2)
6.0
4.0
7 Days
2.0
28 Days
0.0
100 150 200 218 250 300
Cement Content (kg/m3)
FIG. 3. Relationship between unconfined compressive strength and cement

content.
APPLICATION OF AIR FOAM STABILIZED SOIL FOR SUPPORT OF

THE BRIDGE APPROACH
The application of air foam stabilized soil for settlement reduction was
implemented at a bridge approach of a bridge crossing over Khlong Donjun in
Samutsongkram province. The site was located on the inbound lane on highway no.
35 at the kilometer post from 72+712.5 to 72+775.0. The total amount of air foam
stabilized soil installed was approximated at 1,600 m3, of which 550 and 1,050 m3
was used for the first and second phases, respectively. The embankment was
constructed in 3 layers. The thickness of each layer was 0.5 meter in height.
On-site manufacture of air foam stabilized soil
Manufacturing air foam stabilized soil by using mixing plant follows the processes
below.
1. Preparation of slurry
This process uses in-situ clay, in which the sand, gravel, or organic materials are
screened. Then, the clay is mixed with water in the mixing chamber until
homogeneous slurry is achieved. The unit weight of the prepared slurry is then
checked to ensure it meets the requirements, and the prepared slurry was pumped
into a storage tank with stirring blade to prevent sedimentation.
2. Mixing slurry with Portland cement
The prepared slurry is weighted and then pumped into a mixer tank. At the same
time, Portland cement is weighted on the weighing device installed inside the
machine. The cement is then premixed with water before being pumped to the
mixing tank for mixing with the prepared slurry.
3. Preparation of air foam
During preparation of slurry, foaming agent and water are mixed according to the
mix design. The foaming solution id then pumped into a foam generating device. At
the same time, compressed air is inserted into the foaming solution to generate the air
foam mixture.
4. Mixing air foam with mixture of slurry and Portland cement
The mixture of slurry and portand cement is pumped into the air foam mixing
nozzle. The prepared air foam is also pumped into the nozzle in order to mix the air
foam stabilized soil thoroughly.
5. Casting of the prepared air foam stabilized soil
Firstly, the formwork is installed on the construction site. Then, the air foam
stabilized soil is pumped to the construction location, and it is left to cure so as to
gain strength before casting the air foam stabilized soil for the next level.
Construction of air foam stabilized soil embankment
The first phase of construction began with preparation of construction site and air
foam stabilized soil mixing plant, excavation, casting the prepared mixing solution,
curing, and sand fill and compaction. The typical drawings for construction are
shown in Figure 4.
72+712.5
Bearing Slab
ELEVATION
+ 1.60
+ 1.35
+ 1.45
+ 0.962
+ 0.586
+ 1.20
72+726.435 + 1.30
- 0.15 + 0.85
+ 0.35
+ 1.05
Sheet Pile
72+750.0
72+762.5
72+768.75
72+775.0
FIG. 4. Elevation of air foam stabilized embankment.

The amount of air foam stabilized soil used on site was measured by a flow meter
of the mixing plant. In the first phase of construction, the amount of mixing solution
used was 546.3 m3. Samples were taken for quality control of the unit weight and
flowability. Cylindrical samples were taken for unconfined compressive strength test
for both 7 and 28 curing days. During curing, water was sprayed over the surface to
protect moisture loss in the air foam stabilized soil. Compacted sand fill was then
placed on top of the harden air foam stabilized soil.
The second phase of construction of the air foam stabilized soil was made for
embankment on the original lanes. During the second phase, the effect of traffic and
stability of embankment constructed from the first phase were taken into
consideration because it had to temporarily open for the traveling public. In order to
deliver the air foam mixture to the construction site, a delivery pipe was installed
underneath the previously constructed embankment without having any effect to the
traveling public.
In addition, in the second phase of construction, previously installed sheet piles had
to be removed so that the previously and newly installed air foam mixed material
were fully contacted with the previously constructed embankment. The counter
weight berm was also placed at the toe of the previously installed air foam material
for embankment failure protection. During the construction, the volume of the air
foam mixed material installed was 1,154.75 m3. The air foam stabilized material used
each day was sampled and tested for unit weight, flowability and unconfined
compressive strength in order to control quality of the mix.
Unconfined compression test on the air foam material samples
The unconfined compression test was used for quality control of the strength of the
air foam stabilized soil at 7 and 28 curing days. The size of the sample was 50 and
100 mm in diameter and height, respectively. At 28 curing days, the unconfined
compressive strength of the sample, collected from both construction phases, were all
greater than 4 ksc, and the unit weight ranged within 1.0 ±0.07 ton/m3 as shown in
Figure 5.
AIR FOAM STABILIZED SOIL EMBANKMENT MONITORING
Geotechnical instrument was installed at the construction site for short and long
term performance of the constructed embankment. The types of geotechnical
instrumentation and installed locations are shown in Figure 6. After the construction,
all the geotechnical instrumentations were monitored so as to observe the
embankment behavior and performance. The data was obtained and recorded
periodically (every week during construction, and every month after construction);
and is shown in Figures 7-10 for inclinometers, settlement cells, and pressure cells,
respectively.
1 .4
1 .3
1 .2
1 .1
Density (t/m 3)
1 Density = 1.0 ± 7% t/m 3
0 .9
0 .8
UCS ≥ 4 ksc
0 .7 Phase 1
Phase 2
0 .6
0 2 4 6 8 10 12 14
Unconfined Compressive Strength (ksc)
FIG. 5. Unconfined compressive strength of collected samples of 1st and 2nd

stages at 28 days.
FIG. 6. Installed locations of monitoring equipments.

Cum ulative Displacem ent (m m ) Cum ulative Displacem ent (m m )

20.00 10.00 0.00 -10.00 -20.00 20.00 10.00 0.00 -10.00 -20.00
0 0
5 5
10 10
Depth (m)
Depth (m)
15 15
20 20
B+ B+
Embankment A
+
A- Embankment A
+
A-
B- B-
25 25
19-Feb-08 23-Feb-08 26-Feb-08 19-Feb-08 23-Feb-08 26-Feb-08
28-Mar-08 30-Apr-08 30-May-08 28-Mar-08 30-Apr-08 30-May-08
30-Jun-08 28-Jul-08 30-Jun-08 28-Jul-08
a) a-axis b) b-axis
FIG. 7. Inclinometer (IN-1) horizontal movement.
Cum ulative Displacem ent (m m ) Cum ulative Displacem ent (m m )

20.00 10.00 0.00 -10.00 -20.00 20.00 10.00 0.00 -10.00 -20.00
0 0
5 5
10 10
Depth (m)
Depth (m)
15 15
20 20
B+ B+
+ +
Embankment A A- Embankment A A-
B- B-
25 25
14-Dec-07 25-Jan-08 20-Feb-08 14-Dec-07 25-Jan-08 20-Feb-08
23-Feb-08 26-Feb-08 28-Mar-08 23-Feb-08 26-Feb-08 28-Mar-08
30-Apr-08 30-May-08 30-Jun-08 30-Apr-08 30-May-08 30-Jun-08
28-Jul-08 28-Jul-08
a) a-axis b) b-axis
FIG. 8. Inclinometer (IN-2) horizontal movements.

As shown in Figure 7, a vertical inclinometer, IN-1 showed horizontal movement

inwards to the excavation area. Movement in the a-axis mostly occurred after
completion of excavation. The horizontal movement was approximately 6 mm at the
depth of 4.00 m below the top of inclinometer. This movement decreased after
completion of the embankment construction. Currently, at the same location, the
movement of approximately 3 mm was recorded. In b-axis, slightly and insignificant
horizontal movement was detected.
The horizontal movement occurring in the a-axis of another vertical inclinometer,

IN-2 was higher than that of IN-1 (see Figure 8). The maximum horizontal
movement of 15 mm outwards from the embankment was recorded at the depth of
approximately 4 m from the top of the inclinometer. The movement mostly occurred
after completion of the first phase. This may be the result of the traffic load applied
close to the inclinometer location. In the b-axis, the movement of approximately 5
mm was recorded.
In order to measure the settlement of the air foam mixed soil embankment, four
settlement cells were installed underneath the embankment. The settlement was
observed during construction and after opening to traffic as shown in Figure 9. After
the excavation, the data showed swelling behavior of the soft clay, which is its nature
under unloading conditions. The swelling was overcome by settlement after the air
foam mixed soil embankment construction was completed (at the end of February
2008). After the embankment was opened to traffic, the settlement of this section was
regularly monitored. A settlement of approximately 250 mm was observed at SC-2.
The vertical movement at this location occurred mostly after completion of the
construction of the highway structure at the end of March 2008.
A total pressure cell was placed underneath the embankment to observe changes
during and after the construction processes. The data captured is plotted as shown in
Figure 10. The highest pressure was observed after completion of the embankment
construction at the beginning of June 2008. Recently, a pressure of approximately 40
kPa was recorded, which was the vertical stress due to the loading of the highway
structure.
-150
-100
SC-1 SC-2 SC-3 SC-4
-50
0
Settlement (mm)
50
100
150
200
250
20/2/2008 23/2/2008 26/2/2008 28/3/2008
300
30/4/2008 30/5/2008 30/6/2008 28/7/2008
350
7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 -1.00 -2.00
Distance from Centre of the AMS (m )
FIG. 9. Settlements under the AMS Embankment.
45.00 3
40.00
2.5
35.00 Modified Asphalt Layer
Height of Embankment (m)
Asphalt Layer
Total Pressure (kPa)
30.00 2
25.00 Sand Embankment Layer

1.5
20.00 3 rd AMS Layer
15.00 2 nd AMS Layer 1
10.00 1st AMS Layer

0.5
5.00
Total Pressure Embankment Height
0.00 0
15/2/2008 16/3/2008 15/4/2008 15/5/2008 14/6/2008 14/7/2008 13/8/2008
Date
FIG. 10. Changes in Total Pressure.
SUMMARY
Application of air foam material for settlement reduction of a bridge approach was
used on highway number 35 section number 4 at the kilometer post from 72+712.5 to
72+775.0. The material used was in-situ soft clay. In order to attain the compressive
strength of 6 ksc, the amount of Portland cement used was 218 kg/m3. The two stages
of constructions were used in order to allow traffic to flow freely.
Short and long term performances of the installed air foam soil embankment were
monitored through the installed geotechnical instrumentations. To date, the
monitored results show good performance of the constructed embankment, with very
minimal settlement recorded. In addition, the performance monitoring program is
still ongoing so as to observe the long term performance of the air foam stabilized
embankment.
ACKNOWLEDGEMENTS
The authors would like to thank the Public Work Research Institute (PWRI), the
Public Work Research Center (PWRC), and the High Grade Soil Consortium (HGS)
for their collaboration and support in terms of the provision of the air foam mixing
plant and technical experts. In addition, we are also grateful to the Bureau of Road
Construction III, Department of Highways, Thailand, for their cooperation and their
help in the construction of the tested AMS embankment. Lastly, special thanks go to
Thai Nishimatsu Construction for their great work in casting the air foam stabilized
soil on highway number 35 section number 4.
REFERENCES
Balasubramanium, A.S., Bergado, D.T., Chandra, S. and Ting, W.H. (1985). “State-
of-Art Report on Enginering Behavior of Soil in Southeast Asia:
Commemorative Volume of SEAGS”, the Netherlands AA Balkema Puplishing,
pp. 25-26
Cox, J.B. (1968). “A Review of the Engineering Characteristics of the Recent Marine
Clays in Southeast Asia”, Research Report No. 6, Asian Institute of Technology,
Bangkok.
Dechasakulsom, M. (2006). Application of air foam stabilized soil for embankment

settlement reduction, Research and development Report no. 231, Bureau of Road
Research and Development, Department of Highways, Bangkok, Thailand
(written in Thai)
High Grade Soil Research Consortium (2005). Technology Using Soil, High Grade
Soil-Foam Mixed Stabilized Soil Method (Technical Document).
Miki, H., Mori, M., and Chida, S. (2003). “Trial Embankment on soft ground using
lightweight-foam-mixed in-situ surface soil”. In Proc. XXIInd PIARC World
Road Congress, Durban
Simulation of NATM Tunneling Construction in Gravel Formation
– Lessons Learned from Pakuashan Highway Tunnel Project in Taiwan
Chih-tsang Lin1, Wen-ta Hsiao2, Tao-ming Cheng3, and Hsien-tang Wu4

1
Ph. D. Student, Department of Civil Engineering, National Chung Hsing University; Lecturer,
Department of Construction Engineering, Chaoyang University of Technology, Taichung, Taiwan.
e-mail: [email protected]
2
Lecturer, Department of Construction Engineering, Chaoyang University of Technology, Taichung,
Taiwan. e-mail: [email protected]
3
Professor, Department of Construction Engineering, Chaoyang University of Technology, Taichung,
Taiwan. e-mail: [email protected], Corresponding author
4
Assistant Professor, Department of Construction Engineering, Chaoyang University of Technology,
Taichung, Taiwan. e-mail: [email protected]
ABSTRACT: Construction processes are complex operations that include the uses of
equipment, materials, crews and geological conditions. Simulation can be used to
study these processes in order to make competent decisions that will lower costs and
shorten the duration of the project before it has started. Tunneling is a highly repetitive
construction operation and simulation can be used to analyze the process. A model
used to simulate the NATM tunneling construction operations for Pakuashan highway
tunnel project is developed in this paper. This project is unique because of the soil
conditions. A discrete event simulation methodology, CYCLONE, is used to build the
operation model and simulate the tunneling processes. Statistical analysis indicates
that duration of most tunneling operations can be modeled by Beta probability
distribution. Results also show the model predicting the advancement rate of such
tunneling project in accurate.
INTRODUCTION
Simulation can be used in the planning and scheduling of highly repetitive processes
in construction projects. The tunneling process is a good example of construction
projects that are highly repetitive. In tunneling projects the most important activity is
the actual advancement rate of the tunnel. The total costs are directly related to the
advancement rate. Simulation can be used to assess the advancement rate and find the
problems of a project before it has started (AbouRizk et al. 1999; Nido et al. 1999).
Though a few articles have been written concerning simulation and tunneling but most
of these articles discuss the tunnel boring machine (TBM) method (Touran and Asai
1987; Sinfield and Einstein 1996; Touran 1997; Chung et al. 2006) used to build
tunnel in soft rock.
194
However, the Pakuashan highway tunnel located in central part of Taiwan is unique
in soil condition and therefore, the New Austrian Tunneling Method (NATM) is the
chosen method of excavation at the Pakuashan Tunnel. It is worthy to have a model in
predicting the advancement rate when similar tunnel is built. As a result, this paper
aims at presenting a discrete event simulation model to be used to measure the
production rate before a similar project can be started.
GEOLOGICAL CONDITIONS OF PAKUASHAN TUNNEL
The expressway tunnel is passing through the Pakuashan ridge, which is an anticline
with the axis of the fold oriented in the north-south direction. The maximum height of
the ridge or terrace is about 340m. The geological formations at the Pakuashan terrace
consist of a thin layer of laterite, which is Quaternary deposit, near the ground surface
and a thick layer of gravel formation underneath. The gravel formation belongs to the
Toukoshan Formation of Pliocene Epoch, and it is composed of gravel, interlayers of
soft sandstone, mudstone, and sand lenses. Thus, the gravel formation is very
heterogeneous and is a composite stratum, which possesses the properties between
soils and rocks.
The percentage of gravel content, the grain size distribution, the cementation of the
matrix, and the groundwater situation control and affect the mechanical behavior of
the gravel formation. The percentage of gravel content is as high as 86% and the rock
size is as large as 2m for these gravel formations. The shear strength of the formation
will be dominated by the fine content if the gravel content is less than 70%. In general,
the ground condition along the tunnel ranges in character from gravel and loose
running sands to heavily compacted deposits of more cohesive material.
Groundwater can be found at two different sources. One is the perched groundwater,
which is above the impermeable layer near the surface and the other one is the
confined groundwater, which is between the folded belts of clayey layers inside the
gravel formation. The progressing rate of the tunnel was slowed drastically in
September 1998 as groundwater ingress was encountered. The ingress had a dramatic
effect on the stability of the material. Therefore, gravity or vacuum wells were
installed for dewatering around the tunneling face. Also a forepoling scheme, a jet pile
umbrella, was adopted when tunneling in faces with loose materials and groundwater
seepage to prevent overbreaks and the flowing out of materials.
The conventional rock mass classifications, such as rock mass rating (RMR),
Q-system, and rock structure rating (RSR), do not work well for the tunnels in gravel
formation. Therefore, a different ground classification is adopted and used based on
the encountered formations and in situ conditions in Taiwan. When the first subsoil
exploration was conducted, there were five types of ground conditions encountered.
However, once the construction was started, only three types of ground classifications
were used: C2, D1, and D2. Table 1 shows cross sections of the three types of ground
classifications.
TUNNELING PROCESSES
The tunneling processes are similar in each ground. However, there is no forepoling
process in C2 and D1 grounds. Hence, only tunneling process for D2 ground is
described as follows in this paper.
1. The forepoling machine is placed in position and the holes are marked. The holes
are then drilled and cleaned and the steel rods are inserted.
2. In the forepoling process, there will be 28-49 holes drilled into the tunnel face. The
forepoling will advance 8-10 sections. Each forepoling section is 0.8-1.2m long.
3. Once the steel rods are placed, holes can be grouted and the tunnel face can be dug.
4. Shotcrete is always applied to the tunnel face (see Shotcrete Sealing on the flow
chart). However, depending on the geological condition, shotcrete is added to the
sidewall only 70% of the time. The application of shotcrete is 5cm thick.
5. After the shotcrete is applied, the wire mesh is placed.
6. The steel rib cage is assembled and placed. There are different kinds of steel rib
cages used depending on the soil type.
7. After the steel rib is placed, the second application of shotcrete is applied. The
second application of shotcrete is 10-15cm thick.
8. Wire mesh is placed for the second time.
9. A final application of shotcrete is applied in order to finish the surface of the face.
10. Repeat Steps 2-9 for 4-5 sections before bolting is started.
11. After Steps 2-9 have been repeated for 4-5 sections, a temporary heading invert is
needed. The next step is to dig the heading invert.
12. The bolts are placed and grouted.
13. Wire mesh is placed at the heading invert. Then an application of shotcrete,
which is 5cm thick, is applied. The wiremesh is placed a second time and the
second application of shotcrete is applied which is 10-15cm thick.
14. Replacement earthwork.
15. Restart a new section.
Table 1. Ground Classifications.

Classification Description Cross Section
Mudstone
Mudstone containing
layers of soft sandstone, Soft
C2 Sandstone,
siltstone, or gravel siltstone, or
gravel
Mainly gravel with

pockets of sand lenses. Gravel
Mostly coarse and grainy
D1
material with better Sand Lenses
cementation and
interlocking strength.
Loosely
Loosely packed gravel Packed
Gravel
D2 with layers of silty sand,
sand, or clay in between. Silty sand,
sand
or clay
MODEL AND DATA ANALYSIS
The CYCLONE (CYCLic Operation NEtwork) discrete event simulation

methodology is used in developing the simulation model (Halpin 1977). The model
used for the sake of estimating tunneling advancement rate based on aforementioned
tunneling processes is presented in Fig. 1. The name and duration for each activity
used in the model are provided in Table 2. Pakuashan tunnel is a twin tunnel. The data
collected in one tunnel was used to build to simulate the tunneling advancement rate.
Data collected from the other then was used to verify the accuracy of the model. At
least 30 cycles of tunneling operations for most activities were observed from March
to August in year 2000.
Table 2. Activity name and duration for D2 ground.

Activity Name Duration (min)
1 (COMBI) Positioning of Forepoling Machine Constant (30)
4, 14, 24, 34, 44 (COMBI) Mark Holes Beta (5, 19, 10.2, 3.9)
5, 15, 25, 35, 45 (COMBI) Drill Holes Triangular (11, 13, 23)
6, 16, 26, 36, 46, (COMBI) Clean Holes Beta (4, 25, 9.3, 4.3)
7, 17, 27, 37, 47 (COMBI) Insert Steel Rods Triangular (5, 8, 15)
8, 18, 28, 38, 48 (COMBI) Grout Holes Beta (2, 19, 6.7, 4.7)
9 (NORMAL) Wait For Solidification Constant (480)
10, 11 (COMBI) Excavate Tunnel Face Triangular (110, 120, 190)
41 (COMBI) Shotcrete Sealing Beta (20, 50, 36.1, 9.4)
52 (COMBI) Shotcrete Sidewall (5cm) Triangular (25, 30, 45)
53 (COMBI) Install Wire Mesh Beta (20, 45, 28.9, 6.6)
54 (COMBI) Steel Rib Gathering Constant (10)
55 (NORMAL) Assembly Triangular (15, 20, 25)
56 (NORMAL) Installation Beta (30, 50, 38.6, 6.8)
57 (NORMAL) Correction and Fix Beta (15, 35, 29.2, 4.5)
58 (COMBI) Reapply Shotcrete (15cm) Beta (45, 85, 58.1, 12.0)
59 (COMBI) 2nd Wire Mesh Beta (20, 45, 28.9, 6.6)
nd
60 (COMBI) 2 Shotcrete Application (15cm) Beta (45, 85, 58.1, 12.0)
63 (NORMAL) Wait for Solidification Constant (60)
65 (COMBI) Bolting Triangular (200, 320, 510)
66 (COMBI) Grouting Triangular (100, 160, 255
Temporary Heading Invert Beta (110, 195, 146.2,
71 (COMBI) Excavation 25.5)
72 (COMBI) Install Wire Mesh Triangular (30, 35, 55)
73 (COMBI) Reapply Shotcrete Beta (55, 100, 78.6, 13.8)
74 (NORMAL) Cure Shotcrete Constant (120)
75 (COMBI) Dirt Refill Beta (50, 100, 67.3, 14.8)
FIG. 1. Tunnel advancing operations for D2 ground.

The collected data were first used as the resources for duration fitting in 95%
confidence level. Beta and triangular probability distributions are suitable in modeling
the duration required in completing those activities listed in Table 2. For example, the
duration of activity-mark hole, can be modeled by Beta probability density function
whose minimum, maximum, mean, and standard deviation values are 5, 19, 10.2, and
3.9 minutes, respectively. In addition, the duration of activity-insert steel rods, is
suitable to be presented by triangular probability density function whose minimum,
mode, and maximum values are 5, 8, and 15 minutes, respectively. On the other hand,
the activity's duration marked with constant value was acquired based on
superintendent's notes rather than field observations.
The simulation results (see Fig. 2) show the variation of advancement rate in C2,
Dl, and D2 grounds. The average advancement rates for these grounds are 14.34,
12.02, and 18.21 hours per meter, respectively. The least advancement rate is found in
D2 ground because forepoling and temporary heading invert were required when
tunnel was constructed. However, only temporary invert was needed and hence, a
better advancement is found in C2 ground. The best simulated advancement rate is
found in Dl ground and is in great fluctuation because there was 50% probability that
the temporary heading invert is needed when Dl ground is encountered. Figure 3
shows the comparison of advancement rates obtained by simulation and actual
observations. Simulation was run for 10 cycles. The simulated accumulative times for
10 cycles are compared to actual data. The estimated time required to complete 10
cycles of tunneling operations is around 1,630 hours and about 1,661 hours were
observed in finishing these cycles. The accuracy of the prediction is 98.8%
(1,630/1,661*100%).
FIG. 2. Tunneling advancement rate in different grounds.

SUMMARY
This paper presents a discrete event simulation model used to estimate the
advancement rate when NATM tunneling method is adopted in building tunnel
penetrating the gravel formation. The duration of most tunneling activities can be
modeled by Beta or triangular probability density functions in 95% confidence level
using historical data. The simulation illustrates the developed simulation model that is
reliable to be used to measure the production rate before a similar project can be
started.
FIG. 3. Comparison of actual and predicted tunneling advance rate.
REFERENCES
Chung, T.H., Mohamed, Y., and AbouRizk, S. (2006) "Bayesian updating application
into simulation in the north Edmonton sanitary trunk tunnel project." J. Constr.
Eng. Mange., Vol. 132(8): 882-894.
Halpi, D.W. (1977) "CYCLONE-method for modeling job site processes." J. of the
Constr. Div., Vol. 103(CO3): 489-499.
Touran, A. (1997) "Probabilistic model for tunneling project using Markov chain." J.
Constr. Eng. Mange., Vol. 123(4): 444-449.
Nido, A.A. and Abraham, D.M. (1999) "Role of operation simulation in the analysis
and improvement of microtunnelling projects." Trenchless Technol. Res., Vol.14
(1):1-19.
AbouRizk, S.M., Ruwanpura, J.Y., Er, K.C., and Fernando, S. (1999) “Special
purpose simulation template for utility tunnel construction.” Proceedings of the
1999 Winter Simulation Conference, Phoenix/AZ: 948-955.
Sinfield, J.V. and Einstein, H.H. (1996) “Evaluation of tunneling technology using the
“Decision Aids for Tunneling”.” Tunnelling and Underground Space Technology,
Vol.11(4): 491-504.
Touran, A. and Asai, T. (1987) “Simulation of tunneling operation.” J. Constr. Eng.
Mange., Vol.113(4): 554-568.
Author Index
Page number refers to the first page of paper
Cao, Liping, 9, 32 Ni, Fujian, 49

Chen, Chien-Ta, 166
Chen, Fengchen, 72 Qiu, Xin, 15
Chen, Tung-Tsan, 173
Chen, Yu, 107 Steyn, W. J. vdM., 98
Chen, Yu-liang, 39, 56 Strauss, P. J., 86
Cheng, Tao-ming, 194 Sukolrat, Jiraroth, 181
Sukumaran, Beena, 65
Dechasakulsom, Montri, 181 Sun, Lijun, 32
Denny, Kyle, 65
Dong, Zejiao, 9, 32, 72 Tan, Yiqiu, 9, 72
du Plessis, L., 86 Tao, Zhuohui, 49
Tian, Xiaoge, 133
Fu, Xiaohong, 146
Wang, Chun-Yuan, 173
Han, Chunhua, 127 Wang, WeiDong, 138
Hsiao, Wen-ta, 194 Wong, Winggun, 15
Hsu, Yao T., 173 Wu, Hsien-tang, 194
Hu, Changbin, 15
Huang, Xiaoming, 1 Xu, Tao, 1
Hung, Ching-Tsung, 166
Yan, Jinhai, 49
Jamnongpipatkul, Pichit, 181 Yang, Hai-fei, 159
Jia, Jonathan, 49 Yang, Jun, 22
Yang, Shi-jiao, 159
Kilian, A., 86 Yeh, Ming-Chen, 166
Yin, Chaoen, 22
Lambert, Nicholas, 65 Yuan, Jianbo, 116
Levenberg, Eyal, 79
Li, Qingbo, 146 Zeng, Sheng, 159
Li, Xue-lian, 39, 56 Zhang, Qisen, 39, 56, 107, 116, 127,
Lin, Chih-tsang, 194 153
Lin, Jyh-Dong, 166 Zhang, Yingxue, 127, 153
Lu, Guanqun, 146 Zhou, ChengHan, 138
Zhou, Zhi-gang, 56
Mehta, Yusuf, 65
203
Subject Index
Page number refers to the first page of paper
Airport runways, 65 Highway and road construction, 1,

Anisotropy, 79 159
Asphalt pavements, 32, 39, 49, 79, 98, Highways and roads, 127, 133
166
Asphalts, 1, 9, 15, 22 Imaging techniques, 138
Investments, 153
Base isolation, 56
Binders, materials, 15 Kinetics, 1
Bitumen, 22
Bridge construction, 138 Laboratory tests, 22
Bridges, highway, 138, 146, 153, 173, Life cycles, 166
181 Load factors, 56
Load tests, 72
Case reports, 173
China, 1, 9, 39, 49, 56, 72, 133, 153 Mechanical properties, 56
Cold regions, 9 Mixtures, 22, 32
Concrete, 15
Concrete pavements, 56, 86, 107 New Jersey, 65
Construction management, 116, 138,
146, 153 Optimization, 159
Costs, 166
Pavements, 72, 133
Damage, 173 Polymers, 15
Decision making, 127 Porous media, 15, 22, 166
Deformation, 32
Quality control, 116
Economic factors, 159
Embankments, 181 Recycling, 49
Entropy, 127 Rehabilitation, 49
Reinforced Concrete, 86
Finite element method, 72 Research, 9, 32
Foam, 181 Rheology, 9
Friction, 15 Risk management, 116, 159
Full-scale tests, 65 Rubber, 22
Geographic information systems, 133 Scheduling, 138, 146

Government, 153 Simulation, 138
Gravel, 194 Simulation models, 146
205
Skid resistance, 107 Tests, 107

Slope stability, 159 Thailand, 181
Soil compaction, 65 Thermal stress, 56
Soil stabilization, 181 Time factors, 79, 116
South Africa, 86 Tunnel construction, 194
Specifications, 49 Tunneling, 194
Stress, 98
Subgrades, 65 Vehicles, 86, 98
Taiwan, 173, 194 Weight, 127

Temperature effects, 9, 32, 39

Asphalt Material Characterization, Accelerated Testing, and Highway Management

Uploaded by

Copyright:

Available Formats

Asphalt Material Characterization, Accelerated Testing, and Highway Management

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Asphalt Material Characterization, Accelerated Testing, and Highway Management

Uploaded by

Copyright:

Available Formats

GEOTECHNICAL SPECIAL PUBLICATION NO.

ASPHALT MATERIAL CHARACTERIZATION,

August 3–6, 2009

Published by the American Society of Civil Engineers

Asphalt material characterization, accelerated testing, and construction management :

American Society of Civil Engineers

Photocopies and reprints.

This Geotechnical Special Publication (GSP) of the ASCE is a compilation of 25

(1) Asphalt Binder and Mixture Characterization,

Bhaven Naik, University of Nebraska at Lincoln, NE, USA.

Asphalt Binder and Mixture Characterization

Recycled Asphalt Materials and Pavements

Accelerated Testing of Pavement Structures and Materials

Economy and Management of Highway Engineering

Embankment Stabilization and Tunnelling

Xu Tao1 and Huang Xiaoming2

Abstract: Thermogravimetry experiments and thermal analysis kinetics are used to

Magnesium hydroxide (abbr.MH), as a kind of environmental protection type

flame-retarded SBS modified asphalt (6).

RESULTS AND DISCUSSION

Influence of MH on pyrolysis characteristics of asphalt

FIG. 1. TG-DTG curves of asphalt pyrolysis under different MH doses.

Kinetics analysis of asphalt pyrolysis

Pyrolysis mechanism models

where: A –frequency factor(or pre exponential factor), s–1; E – activation energy,

Where: g (Į) – integral function; T0 – initial temperature.

Table 1. Common Pyrolysis Mechanism and Integral Mechanism Functions.

Table 2. Correlation Coefficient of Asphalt Pyrolysis Mechanism Functions Fitting

Pyrolysis kinetics parameters analysis

When pyrolysis mechanism function of asphalt is determined, kinetics parameters are

Table 3. Calculation Results of Asphalt Pyrolysis Kinetics Parameters in N2

This study was funded by the Research Program of Guangdong Provincial

Cao Liping1, Dong Zejiao2, Tan Yiqiu3

Low-temperature cracking can be classified as one of the main distresses for

MATERIALS AND EXPERIMENT

Table 1ˊ Properties of 9 kinds of SBS modified asphalts

Base asphalt’s SBS content, Penetration, 0.1mm Softening Ductility, cm

INFLUENCE OF MODIFIER’S CONTENTS ON THE LTRP

INFLUENCE OF BASE ASPHALT’S GRADE ON THE LTRP

Another important influencing factor of the LTRP of SBS modified asphalts is

30.0 logGą 5.80

SUMMARY AND CONCLUSIONS

Xin Qiu1, Winggun Wong2,Changbin Hu3

BINDER TESTING AND EVALUATION

FIG.1. Test Procedure on Binders.

RESULTS AND ANALYSIS

Softening Point Test

Dynamic Shear Rheometer (DSR) Test

Axial Strain(%) 0.40

Number of Load Repetition

Moisture Damage Test

Table 8. Testing Results of ITSR.

R. Golebiewski, R. Makarewicz, M. Nowak, A. Preis. (2003). “Traffic noise reduction

YANG Jun and YIN Chaoen

MATERIAL DESIGN AND PERFORMANCE TEST

Asphalt binders Characterization

Table 1. Crumb rubber sieve gradation.

The density of crumb rubber is 1.16g/cm3. It is uncontaminated from steel or other

Table 2. Properties of different bitumen.

4cm AK-16A E1 =0.25

6XUIDFHć 0LGGOHOD\HUć %RWWRPOD\HUć

DSsurface: Level 1İ1000 times /mm, 1000<Level 2<3000 times/mm,