AP-T33-04 Austroads 1987

AP-T33
TECHNICAL BASIS OF AUSTROADS

PAVEMENT DESIGN GUIDE
AUSTROADS
TECHNICAL BASIS OF AUSTROADS PAVEMENT
DESIGN GUIDE
Technical Basis of Austroads Pavement Design Guide
First Published 2004
© Austroads Inc. 2004
This work is copyright. Apart from any use as permitted under the Copyright Act 1968,
no part may be reproduced by any process without the prior written permission of Austroads.
National Library of Australia

Cataloguing-in-Publication data:

ISBN 0 85588 709 5
Austroads Project No. PUB.PT.C.007
Austroads Publication No. AP-T33/04
Project Manager
Steve Brown, VicRoads
Prepared by
Geoff Jameson and Kieran Sharp
ARRB Transport Research Ltd
Published by Austroads Incorporated

Level 9, Robell House
287 Elizabeth Street
Sydney NSW 2000 Australia
Phone: +61 2 9264 7088
Fax: +61 2 9264 1657
Email: austroads@austroads.com.au
www.austroads.com.au
Austroads believes this publication to be correct at the time of printing and does not accept responsibility for
any consequences arising from the use of information herein. Readers should rely on their own skill and
judgement to apply information to particular issues.
TECHNICAL BASIS OF AUSTROADS PAVEMENT
DESIGN GUIDE
Sydney 2004
Austroads profile
Austroads is the association of Australian and New Zealand road transport and traffic authorities whose
purpose is to contribute to the achievement of improved Australian and New Zealand road transport
outcomes by:
♦ undertaking nationally strategic research on behalf of Australasian road agencies and communicating
outcomes
♦ promoting improved practice by Australasian road agencies
♦ facilitating collaboration between road agencies to avoid duplication
♦ promoting harmonisation, consistency and uniformity in road and related operations
♦ providing expert advice to the Australian Transport Council (ATC) and the Standing Committee on
Transport (SCOT).
Austroads membership
Austroads membership comprises the six state and two territory road transport and traffic authorities and the
Commonwealth Department of Transport and Regional Services in Australia, the Australian Local
Government Association and Transit New Zealand. It is governed by a council consisting of the chief
executive officer (or an alternative senior executive officer) of each of its eleven member organisations:
♦ Roads and Traffic Authority New South Wales
♦ Roads Corporation Victoria
♦ Department of Main Roads Queensland
♦ Main Roads Western Australia
♦ Department of Transport and Urban Planning South Australia
♦ Department of Infrastructure, Energy and Resources Tasmania
♦ Department of Infrastructure, Planning and Environment Northern Territory
♦ Department of Urban Services Australian Capital Territory
♦ Commonwealth Department of Transport and Regional Services
♦ Australian Local Government Association
♦ Transit New Zealand
The success of Austroads is derived from the collaboration of member organisations and others in the road
industry. It aims to be the Australasian leader in providing high quality information, advice and fostering
research in the road sector.
FOREWORD
The Austroads publication “Pavement Design – A Guide to the Structural Design of Road Pavements” is
intended to assist those responsible for the structural design of pavements for highway traffic. It was
originally produced in 1987 as a result of a review of the NAASRA “Interim Guide to Pavement Thickness
Design”, which was issued in 1979. In 1992, the Guide was revised to include an updated procedure for the
design of rigid pavements and revisions to Chapter 6 (Pavement Materials) and Chapter 7 (Design Traffic).
An essential element in the use of the Guide is a thorough understanding of the origins of the design
procedures, their scope and limitations. Accordingly, this report contains the following four technical reports
which detail the technical basis of both the 1992 and 2004 editions of the Guide:
• Part 1: 1992 Guide procedures for the design of flexible pavements, by David Potter
• Part 2: 1992 Guide procedures for the design of rigid pavements, by George Vorobieff and
John Hodgkinson
• Part 3: 2004 Guide procedures for the design of flexible pavements, by Geoff Jameson
• Part 4: 2004 Guide procedures for the design of rigid pavements, by George Vorobieff
The four reports are augmented by several Appendices which explain, in greater details, some of the
background to the material presented in the Guide. A comprehensive list of References also accompanies
each report.
The material presented here represents almost 40 years of work conducted in Australia and overseas. A large
number of people – representing Austroads Member Authorities, ARRB Transport Research, local
government, industry and consultants – have input into the development of the various editions of the Guide
and their contributions are gratefully acknowledged. I also acknowledge Geoff Jameson for leading this
project and coordinating the work and Kieran Sharp for editing the reports to ensure consistency of
expression and style,
Steve Brown
Manager GeoPave, VicRoads
Project Manager
AUSTROADS 2004
—i—
Technical Basis of Austroads Pavement Design Guide: Part 1

Part 1: 1992 Guide Procedures for Design of Flexible Pavements
David Potter
June 1999
SUMMARY
This report records the work undertaken in the development of Pavement Design – A Guide to the Structural
Design of Road Pavements, initially published by the National Association of Australian State Road
Authorities (NAASRA) in 1987 and subsequently revised and re-issued by Austroads in 1992. It briefly
describes the predecessor and progenitor – the Interim Guide to Pavement Thickness Design (NAASRA
1979) – and then proceeds to review the technical issues encountered, and the solutions adopted, in the
formulation of the Guide.
This material presented in the Guide represented many years of development in Australia and overseas in
design procedures for flexible pavements for highway traffic. The Guide was developed by a series of (then)
NAASRA Working Groups representing both the members of NAASRA and industry. Note that the names
of the various Road Authorities relevant at the time (rather than the current names) are used throughout this
report.
This report does not address the origins of Chapter 9 of the 1992 Guide – the Design of Rigid Pavements –
which is the subject of another report in this document (Part 2).
AUSTROADS 2004
— 1. i —
TABLE OF CONTENTS
Page
1. INTRODUCTION............................................................................................................................. 1-1
2. SETTING THE SCENE ................................................................................................................... 1-2

2.1 The Interim Guide to Pavement Thickness Design................................................................. 1-2
2.2 Establishment of a Working Group to Revise the IGPTD ....................................................... 1-8
3. GRANULAR PAVEMENTS WITH THIN BITUMINOUS SURFACINGS ...................................... 1-11

3.1 Origins of the CBR-Thickness-Traffic Chart.......................................................................... 1-11
3.2 Terminal Condition ................................................................................................................ 1-16
4. DEVELOPMENT OF THE MECHANISTIC PROCEDURE........................................................... 1-19

4.1 Broad Issues ......................................................................................................................... 1-19
4.2 Elastic Characterisation ........................................................................................................ 1-22
4.3 Performance Relationships ................................................................................................... 1-27
4.4 Design Traffic ........................................................................................................................ 1-33
4.5 Incorporation of Location-Specific Temperature Regime ..................................................... 1-35
5. DEVELOPMENT OF OVERLAY DESIGN PROCEDURE ........................................................... 1-36

5.1 Design Deflection Curves...................................................................................................... 1-36
5.2 The Curvature Function ........................................................................................................ 1-38
5.3 Temperature Correction for Deflection and Curvature.......................................................... 1-38
5.4 Reduction in Deflection Parameters due to Overlay Placement .......................................... 1-41
5.5 Adjustment of Asphalt Overlay Thickness to Allow for Locality Temperature ...................... 1-42
REFERENCES CITED IN THE TEXT ..................................................................................................... 1-43

REFERENCES REVIEWED BY WG IN ASSESSING RELEVANCE OF ANISOTROPY ...................... 1-46
BIBLIOGRAPHY FOR FATIGUE OF ASPHALT ..................................................................................... 1-47
APPENDIX A: ORIGINS OF UNBOUND GRANULAR THICKNESS CHART ........................................ 1-48
APPENDIX B: DMR, NSW PROCEDURE FOR PAVEMENT THICKNESS DESIGN
IN 1947 – EXCERPT FROM BRITTON (1947)........................................................... 1-56
APPENDIX C: JAMESON ON THE ORIGINS OF DESIGN DEFLECTION CURVES
– EXCERPT FROM JAMESON (1996)....................................................................... 1-60
AUSTROADS 2004
— 1. ii —
TABLES
Page
Table 1: Methodology and Rationale for Procedures in IGPTD ......................................................... 1-5
Table 2: Revision of IGPTD - Initial Proposal ..................................................................................... 1-9
Table 3: Distribution of Axle Group Type by State/Territory ............................................................. 1-35
Table 4: RORVL Load Distributions on Axle Groups According to Axle Group Type ...................... 1-36
Table 5: Number of Standard Axles for Same Distress as Axle Groups with
(non-rounded) Load Distributions ....................................................................................... 1-34
FIGURES
Page
Figure 1 Presumed CRB, Victoria thickness design chart (George and Gittoes 1959) .................... 1-13
Figure 2 Plot of roughness/initial roughness against cumulative ESAs for “standard”
design ESAs of 105, 106 and 107 (based on procedure F1 thickness
correction factors) ............................................................................................................... 1-18
Figure 3 Predictive ability of deflection parameters (Anderson 1984a) ............................................ 1-40
AUSTROADS 2004
— 1. iii —
1. INTRODUCTION
This report records the work undertaken in the development of Pavement Design – A Guide to the Structural
Design of Road Pavements, initially published by the National Association of Australian State Road
Authorities (NAASRA) in 1987 and subsequently revised and re-issued by Austroads in 1992. It briefly
describes the predecessor and progenitor – the Interim Guide to Pavement Thickness Design (NAASRA
1979) – and then proceeds to review the technical issues encountered and the solutions adopted in the
formulation of the Guide.
This report does not address the origins of Chapter 9 of the Guide – the Design of Rigid Pavements – which
is be the subject of a separate report in this document.
The technical basis of the 2004 Guide is also the subject of a separate report in this document.
AUSTROADS 2004
— 1.1 —
2. SETTING THE SCENE

2.1 The Interim Guide to Pavement Thickness Design
The first document relating to pavement design to be produced conjointly by Australian Road Authorities
was the Interim Guide to Pavement Thickness Design – known by its acronym IGPTD. The document was
produced by the National Association of Australian State Road Authorities (NAASRA), the (then) umbrella
organisation of the Australian State Road Authorities (SRAs). It was drafted by a sub-committee of
NAASRA’s Materials Engineering Committee (MEC) and vetted prior to publication by its Principal
Technical Committee (PTC).
In 1975, a recommendation by the NAASRA Materials Engineering Committee (MEC) resulted in a
direction from the Principal Technical Committee (PTC), at its 45th (September 1975) meeting, that a manual
on pavement thickness design be prepared and that a Sub-Committee of MEC be formed to pursue that task,
with the manual to be vetted by PTC prior to its publication. The members of the Sub-Committee were:
Alan Leask Department of Main Roads, NSW (DMR, NSW) Convenor
Peter Lowe Country Roads Board, Victoria (CBR, Vic.)
Ray Elliott Department of Construction (DoC)
Zandor (Vlas) Vlasic Main Roads Department, Qld (MRD, Qld)
Lester Goodram Main Roads Department, WA (MRD, WA)
John Scala Australian Road Research Board (ARRB)
The initiation of the project was possibly stimulated by the attendance of John Scala, Vlas Vlasic, Len
Chester (Highways Department, SA (HD, SA)) and Alan Leask at the Fourth International Conference on the
Structural Design of Asphalt Pavements, which was held in London in 1972. In addition, there was an
increasing desire by the MEC members for increased National cooperation and the achievement of a uniform
National approach on the subject.
As instigator and Chairman of the project, Alan Leask was considered to be the driving force behind the
development of the IGPTD. However, all MEC members provided full support.
In formulating the sections on flexible and semi-rigid pavements, the Sub-Committee members drew
particularly on the research conducted by John Scala, while Ray Elliott contributed substantially to the
section on rigid pavements. Meetings were frequent, usually lasting about four days at a time, and, as the
Sub-Committee reported to MEC, input from other members of MEC was also significant. Moreover, as
time went by there were changes in the composition of MEC and the Sub-Committee. Other participants
who made valuable contributions to the refinement of the document included Len Chester, Eric Brown
(DoC), Ralph Rallings (Department of Main Roads, Tasmania (DMR, Tas)), Rod Payze (HD, SA), Peter
Rufford (DMR, NSW) and Ed Haber (DMR, NSW).
In principle it was postulated that an ideal pavement design procedure should be one which would predict a
thickness and composition which, without being conservative, would ensure that the pavement would not
deteriorate beyond a tolerable level of serviceability in less than the chosen design period. In practice, this
would facilitate the planning of a maintenance and rehabilitation regime commensurate with the selected
design period, and thus would permit the comparison of designs comprised of different compositions on the
basis of total whole-of-life costs, as distinct from initial cost only.
The August 1978 draft was considered by the 33rd (September 1978) MEC, which agreed that the document
would be ready for publication after some editorial and technical amendments, which had been agreed to at
the meeting, were made. The document was then forwarded to PTC seeking their approval to publish.
AUSTROADS 2004
— 1.2 —
Although the document records 1979 as its publication date (the Foreword is dated April 1979), it was mid-
1980 before it was released – and then only on a very restricted basis. Individual copies were numbered and
it was not available for sale. The reason for this approach may be found in the following sentence in the
Foreword:
“It is stressed that the document is interim in nature and that it is proposed that it be reviewed
after about two years, in the light of comments received, experience obtained and further
research”
Prior to its release, NAASRA PTC had taken a formal decision to review the status of the document before
1982. Had the decision to publish, albeit with this proviso, not been taken, then it is doubtful whether
another attempt would have been made for a considerable period of time. In this regard, the attitude was
taken that the document, though admittedly lacking in many respects, had to be exposed in order to ensure
that feedback was obtained and that further research was conducted to ensure that the necessary knowledge
was acquired to allow these refinements to take place.
The design systems included in the IGPTD were initially based on the approaches adopted by the various
SRAs. However, the compilation process involved a considerable amount of definition, interpretation,
rationalisation, compromise and innovation. Because the document was intended to be applicable over the
whole of Australia, involving diverse materials and environments, and because it might be used by other
organisations having varying degrees of expertise and resources, same parts were deliberately broadly based
in order to allow a hierarchical approach to be taken to the evaluation of the input parameters and the design
procedures.
The task of compiling the IGPTD was enlightening in that it emphasised the inadequacies of the traditional
systems, particularly with respect to pavements incorporating bound layers. Of special concern was the
paucity of performance-related data substantiating the criteria used, and the lack of guidance regarding the
evaluation of the parameters required in the thickness design process. By the same token, it was emphasised
that there was little alternative than to implement newer methods, some of which were innovative and which
would need refinement when compatible performance data became available.
Identification of the deficiencies which inhibited the unqualified implementation of the IGPTD did stimulate
awareness of the need for more precise criteria and data, and resulted in a number of research projects.
2.1.1 Format of the IGPTD

The format of the IGPTD was as follows:
Part 1 Information on factors affecting pavement design and performance.
Part 2 Flexible pavement design procedures.
Part 3 Rigid pavement design procedures.
Part 4 Summary providing all the necessary information to carry out a design.
Whilst the detailing of the procedures for the design of the various types of pavement described was a
primary objective of the Guide, another significant feature of the document was related to the definition of
the factors affecting the design procedures, especially the elaboration, in considerable detail, on the
appraisal, in quantitative terms, of the evaluation of subgrade strength, traffic loading, etc.
For the design of new flexible pavements, the following four procedures were developed:
F1: Granular pavement with thin (< 25 mm) bituminous surfacing
F2: Granular pavement with asphalt surfacing up to 100 mm
F3A: Pavement containing a bound layer or with asphalt > 100 mm) see below
F3B: Pavement containing a bound layer or with asphalt > 100 mm)
AUSTROADS 2004
— 1.3 —
In addition, the following procedures were developed for the design of overlays and rigid pavements:
F4: Overlay design
R1: Rigid Pavement Design
R2: Rigid Pavement Design
Table I (after Potter 1981) lists, for procedures F1 to F4, the essential steps, together with a note on the
rationale behind each step.
The core of procedure F1 was the (now) familiar CBR-thickness-traffic chart. Its development is discussed
in detail later.
The basis of procedure F2 was estimation of the candidate pavement’s maximum deflection (from a quasi-
elastic analysis), this being checked against a specified tolerable value which was a function of both asphalt
thickness and design traffic.
In the context of this brief overview of the design procedures, discussion of procedure F3B logically
precedes discussion of procedure F3A. Procedure F3B involved determining, firstly, the required thickness
of granular cover (from procedure F1) and then substituting asphalt and/or cemented material for the
granular material on the basis of a table of layer equivalencies (thicknesses of granular material equivalent to
unit thickness of the bound material). For asphalt, the layer equivalence depended on the climatic zone,
design traffic, and depth below the surface. For cemented material, it depended only on depth below the
surface1.
Procedure F3A contained elements of both F1 and F2. The first step was to determine (using procedure F1)
the total (granular) cover requirement. Tentative thicknesses were then assigned to each layer in the desired
configuration. A check was then made, for each GRANULAR layer in this tentative structure, to ensure that
there was sufficient THICKNESS of cover over it (regardless of TYPE of cover) – the requirement being
that specified in F1. Layer thicknesses were adjusted (if necessary) to satisfy this requirement. A full elastic
layer analysis was then carried out on the pavement to estimate its maximum deflection under a Standard
Axle.
1
With the load equivalence value being dependent on the depth below the surface, the issue that has never been clear
to the author is does this depth below the surface refer to the depth within the granular pavement where the granular
material is being replaced, or does it refer to the depth at which the substituted material finds itself in the final
pavement configuration? The different interpretations result, in some situations, in non-trivial differences in final
configurations (e.g. 265 mm cf. 300 mm for full-depth asphalt – enough to be significant in alternate tender situations).
AUSTROADS 2004
— 1.4 —
Table 1
Methodology and Rationale for Procedures in IGPTD
Procedure Method Rationale

F1 Determine subgrade design CBR
Estimate cumulative traffic over design life in terms of ESAs AASHO Road Test (relative damage)
ERVL Study (Stevenson 1976)
(axle vehicle loads)
Determine thickness of cover (T) required for subgrade California Highways Department (Porter
condition and estimated traffic 1938)
TRRL, SRA experience
Modify T for desired terminal roughness Scala (1977); (AASHO Road Test data)
Check minimum base requirement SRA experience
F2 Determine thickness of cover using F1
Determine pavement stiffness factor Limited modular ratio of adjacent layers
(Heukelom and Klomp 1962)
Determine deflection under a Standard Axle Elastic layer theory
Reduce deflection by 5% for each 25 mm of AC surfacing SRA experience, Scala (1973)
Check reduced deflection against tolerable deflection SRA experience; Scala (1973)
F3A Determine thickness of cover (T) using F1 See procedure F1
Select a pavement composition having total thickness T –
Adjust pavement layer thicknesses to satisfy cover –
requirements (from F1) for each unbound layer
Assign values to elastic parameters (Young’s modulus, Laboratory, field investigations
Poisson’s Ratio) for each layer
Estimate deflection under a Standard Axle Elastic layer theory
Check estimated deflection against tolerable deflection and, if SRA experience; Scala (1973)
less, reduce thickness of bound layers
F3B Determine thickness of cover (T) using F1 or F2 (as See procedures F1, F2
appropriate)
Determine individual layer thicknesses (ti) from suggested
layer equivalencies (ai) and relationship: T = Σ ai ti
F4 Determine representative deflection under a Standard Axle
Estimate traffic over design life in ESAs AASHO Road Test (relative damage)
ERVL Study (Stevenson 1976)
(axle vehicle loads)
Check representative defection, dREP, against tolerable SRA experience; Scala (1973)
deflection, dTOL
If dREP < dTOL, adopt nominal overlay –
If dREP > dTOL, design structural overlay –
For granular overlay, determine thickness on the basis of a SRA experience
10% deflection reduction for each 45 mm of overlay material
For AC overlay, determine thickness using recommended SRA experience
percentage reduction in deflection per 25 mm for the specific
climate zone
AUSTROADS 2004
— 1.5 —
The use of CIRCLY (Wardle 1977) was recommended for this analysis, provided adequate computer
facilities were available. Otherwise, the use of the 2-layer or 3-layer tables developed by CSIRO (Gerrard
1969; Gerrard and Wardle 1976) was recommended. The estimated deflection was assessed against a
specified maximum tolerable value – which was a function of pavement composition and design traffic. The
thickness of the bound layer was then adjusted (down) and the maximum deflection re-calculated until the
requirement was (just) satisfied.
The interim nature of the IGPTD is well attested by the qualifying comments within the document regarding
the use of procedure F3A. In introducing the alternative design procedures F3A and F3B, the document
includes the following text (under the heading “Qualification”):
“Procedure F3A is suggested as a method which can be developed to satisfy the need for a
completely satisfactory means of selecting the thickness of a flexible pavement, the composition
of which includes one or more layers of bound materials. The basis of the procedure is a
comparison between the surface deflection of a proposed pavement with the deflection that is
assumed to be indicative of the pavement capacity to produce the design performance. This
‘tolerable deflection’ is recommended on the basis of the recorded performance of similar
pavements. The deflection of the proposed pavement is estimated by an elastic analysis of its
behaviour under the action of a standard wheel load.
The present recommended values of tolerable deflection are based on the best information
currently available for Australian conditions. However, whilst they are considered adequate as
a secondary control over a primary design criterion, e.g. subgrade CBR, they cannot be
regarded as sufficiently well established to warrant their acceptance as a single or primary basis
for design. Thus, Procedure F3A, in its present state of development, is recommended only as
a conservative control over those methods currently used, i.e. the substitution of bound for
unbound materials on the basis of empirically established equivalency factors. Such a
comparative use will encourage the accumulation of performance data which can be expected
to improve the accuracy of the method to a stage at which it provides an acceptable degree of
confidence in the ability of the selected pavement to perform as designed.”
The specific introduction to procedure F3B includes the following text:
“The use of Procedure F3A is inhibited at the present time principally by the difficulties of
determining the appropriate values of moduli and also, by the current lack of verification of the
proposed tolerable deflection criteria. As an interim measure, these restrictions are avoided in
practice by the use of empirical equivalency factors.”
Procedure F3A was, in essence, (what is now called) a mechanistic design procedure. It involved the use of
a Response Model (CIRCLY or tables) to determine a single critical response (maximum surface deflection)
of the candidate pavement to Standard Axle loading. The performance of the candidate pavement was
estimated from the (configuration-dependent) plot of tolerable maximum deflection versus cumulative traffic
(ESAs). The IGPTD recognised the relevance of the maximum values of both vertical strain at the top of the
subgrade and horizontal strain at the bottom of bound layers. However, it stated that:
“limiting values for these critical design criteria are not well established for Australian conditions,
and it would be inappropriate to adopt overseas criteria without verification. Moreover, these
particular criteria are difficult to measure and it is unlikely that they would be monitored against
pavement performance in order to establish such relationships”.
In support of maximum surface deflection as an estimator of horizontal tensile strain at the bottom of a
bound layer, it made the following statement:
“ limited performance data from the ARRB and some state road authorities has indicated that
the vertical surface deflection provides a reasonable premise for design in most practical
situations. It is considered that a limiting vertical surface deflection criterion does, for all
practical purposes, control tensile strain at the bottom of a layer. This is because the tensile
strain is governed by the radius of curvature of the deflection bowl on loading which has been
shown, for a given pavement material and thickness, to correlate reasonably well with the
maximum surface deflection”.
AUSTROADS 2004
— 1.6 —
It also stated:
“The further advantage of the vertical surface deflection criterion is that it is easily measured in
the field and is, therefore, more suitable for verification of the design procedure than would be
other criteria”.
With respect to the vertical strain on the top of the subgrade, the document states that, to inhibit loss of
surface shape and ancillary surface cracking:
“It is, therefore, essential to limit the subgrade deformation and this is achieved by limiting the
vertical compressive strain at the top of the subgrade. It should be noted in this regard that
such an approach supports the rationale of the CBR method of design, which, by requiring a
minimum thickness of cover over any material (characterised by its CBR value) at a given level
in the pavement, directly ensures that the vertical compressive strain at the level does not
exceed an implicitly defined acceptable value”.
It further states that:
“...as discussed above, if the design procedure also satisfies CBR design thickness
requirements, it provides a basis for controlling vertical compressive strain in the subgrade”.
The overlay design procedures (F4) involved calculating the thickness of overlay required to reduce the
representative deflection of the section to a tolerable level. In the case of a granular overlay (unbound), the
thickness required was based on each 45 mm thickness of overlay reducing the representative deflection by
10%. The reduction in surface deflection for a given thickness of asphalt overlay is dependent, among other
things, on the operating temperature of the material in place. As this varies according to climatic conditions,
five separate deflection reduction factors were presented, ranging from 12% in the coldest climatic zone to
6% in the warmest. Since the tolerable deflection also decreases as the thickness of asphalt overlay
increases, an iterative process was used to estimate the thickness required to reduce the representative
deflection to a sufficient extent.
The procedures for rigid pavement design, R1 and R2, were virtually the same as those adopted by the US
Portland Cement Association (PCA). The R1 procedure was appropriate when the quality of the traffic data
was such that both axle type and load distribution, and the number of repetitions of each axle/load
combination, could be predicted. The procedure was based on the fatigue concept, in that it was assumed
that, as the flexural stress ratio (ratio of flexural stress caused by the wheel load to the concrete design
flexural strength) decreases, then the number of load repetitions to failure increases. When the stress ratio
was less than 0.5 it was assumed that the concrete could sustain unlimited stress repetitions without loss of
load-carrying ability. Conversely, at high stress ratios, only a limited number of the heavier loads could be
sustained before the concrete failed. Therefore it was essential that projected traffic estimates, especially
with respect to the mass and numbers of heavier axle loads, were reliable, since these virtually controlled the
pavement thickness design.
The more common rigid pavement design procedure, R2, was used when the total traffic composition could
not be predicted and the loading had to be estimated based on the number of commercial vehicles. The
procedure was derived from procedure R1 by utilising data from the NAASRA (1976) Economics of Road
Vehicle Limits (ERVL) study. The effect of traffic was accounted for by calculating the thickness of
concrete required for unlimited stress applications of the most critical axle type at its design load, and then
reducing that thickness to account for the expected number of repetitions of the design axle load, which was
assumed to be a fixed proportion – about one design axle per 1,000 commercial vehicles – of the total
commercial vehicle traffic. The effect of concrete strength and subgrade support on pavement thickness was
accounted for by applying suggested percentage variations in pavement thickness.
Details of reinforcing and jointing techniques applicable to both procedures were also included in the
IGPTD.
AUSTROADS 2004
— 1.7 —
2.1.2 Release of the IGPTD

The IGPTD was launched at a two-day NAASRA/ARRB Seminar on Heavily-Trafficked Flexible
Pavements, held at ARRB in June 1980. (The document was actually released three weeks after the
“launch”). Attendance was strictly by invitation only.
Of the 54 delegates, the only non-SRA/ARRB attendees were as follows:
R. Brewis, K. Kiesel Australian Asphalt Pavement Association (AAPA)
H. Luckhurst-Smith, J. Sutton Shell Company of Australia Ltd
R. Elliott Cement & Concrete Association of Australia
B. Heaton University of Newcastle
I. Lee University of New South Wales
J. Morgan Golder Associates
R. Sharp University of Sydney
K. Wallace James Cook University of North Queensland
It is of interest to note (in light of subsequent developments) that the following all participated actively in the
seminar (presentation of papers, chairing of syndicate sessions, panel discussion, etc.): Alan Leask, Peter
Lowe, John Bethune, David Anderson, Ron Gordon, David Potter and Geoff Youdale. Jack Morgan, in his
concluding remarks, made the following statement:
“I believe this Guide is a document that we can justifiably be proud of as an Australian product.
There are many sound and even innovative approaches described, and when we recognise that
this forms a consensus of SRA thought on this topic there must be high hopes that the standard
of pavement design will continue to improve. There is obviously work to be done in filling in the
blanks, and even deleting some material I believe to be now of only historical interest. However
the Guide provides a framework for collecting information and highlighting significant areas of
needed research”.
In commenting on Geoff Youdale’s excellent paper summarising the (then) state-of-the-art of repeated load
testing of pavement materials (Youdale 1981), Morgan stated:
“Some 15 years ago academics were trying to promote repeated load testing to the SRAs. Now
at least one SRA has taken it up and who knows but in another 15 years the results may even
be used!”
Very prescient of him!
The Seminar Proceedings were printed some 10 months later (Sparks 1981), with availability being restricted
to SRAs and Seminar attendees. The delay in printing the Proceedings was probably attributable to the usual
delay in collating written reports from syndicate Chairmen, etc., coupled with concerns about “the sensitive
nature of some of the data presented” – presumably an allusion to the (then) recent significant pavement
failures. The reasons behind the delay in releasing the IGPTD are not fully clear. However, it is understood
that AAPA (represented by Ken McKenzie, Ollie O’Flynn, Ron Ekberg, etc.) had serious concerns about the
layer equivalency values for asphalt incorporated in procedure F3B, and expressed their concerns to
NAASRA.
2.2 Establishment of a Working Group to Revise the IGPTD

The initial steps taken to institute a revision of the IGPTD were as follows:
• In March 1981, NAASRA MEC decided to submit to PTC a detailed proposal recommending that
revision of the IGPTD commence in 1982.
• In August 1981, a Sub-Committee of MEC met and formulated the detailed proposal which was
subsequently submitted to PTC. The author attended the meeting by invitation. It is of interest to note
MEC’s initial intentions. Table 2 presents the proposed layout of the revised IGPTD, together with the
(proposed) responsible individual(s).
AUSTROADS 2004
— 1.8 —
• The Working Group which was proposed to undertake the components requiring major revision
comprised:
David Potter ARRB (Convenor)
David Anderson CBR, Victoria
Ron Gordon MRD, Queensland
Geoff Youdale DMR, NSW
Gavin Donald (DMR, NSW) was to undertake overall editing.
• The September 1981 meeting of the PTC approved the proposal.
• The Working Group first met for the first time in December 1981.
Table 2
Revision of IGPTD – Initial Proposal
Proposed Format Extent of Person(s) Responsible

Revision
1. Introduction total re-write R. Payze
2. Scope total re-write R. Payze
3. Terminology total re-write A. Leask
4. Choice of pavement type total re-write L. Goodram
5. Basis for design total re-write P. Lowe
6. Evaluation of design parameters
(i) moisture total re-write R. Payze
(ii) subgrade CBR total re-write R. Payze
(iii) subgrade modulus total re-write R. Payze
(iv) pavement materials major revision Working Group
(v) traffic total re-write R. Payze
(vi) evaluation for rehabilitation total re-write R. Payze
7. Design Procedures for Flexible
Pavements F1 total re-write Working Group
F2 major revision Working Group
F3 total re-write Working Group
8. Design Procedure for Rigid Pavements total re-write A. Leask
supported by
E. Haber, J. Cruickshank,
R. Elliott
9. Design Procedures for Overlays major revision Working Group
10. Observations total re-write P. Lowe
In broad terms, the scope of the task assigned to the WG may be summarised as follows:
• For the design of chip-sealed granular pavements, retain procedure F1 (in the IGPTD), i.e. the CBR-
Traffic-Thickness of cover chart.
• For the design of other flexible pavements, devise a mechanistic procedure consistent with the F1
procedure.
• For the design of overlays, devise a procedure consistent with the above two procedures for new
pavements
Alan Leask was retained as Convenor of the review Steering Committee and remained so until his retirement
for MEC in 1983. He was followed as Convenor by Peter Lowe and then by Gavin Donald, who undertook
the final editing.
AUSTROADS 2004
— 1.9 —
2.2.1 A Guide – not a Manual

The decision to title the revised document a Guide (consistent with the original IGPTD) rather than a Manual
reflected the policy of (the then) NAASRA to foster among its members harmonisation of standards,
practices, etc. rather than foisting upon the members sets of mandatory rules. Its replacement, Austroads,
pursues a similar policy. Further, the intent of the document is to guide, educate and provide assistance in
deciding among alternative options, etc.
AUSTROADS 2004
— 1.10 —
3. GRANULAR PAVEMENTS WITH THIN BITUMINOUS SURFACINGS

3.1 Origins of the CBR-Thickness-Traffic Chart
The WG was directed to retain the CBR-Thickness-Traffic chart for design of chip-sealed granular
pavements because it was the consensus view of pavement designers at that time (early 1980s) that
pavements designed in accordance with it provided field performance consistent with design expectations.
Because it was retained in the 1992 edition of the Guide (as Figure 8.4) and because it constitutes a
foundation stone for the mechanistic procedure in the Guide, a review of its origins is considered
appropriate.
Within Australia, the (then) Country Roads Board of Victoria (CRB) led the way in the development of
thickness design for flexible pavements. Hence, following the developments within the CRB provides an
appreciation of the Australian scene.
Jameson (1996) has produced a most adequate review based predominantly on material assembled by the
author. The following comments are supplementary to his review and need to be read in conjunction with it.
It is attached as Appendix A. The interested reader is also referred to an earlier review by Anderson of the
evolution of pavement design within the CRB (Anderson 1981).
The appropriate “starting point” can be considered to be Porter’s development of the (laboratory) CBR test
(Porter 1938). Using this test to characterise subgrades, the California State Highway Department, in
reviewing the performance of its roads over the period 1929-1938, found that soil having a certain CBR
always required the same thickness of flexible macadam construction on top of it in order to prevent plastic
deformation of the soil (Davis 1949). The curve relating the required thickness of flexible macadam to
subgrade CBR is that labelled “7000-lb. wheel load (Light Traffic)” in Jameson’s Figure 2 (the curve was
originally unlabelled). Davis goes on to state that:
“The curve for the wheel load of 12,000 lb. was added later as the result of further experience in
California of heavier traffic conditions subsequent to 1938. The curve for the wheel load of
9,000 lb. has been obtained by interpolation between the curves for the wheel loads of 7,000
and 12,000 lb. It is an implied assumption of these curves that all kinds of flexible construction
of the macadam type spread the load to approximately the same extent”.
Hence, the first CBR design chart was the single (uppermost) curve in Jameson’s Figure 2 (see Appendix A).
It is to be noted that the CBR value refers to the subgrade only and that the material to provide the thickness
of cover is a flexible macadam. The Figure covers subgrade strengths up to CBR 80, requiring
approximately 3 inches (75 mm) of cover on the CBR 80 subgrade.
Having monitored the rapid developments in California post 1942, the CRB released Jameson’s Figure 3
(CRB 1945). The most spectacular progression was in traffic characterisation. From the 1942
characterisation of traffic in terms of a maximum wheel load with unspecified repetitions, by 1945 it was
characterised by the two-way cumulative number of repetitions over the design period of equivalent 5,000 lb
wheel loads2.
The factors taken into account in estimating cumulative traffic were:
• day(s) of the week on which the relevant survey was conducted (every day had a different weighting),
• pavement width (to account for transverse distribution),
• growth rate,
• design period, and
• climatic factor (a multiplier whose value was average rainfall (inches) x average wet days per year ÷
10,000).
2
For a wheel load of L lb, one repetition was equivalent to 2(L–5000) repetitions of a 5,000 lb wheel load. This
corresponds to a “Power Law” exponent in the range 3.8-4.7.
AUSTROADS 2004
— 1.11 —
The Figure covers subgrade strengths up to CBR 80 (as previously), with no cover required on the CBR 80
subgrade (cf. 3 inches (75 mm) previously).3
However, by 1949, with the release of its Technical Bulletin No. 4 (CRB 1949), characterisation of traffic for
conventional design situations had reverted to 1.5 times the average number of trucks and buses (in both
directions) in a 12-hour day count – essentially average daily commercial vehicles. For unusual traffic
situations the 1945 approach was retained. Technical Bulletin No. 4 also introduced estimation of
(laboratory soaked) CBR from gradings, Atterberg Limits and Linear Shrinkage. The design chart provided
for subgrade strengths up to CBR 20 (c.f. 80 previously), with a minimum cover requirement of (approx.) 3
inches (75 mm).
Ten years after the release of Technical Bulletin No. 4, H.P. George (CRB) and C.A. Gittoes (DMR, NSW)
included in their report to the 1959 PIARC Congress (George and Gittoes 1959) a thickness design chart in a
format very similar to the one currently in the Guide – except that design traffic was expressed as repetitions
of a 5,000 lb wheel load, as shown in Figure 1. Hence, despite the issue of Technical Bulletin No. 4, some
interest remained in the characterisation of traffic in terms of cumulative repetitions of a standard loading.
At this same PIARC Congress, MacLean (1959) presented Jameson’s Figure 5 as the then status within the
UK. The UK development work behind this chart is well described by Jameson. Subsequent to the
Congress, CRB rapidly embraced the thickness design chart presented by MacLean, issuing it in Technical
Bulletin No. 21 in the following year (CRB 1960).
The chart is a series of curves providing required depth of construction according to subgrade CBR for seven
ranges of daily traffic – daily traffic being determined as per Technical Bulletin No. 4 except that the
vehicles to be counted changed from “trucks and buses” to “vehicles exceeding 3 tons loaded weight”. As
Jameson notes, the three curves for mid-range traffic align well with California’s original three curves (for
wheel loads of 7,000, 9,000 and 12,000 lb). The chart covers subgrade strengths up to CBR 150 (c.f. 20
previously) and indicates a minimum cover requirement of 2 inches (50 mm). However, in the text the
minimum thickness requirement (of CBR > 80 material) is stipulated to be from 3 inches to 8 inches (75 mm
to 200 mm), depending on the traffic classification (no minimum thickness requirement was specified for the
highest traffic classification – presumably an oversight). Technical Bulletin No. 21 also saw the introduction
of the static and dynamic cone as a basis for estimating subgrade CBR.
In 1969, CRB issued Technical Bulletin No. 26, which superseded Technical Bulletin No. 21. The curves
were re-drawn to reflect these minimum thickness requirements, while retaining coverage of subgrade
strength up to CBR 150. An additional curve was added for unsealed shoulders. For traffic determination,
the multiplier applied to the 12-hour count data increased from 1.5 (Technical Bulletins Nos. 4 and 21) to 3,
i.e. design traffic was doubled for the same project traffic.
3
For the reader who is interested in where the (then) DMR, NSW stood in relation to pavement thickness design at
about this time, its procedure current in 1947 is attached as Appendix B. It is an extract from a paper by A.T. (Sandy)
Britton to the 1947 Meeting of the Highway Research Board (HRB) (Britton 1947). The pavement design was
primarily based on classification testing of the subgrade and pavement materials. This was supplemented by utilising
insitu CBR tests and CBR tests on samples conditioned to predicted moisture conditions in the 1960s.
AUSTROADS 2004
— 1.12 —
Figure 1: Presumed CRB, Victoria thickness design chart (George and Gittoes 1959)
The CBR-thickness-traffic chart for granular pavements with chip seals in Technical Bulletin No. 26
(Jameson’s Figure 6), together with similar charts then in use in other SRAs, provided the basis for the chart
presented (as Figure 2.2) in the IGPTD (Jameson’s Figure 7). According to Black (1977), the traffic
classifications in Technical Bulletin No.26 (daily two-way volumes of vehicles exceeding 3 tons loaded
weight) were converted to cumulative one-way ESAs over the design period on the following basis:
“.....the following assumptions were made:
(i) One commercial vehicle equalled one equivalent standard axle.
(ii) The traffic was equally divided between the two directions.
(iii) A design period of 20 years was adopted.
(iv) The commercial vehicle traffic category was characterised by the average commercial traffic
volume in the category and this was taken as the average value over the design period.”
Jameson’s interpretation – that the daily traffic volumes were considered to be end-of-life volumes after 3%
per annum growth – appears to be an over-complication of what actually transpired.
Table 2.11 of the IGPTD lists the number of ESAs per commercial vehicle (according to State and Road
Functional Class) which was recommended for use at the time. The Table entries were derived from the
(1974) ERVL Survey data. For Victoria, the values were 1.4, 0.8, 1.2, 0.7 and 0.8 for Functional Classes 1,
2, 3, 6 and 7 respectively. Taken across all States, the Table suggests that (rough) average values for rural
and urban roads are 1.2 and 0.8 respectively. Hence, the value of 1 adopted for the translation is a “good
average value”.
It is the author’s recollection that the formula attached to the IGPTD chart is not in exact agreement with the
chart, i.e. the chart was established prior to, and independent of, the formula.
AUSTROADS 2004
— 1.13 —
3.1.1 Quality of Pavement Material and its Cover Requirements

During the course of evolution, later versions of the thickness design chart included cover requirements, not
only over the subgrade, but also over any placed material (the cover requirement being determined inter alia
by the CBR of the placed material).
With regard to the evolution of quality requirements for cover material, Davis’ comment on the early
California curves that “It is an implied assumption of these curves that all kinds of flexible construction of
the macadam type spread the load to approximately the same extent” has already been noted.
The Road Research Laboratory (1955) Road Note 20 characterises both sub-base and base material by their
CBR values, with base material having a CBR > 80 and the maximum contribution of a sub-base to the
subgrade cover requirement being the subgrade cover requirement minus the sub-base cover requirement.
Hence, the Leigh and Croney (1972) statement quoted by Jameson that the design curves:
“....provided a means for estimating the total thickness of construction necessary for
various traffic and foundation conditions, but gave no guidance on the relative thicknesses
of surfacing, base and subbase”
is somewhat unfair to Road Note 20.
In Technical Bulletin No. 21, materials were allotted a “Design CBR” value. With the exception of fine
crushed rock and macadam (both allotted CBR values of 100), a material was allotted – on the basis of its
grading and PI – a CBR value (in the range 3-50) and also a minimum cover requirement over itself
(independent of design traffic).
The minimum thickness of CBR > 80 material has already been noted. The Bulletin states that the Design
CBR values should be used in the pavement design, presumably in the (now) conventional manner.
Technical Bulletin No. 26 states that:
“The pavement itself should be made up of materials increasing in strength and stability towards
the top. Where the California Bearing Ratio of the pavement material is known the chart
provides a guide to the depth at which the material may be used but other factors such as
grading and plasticity index may be more important than the actual California Bearing Ratio”.
However, further on it states that:
“Where the local materials are cheap but of poor quality they should be used in substantial
thicknesses since:- (i) While the CBR design method makes no specific allowance for the
quality of the pavement material, the design curves may be considered as based on average
quality materials. The ability of the poorer materials to spread the load onto the subgrade will
be less than for good materials.”
The IGPTD is most specific and unambiguous. It states in Section 2.1.4 that:
“The total thickness of a granular pavement may be made up of a base and any number of sub-
bases. The composition of the pavement is made up by providing sufficient cover over the
subgrade and each successive sub-base. The thickness of cover required over a sub-base is
determined from its design CBR. If this CBR value is less than 30, the cover required is
determined as for a subgrade material, namely from Fig. 2.2 or Equation 2.1. For a sub-base
with a design CBR greater than 30, it is necessary to provide a minimum thickness of a base
material with a CBR of 80 or above.”
AUSTROADS 2004
— 1.14 —
3.1.2 Concluding Comments

It is worthy of note that, during the development described above, essentially the only finding from the
(1959-61) AASHO Road Test which was “picked up” was the adoption (in 1979 by the IGPTD) of the ESA
as a basis for quantification of traffic loading. This provided no conceptual advance over California’s in-
place use of the 5,000 lb wheel load coupled with its basis for determining “equivalent repetitions” for
different load magnitudes. The (probable) refinement was the adoption of a “Damage Exponent” of 4
derived by Irick and Hudson (1964) subsequent to the AASHO Road Test4.
The main advantage of a review of this nature is that it assists in “getting things into perspective”. As with
any investigation relating to the performance of roads, one comes away feeling disappointed at the abysmal
lack of performance data (except at the macroscopic level).
Although the CBR design method has been in use for over 50 years, there are still misconceptions of the
philosophy and intent of the procedure. For example, Rodway (1997), in an overview of mechanistic
pavement design, makes the following statements in relation to the method:
• The CBR design method is based on a failure mode that involves loss of shape of the
pavement surface caused by overstressing the subgrade. Pavement containing a bound
layer or with asphalt > 100 mm.
• The empirical CBR method implies that loss of surface shape is primarily caused by
overstressing the subgrade. Deformation within the pavement layers is not directly
addressed by the method.
• The CBR design procedure involves increasing pavement life by increasing pavement
thickness to further protect the subgrade, not by improving the pavement materials.
It needs to be clearly understood that the CBR design procedure requires inter alia that:
• the uppermost material be of (relatively) high quality and of thickness in excess of a specified minimum;
• a minimum thickness of cover be provided over each of the other materials comprising the pavement,
with this minimum thickness increasing as material quality decreases; and
• the minimum thickness of uppermost material and minimum thicknesses of cover over the pavement
materials be increased as design traffic increases.
The intents of these requirements are that:
• the stresses imposed on each and every material comprising the pavement be contained in order to limit
the development of permanent deformation within the material; and
• the stresses imposed on each material be concomitant with design life requirements.
Hence, it is contended that the above statements do not do sufficient justice to the CBR design procedure
because they do not acknowledge the provisions within the procedure for limiting the stresses imposed on
(and hence the deformations developed within) the constituent pavement materials.
However, it needs to be borne in mind that, at the time of the development of the procedures, truck tyre
pressures were below 600 kPa. In recent times, pressures of 900 kPa are not uncommon, leading to
considerably increased stress levels with the pavement near the surface. Being empirically based, the CBR
procedure does not have a mechanism for incorporating tyre pressure as a design variable.
4
While granular pavements with a chip-seal surfacing were incorporated in the design of the AASHO Road Test and
were constructed and trafficked, because of their very rapid failure (attributed to poor construction coupled with the
effects of cyclical freezing and thawing), their performance (to the author’s knowledge) was not reported.
AUSTROADS 2004
— 1.15 —
It should also be noted that the CBR test was initially developed as a tool for pavement thickness design,
with a CBR of 100 indicating that a material did not require further cover. The CBR value has since evolved
into many materials specifications where it provides an indicator of shear/bearing strength. As the test
involves pushing a plunger into a rigidly-constrained sample, its use for estimating the shear strength of base
materials, particularly those with large particle sizes, is of limited value.
3.2 Terminal Condition

Implicit in the design procedure for these pavements (Section 8.3 and, specifically, Figure 8.4 of the Guide)
is a terminal condition which is considered to be unacceptable and, hence, signifies the end of life for the
pavement. Quantification of this condition could be expected to be in terms of level of roughness and
severity and extent of rutting.
As noted above, this design procedure was taken across (essentially unaltered) from the IGPTD.
Unfortunately, no statement of terminal condition accompanied it. The view of the MEC Review Committee
at the time was that, in terms of rutting, it probably represented an average rut depth of about 20 mm. In
terms of roughness, the procedure in Section 7.8 of the Guide (which provides a basis for altering design
traffic to achieve a specific terminal roughness) provides some guidance. This procedure (which was also
taken across from the IGPTD) is based on the premise that the terminal roughness of pavements designed in
accordance with Figure 8.4 is three times the initial roughness. This premise was formulated by John Scala
(ARRB) based on inter alia analysis of performance data from the AAHSO Road Test. There appears to be
no record of the development of it.
3.2.1 Modification of Terminal Condition

For the design of flexible pavements whose critical distress mode is permanent deformation of the subgrade
and granular layers, the use of Figure 7.2 allows the designer to select a pavement design which has a
designer-specified terminal roughness (relative to its initial roughness). To achieve this, the designer simply
enters Figure 7.2 with a value for the Design Traffic and also a value for the desired ratio of terminal to
initial roughness and reads from the Figure a value for Modified Design Traffic. This value is then used in
the design procedure in lieu of the Design Traffic.
The relationship plotted in Figure 7.2 was derived in the following manner.
The IGPTD (Section 2.1.3) provides – for granular pavements with thin bituminous surfacing – a basis for
altering the thickness of cover over the subgrade to achieve a designer-specified terminal roughness value.
The thickness modification is expressed in the following relationship:
T = t. [2R1/(R2 - R1)]0.25 (1)
where: R1 is the expected initial roughness,
R2 is the desired terminal roughness,
t is the thickness of cover determined from IGPTD Procedure F1 (Figure 8.4 in the
1992 Guide), and
T is the modified thickness of cover.
This relationship is the foundation for the relationship in Figure 7.2.
AUSTROADS 2004
— 1.16 —
One further piece of information from the IGPTD (Section 2.1.2) which is relevant here is that the CBR-
thickness-traffic relationship in Figure 8.4 of the 1992 Guide is as follows:
t = [219 – 211 (log10CBR) + 58(log10CBR)2]. log10(N/120)
where t (mm) is the thickness cover required over a material of given CBR when the design traffic is N
ESAs.
This equation can be rewritten more succinctly as:
t/f(CBR) = log10 (N/120) (2)
where f(CBR) = 219 – 211(log10CBR) + 58(log10CBR)2.
The Modified Design Traffic – denoted by NM – is simply the Design Traffic associated with the modified
thickness of cover T.
Hence, from equation 2: T/f(CBR) = log10 (NM/120)
i.e. log10 (NM/120) = T/f(CBR)
Now, from equation 1, T =t.[2R1/(R2 - R1)]0.25
Therefore log10 (NM/120) = [t/f(CBR)] * [2R1/(R2 - R1)]0.25
= [t/f(CBR)] * [2/(R2/R1-1)]0.25
Substituting from equation 2:
log10 (NM/120) = log10 (N/120) * [2/(R2/R1-1)]0.25 (3)
i.e. log10 NM = [2/(R2/R1-1)]0.25 log10N + [1 - [2/(R2/R1-1)]0.25] * log10120
This is the relationship which is plotted in Figure 7.2 of the 1992 Guide.
3.2.2 Implicit Model for Roughness Progression

It is of interest to note the model for the development of roughness with traffic that is implicit in Figure 7.2
of the Guide. From equation 3 above, we have:
log10 (NM/120) = log10 (N/120) * [2/(R2/R1-1)]0.25
Rearranging these terms gives: (R2/R1-1)/2 = [(log10 (N/120))/ log10 (NM/120))]4
i.e. R2/R1 = 1 + 2 * [(log10 (N/120))/ log10 (NM/120))]4
This relationship indicates, for a given value of “unmodified” design traffic (N), how the ratio
roughness/initial roughness increases as cumulative traffic NM increases. Hence, it is a statement of the
roughness progression model implicit in Figure 7.2. Representative plots of the model are shown in Figure
2.
AUSTROADS 2004
— 1.17 —
10
7 10
5
6
6
Ratio of Roughness 5 10
to Initial Roughness
4 7
10
0
1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E+08
Cumulative (ESA)
Figure 2: Plot of roughness/initial roughness against cumulative ESAs for “standard” design
5 6 7
ESAs of 10 , 10 and 10 (based on procedure F1 thickness correction factors)
AUSTROADS 2004
— 1.18 —
4. DEVELOPMENT OF THE MECHANISTIC PROCEDURE

4.1 Broad Issues
A major attribute sought in a pavement design procedure is versatility – the ability to assess the likely
performance of a broad range of pavement configurations in a broad range of field conditions (traffic,
environment, etc.).
Its versatility may be conveniently assessed by determining where it lies between two idealised extremes.
These extremes are the fully empirical procedure and the fully mechanistic procedure. A fully empirical
procedure relies entirely on past observations of field performance. It allows no extrapolation beyond the
range of these observations. At the other extreme, the fully mechanistic procedure allows unbounded
extrapolation beyond past observations. It has this capability because, intrinsic to it, is a fundamental
understanding, for any pavement configuration, of:
• the effect on each component material in the configuration of any wheel loading applied to the pavement
surface (such effects are changes to the stress-strain state within the material); and
• the performance of each of the component materials when subjected to the sequence of changes in its
stress-strain state caused by the traffic loading.
The fully empirical procedure is idealised because it is unworkable – no pavement about to be built and
trafficked will correspond exactly with one previously observed. The fully mechanistic procedure is
idealised because it is unattainable – due to the level of complexity involved. Hence, all pavement design
procedures fall between these two idealisations. Between them the complexity increases along with
versatility.
The WG was directed to develop a mechanistic procedure for pavements containing bound materials because
of the lack of versatility of the relevant procedures in the IGPTD.
Bearing in mind that the end users of the (yet to be developed) procedure were pavement designers “out in
the real world”, a major decision confronting the WG was the appropriate compromise between increasing
versatility and increasing complexity (the latter translating ultimately into decreasing likelihood of use of the
procedure).
The starting point for the WG was an assessment of the then state-of-the-art in mechanistic analysis and
design. The basis for mechanistic analysis may be summarised as follows.
The response within a pavement to the passage of a wheel load over it takes the form of changes in the levels
of stress and strain within it. These changes are predominantly transient, i.e. after the passage of the wheel
load, the stress-strain state within the pavement is very close to its state prior to the passage of the wheel
load. Although this residual change is small, it is nevertheless important because it reflects the distress
caused to the pavement by the passage of the wheel load.
This residual change in the stress-strain state may be quite large at some specific locations within the
pavement (locations of localised shear, crack initiation, crack propagation). However, when it is averaged
over the volume of material affected by the passage of the wheel load, the residual change in the stress-strain
state is very small compared to the peak change that occurs during the passage of the wheel load. Such
localised occurrences attributable to localised inhomogenities constitute the essence of distress within the
pavement. It is the accumulation of these occurrences with the passage of traffic loads which eventually
becomes manifest as observable and readily quantifiable distress.
Modelling of these localised inhomogenities and their effects (by the use of statistical mechanics) was in its
infancy at the time of the review.
Models which assume homogeneity of materials and (iteratively) determine the stress-strain state after the
passage of each wheel load in the sequence of traffic loading (and, hence, the development of distress) were
beginning to appear. A leading example of this genre is Yandell’s Mechano-Lattice model (e.g. Yandell
1981).
AUSTROADS 2004
— 1.19 —
The vast majority of modelling work being undertaken and implemented at the time was based on the
following premises:
• The distress within a pavement which is attributable to a single passage of a specific load on a specific
axle configuration can be assessed from the peak levels of the pavement’s transient response (stresses and
strains) to the passage of the axle load – the peak response levels being determined in the early-life
(undistressed) pavement.
• The distress caused by n passages of the axle group:
i) is proportional to n for fatigue cracking, and
ii) for permanent deformation, either increases exponentially with n (conventional models) or
asymptotes to a plateau value (shakedown model).
For mixed traffic loading:
• fatigue damage is determined from Miner’s hypotheses, and
• for permanent deformation damage, there are more than one alternative summation models in use.
Miner’s hypothesis states that, for mixed traffic loading, fatigue failure will occur when:
∑n / N
i
i i
reaches a value of 1, where:

ni = the number of passages (within the mixed traffic loading) of loading type i; and
Ni = the number of passages of loading type i which will cause fatigue failure when loading
type ‘i’ is the ONLY loading applied.
The summation is taken over all loading types present in the mixed traffic.
For the determination of peak transient response, the following two types of model were in use:
• Linear elastic layer models, in which the pavement is assumed to be of infinite extent longitudinally and
transversely and downwards and the materials to be linear elastic with moduli independent of applied
stresses.
• Finite element models, in which the pavement dimensions are finite and material moduli may be stress-
dependent.
While the above modelling work was, at the time, undergoing a vigorous phase of development, all
mechanistically-based design procedures either in use, or assembled but yet to be evaluated, opted for:
• a linear elastic layer model for determination of peak transient responses;
• the peak level of tensile strain at the bottom of an asphalt layer as a predictor of fatigue life of the asphalt;
and
• the peak level of either tensile strain or tensile stress at the bottom of a cemented layer as a predictor of
fatigue life of the cemented material.
Procedures which incorporated the capability of predicting the development of permanent deformation with
the passage of traffic used the peak levels of compressive strain at locations throughout the relevant
materials. None of these models enjoyed broad-based acceptance.
Procedures with the lesser capability of predicting cumulative traffic to produce a pre-assigned level of
surface rutting used the peak level of compressive strain at the top of the subgrade as the predictor.
After due deliberation, the WG opted for a mechanistic analysis model comprising the following:
• a linear elastic layer model to estimate peak levels of transient responses – specifically, the CIRCLY
model;
AUSTROADS 2004
— 1.20 —
• peak tensile strain at the bottom of the layer as the appropriate response element for estimating the fatigue
lives of both asphalt and cemented materials; and
• peak compressive strain at the top of the subgrade as the appropriate response element for estimating
permanent deformation both within the granular material and the subgrade.
The specific sub-models adopted to predict performance from these critical responses are discussed below.
The model adopted incorporates major simplifications of the complex behaviour that actually occurs. The
WG was very aware of the nature and extent of the inherent simplifications at the time this model was
adopted. Its adoption was based on the following rationale.
The utility of an estimation model depends on:
• the accuracy of its estimate when the inputs are known,
• the accuracy of its inputs, and
• the likely extent of its use.
The more complex a model is, the better it scores on the first point and the worse it scores on the other two
points. Hence, the choice of level of model complexity involves compromise and is, in the final analysis,
subjectivity based.
In this context, it is of interest to note the outcome of a recent critical review of the mechanistic procedure in
the Guide (Rallings 1997). The review was quite detailed and encapsulated the views of Australia’s leading
proponents of pavement design and performance prediction. While the shortcomings inherent in the (in-
place) estimation procedures were duly noted along with possible fruitful areas for improvement, the
alternatives offered to replace part or all of the model were:
• The shakedown model,
• The mechano-lattice elasto-plastic model (Yandell 1981),
• The Vesys model (Kenis, Sherwood and McMahon 1981), and
• An adaptation of the Vesys rut depth prediction model (Vuong 1992).
The WG, at the time of formulation of the mechanistic procedure, was cognisant of the Mechano-lattice and
Vesys models and the early stages of the development of the shakedown model. The WG was of the view
that, while all models offered most desirable advances in the area of simulation of actual behaviour, the input
data requirements could not be reasonably expected to be available to the routine pavement designer, nor
could the designer be reasonably expected to achieve and retain both an understanding of the models and
competency in their use. Hence, the result of introducing such a level of analysis complexity would be “to
frighten the horses”.
In the view of this author, the situation has changed little since that time5. Rut depth prediction within Vesys
forms part of its performance model. In common with the Guide, Vesys uses a linear elastic response model
to determine the values of critical responses for use in the performance model (even though the word Vesys
is an acronym for viscoelastic system). Hence, if one chooses to describe the mechanistic procedure in the
Guide as “an elastic design system” (Rallings 1997), then such a descriptor is equally applicable to Vesys
and, further, to all mechanistic design procedures which enjoy routine use.
5
For example, the author is (and has been for 20 years) enamoured by the treatment within Vesys of the effect of
material variability on pavement performance. However, the input data requirements are demanding to the extent that,
to the author’s knowledge, only one such data set has ever been assembled! Again, with regard to the Vesys rut depth
prediction model, its formulation has considerable intuitive appeal. However, in application its predictive capability
has been, at best, fair. The detailed level of its input requirements is well illustrated by Table IV of Vuong (1992).
AUSTROADS 2004
— 1.21 —
4.2 Elastic Characterisation
4.2.1 Isotropic or Anisotropic Characterisation?

This issue relates to the elastic characterisation within CIRCLY of pavement and subgrade materials.
Isotropic materials have the same properties in all directions, whereas anisotropic materials do not. In terms
of the stress-strain behaviour of the material, the difference is illustrated by considering the simple stress
state of equal principal stresses (σ1 = σ2 = σ3). For an isotropic material, the resultant strain state is ∈1 = ∈2 =
∈3; γij = 0. For an anisotropic material, these relationships do not hold. Further, the values of ∈i, γij depend
on the orientations of the symbol σi with respect to the material.
Isotropic materials are characterised by two parameters, usually Young’s Modulus (E) and Poisson’s Ratio
(ν). For comparison, the different “classes“ of anisotropy require the following number of parameters to
characterise the material:
general case 36
strain energy conserved 21
orthorhombic 9
cross-anisotropic 5
The orthorhombic case is where there are three principal axes of anisotropy (conceptually appropriate for
material which is mixed, placed and roller-compacted in one direction – the axes being vertical, in the
direction of rolling, and transverse to the direction of rolling).
The cross-anisotropic case is where there is an axis of symmetry of rotation, with the properties being equal
in all directions perpendicular to this axis (but different to those in the direction of the axis). This is the case
modelled in CIRCLY.
This case is considered to be appropriate for naturally deposited material. Assuming the axis of symmetry to
be vertical, the five parameters required are (Wardle 1977):Eν, Eh, Fν, νh, ννh. The WG, having adopted
CIRCLY as the response model for the mechanistic procedure, was faced with the decision as to whether
cross-anisotropic characterisation was appropriate for (any or all of ) pavement and subgrade materials.
An additional consideration was the observation that measured deflection bowls were narrower than those
estimated from elastic layer analysis when isotropic characterisation was adopted. The view was taken by
the WG that, even if no evidence of anisotropic behaviour was discovered, anisotropic characterisation may
be appropriate within the constraints of the CIRCLY model to act – in the case of granular and subgrade
materials – as a surrogate for the (real) stress-dependence of modulus.
The issue was approached in two ways. Firstly, a literature review was carried out to determine (for each
material type) evidence of anisotropic behaviour in laboratory stress-strain measurements (appropriate field
measurements were non-existent). Approximately 30 references were located and reviewed, leading to the
conclusion that there was little evidence of anisotropic behaviour in the cases of asphalt and cemented
materials, while there was definite evidence in the cases of granular and subgrade materials. However, there
was considerable variability in the values of modular ratio (Eν/Eh) reported. For granular materials, they
were predominantly greater than 1 (ranging up to values as high as 4), while for fine-grained subgrade
materials they ranged from less than 1 to greater than 1.
The second approach involved response-to-load analyses (Youdale 1984a) to determine:
• the effect on CIRCLY-estimated responses of different modular ratios for granular material; and
• the effect on Finite Element Model (FEM)-estimated responses of incorporating (both independently and
conjointly) the stress-dependency of modulus of granular material and a modular ratio of 2.
AUSTROADS 2004
— 1.22 —
It was found that similar effects on FEM-estimated responses were obtained by :

(i) modelling the granular material as isotropic and incorporating stress-dependency, AND
(ii) modelling the granular material as anisotropic (Eν/Eh > 1) without stress-dependency.
In both cases, there was (cf. isotropic, no stress-dependency) a narrowing of the deflection bowl, an increase
in maximum deflection, and an increase in vertical compressive strain on top of the subgrade.
The WG deduced from this that, regardless of the degree of anisotropy pertaining to granular materials,
adoption of anisotropic characterisation of granular materials (with Eν/Eh > 1) within the CIRCLY model
would be “a step in the right direction” towards encompassing their known stress-dependency of modulus
together with their reported anisotropy. Further, to obtain a closer fit between observed and CIRCLY-
estimated deflection bowls, it was decided to adopt anisotropic characterisation (with Eν/Eh > 1) for
subgrades.
A value of 2 for the modular ratio (Eν/Eh) was adopted for both granular and subgrade materials as a “best
estimate compromise”.
The value of Eν (for use within CIRCLY) was taken to be the value of E determined from triaxial test results
under the assumption of isotropy. The values of νh, ννh were taken to be the value of (isotropic) ν. The
remaining cross-anisotropic parameter – Fν – was set equal to Eν /(1+ ν).
With these additional assumptions, the cross-anisotropic characterisation of granular and subgrade materials
was specified by values for Eν and ν. For asphalt and cemented materials, isotropic characterisation was
considered to be adequate and, hence, was adopted.
4.2.2 Values for Poisson’s Ratios

In estimating the levels of critical responses within the pavement, the pavement materials are modelled in the
Guide as being linear elastic. Bound materials (asphalt and cemented materials) are modelled as isotropic
and, as such, are fully characterised by values for the (Young’s) Modulus (E) and Poisson’s Ratio (ν),with
the latter being constrained (by the principle of conservation of energy) to the range 0-0.5. A material with
value 0, when subjected to uniaxial stress, exhibits no strain in any direction transverse to that of the applied
stress. With the same stress conditions, a material with value 0.5 exhibits a level of transverse strain such
that the volume of the material in the stressed state is equal to its volume in the unstressed state. Hence, an
isotropic material is either compressible or incompressible. It cannot increase in volume (dilate) under
compression, such response being associated with ν > 0.5. While it is not uncommon to find values of ν >
0.5 reported in the literature, these values were observed from laboratory tests where the level of
confinement of the material was less than that in a pavement (except at a pavement edge). Such values are
attributable to (non-recoverable) re-orientation of particles within the material.
Granular materials, be they pavement or subgrade materials, are modelled as cross-anisotropic, with stiffness
in the vertical direction being twice that in (all) horizontal directions. A cross-anisotropic material is
characterised by 5 parameters – two direct moduli (Ev, Eh), one shear modulus (Fv), and two Poisson’s Ratio
(νh, νvh). In the Guide, the designer specifies a single modulus (Ev) and a single Poisson’s Ratio (ν). The
following relationships are used to generate the parameter values:
νh = νvh = ν
Eh = Ev/2
Fv = Ev / (1 + ν)
AUSTROADS 2004
— 1.23 —
For pavement materials, representative values and ranges for ν are provided in Tables 6.4 (a) and (b) of the
Guide and, for subgrades, representative values, distinct for cohesive and non-cohesive soils, are provided in
Section 5.7. These numerical values are identical to those tabulated in the IGPTD (Table 2.8). While the
WG, during the course of its various literature reviews, noted values for Poisson’s Ratio reported in the
literature, it saw no evidence to support changing the values published five years earlier.
4.2.3 Relationship between Subgrade Modulus and CBR

The establishment of a relationship between modulus and CBR is intrinsically fraught with difficulty. In the
first instance, the two properties are markedly different. Resilient modulus is determined under stress-strain
conditions wherein the permanent strain in the material (after removal of the applied stress) is but a small
proportion of the total strain induced by the applied stress. In contrast to this, the CBR value is associated
with the peak resistance developed to a progressive shearing failure of the material. Hence, the two
properties are associated with the opposite ends of the stress-strain plot for the material. There is the further
complication that the value of resilient modulus depends on the values of stresses applied.
While the above difficulties pertain to one material in one condition (specified by, say, dry density and
moisture content), a relationship was sought which was relevant for all naturally-occurring fine-grained
materials and, for each material, over a wide range of conditions (even this is a considerable over-
simplification, as it leaves aside the effects of how the material arrived at the specified condition).
However, because the CBR test (or some surrogate for it) was (and essentially still is) the only test
commonly used to characterise subgrade materials, a relationship between CBR and modulus was an
essential component of the mechanistic design procedure.
The WG started with consideration of the information presented in Sparks and Potter (1982). The report
contained a review and comparison of relationships then in use (or proposed), together with relationships for
two Melbourne clays developed from results of in-house testing.
The test results reported, together with relationships developed, well illustrated the intrinsic variability in
subgrade material and the effect of the stress-dependency of the modulus (more significantly, the strain-
dependency).
After reviewing the information in this report, the WG tentatively adopted the relationship in the report for
the two Melbourne clays (E = 40.7 CBR0.37) on the grounds that “at least it is for Australian soils and it is in
amongst other relationships in use”.
However, upon later reflection, the WG opted for the relationship E = 10 CBR (developed by Heukelom and
Klomp (1962) from results of Rayleigh Wave and Dynamic Impedance testing in the Netherlands and
Rayleigh Wave testing in the UK) on the following grounds:
• the relationship encompassed more soil types;
• it improved the fit of the subgrade strain relationship derived from mechanistic analysis of granular
pavements;
• it had a simpler form; and
• it had been widely adopted.
The comment in the Guide that “a maximum value of 150 MPa is normally used” is in reference to cohesive
soils.
AUSTROADS 2004
— 1.24 —
4.2.4 Typical Modulus Values

The typical modulus values for unbound granular materials listed in Table 6.4(a) of the Guide are based on
the results of triaxial testing undertaken by Youdale on a range of materials. The values for modulus
achieved over granular material warrant a footnote indicating that they are values achieved in the upper
portion of the material when it is covered by a thin surfacing.
The typical modulus values for cemented material are based on limited testing undertaken by the Main Roads
Department, Queensland (1982), supplemented by test results reported in the literature.
4.2.5 Stress Regimes for Triaxial Testing of Granular Materials

The recommended stress conditions for triaxial testing indicated in Table 6.7 were determined from:
• triaxial testing of granular materials over a broad stress regime to determine a stress-dependent
relationship for the modulus of the material;
• finite element analyses of representative pavement structures incorporating the material; and
• the selection of representative stress conditions within the pavement from the results of the analyses.
The laboratory investigations are reported in Youdale (1978).
4.2.6 Sub-Layering of Granular Materials and Assignment of Moduli

With the knowledge that:
• the modulus of a granular material depends on the stress conditions it is subjected to;
• these stress conditions depend on the stiffness(es) of the underlying material(s);
• the stress levels decrease within a pavement structure as the depth below the surface increases; and
• in the response model CIRCLY, layers are constrained to having a fixed modulus value
then it was obvious that the simulation within CIRCLY of actual stress-strain conditions would be improved
for many pavement configurations by considering the granular material to be comprised of more than one
layer, each with a distinct modulus value.
The rules for determining when the sub-layering is necessary and, if so, the appropriate number of sub-
layers, together with appropriate modulus for each sub-layer, are presented in Section 8.2.2 of the Guide.
The rules were formulated on the basis of the results of finite element analyses of representative pavement
structures undertaken by Youdale (1983). The results of these analyses were used to determine the
thicknesses of bands (sub-layers) within the granular material wherein the variation in its modulus in the
region beneath the load was acceptably low. This provided a basis for determining when sub-layering was
appropriate and also the appropriate range of sub-layering thicknesses. Consideration of representative
values for the modulus of each band provided the basis for establishing a rule for the assignment of moduli to
each of the sub-layers. The following section describes how appropriate values were determined for the top
sub-layer.
The rule as finally adopted encapsulated the influences of both the cover over the granular material and the
stiffness of the supporting subgrade while providing a (reasonably) smooth transition of modulus with depth.
In the course of the development of these rules, the rules adopted by Shell in its Pavement Design Manual
(Shell 1978) were scrutinised. It was considered that the Shell rules, while presumably quite adequate for
their intended role, were too coarse (minimum sub-layer thickness of 150 mm) for the situation where the
granular material is covered with a thin surfacing.
AUSTROADS 2004
— 1.25 —
4.2.7 Modulus of Top Sub-Layer of Granular Material

The entries in Table 6.6 of the Guide were derived from results of a study by Youdale (1983) wherein a finite
element package (PAVAN1) was used to determine modulus values in a granular material which was
covered with asphalt and subjected to surface loading representing one side of a Standard Axle. The
following points are relevant:
• The granular material modelled was Hornsby Breccia modified with 2% hydrated lime, representing a
good quality crushed rock base material, well compacted at about 2% below optimum moisture content
(OMC).
• Its stress-dependency – based on repeated load triaxial testing – was given by
MR = 7.46 σ00.022τ00.648
where MR = resilient modulus,
σo = mean direct stress (kPa), = 1/3(σ1 + 2σ3),
τo = octahedral shear stress (kPa), = √3/2(σ1 –σ3),
σ1 = vertical stress,
σ3 = horizontal stress, and
Poisson’s ratio = 0.35.
• Because PAVAN1 is an axi-symmetric model, and because the granular material was stress-dependent, it
was not possible to model one side of the Standard Axle as two distinct tyre prints. A “composite”
loading was therefore adopted, being a single circular area of radius 300 mm loaded to 40 kN.
• Six pavement configurations were analysed – asphalt thicknesses of 0, 50, 100 and 150 mm with an
asphalt modulus of 2,800 MPa, and asphalt thicknesses of 100 and 150 mm with an asphalt modulus
1,000 MPa. The subgrade modulus was fixed at 30 MPa. All materials were modelled as isotropic.
• Contours of modulus within the granular material were plotted.
• The six values deduced from the contour plots form the entries in Table 6.6 for material compacted using
Standard compactive effort, covered by asphalt at 25ºC of < 25, 50, 100 and 150 mm and covered by
asphalt at 30ºC of thicknesses 100 and 150 mm.
• All the other Table 6.6 entries were determined by extrapolating from these values.
4.2.8 Modulus of Granular Material Overlying a Cemented Layer

Section 8.2.2(1) of the Guide states, “For granular materials placed directly on a stiff cemented sub-base, no
sub-layering is required”. This statement is based on results of a finite element analysis undertaken by
Youdale (1984b). The modelling and analysis was as outlined above (Section 4.2.7). Six pavements were
modelled: granular thicknesses of 150, 200 and 250 mm on cemented material having stiffnesses of 2,000
and 5,000 MPa.
4.2.9 Relationships between Modulus and UCS for Cemented Materials.

Two relationships are presented in Section 6.3.2.3 of the Guide for estimating the modulus of cemented
material from its UCS value – one for cemented crushed rock and one for cemented natural gravel.
Reference is given in the Section to a MRD Queensland (1982) report. The relationships stem from two
sources.
AUSTROADS 2004
— 1.26 —
Otte (1978) determined relationships between flexural strength and flexural stiffness for the two classes of
materials, based on an extensive program of laboratory testing (70 field samples for each material, six
specimens from each sample). He determined the following relationships:
Eb = 8.15σb + 3,485 (cemented crushed rock)
Eb = 10.06σb + 1,098 (cemented gravel)
where Eb (MPa) is the bending stiffness and σb (kPa) is the bending strength.
Otte refers in his thesis to a review by Walker (1976) of the relationship between UCS and bending strength
wherein Walker suggests that the following relationship is appropriate:
σb = 0.51.UCS0.88
where UCS is in units of kPa.
Otte substituted this relationship for σb in the above two relationships. The resulting relationships (with UCS
in units of MPa) are as given in the Guide.
4.2.10 Characterisation of Cracked Cemented Materials

Section 8.5 of the Guide discusses the possibility of significant life remaining in a pavement subsequent to
the cracking of cemented material, indicating that it is appropriate to characterise the cracked material as
granular material. The intention of this latter statement is that the cracked material be considered as a
granular material in all respects (2:1 anisotropy, Poisson’s Ratio = 0.35, sub-layering as per Section 8.2.2).
Due to an oversight, no guidance was offered on which of the two sets of entries in Table 6.6 (for Modified
and Standard compactive efforts) was appropriate. The intention was that materials whose initial stiffnesses
were 5,000 MPa or more be assigned values corresponding to Modified compactive effort, and materials
with lower stiffnesses be assigned values corresponding to Standard compactive effort.
4.3 Performance Relationships
4.3.1 Subgrade Strain Criterion

In formulating procedures for the design of flexible pavements, one very early decision (by the MEC Review
Committee) was that the existing procedure for the design of granular pavements with thin bituminous
surfacing (Procedure F1 of the IGPTD) be retained. The procedure (unaltered except for a minor change in
the minimum thickness of base material required) forms Section 8.3 of the current Guide.
At a later stage, in the course of the development of the mechanistic procedure, the WG decided to follow the
approach adopted by Shell by selecting the maximum vertical strain generated at the top of the subgrade (by
a Standard Axle) as the pavement response best suited for the prediction of allowable traffic before
permanent deformation in the granular layers and subgrade reached an unacceptable level.
The prediction relationship takes the general form:
N = a ∈-b
where N is the allowable number of repetitions of strain ∈, and a and b are (+ve) constants.
This relationship is commonly referred to as the Subgrade Strain Criterion.
AUSTROADS 2004
— 1.27 —
As the failure mechanism adopted for granular pavements with thin bituminous surfacing is permanent
deformation in the granular layers and subgrade (manifest at the surface as rutting and roughness), to achieve
consistency between designs produced by the Section 8.3 procedure and designs produced by the (yet to be
developed) mechanistic procedure, it was necessary to derive the subgrade strain criterion in such a manner
that it would predict the performance implicit in Figure 8.4 of Section 8.3 (the CBR-Thickness-Traffic
Chart).
Derivation of the relationship was carried out by Youdale (1984c) and Jameson (1996) undertook a thorough
appraisal of the derivation process which is reproduced here with minor amendments.
The subgrade strain criterion which was adopted for use in the Guide is:
N = (8511/µ∈)7.14
where N is the allowable number of strain repetitions before an acceptable level of rutting, and µ∈ is the
vertical compressive strain (microstrain).
This relationship was derived from back-analyses of 25 pavements selected from Figure 8.3 of the Guide.
For each pavement, CIRCLY was used to calculate the compressive strain at the top of the subgrade between
the dual wheels of the Standard Axle. The following procedures were used in the modelling:
• Each pavement was modelled as consisting of a single base layer, with one or more sub-base layers.
• Bases, sub-bases subgrades were considered as cross-anisotropic, with the vertical modulus being twice
the horizontal modulus. This anisotropy was regarded (Potter and Donald 1985) as a device to
compensate for the absence of a lateral stress dependent mechanism for elastic modulus.
• The base thickness was made equal to 150 mm except where the total pavement thickness was less than
250 mm, in which case the base thickness was reduced to 100 mm. The base vertical modulus used was
350 MPa.
• Sub-base layers were sub-divided such that the sub-layer thicknesses did not exceed 150 mm and the ratio
of the moduli of any two layers was less than 2.
• The Poisson’s ratio of all the granular pavement layers was 0.35.
• The subgrade vertical modulus (MPa) was taken as 10 times the subgrade CBR and the Poisson’s ratio
was assumed to be 0.45.
• The Standard Axle loading consisted of two 110 mm radii circular loads separated by 330 mm centre to
centre, with a tyre pressure of 550 kPa.
The calculated subgrade strain for each pavement was plotted against its design traffic. The plot indicated
that the results for pavements with a subgrade CBR of 20 were not consistent with other results (their strain
values being somewhat lower for the same design traffic). This was not considered to be of great importance
(Youdale 1984c) because the correlation between CBR and modulus is questionable at high CBR values. In
addition, it was suggested that subgrades with high CBR values would generally have low plasticity and
hence would not tend to deform plastically.
For these reasons a linear regression analysis was carried out on the results excluding the pavements with
subgrade CBR of 20 and the following relationship was obtained:
log µε = 3.93 – 0.14 log N
where N is the allowable number of Equivalent Standard Axles of loading before an unacceptable level of
rutting; and µε is the vertical compressive strain under a Standard Axle (microstrain).
By rearranging this equation, the Austroads subgrade strain relationship (above) was obtained.
AUSTROADS 2004
— 1.28 —
Jameson noted the following inconsistencies between the modelling procedure adopted by Youdale and the
modelling procedure adopted in the Guide:
• The manner in which the moduli for the granular layers was estimated. In the Guide the total thickness of
granular material is sub-divided such that layer thicknesses are in the range 50-150 mm and the ratio of
the moduli of adjacent layers is less than two. As discussed above, Youdale used a different procedure.
• The subgrade strains between the dual wheel loads were calculated by Youdale rather than the maxima of
the strains between and under the wheels.
This author’s recollection is that Youdale’s approach was consistent with the views of the WG at the time he
undertook the derivation.
Jameson notes as a third inconsistency: Youdale’s use of 550 kPa tyre pressure. The author begs to differ in
this matter. Firstly the Guide (Table 8.1, Step 11) recommends that the designer use a tyre pressure in the
range 550-700 kPa. Secondly, the Example Charts in the Guide are based on a tyre pressure of 550 kPa.
Thirdly, and more importantly, is the consideration that the subgrade strain criterion was derived to reflect
the performance (as encapsulated in Figure 8.4 of the Guide) of granular pavements with thin bituminous
surfacings. Figure 8.4 was adopted by the WG (at the behest of the MEC Review Committee) as being a fair
reflection of observed field performance. This observed field performance was performance under truck
traffic when tyre pressures of around 550 kPa were the norm. Hence, something akin to 550 kPa was the
appropriate tyre pressure to adopt at that time for estimating the strains generated in those pavements whose
performance is reflected in Figure 8.4. (The corollary to this is that, with the current tyre pressures in excess
of 700 kPa, the validity of Figure 8.4 should be re-assessed.)
Jameson points out that a regression of ESAs on strain is more appropriate for use in the mechanistic
procedure than the regression of strain on ESAs undertaken. He further points out that, because the
relationship is used “both ways”, the appropriate relationship is the bisector of the two. The author fully
concurs.
Jameson’s discussion relating to the fact that the relationship in the Guide is a “mongrel” relationship in that
it predicts the cumulative number of ESAs before the terminal condition is reached on the basis of strain
produced by a Standard Axle is most apposite. What was sought was a general relationship which predicts,
for a given level of strain, the allowable number of repetitions of that strain. The existing relationship does
not achieve this because the cumulative number of ESAs represent mixed traffic loading and, hence, a broad
range of strain values.
The WG was aware of this dilemma at the time of adoption of the relationship, but could not discover the
appropriate solution. After reading Jameson’s discussion the author is of the view that a satisfactory solution
can be reached in the following manner.
• For a pavement which produces a strain level, µ∈, when loaded by a Standard Axle, accept that the
allowable cumulative number of ESAs is given by the term (8511/µ∈)7.14. The traffic loading which is
reflected in this cumulative number of ESAs is the traffic loading that was on the road network when the
performance encapsulated in Figure 8.4 was verified/accepted.
The problem was that there was no representative distribution of loads on axle groups for the traffic of that
time. In the absence of this information, let us adopt the information in Table 8.3 of the Guide. For this
mixture of traffic loading, it is known that N ESAs produce the same level of permanent deformation as
1.1*N repetitions of a Standard Axle. Hence (8511/µ∈)7.14 ESAs produce the same level of permanent
deformation as 1.1*(8511/µ∈)7.14 repetitions of a Standard Axle.
AUSTROADS 2004
— 1.29 —
• Hence, the relationship can be re-written as follows:

N = 1.1 * (8511/µ∈)7.14
where µ∈ is the strain produced by a Standard Axle and N is the number of repetitions of the Standard
Axle.
This is the general relationship sought. Simplifying the relationship, and generalising the
interpretation of the terms, gives:
N = (8625/µ∈)7.14
where N is the allowable number of repetitions of strain µ∈.
• Better estimation of the “early days” traffic load distribution would lead to a better relationship.
The above approach is similar to Jameson’s except that he opts (for reasons unclear to the author) to use the
more recently derived traffic distribution in Table I-1.
4.3.2 Fatigue Cracking of Cemented Material

At the time of developing the Guide, the leading proponent within Australia of extensive use of cemented
materials was the Main Roads Department, Queensland (MRDQ). Apace with this increased use, an
increased appreciation of its performance developed within MRDQ. The WG availed itself of this
knowledge base and was guided by them in the formation of fatigue relationships for cemented materials.
The performance relationship in both the 1987 and 1992 versions of the Guide is:
N = (K/µ∈)18
where N is the number of repetitions of tensile strain at the bottom of the cemented layer before fatigue
failure occurs, i.e. when the level of this strain is µ∈ microstrain.
The numerator K depends on the stiffness of the material as follows:
Modulus of Cemented Value of K
Material (MPa)
2,000 280
5,000 200
10,000 150
In November 1997, a revision to the Guide was issued by Austroads which replaced the above relationships
by the following:
N = (K/µ∈)12
and the numerator K depended on the stiffness of the material as follows:
Material (MPa)
2,000 440
3,500 350
5,000 310
10,000 260
15,000 240
The development of these two sets of relationships can be summarised as follows.
AUSTROADS 2004
— 1.30 —
In the development of pavement thickness design curves for Queensland conditions based on elastic analysis
methods, Baran and Aubrey (1978) noted the following relationship developed by Pretorius (1969):
N = (142/µ∈)20.3
The modulus of the material was not known, but assumed to be > 10,000 MPa (later confirmed in Pretorius
and Monismith (1972) to be 28,000 MPa). They also noted a (graphical) relationship between strain-at-
break and modulus for cement-treated natural weathered gravel in Walker et al. (1977).
The Pretorius relationship gave a tolerable strain level of 72 µ∈ for 106 repetitions which, from the Walker et
al. plot corresponded to 65% of the strain at break for materials stiffer than 10,000 MPa. This ratio
(tolerable strain for 106 repetitions)/(strain at break) = 0.65 was adopted as being applicable to materials with
moduli down to 2,000 MPa. For a given modulus, the corresponding strain at break was determined from the
Walker et al. plot and then multiplied by 0.65 to give the tolerable strain for 106 repetitions. In a similar
manner, values of this ratio for 105 and 107 repetitions were determined for the Pretorius material and applied
to less stiff material.
On this basis, fatigue relationships were developed for materials of moduli 2,000, 5,000, 7,000, and ≥10,000
MPa over the range 105 to 107 strain repetitions. These relationships were then used in the development of
thickness design charts by Baran and Aubrey. Angell (1988) reported that the relationships for the materials
with moduli 2000, 5000 and ≥ 10,000 MPa were of the form:
N = (K1/ε)K2
with the values of K1 and K2 as follows:
Modulus of Cemented K1 K2
Material (MPa)
2000 259 19.9
5000 244 14.5
≥ 10000 152 18.3
Litwinowicz (1982) undertook a review of the basis for these relationships and found that (in Angell’s
words):
“the general level of these relationships appeared to be appropriate but that their exact form
and slope still required further investigation”
The WG, in reviewing these relationships, expressed some surprise that the value of the exponent (K2) did
not change monotonically with the material modulus.
Further investigations were undertaken and the relationships eventually adopted by the WG for inclusion in
the 1987 Guide were recommended by Litwinowicz (1984) on the basis of his investigations.
Subsequent to the WG’s adoption of the relationships with exponent 18, Angell (1988), in the course of
development of a pavement design manual for MRDQ, undertook a further review of the literature and
reported fatigue exponents of 32, 9, 12.7, 12.2 and 12 for relationships developed by workers in four
countries.
In light of this, together with his proposition that, if the true exponent were 18, then cement-treated
pavements in Queensland would be failing very early in their design life because of vehicle overloading,
Angell opted for an exponent of 12 and derived numerator (K1) values such that the revised relationships (in
his words) “allow approximately the same levels of strain as the relationships previously used”. Angell
developed the following relationship:
N = (K/µ∈)12
AUSTROADS 2004
— 1.31 —
with values of K as follows:

Material (MPa)
2,000 440
5,000 310
15,000 240
The WG was apprised of MRDQ’s intention to adopt these revised relationships while the original (1987
version) Guide was in press. The Austroads revision note (November 1997) adopted these relationships,
together with additional relationships for materials with moduli of 3,500 and 10,000 MPa. The additional
relationships were determined as follows:
• From Angell’s three relationships, Jameson, Sharp and Yeo (1992) derived (by linear regression) the
following general relationship:
log10 N = 43.21 – 3.58log10E – 12log10 µ∈
• Substituting for E the values 3,500 and 10,000 gives the additional relationships.
Issue of the revision note was prompted by a further literature review conducted by Jameson (1995) and by
the findings of a recent ALF trial of cemented materials (Jameson et al. 1995).
4.3.3 Fatigue Relationship for Asphalt

To assist the WG in arriving at a recommended fatigue relationship for asphalt, Anderson (1982a) undertook
a comprehensive review of the international literature, comparing published relationships from eight sources
encompassing South Africa, the USA, and Europe. The Bibliography is attached (see page 1-50). Direct
comparison between relationships proved difficult because of the differences in materials (aggregates, binder
sources), mix compositions, laboratory test methods, and definitions of fatigue test failure. Desirable
attributes sought in a relationship were that:
• it be developed from controlled strain testing (appropriate for thin asphalt courses);
• allowance be made for crack propagation;
• some cracking be tolerated during field fatigue life; and
• it was appropriate for mixes with stiffnesses ranging from 500 to 20,000 MPa.
Three relationships were considered to satisfy the above requirements reasonably well – Shell, Paterson and
Maree, and Santucci (see Bibliography). Among these, Shell was the most versatile in terms of mix
properties, temperatures and loading times. After due consideration, therefore, the WG adopted the Shell
relationship for recommended use in the Guide.
It is now well understood that the Shell relationship is for fatigue failure in the laboratory and not in the
field. This fact was not appreciated by the WG at the time. Commenting on the Shell relationship, Anderson
stated that it is “supposedly directly applicable to design, allowing for crack propagation through the
asphalt.” In discussing desirable attributes of a relationship, Anderson states:
“Consideration should be given to crack propagation and tolerable cracking as well as the
possibility of crack healing occurring under intermittent loading. Shell (1978) report field to
laboratory fatigue life correction factors of 10-20.”
Further, it has been noted above that the Shell relationship, as adopted, was considered by the WG to provide
allowance for crack propagation and to tolerate some cracking in the field. Hence, while it was the desire
and the intention of the WG that the adopted relationship incorporate laboratory-to-field corrections, the
relationship ultimately adopted contains no such correction. A detailed review of the development of the
Shell laboratory fatigue relationships (including materials tested, test conditions, etc.), together with Shell’s
use of correction factors in its design procedure, is provided in Potter (1997).
AUSTROADS 2004
— 1.32 —
4.4 Design Traffic

4.4.1 Axle Loads Which Cause Equal Damage
In assessing the damaging effect of design traffic, an essential requirement is the ability to compare
damaging effects of loads on different axle configurations. In the Guide, the basis for this comparison is
provided in Table 7.1, which lists, for the common axle configurations, the load on the axle configuration
which will produce the same damage as a Standard Axle (a Standard Axle is defined as a dual-tyred single
axle transmitting a load of 80 kN to the pavement).
In the mid-1960s, several independent analyses of AASHO Road Test data were reported which inter alia
provided estimates of the relative damaging effects of dual-tyred single axles and dual-tyred tandem axles.
Because these estimates were based on the performance of pavements with relatively thick asphalt surfacings
which were subject to freeze-thaw cycles, they were considered to be not directly applicable to the bulk of
the Australian road network – with its surfacing of chip seal or thin asphalt and not subject to freeze-thaw
cycles6.
For these reasons, Scala undertook a field study based on the premise (reasonably well supported by limited
AASHO data) that axle groups which cause equal maximum deflection in the pavement cause equal
pavement damage. The study was undertaken on a range of pavements, with both chip seal and thin asphalt
surfacing, in the Altona-Williamstown area of Melbourne. A “scaled up” version of the Benkelman Beam
was used to record peak deflections for steer axle and triaxle deflections, a pad (approx. 50 mm thick, made
from industrial rubber conveyor belting) with a transverse slit in it was placed on the road and a conventional
Benkelman Beam was positioned transversely with its tip in the slit. Maximum deflection was recorded as
the axle (group) passed over the pad.
Scala recorded the data and his analysis in internal ARRB reports. The author is not aware of any extant
copies of these reports. However, Scala’s findings and some of the data are reported in Scala (1970a).
With regard to load on a single-tyred single axle which produces the same maximum deflection as a Standard
Axle, Scala states (in the paper): “The equivalent load by deflection tests in about 11.6 kips” (51.6 kN) and
“In this paper 12 kips (53.4 kN) is used mainly for ease of computation”.
The only data in the paper relates to maximum deflections recorded for dual-tyred single and tandem axles.
It is presented in the form of a plot of the ratio (tandem axle deflection)/(single axle deflection) versus the
ratio (tandem axle load)/(single axle load). A broad range of deflection ratios is plotted for each of six load
ratios (corresponding to six days of testing). For three of the six load ratios, the reader is cautioned that the
data “may be affected by water penetration”.
With regard to the load on a tandem axle group which produces the same maximum deflection as a Standard
Axle, the paper provides two values – 28.9 kip (128.6 kN) and 29.2 kip (129.9 kN) – in a summary table
(Table VIII), together with the statement “Assuming that a 30 kip tandem axle load gives a deflection of the
same magnitude as an 18 kip single axle (dual tyre) load, ....”.
With regard to the load on a triaxle group which produces the same maximum deflection as a Standard Axle,
information in the paper is restricted to the statement; “.....it is expected that the three axle group with a load
of 40.7 kips (181.0 kN) would be equivalent (in terms of maximum deflection) to a single axle of 18 kip
(80.1 kN).”
6
In addition, the analyses did not encompass single-tyred single axles (steer axles) or triaxles. Steer axles were
considered to have caused minimal damage at the AASHO Road Test and, hence, were not included in the analyses.
Triaxles were not included in the AASHO performance studies.
AUSTROADS 2004
— 1.33 —
In an ARRB internal report (Scala 1970b), written approximately five months after the above paper, Scala
was much more focussed, stating that:
“Using an 18 kip single axle load (dual wheel) as the standard axle load, equivalent repetitions
of other axle loads are given by:
(i) Single axle (single wheel) (w/12)4
(ii) Single axle (dual wheel) ((w/18)4
(iii) Tandem axle (dual wheels) (w/30)4”
For the NAASRA Economics of Road Vehicle Limits (ERVL) Study, Stevenson (1976) adopted the
following values, based on the above two Scala references and discussions with him:
single-tyred single axle 5.4 t (53.0 kN)
dual-tyred single axle 8.2 t (80.4 kN)
dual-tyred tandem axle 13.6 t (133.4 kN)
dual-tyred triaxle 18.5 t (181.5 kN)
The IGPTD (1979) adopted the first three of the above values in its Table 2.15. It did not cater for triaxles –
most probably an oversight.
The WG, in its formulation of Table 7.1 in the Guide, reviewed the above material and, in addition, values in
use overseas. The largest discrepancy between the above values and those in use overseas was for the dual-
tyred tandem axle (see, for example, the values for AASHO and Asphalt Institute in Table VII of Scala
(1970a)). Further, the WG noted Scala’s later adoption of 13.7 t for tandem axles (Scala 1977). On these
grounds, the WG opted for the values presented in Table 7.1, as follows:
single-tyred single axle 53 kN
dual-tyred single axle 80 kN
dual-tyred tandem axle 135 kN
dual-tyred triaxle 181 kN
4.4.2 Derivation of “Standard Axle (or Traffic) Factors”

The derivation of “Standard Axle (or Traffic) Factors” in Sections 7.5.2 and 7.5.3 of the Guide and also of
the data in Table 8.3(a) and (b) of the Austroads (1992) Pavement Design Guide is documented in notes
prepared by the author and faxed to David Angell (formerly of MRDQ) on 1st April 1987.
Derivation of Tables 8.3(a) and 8.3(b)

ABS data was used to compile the following Table (Table 3) of axle group-km x 106 travelled annually (year
unknown) in non-urban areas according to type of axle group and State/Territory:
AUSTROADS 2004
— 1.34 —
Table 3
Distribution of Axle Group Type by State/Territory
Axle Group Type
State/Territory Single Axle Single Axle Tandem Axle Triple Axle

Single Tyres Dual Tyres Dual Tyres Dual Tyres
NSW 3,802 2,875 1,551 571
Vic 2,589 1,972 1,105 366
Qld 1,704 1,271 711 269
WA 1,309 1,044 517 167
SA 868 573 488 174
Tas 270 186 160 18
ACT 76 50 36 12
NT 146 80 132 88
Total 10,764 8,051 4,700 1665
The column totals give the proportions in Table 8.3(a) of the Austroads Guide.
Using RoRVL data, the data in Table 4 was compiled of percentage distribution of loads on axle groups
according to type of axle group and State/Territory.
Derivation of “Standard Axle Factors”

For axle group-km travelled by Single Axles with Single Tyres, that percentage which is travelled in NSW
with a 1 tonne load is: (3,802/10,745) x 0.5. (The first two numbers are given in italics in the left-hand
column of Table 3, whist the third number is shown in italics in the left hand column of Table 4 (AG1).)
Similarly, for Victoria, the percentage is (2,589/10,764) x 5.7. And so on for all States/Territories.
Summing these quantities for all States/Territories gives that percentage of the axle group-km travelled in
non-urban Australia by Single Axle Single Tyres for which the axle load is 1 tonne. The answer to this
summation is approximately 3.633%. Converting that to a proportion gives approximately 0.03633. The
number (but abbreviated to 0.036) forms the first entry in the first column of Table 8.3(b) of the Austroads
Guide.
Repeating this process for all loads on all axle group types for all States/Territories, and summing the results
(after rounding to 3 decimal places) gives the entries in Table 8.3(b).
The distribution of loads on Single Axle Single Tyres (column 1 of Table 8.3(b)) produces the same
subgrade distress as the following number of Standard Axles:
0.036(1/5.4)7.14 + 0.116(2/5.4)7.14 + 0.004(7/5.4)7.14 + 0.001(8/5.4)7.14 = 0.4091 (approx.)
If the entries in column 1 of Table 8.3(b) are not rounded to 3 decimal places, then this sum = 0.4227
(approx.).
A similar procedure is adopted for asphalt fatigue and cemented material fatigue, with the exponent used
being 5 and 18 respectively rather than 7.14 (as per the performance models in the Austroads Guide).
Similarly, ESAs would be derived using an exponent of 4.
AUSTROADS 2004
— 1.35 —
Table 4
RORVL Load Distributions on Axle Groups According to Axle Group Type
AG1: Single Axle, Single Tyres
Load (t) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
NSW 0.5 5.3 11.8 25.2 50.3 6.4 0.5 0.1
VIC 5.7 16.5 17.2 24.5 30.2 5.5 0.2 0.1 0.0
QLD 8.0 14.7 15.8 28.9 27.8 4.6 0.2 0.0
WA 1.7 14.8 23.9 30.7 23.9 4.1 0.5 0.1 0.1 0.1
SA 3.1 12.4 14.7 21.8 36.2 10.8 0.9
TAS 8.5 14.0 14.7 33.8 26.8 1.7 0.1 0.1 0.1 0.1
ACT 5.7 21.0 25.5 24.8 19.7 3.2
NT 8.0 11.8 13.5 31.2 29.5 5.6 0.5
TOT 5.1 13.6 16.7 27.6 31.1 5.4 0.4 0.0 0.0 0.0
AG2: Single Axle, Dual Tyres

Load (t) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
NSW 0.7 6.9 11.8 15.7 13.6 14.9 12.4 14.1 7.4 1.6 0.5 0.4
VIC 4.8 21.2 20.1 13.8 9.7 7.7 6.5 8.0 5.8 1.9 0.4 0.1 0.1
QLD 8.4 18.6 21.6 12.8 10.3 6.4 7.2 8.8 4.8 0.9 0.3 0.0
WA 2.7 14.4 21.8 18.8 10.3 9.6 6.8 7.3 4.9 2.6 0.5 0.4
SA 5.4 17.5 15.5 13.6 12.4 10.9 8.7 10.9 4.3 0.6 0.2
TAS 10.6 23.3 16.0 14.1 9.6 5.7 5.7 8.4 5.4 0.9 0.3
ACT 5.9 30.6 24.7 10.6 10.6 7.1 2.4 2.4 3.5 1.2 1.2
NT 12.8 18.5 21.9 14.6 9.7 7.1 5.4 4.8 2.8 1.5 0.7 0.1 0.1
TOT 6.0 18.0 19.5 14.6 10.5 8.4 7.2 8.6 5.1 1.5 0.4 0.2 0.0
AUSTROADS 2004
— 1.36 —
Table 4 (con’t)
AG3: Tandem Axle, Dual Tyres
Load (t) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
NSW 0.7 2.2 4.4 5.3 5.0 3.6 2.9 3.5 5.5 6.4 7.8 13.6 21.2 12.2 4.3 1.0 0.2 0.1 0.1
VIC 0.0 0.2 2.0 5.6 8.9 12.9 8.1 5.4 3.8 4.6 4.1 4.7 6.2 9.0 10.9 9.3 3.3 0.6 0.2 0.1 0.1
QLD 0.0 0.1 1.1 4.6 11.0 11.5 8.4 4.8 4.2 3.0 3.5 4.7 8.4 10.7 12.9 6.9 2.3 0.9 0.6 0.2 0.1
WA 0.1 0.9 6.7 12.0 8.1 8.4 6.0 4.1 2.9 3.2 3.6 5.1 5.8 7.2 7.9 8.4 5.6 2.5 0.9 0.4 0.1 0.0
SA 1.5 4.8 6.7 8.1 5.9 4.9 5.1 5.4 5.2 7.3 6.8 7.6 10.4 14.3 5.1 0.7 0.2 0.1
TAS 0.2 1.7 5.7 7.1 9.9 5.2 3.7 7.5 4.3 3.2 4.6 6.7 7.7 15.6 12.5 3.2 1.1 0.2
ACT 3.6 4.5 11.8 23.6 9.1 6.4 1.8 1.8 1.8 2.7 1.8 3.6 7.3 3.6 2.7 2.7 1.8 4.5 2.7 0.9 0.9
NT
TOT
AG4: Triple Axle, Dual Tyres

Load (t) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
NSW 0.1 1.7 3.6 2.3 2.1 2.1 1.2 1.8 2.8 1.7 3.3 4.9 5.5 5.5 9.9 20.7 14.7 7.7 4.6 1.3 0.5 0.6 0.6 0.6 0.2
VIC 1.1 10.5 12.5 6.5 4.5 3.4 4.0 1.5 2.2 2.5 2.4 3.4 6.2 5.5 12.0 11.0 6.7 2.9 0.8 0.1 0.1
QLD 0.1 1.9 8.9 10.8 7.1 4.6 3.8 2.5 3.1 1.7 2.3 2.5 3.8 4.8 7.5 9.9 11.1 5.6 3.6 1.8 1.2 0.7 0.2 0.3 0.2 0.1
WA 4.8 11.0 3.8 10.0 8.4 6.7 4.1 3.6 2.2 3.8 2.6 5.7 4.8 6.0 6.0 6.0 4.5 3.1 1.4 1.0 0.5
SA 2.0 5.7 4.4 4.8 2.7 2.7 3.0 2.0 2.7 5.5 4.4 4.8 10.1 8.3 11.9 10.7 8.5 3.6 1.1 0.9 0.2 0.2
TAS 0.5 2.3 13.2 6.8 6.8 2.7 1.8 1.4 3.2 2.3 3.2 5.9 7.8 5.0 10.0 11.9 10.0 4.6 0.5
ACT 17.6 11.8 11.8 5.9 5.9 5.9 11.8 5.9 11.8 5.9 5.9
NT 0.1 1.1 6.5 12.3 6.7 5.4 5.3 4.4 3.1 2.6 2.5 2.5 3.5 3.5 5.3 6.1 7.7 6.0 6.0 5.0 2.3 0.8 0.8 0.3 0.1
TOT 0.1 1.5 7.4 9.1 5.5 4.5 3.7 3.4 2.7 2.3 2.9 3.4 4.1 5.7 7.0 11.4 10.5 6.5 4.0 2.1 1.0 0.5 0.3 0.2 0.1 0.1
AUSTROADS 2004
— 1.37 —
By using non-rounded entries for Table 8.3(b) in conjunction with Standard Axle loads of 5.4 t, 8.2 t, 13.6 t
and 18.5 t for the four axle group types, and going through the above procedure for each axle group type and
each distress mode (and also for the exponent 4), the following tabulation (Table 5) of the number of
Standard Axles for the same distress as axle groups with (non-rounded) Table 8.3(b) load distributions is
derived.
Table 5
Number of Standard Axles for Same Distress as Axle Groups with (non-rounded) Load Distributions
Distress Exponent
Axle/Tyre Subgrade Asphalt Cemented ESA
(7.14) (5) (18) (4)
Single Axle Single Tyre 0.4227 0.4396 12.8507 0.4709
Single Axle Dual Tyres 0.3872 0.3421 4.9029 0.3457
Tandem Axle Dual Tyres 1.2742 0.9473 16.8339 0.8526
Triple Axle Dual Tyres 0.7630 0.6635 7.4317 0.6446
Now, if for each column of Table 5, the entries are weighted by the appropriate proportions in Table 8.3(a)
of the Guide, then the following is derived:
Distress Mode Number of Standard Axles per Axle

Group which Produce Same Distress
subgrade 0.5884
asphalt 0.5155
cemented 10.6451
and the number of ESAs per axle group = 0.5117. Dividing the above entries by 0.5117 gives:
No. of Standard Axles for same Subgrade distress: 1.1499 x no. of ESAs
No. of Standard Axles for same Asphalt distress: 1.0074 x no. of ESAs
No. of Standard Axles for same Cemented distress: 20.8034 x no. of ESAs
On the basis of these calculations, it was decided to adopt 1.1, 1.1 and 20 respectively for the above factors.
4.4.3 Cumulative Growth Factor for Estimation of Design Traffic

In the Guide, cumulative traffic over a design period is determined by multiplying the first-year traffic by the
Cumulative Growth Factor for the specific design period and annual growth rate of traffic.
The first-year traffic is 365 x initial AADT. (This is simply a re-statement of the definition of AADT.)
Values of Cumulative Growth Factor are provided in Table 7.2 of the Guide. For the case where the annual
growth rate is zero, the Cumulative Growth factor is equal to the design period (years). For the case where
the annual growth rate is non-zero, the value of the Cumulative Growth Factor is:
[(1 + GR/100)DP - 1] / (GR/100)
where GR is annual growth rate (%), and
DP is the design period (years)
The formula is simply the summation of the series:
1 + (1 + GR/100) + (1 + GR/100)2 + + (1 + GR/100)DP -1
The terms in this series are simply the ratios of (traffic in year i) / (traffic in year 1).
AUSTROADS 2004
— 1.34 —
4.5 Incorporation of Location-Specific Temperature Regime

There is (understandably, in the author’s view) often confusion about the roles and merits of using Pavement
Life Multipliers (PLMs) or, alternatively, Weighted Mean Annual Pavement Temperature (WMAPT) to
account for the influence of the location-specific temperature regime on the performance of asphalt
pavements. The reason for this confusion is very simple. There is no guidance provided on how to
determine a design temperature for a specific thickness of asphalt in a specific location when it is subjected
to a specific pattern of daily traffic. (While asphalt-surfaced granular pavements are catered for by the use of
PLMs (see below), asphalt-surfaced pavements containing cemented material are not catered for.)
WMAPT was included in the Guide as a necessary input factor in the design procedure for asphalt overlays.
It is to be interpreted as a “representative” temperature for a thin (50 mm) asphalt surfacing. Implicit in its
“representative” nature is an assumed diurnal distribution of traffic. Hence, WMAPT can be considered as
an appropriate design temperature for thin asphalt when subjected to normal variations in traffic intensity
throughout the day. Because of the lack of any alternative, it is commonly used as the location-specific
design temperature regardless of asphalt thickness and daily traffic pattern.
PLMs provide a means whereby the performance of an asphalt-surfaced granular pavement in a given
location, and with a specific day/night traffic split, can be readily assessed from the performance of the same
pavement when the design asphalt temperature is 25ºC.
For use in the General Mechanistic Procedure, the designer simply adjusts the value of design number of
Standard Axles for asphalt (NSA x 365 x GF in Section 7.5) by dividing it by the value for PLM determined
from Appendix B of the Guide. Having made this adjustment, the designer then adopts 25ºC as the design
asphalt temperature and proceeds in the standard manner. For a conventional mix subjected to highway-
speed traffic, the Guide has adopted 2,800 MPa as a representative design asphalt modulus when the design
asphalt temperature is 25°C. In essence, use of the PLM brings into the design procedure the fatigue life of
the asphalt applicable to the specific location and day/night split of traffic. The PLMs are not applicable to
pavements containing cemented material because the dominant distress mode for such pavements is, for the
vast majority of situations, fatigue cracking of the cemented material (see relevant Example Design Charts).
Hence, adjustments to asphalt fatigue life are of little relevance.
Because the Example Design Charts for asphalt-surfaced granular pavements have a design asphalt
temperature of 25ºC (modulus 2800 MPa), PLMs may be readily used in conjunction with these charts.
More specifically, they are applicable to those portions of the charts where asphalt thickness is less than
(say) 200 mm AND the dominant distress mode is fatigue of asphalt. In use, the designer simply enters the
Chart with a revised value for traffic given by the design traffic/PLM. The development and application of
PLMs is fully covered in Youdale (1984d).
AUSTROADS 2004
— 1.35 —
5. DEVELOPMENT OF OVERLAY DESIGN PROCEDURE

The overlay design procedure in the Guide (Chapter 10) is comprised of the following elements:
• A limit on the design value of the maximum deflection to contain permanent deformation during the
design period to within an acceptable range. This design maximum deflection decreases with increasing
design traffic and is in the Guide as Curve 1 of Figure 10.3. It performs the same role as the subgrade
strain criterion (for new pavements) and, in essence, acts as a surrogate for the strain criterion.
• A somewhat tighter limit on design maximum deflection – to be applied if the pavement contains a layer
of cemented material. Again, this design maximum deflection decreases with design traffic (Curve 2 of
Figure 10.3 in the Guide). Its role is to prevent fatigue cracking of the cemented layer during the design
period and, as such, it acts as a surrogate for maximum tensile strain at the bottom of the cemented layer.
• A limit on the design value of the Curvature Function (maximum deflection minus deflection 200 mm
distant from the location of the maximum deflection) – to be applied when the pavement is surfaced with
asphalt or is to be surfaced with asphalt. This design Curvature Function decreases with design traffic
(Figure 10.4 of the Guide) and has the role of preventing fatigue cracking in the asphalt during the design
period. As such, it acts as a surrogate for maximum tensile strain at the bottom of the asphalt.
• A basis for estimating the values of maximum deflection and Curvature Function for the pavement when
it is at its representative operating temperature (WMAPT) – from values measured at another temperature
(Figure 10.2 of the Guide).
• Bases for estimating the reduction in values of deflection parameters resulting from placement of an
overlay, specifically:
a) reduction in maximum deflection of a granular pavement due to placement of a granular overlay
(Figure 10.5);
b) reduction in maximum deflection of a chip-sealed or asphalt-surfaced granular pavement due to
placement of an asphalt overlay – the reduction being that achieved when the pavement temperature is
25°C (Figure 10.7); and
c) reduction in Curvature Function of a chip-sealed or asphalt-surfaced granular pavement due to
placement of an asphalt overlay –- applicable for the normal range of pavement temperatures (Figure
10.8).
• Having calculated the thickness of asphalt overlay necessary to reduce the maximum deflection to its
design value when the pavement temperature is 25°C, an adjustment is applied to this thickness so that
the required deflection reduction is achieved when the pavement is at its representative operating
temperature (WMAPT) (Figure 10.9).
The development of each of these elements is now discussed in turn.
5.1 Design Deflection Curves

Both Curves 1 and 2 in Figure 10.3 of the Guide have been taken directly from the IGPTD – with minor
extensions to the upper limit of design traffic from 3x107 to 108 ESAs. Hence, a review of their origins
entails a review of the basis for their incorporation in the IGPTD.
AUSTROADS 2004
— 1.36 —
The procedure in the IGPTD for design of overlays was based on the premise that all modes of distress
(permanent deformation within granular layers and the subgrade, fatigue cracking of asphalt and cemented
material) could be efficiently controlled by requiring that the maximum surface deflection (as measured by
the Benkelman Beam) be less than a specified tolerable value. The tolerable value decreased as both the
design traffic and composite stiffness of the pavement structure (excluding the subgrade) increased. Four
curves for tolerable deflection versus design traffic (ESAs) were presented in the IGPTD for four classes of
composite stiffness of the pavement layers. The origins of these four curves, and the carry-over of two of
them (the two extreme cases) to the Guide, is well documented in Jameson (1996) – the relevant excerpt of
which is attached as Appendix C. Figure C2 of Appendix C shows the four curves.
In the IGPTD, the labels attached to curves 2 and 4 are somewhat incomplete. A fuller account of the
intended use of the curves is provided in Table 2.2 of the IGPTD. This Table indicates that:
• With respect to deflection requirements when a stabilised layer was used, no distinction was made on the
basis of the type of stabilising agent, be it cement or lime or bitumen.
• Incorporation of a stabilised sub-base (cement or lime or bitumen) in a chip-sealed granular pavement
moved the deflection requirement down from Curve 1 to Curve 2. (For asphalt surfaced pavements,
incorporation of a stabilised sub-base did not affect the relevant curve.)
• For a chip-sealed pavement, if this stabilised material was placed over, instead of under, the granular
material (hence forming the base instead of the sub-base), then the deflection requirement moved down
from Curve 1 to Curve 4.
Hence, the Curve 2 label would be improved by adding the extension “OR stabilised sub-base with unbound
base and thin bituminous surfacing” and the Curve 4 label would be improved by replacing the word
“cemented” with “stabilised”. Also, a note indicating that “stabilised” refers to the addition of cement OR
lime OR bitumen is warranted.
5.1.1 Adoption of IGPTD Curve 1

In the Guide, the IGPTD Curve 1 has been adopted (with a minor extrapolation from design traffic of 3 x 107
ESAs to design traffic of 108 ESAs) as the basis for restricting the accumulation of permanent deformation in
the pavement and subgrade to within acceptable limits during the design life of the overlay.
In tracing the development of Curve 1, Jameson notes that Scala (1965) proposed a family of curves, with
the “family” consisting of distinct curves for each subgrade CBR (applicable for all cover thicknesses) OR
distinct curves for each thickness of cover (applicable for all subgrade CBRs) (Appendix C, Figures C6 and
C7). However, although being aware of Scala’s work, the IGPTD drafting committee opted for a single
curve (Curve 1).
Jameson further notes Youdale’s report to the WG (Youdale 1984c) indicating that a mechanistic analysis
supported the appropriateness of such a “family”, and also Anderson’s report (Anderson 1983b) to the WG
reviewing UK experience. Anderson drew the following conclusions:
“It can be concluded that the work of Lister et al. indicated no dependence of tolerable
deflections on pavement thickness for the pavements studied. It should be noted that all
pavements had asphalt surfacing and bases, but sub-base materials included granular and
cemented materials.
Although Lister’s data indicates that the tolerable deflections for pavements containing asphalt
surfacing and base are independent of pavement thickness, no clear conclusion can be drawn
for other pavements, particularly unbound pavements with thin bituminous seals (F1 type).
Lister’s data (ref. Figure 2) plot somewhere between curves 3 and below curve 4 of the
NAASRA Interim Guide Figure 2.8 and between the curves for 300 mm and 500 mm thick
pavements attached to G. Youdale’s report dated 20th December 1982. (Lister’s data was for
pavements of thickness 300 mm-600 mm).”
AUSTROADS 2004
— 1.37 —
Jameson, after clarifying that Lister’s data had been corrected to account for variations in subgrade CBR
(and hence related to essentially a single subgrade CBR), made the following statement:
“Consequently, contrary to the conclusions of the NAASRA Working Group, Lister’s data does
not support the use of a single design deflection for all granular thicknesses and CBRs”.
Bearing in mind that the intention of Anderson’s review was to shed light on appropriate tolerable
deflections for granular pavements with thin bituminous surfacings, and that part of his conclusions was “...
no clear conclusion can be drawn for other pavements, particularly unbound pavements with thin bituminous
seals (F1 type).”, the author cannot recall the WG drawing the imputed conclusion.
Although the WG had (and, most probably, still have) strong reservations about the appropriateness of
adopting the IGPTD Curve 1 to control permanent deformation in granular and subgrade materials –
regardless of the subgrade CBR value and cover thickness – the Curve was regarded by the then MEC
Review Committee as well-supported by field experience and hence warranted continued use. (This stance
continues to find support.) For this reason, the Curve was adopted by the WG for the role it plays in the
Guide.
5.1.2 Adoption of IGPTD Curve 4

The role of Curve 4 in the IGPTD was to prevent premature fatigue cracking of a substantial layer of bound
material, be it asphalt or cemented material.
It is the author’s understanding that the curve had its origins in mechanistic analyses of pavements with
cemented bases, supplemented by limited observations of field performance.
The curve was adopted by the WG (as Curve 2 in Figure 10.3 of the Guide) as a control against premature
cracking in cemented material for the reason (as stated in the Guide) that no satisfactory alternative measure,
derived from a deflection bowl, had been established at that time. There appears to have been little if any
improvement in this area since that time.
5.2 The Curvature Function

As noted above, the IGPTD adopted the stance that control of maximum deflection would prevent premature
fatigue cracking of asphalt. This is well evidenced by the role of its Curves 1-4 for tolerable deflection. For
asphalt-surfaced granular pavements, the relevant curve depends solely on asphalt thickness. Although there
were strong views at the time that a deflection ratio (commonly d150/d0, less commonly d200/d0) was superior
to maximum deflection (d0) as a predictor of tensile strain at the bottom of the asphalt layer, the prevailing
view was that its adoption at that time was not warranted.
In the late 1970s, Anderson (1979) undertook a comprehensive analysis of a broad range of asphalt-surfaced
granular pavements, using elastic layer analysis, to assess the relative merits of a variety of deflection bowl
parameters in estimating asphalt strain. The work formed part of a Master’s Degree undertaken at the
University of California at Berkeley and was supervised by Professor Carl Monismith. Anderson concluded
that the difference between the maximum deflection and the deflection at an offset of 200mm (d0-d200),
which he termed the Curvature Function, was considerably superior to other candidate parameters in
estimating asphalt strain.
A comparison of the relative merits of d0, d200/d0 and d0-d200 is shown in Figure 3 (after Anderson 1984a).
Anderson’s study also provided corroboration of maximum deflection as an adequate predictor of subgrade
strain. His Master’s thesis entailed assembling an overlay design procedure on these bases which was
subsequently implemented by the (then) Country Roads Board (CRB 1982). The procedure and its
development are comprehensively reported in Anderson (1984a) and Anderson and Kosky (1987).
The WG considered that the structure of this procedure was most appropriate for the design of asphalt
overlays, and adopted it for this purpose.
AUSTROADS 2004
— 1.38 —
5.3 Temperature Correction for Deflection and Curvature

The following is the author’s understanding of the evolution of Figure 10.2 in the Guide – for correction of
measured deflections and curvatures to account for differences between temperature at the time of
measurement and WMAPT:
• The format of the Figure is as initially developed by Anderson.
• Anderson initially proposed three curves – for asphalt thicknesses of 25, 50 and 100 mm – with the
curves to be used for the correction of both maximum deflection and curvature. These curves are
included as Figure 11 in Anderson (1984a). They were based on elastic layer analysis of pavements, with
the dependency of asphalt stiffness on temperature determined from the Shell nomographs. A loading
time of 0.2 sec was adopted to represent a test speed of 2 km/h. Details are provided in Anderson
(1982b). Anderson notes in this report that the resulting adjustments were compatible with those obtained
from the procedures in CRB Technical Bulletin 29 (CRB 1975) and in the IGPTD. These curves were
adopted for use by the CRB in its Interim Technical Bulletin on pavement strength evaluation and
rehabilitation (CRB 1983).
AUSTROADS 2004
— 1.39 —
Figure 3: Predictive ability of deflection parameters (Anderson, 1984a)
AUSTROADS 2004
— 1.40 —
• The WG, in reviewing the correction factors associated with these curves, undertook to make some
comparisons with SRA field experience in Queensland and New South Wales (Gordon 1984). Also, the
results of further elastic layer analyses (MRDQ 1984) were considered. The information from both these
sources related only to maximum deflection.
• Following a review of this additional information and a re-examination of his original data by Anderson
(1984b), the WG developed revised curves for the correction of maximum deflection while retaining the
original curves for correction of curvature. The revised curves are those for asphalt thicknesses of 50,
100 and 150 mm in Figure 10.3 of the Guide. The adjustments determined from the revised curves are
substantially less than those derived from the original curves.
• Jameson (1985), in an in-house review for the RCA (formerly CRB) of its then temperature correction
procedures (Anderson’s original chart for both maximum deflections and curvature), undertook an
extensive field study of the effect of temperature on both maximum deflection and curvature for
pavements with asphalt thickness ranging from 55 to 300 mm. His conclusions were:
• for correction of maximum deflection, the revised curves produced by the WG were appropriate for
the asphalt thicknesses they encompassed (up to 150 mm);
• for correction of maximum deflection for thick asphalt pavements, Anderson’s original curve for 100
mm of asphalt fitted the data well; and
• correction factors for curvature were similar to those for maximum deflection – hence, the one
correction chart would satisfy both requirements
Jameson’s recommendations were in line with his conclusions and were supported by Colin Kosky (the then
Pavements Engineer). The WG reviewed Jameson’s report and endorsed its conclusions and
recommendations, resulting in the chart as presented in Figure 10.3 of the Guide.
5.4 Reduction in Deflection Parameters due to Overlay Placement

5.4.1 Reduction in Maximum Deflection due to a Granular Overlay
Figure 10.5 of the Guide provides a basis for determining the maximum deflection of a granular pavement
after placement of a granular overlay, given the deflection before overlay and the overlay thickness. The
family of lines plotted are represented by the equation:
da/db=0.94(T/25)
where da = deflection after overlay,
db = deflection before overlay, and
T = overlay thickness(mm).
As such, they represent a deflection reduction of 6% per 25mm of overlay thickness.
The author’s reconstruction of events leading to the adoption of 6% per 25mm is as follows:
• The IGPTD adopted 10% per 45 mm. This is equivalent to 5.7% per 25 mm.
• DMR, NSW adopted the IGPTD chart for its pavement design manual (Form 76).
• Youdale (1980a), based on CIRCLY analyses, recommended that DMR change to 3.5% per 25 mm.
• Wyman (1981), based on elastic layer analyses (Chevron), reported to MRD, Queensland that 4% per 25
mm was appropriate.
• The May 1982 Meeting of the WG generally agreed that 3.5% per 25 mm was appropriate.
• Youdale (1982b) undertook further CIRCLY analyses, concluding that 3.5% per 25 mm be adopted for
the Guide.
• Potter and Donald (1984) reported the forthcoming Guide would adopt 4% per 25 mm.
AUSTROADS 2004
— 1.41 —
It is the author’s (somewhat hazy) recollection that:

• RCA, Victoria, remained uneasy with the values of 3.5-4% per 25 mm, considering them to be too low;
and
• the WG later accepted the RCA stance, opting initially to retain the IGPTD relationship and, finally,
“rounding” it to 6% per 25 mm.
5.4.2 Reduction in Maximum Deflection (at 25°C) due to an Asphalt Overlay

Figure 10.7 of the Guide provides a basis for estimating – when the asphalt temperature is 25°C – the
deflection after placement of an asphalt overlay, from deflection before overlay and overlay thickness. The
Figure is based on a deflection reduction of 10% per 25 mm of overlay. The information available to the
WG which led to its adoption of this percentage reduction is as follows:
• In the IGPTD, percentage reductions per 25 mm ranged from 12 to 6 depending on climatic region. The
percentages applicable to Sydney and Melbourne were 8 and 10 respectively (WMAPTs for Sydney and
Melbourne are 28.0 and 24.2 respectively).
• Youdale (1980b), on the basis of elastic layer analyses, recommended 8% for areas of typical usage of
asphalt within NSW and 10% in cooler areas.
• Anderson (1982b) adopted 10% on the basis of elastic layer analyses supported by field data. Figure 11
of Anderson (1984a) refers.
• Wyman (1982), on the basis of field observations, recommended percentage reductions per 25 mm
ranging from 12 (for 50 mm overlay thickness) to 20 (for 100 mm overlay thickness).
• Gordon (1982) presented results of elastic analyses which indicated that a value of 5.5% was appropriate.
• The WG sought field data from HD SA, MRD WA, MRD Tas, and DTW NT (Anderson 1982c). HD SA
responded with data which supported a value of 9% (Cops 1982).
After consideration of this assembled information, it was apparent to the WG that adoption of a value up to
10% reduction in deflection (at 25°C) for each 25mm of asphalt overlay was consistent with the available
data.
5.4.3 Reduction in Curvature Function due to an Asphalt Overlay

Figure 10.8 of the Guide provides a basis for estimating the reduction in Curvature Function due to
placement of an overlay. The Figure is based on a 20% reduction per 25 mm of overlay. The value was
determined from elastic layer analyses undertaken by Anderson during the development of the overlay
design procedure. (Anderson 1984a).
5.5 Adjustment of Asphalt Overlay Thickness to Allow for Locality Temperature

Figure 10.9 of the Guide provides a basis for correcting the value of overlay thickness which provides the
required reduction in maximum deflection when the asphalt is at 25°C to a thickness value which will
provide this same reduction when the asphalt is at its operating temperature (WMAPT). The Figure
incorporates the temperature dependency of deflection reduction in a more quantified manner than Table 2.3
in the IGPTD. The Figure was originally developed by Anderson (1984a) for a base WMAPT of 20°C and
subsequently re-drawn for a base WMAPT of 25°C. The author has no record of the underlying basis.
For reasons indicated in the Guide, no correction for locality temperature is applied to the value of overlay
thickness which provides the required reduction in curvature when the asphalt is at 25°C.
AUSTROADS 2004
— 1.42 —
REFERENCES CITED IN THE TEXT

Anderson, D.T. (1979). The Design of AC Overlays for Flexible Highway Pavements. M.Sc. Thesis,
University of California at Berkeley.
Anderson, D.T. (1981). The Basis of the Country Roads Board Pavement Design Process. Proc. LGE
Conference. Melbourne, Australia. March.
Anderson, D.T. (1982a). Study of Asphalt Fatigue Relationship. Report to WG. October 16. 17pp.
Anderson, D.T. (1982b). Pavement Analysis and Asphalt Overlay Design Procedures. Report to WG. April
26, 28 pp.
Anderson, D.T. (1982c). Results of a Survey of SRAs Seeking Information on the Effect of Asphalt Overlays
on Pavement Deflections. Report to WG, December 29.
Anderson, D.T. (1983a). Review of NAASRA IGPTD. Report to WG. November 16. 5pp.
Anderson, D.T. (1983b). Notes on the Effect of Pavement Thickness on Tolerable Deflection. Report to WG.
Anderson, D.T. (1984a). A New Approach to Asphalt Overlay Design. Proc. 12th ARRB Conference 12 (3),
pp. 54-67.
Anderson, D.T. (1984b). Effect of Asphalt Temperature on Pavement Deflection and Asphalt Overlays.
Report to WG February 29, 4 pp.
Anderson, D.T. and Kosky, C.K. (1987). Advances in Asphalt Overlay Design Procedures. Proc. 6th Int.
Conf. on Structural Design of Asphalt Pavements. Vol. 1, pp. 748-61.
Angell, D. (1988). Technical Basis for the Pavement Design Guide. Report RP 1265. Pavements Branch.
Department of Main Roads, Queensland.
Austroads (1992). Pavement Design – A Guide to the Structural Design of Road Pavements. Austroads,
Sydney.
Austroads (1997). Revision Note for Pavement Design – A Guide to the Structural Design of Road
Pavements. APRG, November.
Baran, E. and Aubrey, S.R. (1978). Preliminary Report on Pavement Thickness Design Curves for
Queensland Conditions Based on Elastic Layer Methods. Report RP531. Materials Branch, Main Roads
Department., Qld. October.
Black, D.J. (1977). Proposed NAASRA Publication – Manual of Pavement Thickness Design. Letter from
NAASRA Engineer-Secretary to ARRB, May.
Britton, A.T. (1947). Flexible Pavements – Design and Selection. Proc. 27th Meeting Highway Research
Board.
Cops, P.L. (1982). Information Forwarded in Response to Anderson (1982c). In Anderson (1982c).
Country Roads Board, Victoria (1945). Annual Report. Chief Engineer’s Report. Country Roads Board,
Melbourne, Australia.
Country Roads Board, Victoria (1949). The Design of Flexible Pavements. Technical Bulletin No. 4.
Country Roads Board, Melbourne, Victoria.
AUSTROADS 2004
— 1.43 —
Country Roads Board, Victoria (1975). Deflection Testing using the Benkelman Beam. Technical Bulletin
No. 29.
Country Roads Board, Victoria (1983). Pavement Strength Evaluation and Rehabilitation. Interim Technical
Bulletin. Country Roads Board, Melbourne, Victoria
Davis, E.H. (1949). The California Bearing Ratio Method for the Design of Flexible Roads and Runways.
Géotechnique 1(4), pp. 249-63.
George, H.P. and Gittoes, C.A. (1959). Report to XIth PIARC Congress on Section 2, Question VI, Part A.
PIARC.
Gerrard, C. M. (1969). Table of Stresses, Strains and Displacements in Two-layer Elastic Systems under
Various Traffic Loads. Australian Road Research Board. Special Report, SR No. 3.
Gerrard, C. M. and Wardle, L.J. (1976). Tables of Stresses, Strains and Displacement in Three-layer Elastic
Systems under Various Traffic Loads. Australian Road Research Board. Special Report, SR No. 4.
Gordon, R.G. (1982). Figure 15 from Report RP 649. Materials Branch. Main Roads Department, QLD.
Gordon, R.G. (1984). Summary Plots submitted to the WG for its Consideration. Report to WG.
Heukelom, W. and Klomp, A.G.J. (1962). Dynamic Testing as a Means of Controlling Pavements During
and After Construction. Proc. Int. Conf. on the Structural Design of Asphalt Pavements, Univ. Michigan,
Ann Arbor, pp. 667-79.
Irick, P.E. and Hudson, W.R (1964). Guidelines for Satellite Studies of Pavement Performance. NCHRP
Report No. 2, Highway Research Board.
Jameson, G.W. (1985). Temperature Adjustments of Pavement Deflections and Curvatures. Report No.
58M171 to Group Manager-Materials, April 16. RCA.
Jameson, G.W. (1995). Response of Cementitious Pavement Materials to Repeated Loadings. ARRB
Transport Research, Contract Report RI 949, March.
Jameson, G.W. (1996). Origins of Austroads Design Procedures for Granular Pavements. ARRB Transport
Research, Research Report ARR No. 292.
Jameson, G.W., Sharp, K.G. and Yeo, R. (1992). Cement-Treated Crushed Rock Pavement Fatigue Under
Accelerated Loading: The Mulgrave (Victoria) ALF Trial, 1989/1991. Australian Road Research Board.
Research Report ARR No. 229.
Jameson, G.W., Dash, D.M., Tharan, Y. and Vertessy, N.J. (1995). The Performance of Deep-lift In-situ
Pavement Recycling under Accelerated Loading: the Cooma ALF Trial 1994. APRG Report No. 11.
Austroads
Kenis, W.J., Sherwood, J.A., and McMahon, T.F. (1981). Verification and Application of the VESYS
Structural Subsystem. Proc. 5th Int. Conf. on Structural Design of Asphalt Pavements, Ann Arbor, Michigan.
Litwinowicz, A. (1982). Fatigue Characteristics of Cement Treated Materials: an Overview. Report RP 752.
Materials Branch. Main Roads Dept., Qld, October.
Litwinowicz, A. (1984). Private Communication to Author. March.
MacLean, D.J. (1959). Report to XIth PIARC Congress on Section 1, Question 1, Part A. PIARC.
AUSTROADS 2004
— 1.44 —
Main Roads Department, Queensland (1982). Fatigue Characteristics of Cement Treated Materials – an
Overview. Materials Branch Report RP 533.
Main Roads Department, Queensland (1984). Report on Temperature Correction Factors for Rebound
Benkelman Beam Deflections. Materials Branch Report RP 826.
Mitchell, J.K. (1976). The Properties of Cement Stabilised Soils. Workshop on Materials and Methods for
Road, Rail and Reclamation Works, Leura, NSW. University of NSW, September.
National Association of Australian State Road Authorities (1979). Interim Guide to Pavement Thickness
Design. NAASRA. Sydney.
National Association of Australian State Road Authorities (1987). Pavement Design – A Guide to the
Structural Design of Road Pavements. NAASRA. Sydney.
Otte, E. (1978). A Structural Design Procedure for Cement-Treated Layers in Pavements. D.Sc.(Eng.)
Thesis, University of Pretoria, May.
Porter, O.J. (1938). The Preparation of Subgrades. Proc. Highway Res. Board 18(2), pp. 324-31.
Potter, D.W. (1981). Suggested Research Areas in Flexible Pavement Design. Proc. NAASRA/ARRB
Seminar on Heavily Trafficked Flexible Pavements. ARRB Internal Report, AIR 000-168.
Potter, D.W. (1997). Appropriate Laboratory Fatigue Testing of Asphalt for Australia and its Role in
Australian Pavement Design. ARRB TR Working Document R97/021, April.
Potter, D.W. and Donald, G.S. (1984). Revision of NAASRA Interim Guide to Pavement Thickness Design.
Paper to Workshop on Structural Design of Road Pavements, 12th ARRB Conference.
Pretorius, P.C. (1969). Design Considerations for Pavements Containing Soil-Cement Bases. Ph.D.
Dissertation. University of California, Berkeley.
Pretorius, P.C. and Monismith, C.L. (1972). Fatigue Crack Formation and Propagation in Pavements
Containing Soil-Cement Bases. Highway Research Record No. 407.
Rallings, R.A. (1997). APRG Workshop on Structural Behaviour of Unbound Granular Pavements. Report
APRG 97/03(DA).
Road Research Laboratory. (1955). Construction of Housing Estate Roads Using Granular Base and
Subbase Materials. Road Note 20. HMSO, London.
Rodway, B. (1997). Going ‘Round in Circlies. Geomechanics Society Pavements Symposium, April.
Scala, A.J. (1965). CBR Design Method Deflection Dependency. ARRB Internal Report AIR 010-2.
Scala, A.J. (1970a). Comparison of the Response of Pavements to Single and Tandem Axle Loads. Proc. 5th
ARRB Conf.5(4), pp. 231-52.
Scala, A.J. (1970b). Prediction of Repetitions on Roads. ARRB Internal Report, November.
Scala, A.J. (1977). Preliminary Study of a Pavement Management System. ARRB Internal Report AIR 175-
1, April.
Sparks, G.H. (Ed.) (1981). Proceedings of NAASRA/ARRB Seminar on Heavily Trafficked Flexible
Pavements. ARRB Internal Report AIR 000-168.
Sparks, G.H. and Potter, D.W. (1982). An Investigation into the Relationship Between California Bearing
Ratio and Modulus for Two Clays. ARRB Internal Report AIR 295-1.
Stevenson, J. McL. (1976). Pavements. Study Report T4, ERVL. NAASRA.
AUSTROADS 2004
— 1.45 —
Vuong, B. (1994). Prediction Versus Performance of a Granular Pavement Tested with the Accelerated
Loading Facility (ALF). Proc. Symp. on Prediction Versus Performance in Geotechnical Engineering,
Bangkok. A.A. Balkema, Rotterdam.
Walker, R.N., Paterson, W.D.O., Freeme, C.R., and Marais, C.P. (1977). The South African Mechanistic
Pavement Design Procedure. Proc. 4th Int. Conf. on Structural Design of Asphalt Pavements, Vol. 2.
Wardle, L. J. (1977). Program CIRCLY User’s Manual. CSIRO Division of Applied Geomechanics.
Wyman, A.C. (1981). The Development of Asphalt and Granular Overlay Design and Tolerable Deflection
Levels by Elastic Analysis. Report RP 649. Materials Branch, Main Roads Department, Queensland.
Wyman, A.C. (1982). Empirical Investigation into the Development of an Asphalt Overlay Design Method
for Local Conditions. Report RP 716. Materials Branch, Main Roads Department, Queensland.
Yandell, W.O. (1981). Applications of the Mechano-lattice Analysis in Materials Engineering. Proc. 2nd
Aust. Conf. on Engineering Materials, Sydney. pp. 401-19. University of New South Wales.
Youdale, G. P. (1978). Repeated Load Triaxial tests on Granular Pavement Materials. Materials Research
Laboratory Test Report No. RS21 PTII. DMR, July 26.
Youdale, G.P. (1980a). MRD Form 76 – Pavement Thickness Design, January 1980. Reduction of Surface
Deflection due to Overlay with Granular Material. Report to Materials and Research Engineer. DMR NSW.
Youdale, G.P. (1980b). MRD Form 76 – Pavement Thickness Design. January 1980. Reduction of Pavement
Surface Deflection due to Overlay with Asphaltic Concrete, or Replacement of Granular Material with
Asphaltic Concrete. Report to Materials and Research Engineer. DMR NSW.
Youdale, G.P. (1981). Materials Testing for the Analysis of Heavily Trafficked Flexible Pavements. Proc.
NAASRA/ARRB Seminar on heavily trafficked flexible pavements. ARRB Internal Report, AIR 000-168.
Youdale, G.P. (1982a). Investigation into the Deflection Design Criteria for Granular Pavement Overlays.
Report to WG, December, 5 pp.
Youdale, G.P. (1982b). Investigation of Granular Overlay Deflection Reduction Factors. Report to WG,
November 11, 8 pp.
Youdale, G.P. (1983). Investigation of the Variation of Stiffness with Depth of a Granular Layer under
Variable Thickness of Asphaltic Concrete. Report to WG, July 8, 8 pp.; October 25, 5 pp.
Youdale, G.P. (1984a). Investigation of the Effects of and the Interaction Between the Stress Dependency of
Moduli and the Anisotropy of Granular Pavement Materials on the Results of Pavement Analysis using
CIRCLY. Report to WG, April 13, 7 pp.
Youdale, G.P. (1984b). Investigation of the Variation of Stiffness with Depth of a Granular Layer over a
Bound Subbase Layer. Reports to WG, March 9, 9 pp.
Youdale, G.P. (1984c). Review of Limiting Subgrade Strain Criteria. Report to WG, April 13, 7 pp.
Youdale, G.P. (1984d). The Design of Asphalt Pavements for Particular Temperature Environments. Proc.
12th ARRB Conf. 12(3), pp. 78-88.
AUSTROADS 2004
— 1.46 —
REFERENCES REVIEWED BY WORKING GROUP IN ASSESSING

RELEVANCE OF ANISOTROPY
Allen, J.J. and Thompson, M.R. (1974). Resilient Response of Granular Materials Subjected to Time-
Dependent Lateral Stresses. Transportation Research Record 510, pp. 1-13.
Barrett, J.R. and Smith, D.M. (1976). Stress History Effects in Base-course Materials. Proc. 8th ARRB Conf.
8(7), pp. 30-39.
Dehlen, G.L. and Monismith, C.L. (1969). Effect on Non-linear Material Response on the Behaviour of
Pavements under Traffic. NITRR Res. Rep. No. 122.
Freeme, C.R. (1970). Anisotropic and Non-Linear Characterization of Bituminous Mixtures (Part 1). NITRR
RB/17/70.
Freeme, C.R. (1971). Application of Sonic and Ultrasonic Test Methods to the Measurement of Elastic
Moduli of Pavement Materials. NITRR RB/2/71.
Gerrard, C.M. and Wardle, L.J. (1973). Some Aspects of the Design of Surface Pavement Layers. Research
Paper 208, CSIRO, Melbourne, 36 pp.
Gerrard, C.M. and Wardle, L.J. (1980). Design of Surface Pavement Layers. Australian Road Research
10(2), pp. 3-15.
Moore, P.J. (1980). Behaviour of Layered Pavements. Australian Road Research Board, Special Report, SR
No. 19.
Morgan, J.R. and Gerrard, C. (1973). Anisotropy and Non-linearity in Sand Properties. Proc. 8th Int. Conf.
on Soil Mechanics and Foundation Engineering. 1(2), Paper 1/43, pp. 287-92.
Paterson, W.D.O. (1974). Evaluating the Structural Response of a Pavement and the Moduli of Materials: a
Review. NITRR RP/1/74.
Richards, B.G. (1974). The Analysis of Flexible Road Pavements in the Australian Environment – Stresses,
Strains and Displacements under Traffic Loading. Division of Applied Geomechanics Technical Paper 20,
CSIRO, Melbourne, 20 pp.
Thrower, E.N. (1978). Stress Invariants and Mechanical Testing of Pavement Materials. Transport and Road
Research Laboratory, Laboratory Report LR 810, 48 pp.
Wallace, K. and Monismith, C.L. (1980). Diametral Modulus Testing on Non-Linear Pavement Materials.
Proc. AAPT. Vol. 49, pp 633-52.
White, P.D. (1979). An Investigation of Repeated Triaxial Loading Tests in Relation to Pavement
Performance. Unpublished M.Eng. Thesis, Univ. Melbourne, 194 pp.
AUSTROADS 2004
— 1.47 —
BIBLIOGRAPHY FOR FATIGUE OF ASPHALT

Brown, S.F., Pell, P.S., and Stock, A.F. (1977). The Application of Simplified, Fundamental Design
Procedures for Flexible Pavements. Proc. 4th Int. Conf. on the Structural Design of Asphalt Pavements, Ann
Arbor, Michigan.
Deen, R.C. et al. (1972). Structural Analysis of Bituminous Concrete Pavements. Highway Research Record
No. 407.
Kingham, R.I. (1971). Fatigue Criteria Developed from the AAHSO Road Test Data. Asphalt Institute RR
71-1.
Monismith, C.L. (1978). Design and Construction of Flexible Pavements. Notes of residential course, Leura.
University of New South Wales. (Data used in Asphalt Institute Design method for full depth asphalt MS-
11.)
Paterson, W.D.O. and Maree, J.H. (1978).An Interim Mechanistic Procedure for the Structural Design of
Asphalt Pavements. NITRR RP/5/78.
Pell, P.S. and Taylor I.F. (1969). Asphaltic Road Materials in Fatigue. Proc. AAPT February.
Santucci, L.E. (1977). Thickness Design procedure for Asphalt and Emulsified Asphalt Mixes. Proc. 4th Int.
Conf. on the Structural Design of Asphalt Pavements, Ann Arbor.
Shell (1978). Pavement Design Manual.
AUSTROADS 2004
— 1.48 —
APPENDIX A
ORIGINS OF UNBOUND GRANULAR THICKNESS CHART

The following is an excerpt from Origins of the Austroads Design Procedures for Granular Pavements
(Jameson 1996). The Figure and Table numbers have been to changed to reflect that the excerpt is Appendix
A of this report.
The Austroads (1992) thickness design chart for granular pavements with thin bituminous surfacings is
shown in Figure A.1. The origins of this chart can be traced back to the Californian State Highways
Department CBR method of pavement design (Porter 1942). From 1928-1942, the Department examined the
quality and thicknesses of base, sub-base and subgrade materials under both failed and sound sections of
flexible pavements throughout the California highway system. From these data, curves were formulated for
determining the total depth of construction (base, sub-base and imported fill) required to carry the anticipated
traffic. The resulting design curves are given in Figure A.2.
In 1945 the Victorian Country Roads Board (CRB) proposed a tentative thickness design chart (Figure A.3)
which seems to have been based on the Californian design curves (Gawith and Perrin 1962). This design
procedure was an improvement on the 1942 Californian procedure in that it quantified the traffic, provided
factors which allowed for transverse distribution of traffic, and a factor to correct thickness for rainfall. This
method was refined further when the CRB issued Technical Bulletin 4 in 1949. This was used by the CRB
until Technical Bulletin 21 was issued in 1960, as discussed below.
Figure A.1: Austroads design chart for granular pavements with

thin bituminous surfacings (Figure 8.4 of Austroads 1992)
AUSTROADS 2004
— 1.49 —
Figure A.2: California State Highway Department 1940's CBR method thickness design curve
(RRL 1952; Porter 1942)
Figure A.3: 1945 Victorian Country Roads Board tentative thickness design curves
(Gawith and Perrin 1962)
AUSTROADS 2004
— 1.50 —
In the late 1940s the U.K. Road Research Laboratory (RRL) compared the total pavement thicknesses
required by the Californian CBR method with actual thicknesses of roads of various condition (Davis 1949).
The results are shown in Figure A.4. Data was examined from seven sites where at least part of the road was
distressed due to deformation of the subgrade. In making the comparison with the Californian curves it was
considered that:
• the Californian design curve for a maximum wheel load of 7,000 lb was equivalent to less than 50
commercial vehicles per day (light traffic);
• the Californian design curve for a maximum wheel load of 9,000 lb was equivalent to medium traffic of
50 to 300 commercial vehicles per day (medium traffic ); and
• the Californian design curve for a maximum wheel load of 12,000 lb was equivalent to more than 300
commercial vehicles per day (heavy traffic).
Davis concluded that:
“Evidence of the validity of the design curve is provided by the fact that all "critical condition"
points lie close to the 45° line, all the 'definite failure' points lie below the line and all the "no
failure" points lie above the line. The number of points in this figure (Fig. 4) is hardly sufficient
to provide conclusive evidence and further investigations of this type are desirable.”
Figure A.4: Early U.K. RRL data regarding the validity of the Californian CBR method of pavement design
(Davis 1949)
Several years later MacLean (1954) reported that the design curves A-F in Figure A.5 were being considered
for use by the RRL. According to MacLean:
“ The form of these curves is based on a consideration of the results of full-scale road
experiments carried out by the Laboratory and of information supplied by county road
authorities who have applied the Californian bearing ratio method of design in normal road
construction. Six curves A, B, C, D, E and F are shown relating the thickness of construction to
the Californian bearing ratio of the sub-soil for roads carrying six different intensities of traffic.
This classification of traffic into 6 groups is based on the number of vehicles using a road per
day having a loaded weight exceeding 3 tonnes.”
AUSTROADS 2004
— 1.51 —
Curve A is a new curve which has been proposed for roads carrying 0-15 vehicles per day
weighing more than 3 tons. It refers to cul-de-sacs on housing estates and to isolated roads
built in connection with limited private housing development.
Curve B is another new curve for roads carrying 15-45 vehicles per day weighing more than 3
tons. It refers to minor through roads on housing estates which carry a fair amount of traffic
during the period of house construction but which carry no heavy lorry traffic or public service
vehicles subsequently.
Curve C for roads carrying 45-150 vehicles per day weighing more than 3 tons, has been in use
for many years. It applies to lightly trafficked county roads and to roads on housing estates
carrying up to 50 public service vehicles per day together with a fair number of tradesman's
vehicles.
Curve D has also been in use for some time and is for roads carrying 150-450 vehicles per day
weighing more than 3 tons. It refers to county roads carrying a medium intensity of traffic and to
main roads in urban areas where form 50-150 public service per day are operating.
Curve E, for roads carrying 450-1500 vehicles per day weighing more than 3 tons, has also
been in use for many years. It refers to principal shopping streets in large towns and to main
county roads.
Curve F is another new curve for heavily-trafficked truck roads carrying 1500-4500 vehicles per
day weighing more than 3 tons. The need for this curve has become apparent as the result of
investigations of structurally weak sections of roads and of full-scale experiments on truck
roads.
Figure A.5: Proposed CBR design curves for different classes of roads
(MacLean 1954 and 1959)
AUSTROADS 2004
— 1.52 —
It should be noted that curves C, D and E were very similar to Californian CBR design curves (Figure A2).
Note that the traffic loadings associated with these three curves differ from those initially adopted by Davis
(1949). In 1959 MacLean reported that curve G was being used for new roads "which may carry traffic in
excess of 4,500 commercial vehicles per day."
In 1955 the RRL published Road Note 20 Construction of Housing-Estate Roads using Granular Bases and
Subbases Materials. For such roads design curves A-E were proposed.
As stated by Leigh and Croney (1972), the Figure A5 design curves:
“...provided a means for estimating the total thickness of construction necessary for various
traffic and foundation conditions, but gave no guidance on the relative thicknesses of surfacing,
base and subbase.”
Accordingly, a series of full-scale experiments of in-service roads was conducted to examine the
performance of roads with variations in materials and layer thicknesses. By 1960 sufficient data was
available to issue preliminary design standards and these were contained in RRL's Road Note 29 A Guide to
the Structural Design of Pavements for New Roads. This document superseded Road Note 20 for roads with
a traffic loading of more than 150 commercial vehicles per day. In 1965 the second edition of Road Note 29
was extended to lightly-trafficked roads and the use in Britain of Road Note 20 presumably ceased.
In 1960, the CRB adopted (CRB 1960) the 1959 RRL design curves (Figure A.5). These curves were
revised in 1969 (CRB 1969) to provide higher minimum pavement thicknesses (Figure A.6). It was also
specified that the curves were only applicable for pavements in rural areas, that is granular pavements with a
sprayed seal surface.
Figure A.6: 1969 Country Roads Board pavement thickness design curves for roads in rural areas
(adapted from MacLean 1959)
AUSTROADS 2004
— 1.53 —
At that time traffic loadings were expressed in terms of:

“ the average number of commercial vehicles (CV) exceeding three tons in weight
(approximately this means vehicles with dual tyres on one or more axles) which the road is
expected to carry in 24 hours at some time towards the end of its life, e.g. in about 20 years
time. This is the total traffic in both directions of a two lane pavement or on both
carriageways of a divided highway." (CRB 1969).”
Using this design commercial vehicle volume, the road was considered to be in one of seven RRL traffic
categories as indicated in Table A.1.
Table A.1
Traffic Loading Characterisation
Traffic Two-Way Two-Way One-Way

Category CV/day CV/20 years ESAs/20 years
A 0-15 7 x 104 3 x 104
B 15-45 2 x 105 1 x 105
C 45-150 7 x 105 3 x 105
D 150-450 2 x 106 1 x 106
E 450-1500 7 x 106 3 x 106
F 1500-4500 2 x 107 1 x 107
G >4500 7 x 107 3 x 107
When the granular thickness chart was adopted by NAASRA (Figure 2.2 of NAASRA 1979), the
characterisation of traffic loading was converted from traffic categories to cumulative equivalent standard
axles (ESAs) over the design period. Based on Black's (1977) explanatory notes it seems the following
conversion procedure was adopted:
• The mid-range values of two-way CVs towards the end of the design period were divided by 1.5 to derive
the two-way CVs on opening (the factor of 1.5 is equivalent to a compound growth rate of 3% over 10-15
years; this factor appears to have obtained from CRB Technical Bulletin 21).
• Assuming a growth rate of 3%, the two-way CVs over 20 years were determined.
• The traffic was equally divided between the two directions to estimate the one-way CVs over 20 years.
• One commercial vehicle equalled one Equivalent Standard Axle.
The cumulative ESA values so determined are given in Table A.1.
It should be noted that the minimum pavement thicknesses of the NAASRA granular thickness chart (Figure
A.7) are greater than those derived from the RRL (1959) curves and different from the CRB (1969) values.
These minimum thicknesses only influence the thickness adopted for the unusual cases where a design
subgrade CBR exceeds 15. The Austroads 1992 granular thickness chart (Figure A.1) is similar to the
NAASRA, except for changes to the minimum pavement thickness and the design traffic range.
A limited survey (Potter et al. 1996) of experienced engineers on design reliability has indicated that
pavements designed with the Austroads (1992) granular thickness chart have a low probability of premature
distress. There was a wide scatter of responses in the survey with an average response being that pavements
designed in accordance with the chart had about a 90% probability of exceeding the design traffic.
AUSTROADS 2004
— 1.54 —
Figure A.7: 1979 NAASRA design chart
AUSTROADS 2004
— 1.55 —
REFERENCES
Austroads (1992). Pavement Design – A Guide to the Structural Design of Road Pavements. (Austroads:
Sydney.)
Black, D.J. (1977). Proposed NAASRA Publication "Manual of Pavement Thickness Design”. Letter from
Country Roads Board of Victoria (1960). The Design of Flexible Pavements. Technical Bulletin 21.
Country Roads Board of Victoria (1969). The Design of Flexible Pavements. Technical Bulletin 26.
Davis, E.H. (1949). The California Bearing Ratio Method for the Design of Flexible Roads and Runways.
Geotechnique I(4) December, pp. 249-63.
Dorman, G.M. and Metcalf, C.T. (1965). Design Curves for Flexible Pavements based on Layered System
Theory. Highway Research Board, Washington, Record No. 71.
Gawith, A.H. and Perrin, C.C. (1962). Development in the Design and Construction of Bituminous Surfaced
Pavements in the State of Victoria, Australia. Proc. Int. Conf. Structural Design of Asphalt Pavements.
Leigh, J.V. and Croney, D. (1972). The Current Design Procedures for Flexible Pavements in Britain. Proc.
3rd Int. Conf. on Structural Design of Asphalt Pavements, pp. 1039-48.
MacLean, D.J et al. (1959). Permanent International Association of Road Congresses XIth Congress, Rio de
Janiero, Section 1, Question 1, Report 7.
NAASRA (1979). Interim Guide to Pavement Thickness Design. (NAASRA Sydney.)
Porter, O.J. (1942). Foundations for Flexible Pavements. Proc. Highw. Res. Board, Washington DC, 22, pp.
100-36.
Potter, D.W, Jameson, G.W, Makarov, A, Moffatt, M.A. and Cropley, S.M. (1996). A Basis for
Incorporating Reliability in the Austroads Pavement Design Procedures. ARRB TR WD TI96/014.
Road Research Laboratory (1952). Soil Mechanics for Road Engineers. (HMSO: London.)
Road Research Laboratory (1955). Construction of Housing-Estate Roads Using Granular Base and
Subbase Materials. Road Note 20. (HMSO: London.)
Youdale, G.P. (1984). Review of Limiting Subgrade Strain Criterion. Submission to NAASRA Working
Group on the Revision of NAASRA Interim Guide to Pavement Thickness Design (IGPTD), April.
AUSTROADS 2004
— 1.56 —
APPENDIX B
DMR NSW PROCEDURE FOR PAVEMENT THICKNESS DESIGN IN 1947

– EXCERPT FROM BRITTON (1947)
PAVEMENT THICKNESS
The effective thickness of pavement required over a given subgrade or of upper courses over a given sub-
base or base course, assuming normal and satisfactory drainage and proper compaction and that a bituminous
surface or upper course is to be provided, is in general, to be computed two ways, as follows:
(a) Grading Rule
Disregard all material retained ¾ in. square sieve.
Compute the following ratios:
Title Ratio (percent) of all Passing To all Passing
R Passing ¾ in. sq. sieve but retained No. 7 B.S. ¾ in. sq
A Passing No. 36 B.S. No. 7 B.S.
B Passing No. 200 B.S. No. 36 B.S.
C Less than 0.0135 mm. No. 200 B.S.
Let D, E and F be departures of A, B and C respectively outside range 40 to 60.

e.g. D = A – 60 if A is greater than 60
= 0 if A is from 40 to 60
= 40 – A if A is less than 40
If neither A, B, nor C, is less than 40; compute the sum of D + E + F.
If neither A nor B is less than 40, but C is; compute the same sum but count not more than 20 for F.
If A is not less than 40 but B is; compute the same sum but count not more than 20 for E + F.
If A is less than 40; compute the same sum if it is less than 20, otherwise count D + E + F as 20.
To the sum so determined add one half the plastic index and subtract one quarter R. Call this final total T.
Effective cover required in inches is then: -
0.15 T for Heavy Loading (Max. wheel = 6 English Tons or, say = 13,500 lb.)
0.12 T for Normal Loading (Max. wheel = 4 English Tons or, say = 9,000 lb.)
(b) Soil Moisture Relations Rule

Let U = Upper Solid Limit
= PL if plastic
= LL if non-plastic
S = Linear Shrinkage from LL of Pass No.7 B.S. Portion in percent of original length.
W = Max. Dry Weight (lb. per cu ft.) by standard Proctor compaction method.
AUSTROADS 2004
— 1.57 —
Effective cover required in inches is then:

32 + 0.16 U + 0.27 S – 0.24 W for heavy loading.
26 + 0.13 U + 0.22 S –- 0.195 W for normal loading.
Tables A and B cover all usual cases.
Table A
Heavy Loading
U 0.16 U S 0.27 S W 32-0.24 W W 32-0.24 W

11 1.76 1 0.27 85 11.60 109 5.84
12 1.92 2 0.54 86 11.36 110 5.60
13 2.08 3 0.81 87 11.12 111 5.36
14 2.24 4 1.08 88 10.88 112 5.12
15 2.40 5 1.35 89 10.64 113 4.88
16 2.56 6 1.62 90 10.40 114 4.64
17 2.72 7 1.89 91 10.16 115 4.40
18 2.88 8 2.16 92 9.92 116 4.16
19 3.04 9 2.43 93 9.68 117 3.92
20 3.20 10 2.70 94 9.44 118 3.68
21 3.36 11 2.97 95 9.20 119 3.44
22 3.52 12 3.24 96 8.96 120 3.20
23 3.68 13 3.51 97 8.72 121 2.96
24 3.84 14 3.78 98 8.48 122 2.72
25 4.00 15 4.05 99 8.24 123 2.48
26 4.16 16 4.32 100 8.00 124 2.24
27 4.32 17 4.59 101 7.76 125 2.00
28 4.48 18 4.86 102 7.52 126 1.76
29 4.64 19 5.13 103 7.28 127 1.52
30 4.80 20 5.40 104 7.04 128 1.28
31 4.96 21 5.67 105 6.80 129 1.04
32 5.12 22 5.94 106 6.56 130 0.80
33 5.28 23 6.21 107 6.32 131 0.56
34 5.44 24 6.48 108 6.08 132 0.32
(c) Application of Rules

If the dispersion fails, the ratios A, B and C are not determined and the grading rule cannot be applied. The
S.M.R. rule only is taken into account.
If the dispersion does not fail:
1. If the adverse constituents are present in considerable quantity the grading rule does not apply and the
S.M.R. rule only is taken into account.
2. In A3 soils7 free from appreciable adverse constituents S.M.R. rule does not apply and grading rule only
is taken into account. (If much adverse constituents present, S.M.R. rule applies but not grading – vide 1
above).
3. In other cases average of the two rules is taken.
7
Passing 200 less than 10 percent of pass 7 or less than 15 percent of pass 36.
AUSTROADS 2004
— 1.58 —
Except in the case of A3 soils the computation by the two rules is usually a simple guide as to presence of
adverse constituents. If grading rule exceeds (S.M.R. rule minus 2 in.) the adverse constituents may be
neglected and the mean of the two rules taken. If grading rule is less than (S.M.R. rule minus 4 in) they are
present in quantity and grading rule is discarded.
Table B
Normal Loading
U 0.13 U S 0.22 S W 26-0.195 W 26-0.195W

W
11 1.43 1 0.22 85 9.42 109 4.74
12 1.56 2 0.44 86 9.23 110 4.55
13 1.69 3 0.66 87 9.04 111 4.36
14 1.82 4 0.88 88 8.84 112 4.16
15 1.95 5 1.10 89 8.64 113 3.96
16 2.08 6 1.32 80 8.45 114 3.77
17 2.21 7 1.54 91 8.26 115 3.58
18 2.34 8 1.76 92 8.06 116 3.38
19 2.47 9 1.98 93 7.86 117 3.18
20 2.60 10 2.20 94 7.67 118 2.99
21 2.73 11 2.42 95 7.48 119 2.80
22 2.86 12 2.64 96 7.28 120 2.60
23 2.99 13 2.86 97 7.08 121 2.40
24 3.12 14 3.08 98 6.89 122 2.21
25 3.25 15 3.30 99 6.70 123 2.02
26 3.38 16 3.52 100 6.50 124 1.82
27 3.51 17 3.74 101 6.30 125 1.62
28 3.64 18 3.96 102 6.11 126 1.43
29 3.77 19 4.18 103 5.92 127 1.24
30 3.90 20 4.40 104 5.72 128 1.04
31 4.03 21 4.62 105 5.52 129 0.84
32 4.16 22 4.84 106 5.33 130 0.65
33 4.29 23 5.06 107 5.14 131 0.46
34 4.42 24 5.28 108 4.94 132 0.26
In intermediate cases the test results should be examined in detail to decide the point (if uncertain there is
little error in taking 1 in less than the S.M.R. rule).
In A3 soils there is a transition zone where application of S.M.R. and grading rules is uncertain, but this is a
rare case in practice and thicknesses are not unduly large. A safe method is to take the higher of the two
rules in this doubtful zone.
AUSTROADS 2004
— 1.59 —
(d) Effective Cover

The effective cover is the sum for all overlying courses counted as follows:
Macadam: Stone; Sandstone: Actual Thickness
Gravel-sand Clay; Sand-clay: Actual Thickness
Bituminous Courses:
Dense and solid: Twice actual thickness
Semi-dense grading: One and one-half times actual thickness
Very open grading: Actual thickness
Surface treatment only Neglect
Soil sub-bases, etc:
A1, A2, A3 (1942 PRA): Actual thickness
Other groups: Two-thirds actual thickness
For unsurfaced pavements the same thickness is required on sandy non-plastic soils as for bituminous
pavements. On plastic and high organic soils the thickness may be reduced by one third if unsurfaced.
AUSTROADS 2004
— 1.60 —
APPENDIX C
JAMESON ON THE ORIGINS OF DESIGN DEFLECTION CURVES –

EXCERPT FROM JAMESON (1996)
The following is an excerpt from Origins of the Austroads Design Procedures for Granular Pavements
(Jameson 1996). The Figure and Table numbers have been to changed to reflect that the excerpt is Appendix
C of this report.
DESIGN DEFLECTIONS
Origins of Austroads Design Deflections
The Austroads (1992) overlay design procedures include a design deflection curve (Curve 1, Fig. 10 – shown
here as Fig. C1) for use in controlling "the rate of permanent deformation in the pavement and subgrade and
may be used for all pavements regardless of surfacing types". Contrary to the mechanistic approach to
overlay design (Austroads 1992), Curve 1 applies to all pavement types; different curves are not provided for
varying subgrade strengths or pavement thicknesses. Consequently, it was of interest to trace the origins of
the Austroads design deflection criterion.
1.6
1.4
1.2
Design Curve 1
Surface 1.0
Deflection
(mm)
0.8
0.6
Curve 2
0.4
0.2
0
105 2 4 6 8 106 2 4 6 8 107 2 4 6 8 108
Design Traffic (ESAs)
Figure C1: Austroads design deflection criteria (Fig. 10.3 of Austroads 1992)
Austroads Curve 1 is the same as NAASRA (1979) Curve 1 (Fig. 11 – shown here as Fig. C2), except that
the NAASRA8 Curve 1 is only applicable to pavements with unbound bases with thin asphalt surfacings.
The other principal difference is that the Austroads Curve 1 is applicable to traffic loadings up 108 ESAs,
whereas the upper limit of NAASRA Curve 1 is 3 x 107 ESAs.
8
Incorrectly stated as “Austroads” in Jameson (1996).
AUSTROADS 2004
— 1.61 —
Note that in adopting the NAASRA Curve 1 for all pavements types, it seems that the NAASRA Working
Group revising the 1979 NAASRA Guide reasoned that NAASRA Curves 2, 3 and 4 were provided to
control fatigue of asphalt and cemented materials. Consequently, the Group adopted NAASRA Curve 1 to
control rutting for all pavement types. NAASRA Curve 4 was retained to inhibit cracking in cemented bases
and a new curvature function was adopted to control asphalt fatigue.
Figure C2: NAASRA (1979) design deflection criteria
According to Black (1977) the 1979 NAASRA design deflection curves:

"were adopted from the information available and appropriate to Australian conditions. In
particular Curve B (Curve 2) was derived from the deflections levels adopted by the 40th
NAASRA (1969) for unbound base to be surfaced with bituminous concrete, and Curves
A, C and D (Curves 1, 3 and 4) were derived from experience in CRB VIC and DMR
NSW."
The basis of the 1979 NAASRA Curve 1 seems to be the design deflections agreed at the 40th meeting of
NAASRA in 1969. These design deflections are given in Table C1 and Fig. C3. Note that the 1969
NAASRA design deflections assumed that pavements have sprayed seal surfaces for traffic up to 450
commercial vehicles per day and asphalt surfaces for higher traffic volumes.
The minutes of the 40th NAASRA meeting give some background to the origin of the 1969 NAASRA
design deflections:
AUSTROADS 2004
— 1.62 —
"The values for tolerable deflections of pavement are based on:

(a) Consideration of recommended values of tolerable deflections for economic
performance of pavements reported from extensive deflections surveys in Europe
and Northern America.
(b) Limited surveys conducted by ARRB in Victoria and Tasmania correlated with later
performance.
(c) Experience of State Road Authorities with deflection surveys."
Table C1
Early Design Deflections for Granular Pavements with a Sprayed Seal Surfacing
Design Traffic Design Traffic 1965 Scala 1969 1969 CRB 1975 CRB 1979
Two-way One-way Allowable NAASRA Design Design NAASRA
End-of-Life ESAs/20 Deflection Design Deflection Deflection Design
CV/day years (mm) Deflection (mm) (mm) Deflection
(mm) (mm)
0-50 8 x 104 1.78 1.78 1.80
50-150 3 x 105 1.40 1.40 1.27 1.30 1.36
150-450 1 x 106 1.02 1.02 1.14 1.10 1.14
450-1500 3 x 106 0.76 0.76* 1.02 1.00 0.98
1500-4500 1 x 107 0.64* 0.89 0.90 0.88
>4500 3 x 107 0.51* 0.76 0.80 0.83

* Design deflections for asphalt-surfaced pavements.
2
Design deflection (mm)
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
1E4 1E5 1E6 1E7 1E8
Design Traffic (ESAs)
NAASRA 1979 CRB 1975 NAASRA 1969
Figure C3: Comparison of design deflections
AUSTROADS 2004
— 1.63 —
The design deflections adopted for the sprayed seal pavements appear to be largely based on research
conducted by Scala (1965). As discussed below, Scala conducted field measurements of 200 Victorian
pavements with sprayed seal surfaces. Using the high deflection data (> 0.8 mm), and taking account of the
findings of the AASHO Road Test, Scala proposed allowable deflections which were identical to those later
6
adopted by NAASRA in 1969 (see Table C1) for traffic loadings up to 1 x 10 ESAs. These are the highest
expected deflections over all subgrade CBRs and granular thicknesses, as discussed in Section 4.2.
The origins of the 1969 NAASRA design deflections for higher traffic loadings (asphalt surfaced pavements)
are less clear. The design deflection for 3 x 106 ESAs is the same as that recommended by Scala (1965). It
is noted that all three design deflections are similar to the Californian design deflection for a granular
pavement with a 50 mm thick asphalt surfacing (Zube and Forsyth 1966). These Californian design
deflections were based on the performance of heavily-trafficked (about 107 ESAs) roads extrapolated to
other loadings using the slope of a laboratory asphalt fatigue line. It seems, then, that these Californian
design deflections were related to asphalt fatigue rather than to rutting.
In 1969 the CRB adopted separate design deflections for sprayed seal surfaced and asphalt surfaced
pavements (Currie 1969). The sprayed seal surfaced values for traffic loadings exceeding 106 ESAs seem to
have been estimated by adding 0.25 mm (0.01 inch) to the 1969 NAASRA asphalt surfaced values. This
adjustment factor was considered to be somewhat conservative based on CRB experience but in line with
specifications used in the United States (Currie 1969). For traffic loadings of 106 ESAs and less, the CRB
(1969) design deflections were slightly less than the NAASRA (1979) values.
In 1975, the CRB issued Technical Bulletin 29 "Pavement Deflection Testing Using the Benkelman Beam".
As seen in Table C1, these design deflections for sprayed seal surfaced pavements were similar to the CRB
(1969) values except for some rounding off in the metrication process.
In summary, for traffic loading below about 106 ESAs the current Austroads design deflections appear to
have been derived from research conducted by Scala taking account of the findings of the AASHO Road
Test. Above 106 ESAs, the design deflections may have been based on the Californian design deflections
for 50 mm thick asphalt pavements adjusted to estimated equivalent sprayed seal values.
Design Deflection Dependence on Granular Thickness and Subgrade CBR

As already mentioned, the Austroads design deflection curve is applicable to all pavements irrespective of
granular thickness or subgrade CBR. The concept of a single design deflection curve for all pavements is at
variance with the mechanistic approach to overlay design (Austroads 1994). Research conducted by Scala,
on which the Austroads design deflections are in part based, is discussed below in relation to this issue. In
addition, the relevant findings of TRRL research are reviewed.
Scala's Findings
In the early 1960s, Scala derived a method of new pavement design based on deflections to complement the
CBR approach to pavement design. As it is impossible to measure deflections on a proposed pavement,
Scala's deflection method involved comparing the predicted deflection for the proposed pavement with the
design deflection. Based on field measurements of about 200 Victorian pavements with sprayed seal
surfaces, Scala (1965) developed relationships to predict Benkelman Beam deflections from subgrade
strength (CBR) and granular thickness. Scala determined two such relationships:
• the line of best fit, and
• the line estimating the maximum expected deflections for any combination of granular thickness and
subgrade CBR.
AUSTROADS 2004
— 1.64 —
As stated by Scala:
"The line of best fit will give a relation to be expected for the average structural condition
of the pavement material and subgrade. If the pavement materials are not compacted as
well as the average condition in Victoria, which is thought to be higher than normal, then
the actual deflections will be higher than expected from the (line of best fit) prediction.
Further, the CBR of a subgrade is a very variable quantity. In this investigation, the
variation is both longitudinally and vertically (with depth). Neglecting any longitudinal
variation the choice of the correct CBR rating for a subgrade which varies with depth is
difficult; this variation must be reflected in total deflection....
To allow for these conditions, or to cover any risk of failure it is probably preferable to fit
the envelop or the line giving the maximum expected deflection for any combination of
CBR and granular thickness."
Mean Deflection Relationship

The following line of best fit was obtained by Scala using multiple regression analysis on all the data:
log(d) = –(0.88+0.0165t) – 0.5 log(CBR) (C1)
where d = Benkelman Beam maximum deflection (inches),
t = thickness of granular material (inches), and
CBR = subgrade CBR.
This relationship, which is plotted in metric units in Fig. C4, gives the expected or mean deflection for a
given granular thickness and subgrade CBR.
1.8
Mean Deflection (mm)
1.6
1.4
1.2
1
0.8
0.6
2 4 6 8 10 12 14 16
Subgrade CBR (%)
200 mm 300 mm 400 mm 500 mm
Figure C4: Measured dependence of mean Benkelman Beam deflection on subgrade CBR for
various granular thicknesses (Scala 1965)
"High" Deflection Relationship

As stated above Scala also developed a relationship to predict the maximum expected deflection for any
combinations of subgrade CBR and granular thickness. Pavements with low deflections (< 0.8 mm) were
excluded from this analysis. This resulted in less than half the original 200 pavements available for analysis.
In order to increase the sample size use was made of the results from an additional 60 sites from another
investigation in which residential streets were tested. The following equation was obtained from regression
analysis of the combined data set:
AUSTROADS 2004
— 1.65 —
log(d) = – 0.34(1+0.1t) – 0.7 log(CBR) (C2)

where d = Benkelman Beam maximum deflection (inches),
t = thickness of granular material (inches), and
CBR = subgrade CBR
This relationship, which is plotted in metric units in Fig. C5, enabled the designer of a new pavement to
determine the granular thickness required in order that, when the pavement was constructed:
" the structure is 95% certain, after testing by the normal beam procedure, to have a
deflection less than the specified standard of deflection." (Scala 1965)
Note that for a given design deflection and subgrade CBR, higher thicknesses of granular materials were
required by the "high" deflection relationship than by the mean deflection relationship. As such the use of
the "high" deflection relationship was considered to be a conservative approach to the design of new
pavements.
3
"High" Deflection (mm)
2.6
2.2
1.8
1.4
1
0.6
2 4 6 8 10 12 14 16
Subgrade CBR (%)
200 mm 300 mm 400 mm 500 mm
Figure C5: Measured dependence of "high" Benkelman Beam deflection on subgrade CBR for
various granular thicknesses (Scala 1965)
Design Deflection
Scala proposed deflection criterion using:
• the "high" Benkelman Beam deflection dependence on granular thickness and subgrade CBR (eqn (8)),
and
• the relationship between design traffic loading for a given subgrade CBR and granular thickness,
obtained from the 1959 RRL granular thickness chart (Fig. 5) and a conversion between traffic category
and design ESAs (see Table 1).
The resulting relationships between "high" deflection and design traffic loading are given in Fig. C6 for
various thicknesses of granular material and Fig. C7 for various subgrade CBRs. Using this data and in view
of the AASHO road test findings, Scala proposed the design deflections given in Table C1 As mentioned
6
above, for traffic loading less than 10 ESAs Scala's design deflections formed the basis of the 1992
Austroads design deflections. These Austroads (1992) design deflections are also shown in Fig. C6 and Fig.
C7. It is apparent that the Austroads design deflections are the maximum expected deflections for a given
design traffic over all subgrade CBRs and granular thicknesses.
AUSTROADS 2004
— 1.66 —
It should be noted that Scala used the "high" deflection relationship rather than the mean deflection
relationship. When the mean deflection relationship (eqn (C1)) is used to derive design deflection curves,
the relationship between deflection and design traffic loading is very different, as illustrated in Fig. C8. In
this case, there is a much stronger dependence of design deflection on granular thickness and the design
deflections are substantially lower than those adopted by Austroads for all granular thicknesses. This
suggests that the Austroads (1992) design deflection curve overestimates the average allowable design traffic
loadings. However, it should be noted that these overestimates of average loading in part offset the
conservatism of the relationship between deflection and allowable loadings resulting from the use of the
conservative granular thickness chart (Fig. 1) in their deviation.
2
1.8
1.6
1.4
1.2
1
0.8
0.6
1E4 1E5 1E6 1E7
Design Traffic Loading (ESAs)
200 mm 300 mm 400 mm

500 mm AUSTROADS
Figure C6: Relationship between "high" Benkelman Beam deflection and

design traffic loading for various granular thicknesses
1.8
1.6
1.4
1.2
1
0.8
0.6
1E4 1E5 1E6 1E7
Design Traffic Loading (ESAs)
CBR=2 CBR=3 CBR=5

CBR=10 CBR=15 AUSTROADS
Figure C7: Relationship between "high" Benkelman Beam deflection and

design traffic loading for various subgrade CBRs
AUSTROADS 2004
— 1.67 —
The above discussion suggests that Scala's data, on which the Austroads design deflection criterion is based,
does not support the use of a single design deflection curve for all granular thicknesses and subgrade CBRs.
The Austroads design deflections are the maximum expected deflections over all granular thicknesses and
subgrade CBRs.
Mean Deflection (mm) 1.8

1.6
1.4
1.2
1
0.8
0.6
1E4 1E5 1E6 1E7
Design Traffic Loading (ESA)
200 mm 300 mm 400 mm

500 mm AUSTROADS
Figure C8: Relationship between mean Benkelman Beam deflection

and mean design traffic loading for various granular thicknesses
TRRL Research Findings

The NAASRA design deflection criterion (Curve 1 NAASRA 1979) was also re-examined by the NAASRA
Working Group Revising the NAASRA Interim Guide to Pavement Thickness Design. Using the
mechanistic design procedures, Youdale (1984) derived a similar dependence on design deflection on
granular thickness to that observed in the derivation of ASMOL, the interim Austroads Simplified Method of
Overlay design. As there was no dependence on granular thickness in NAASRA Curve 1 design deflection
criterion used at that time (the same as Austroads Curve 1), the Working Group (Anderson 1984) reviewed
some of the work of Lister (1972) of TRRL to assess whether the thickness dependence was supported by
British performance data.
Lister used data obtained in the Alconbury Hill experiment to investigate the relationship between deflection,
structural pavement parameters and performance. Test sections with rolled asphalt bases, granular subbases
and relatively uniform subgrades provided data which enabled the influence of asphalt base and sand subbase
on deflection to be estimated. Fig. C9 was derived by Lister after correcting the deflections at individual
points to take account of differences in subgrade CBR from the mean value of 4.5%.
The results indicate that the same design deflection curve can be applied to a range of sand subbase
thicknesses (125-500 mm). However, this is not necessarily in disagreement with either the observed
thickness dependence in mechanistically-based overlay design procedures or the above analysis of Scala's
data. Lister's results only apply to one subgrade strength (CBR = 4.5). As seen from Fig. C7, Scala's data
indicates a single design curve can be used for a single subgrade strength. Consequently, contrary to the
conclusions of the NAASRA Working Group, Lister's data does not support the use of a single design
deflection curve for all granular thicknesses and subgrade CBRs.
AUSTROADS 2004
— 1.68 —
Figure C9: Relationship between deflection , critical life and thickness for pavements
with rolled asphalt bases at Alconbury Hill (Fig. 31 of Lister 1973)
AUSTROADS 2004
— 1.69 —
REFERENCES
Anderson, D (1984). Notes on the Effect of Pavement Thickness on Tolerable Deflection. Submission to
NAASRA Working Group on the Revision of NAASRA Interim Guide to Pavement Thickness Design
(IGTPTD).
Austroads (1992). Pavement Design – A Guide to the Structural Design of Road Pavements. (AUSTROADS:
Sydney.)
Austroads Pavement Research Group (1994). Austroads Pavement Design Guide. Interim Version of
Revised Overlay Design Procedures. APRG Document No. APRG 94/10 (DA), August.
Black, D.J. (1977). Proposed NAASRA Publication "Manual of Pavement Thickness Design. Letter from
Country Roads Board of Victoria (1975). Pavement Deflection Testing Using the Benkelman Beam.
Technical Bulletin 29.
Currie, D. (1969). Use of Benkelman Beam in Pavement Evaluation. Paper presented to the Highways and
Traffic Branch of the Victorian Division of Inst. Eng. Aust., September.
Lister, N.W. (1972). Deflection Criterion for Flexible Pavements. Transport and Road Research Laboratory
(TRRL) Lab. Report LR 375.
NAASRA (1969). Notes of 40th Meeting of NAASRA. (NAASRA Sydney.)
NAASRA (1979). Interim Guide to Pavement Thickness Design. (NAASRA Sydney.)
Scala, A.J (1965). CBR. Design Method Deflection Dependency. Australian Road Research Board. Internal
Report, AIR 010-2.
Youdale, G.P. (1984). Review of Limiting Subgrade Strain Criterion. Submission to NAASRA Working
Group on the Revision of NAASRA Interim Guide to Pavement Thickness Design (IGTPTD), April.
Zube, E. and Forsyth, R. (1966). Flexible Pavement Maintenance Requirements as Determined by Deflection
Measurement Highway Research Record No. 129.
AUSTROADS 2004
— 1.70 —

Part 2: 1992 Guide Procedures for Design of Rigid Pavements
George Vorobieff and John Hodgkinson

June 2001
SUMMARY
This report records the work undertaken in the development of Chapter 9 – Design of New Rigid Pavements
– of the 1992 edition of Pavement Design – A Guide to the Structural Design of Road Pavements, published
by Austroads in 1992.
This material presented in this Chapter of the Guide represented over 30 years of development in Australia
and overseas in design procedures for determining the thickness of concrete pavements for highway truck
traffic. The content of the Chapter was drawn from design procedures and performance of pavements in
service in the USA and France as well as aspects of Australian experience.
AUSTROADS 2004
— 2.i —
TABLE OF CONTENTS
Page
1. GENERAL ....................................................................................................................................... 2-1
2. PAVEMENT TYPES........................................................................................................................ 2-2
3. DESIGN INPUT FACTORS............................................................................................................. 2-3
3.1 General.................................................................................................................................... 2-3
3.2 Concrete Flexural Strength ..................................................................................................... 2-3
3.3 Traffic Loading......................................................................................................................... 2-4
3.4 Foundation Support................................................................................................................. 2-7
3.5 Base and Subbase Debonding ............................................................................................... 2-9
3.6 Concrete Shoulders................................................................................................................. 2-9
3.7 Load Transfer at Joints ......................................................................................................... 2-10
4. DESIGN PROCEDURE METHODOLOGY ................................................................................... 2-12

4.1 General.................................................................................................................................. 2-12
4.2 Flexural Fatigue Analysis ...................................................................................................... 2-12
4.3 Erosion Analysis.................................................................................................................... 2-13
5. REINFORCEMENT FOR CRCP ................................................................................................... 2-16

6. OVERVIEW OF DESIGN RESULTS ............................................................................................ 2-17
6.1 General.................................................................................................................................. 2-17
6.2 Minimum Thickness Limit...................................................................................................... 2-17
6.3 Maximum Wheel Load........................................................................................................... 2-17
6.4 Influence of Concrete Shoulders........................................................................................... 2-18
6.5 Relative Influence of Traffic Volumes and Axle Loads................................................. .........2-17
6.6 Influence of Pavement Type ................................................................................................. 2-20
REFERENCES ........................................................................................................................................ 2-21
AUSTROADS 2004
— 2.i —
TABLES
Page
Table 1: Legal axle loads and extent of overloads for the typical rural and
urban road axle group distributions in Appendix I of the 1992 Guide.............................. 2-7
Table 2: Faulting criteria for major roads (Packard 1977) ........................................................... 2-14
Table 3: Static load equivalence for various commercial vehicle axle groups with
a maximum wheel load of 65 kN and a LSF of 1.2.......................................................... 2-16
FIGURES
Page
Figure 1 Typical longitudinal section of plain concrete pavement (PCP)....................................... 2-2
Figure 2 Typical longitudinal section of jointed reinforced concrete pavement (JRCP) ................ 2-2
Figure 3 Typical longitudinal section of continuously reinforced concrete
pavement (CRCP)............................................................................................................ 2-2
Figure 4 Typical cross-section of Dowelled Plain Concrete Pavement (PCP-D)........................... 2-2
Figure 5 Design model assumption of concrete strength gain with age (PCA 1984) .................... 2-3
Figure 6 Fatigue relationship adopted in design model.. ............................................................... 2-4
Figure 7 Plan of the four most common commercial vehicle axle groups in Australia .................. 2-5
Figure 8 Axle group load distributions for a typical urban site from Appendix I
of the Guide...................................................................................................................... 2-5
Figure 9 Axle group load distributions for a typical rural site from Appendix I
of the Guide...................................................................................................................... 2-6
Figure 10 The base will curl on a daily cycle provided the base and subbase
are debonded ................................................................................................................... 2-8
Figure 11 Equivalent edge stress factor depends on percentage of trucks at
pavement edge (PCA 1984) ............................................................................................ 2-9
Figure 12 The void under the slab allows water to push the fines from the
subbase/subgrade material, and in some cases water has been seen
to spray out of transverse joints ..................................................................................... 2-10
Figure 13 A joint with 100% joint effectiveness will deflect equally across the joint ...................... 2-10
Figure 14 A joint with 0% joint effectiveness will have zero deflection across the joint................. 2-10
Figure 15 Position of axle load group for the critical base flexural stress...................................... 2-12
Figure 16 Position of axle load group for the critical base deflections........................................... 2-13
Figure 17 Base thickness for varying CVAGs for plain concrete pavement with
and without dowels. Note: Chart for PCP, LSF = 1.1, undowelled joints,
effective CBR15% and rural traffic distribution. ............................................................. 2-16
Figure 18 Base thickness for varying CVAGs for plain concrete pavement with
and without dowels. Note: Chart for PCP, LSF = 1.1, undowelled joints,
effective CBR of 15% and Rural axle load distribution .................................................. 2-16
Figure 19 Typical rigid pavement thickness design curve for a specific effective
CBR and concrete flexural strength demonstrating that erosion of the
transverse joints is generally the dominant distress mechanism for
high-volume vehicle traffic ............................................................................................. 2-17
AUSTROADS 2004
— 2.ii —
1. GENERAL
The 1992 Edition of Chapter 9 "Design of New Rigid Pavements " represents over 30 years of development
in Australia and overseas in design procedures for determining the thickness of concrete pavements for
highway truck traffic. The current guide is drawn from principally design procedures and performance of
pavements in service in the USA PCA 1984) and France, and aspects of Australian experience.
The basis for the Austroads procedure is analytical rather than empirical. It is mechanistic and has been so
over its 30-year development period. The mechanistic approach is based on selecting a trial pavement
thickness with the thickness being tested against boundary conditions which places limits on the damage
caused to the pavement under traffic loading.
The procedure analyses the two most probable causes of pavement distress which may occur in a ground-
supported pavement under long term heavy repeated traffic loads. These are as follows.
• Flexural fatigue from repeated flexural tensile stresses at the bottom of the slab. Built into this analysis
is consideration of the locations of wheel paths relative to the outside longitudinal edge of the pavement
base where critical stresses occur.
• Erosion in the pavement foundation in the areas under joints or cracks caused by the accumulated effects
of deflections from repeated traffic loads as they cross them.
It is important to those who may the use of the Austroads Chapter 9 to understand that the procedure uses
criteria applicable to large volumes of heavy commercial road vehicles moving in one direction, with wheel
paths within a reasonably defined zone in a defined traffic lane and at speeds above 60 km/h. This is why in
the opening paragraph to Chapter 9 it is pointed out that Chapter 9 may not be applicable to residential
streets or many industrial and airfield pavements where the design conditions are not similar.
The Austroads Guide does not include design procedures for slab anchors, joint detailing and layouts. Other
documents, such as the RTA, NSW Concrete Pavement Manual (RTA, NSW 1996) or the Australian Road
Concrete Training Manual (Vorobieff 1998)), should be used to finalise the design of rigid pavements.
AUSTROADS 2004
— 2.1 —
2. PAVEMENT TYPES
The thickness design procedure in the Guide allows for:
• Jointed plain (unreinforced) concrete pavements (PCP) – refer to Fig. 1.
• Jointed reinforced concrete pavements (JRCP) – refer to Fig. 2.
• Continuously reinforced concrete pavements (CRCP) – refer to Fig. 3.
Whilst it is noted in the Guide that the spacing of transverse joints in PCP are in the range of 4 to 5 m, recent
experience indicates that 3.5 to 4.5 m is more suitable for PCP.
The Guide is also applicable to dowelled jointed plain (unreinforced) concrete pavements (PCP-D) as shown
in Fig. 4 and steel-fibre reinforced concrete pavement types. In both the jointed reinforced and continuously
reinforced pavements the purpose of the steel reinforcement is to manage crack widths between planned
joints to allow load transfer by aggregate interlock. The Guide could not be used for doubly reinforced
concrete pavements whereby the bottom layer reinforcement is subject to bending stresses.
Induced &
3.5 to 4.5m
sealed joints
Figure 1: Typical longitudinal section of plain concrete pavement (PCP).
Steel mesh Dowels Sawn &

8 to 15 m sealed
joints
Figure 2: Typical longitudinal section of jointed reinforced concrete pavement (JRCP)
0.5 to 1.5m
Steel bars
Figure 3 Typical longitudinal section of continuously reinforced concrete pavement (CRCP)
Dowels Induced &

up to 5 m
sealed joints
Figure 4: Typical cross-section of Dowelled Plain Concrete Pavement (PCP-D)
AUSTROADS 2004
— 2.2 —
3. DESIGN INPUT FACTORS

3.1 General
There are five design input factors to the design procedure:
• Concrete flexural strength.
• Assignment of traffic loads.
• Foundation support and subbase design leading to an 'effective' design California Bearing Ratio (CBR).
• Load transfer mechanism at joints or cracks, i.e. dowelled joints.
• The inclusion or omission of concrete shoulders.
Each of these factors are covered in detail in the following sections.
3.2 Concrete Flexural Strength

When a concrete slab is loaded to the point of rupture, the rupture occurs in flexure rather than compression.
The key stresses are flexural tensile stresses at the bottom of the slab and for this reason the design input
factor is concrete flexural strength. For reasons of convenience, cost, reliability, experience and available
laboratory resources, compressive strength is commonly used to monitor flexural strength for construction
purposes.
Guidance on the correlation between flexural and compressive strength is given in Chapter 6 of the Guide "
Pavement Materials ".
As an element of conservatism, the design procedure (PCA 1984) applies a reduction of one statistical
coefficient of variation to the design concrete strength in the various Tables and nomographs. That is, the
full design strength is not utilised in the design procedure.
The base concrete must meet two criteria;
• provide sufficient strength to produce an economical base thickness,
• have good durability in terms of surface wear and permeability where reinforcement is included.
To satisfy both criteria, AS 3600 Concrete Structures indicates a minimum 28-day characteristic compressive
strength of 32 MPa. This typically corresponds to a 28-day flexural strength of about 4.25 MPa but the
designer should verify the flexural strength properties from local data using similar aggregates and
cementitious materials.
The design model has been based on the average concrete strength at 28 days age and the Guide adopted the
characteristic strength, that is the 95% percentile. The difference between the characteristic and average
strength is about 15%.
It must be understood by designers that the design value is just a number. It is the role of construction
Specifications to determine how the concrete is to be produced and placed to achieve the design requirement.
These considerations are outside the scope of the Austroads Guide.
In the design model it has been assumed that the concrete strength increases with age from 28 to 90-days at
about 10% (Fig. 5). Most concrete mixes for road construction appear to not achieve this growth without the
use of supplementary cementitious materials, such as fly ash. Typically, specifiers make an adjustment for
low strength gain after 28-days by specifying a 10% higher flexural strength than the design calculations.
The lean-mix concrete subbase requirements is to produce a concrete that has a characteristic concrete
compressive strength of 5 to 7 MPa at 28-days. The purpose of the low strength is for the pavement to not
gain high strength in order for the subbase to produce uniform cracks to provide uniform load transfer across
both transverse and longitudinal joints in the base concrete. Also, a minimum strength is required to ensure
low-erodability requirements. No structural gain is provided except by the increase in subgrade stiffness.
AUSTROADS 2004
— 2.3 —
Flexural strength % of 28-day)

130%
120%
90-day design
110%
28-day design
100%
90% 10 100 1000 10000
28 days 90 1 yr 3 5
Age of Concrete
Figure 5: Design model assumption of concrete strength gain with age (PCA 1984)
One important element in the design model is the fatigue relationship adopted for concrete such that the
various axle loads and axle groups that produce flexural stress in the base can be accommodated to establish
the minimum base thickness and concrete strength. Fig. 6 shows the relationship between concrete flexural
stress ratio and number of repetitions.
0.9
0.8
Stress Ratio
0.7
0.6
0.5
0.4
1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07
Load Repetitions
Figure 6: Fatigue relationship adopted in design model
Thermal, elastic and shrinkage properties are not required in the thickness design procedure but are used in
the determination of the amount of longitudinal reinforcement in continuously reinforced concrete
pavements.
3.3 Traffic Loading

The response of a concrete pavement to repeated loading is that rupture will be caused by the heavier
wheel/axle loads in a particular traffic stream. For this reason the use of generalised load equivalencies is
inappropriate and such load equivalencies can lead to a structurally insufficient pavement. The preferred
design input is a matrix of numbers and loads for different axle types as discussed in Chapter 7 " Design
Traffic Loads ".
Cars and light vehicles are not included in the thickness analysis as the stress ratio they produce is low
enough to not affect the summation of damage due to fatigue.
AUSTROADS 2004
— 2.4 —
The thickness design procedure allows the direct input of measured load distributions, such as those from
weigh-in-motion data, and these traffic load distributions are based on four commercial axle groups (Fig. 7),
namely:
• single axle with single wheels (SS)
• single axle with dual wheels (SD)
• tandem axle with dual wheels (TAD)
• triaxle with dual wheels (TRD)
To illustrate the use of axle load distributions in the design procedure, two load distributions labelled "urban"
and "rural' (Figs 8 and 9) are given in Appendix I to the 1992 Guide. The key to these distributions is the
dramatic effect on the number of repetitions to pavement failure from the number of the heavy axle group
loads in each axle type at the “high load end” of the load spectrum. The heavy loads produce substantial
flexural stresses in the slab leading to low allowable axle repetitions to failure (refer to stress ratio diagram in
Fig. 6). Conversely, cars and light vehicles with axle loads less than 1.5 t produce a low stress ratio and
subsequently are not used in the thickness design procedure.
Single axle Single axle Dual axles Three axles

Single wheels Dual wheels Dual wheels Dual wheels
(SS) (SD) Tandem Triaxle
(TAD) (TRD)
Figure 7: Plan of the four most common commercial vehicle axle groups in Australia
It should be noted that these load distributions have no exclusive connection to concrete pavements and
could be used in the design of any pavement type. The distributions in the Appendix were compiled by the
design Working Group as being representative of weigh-in-motion (CULWAY) data from a number of sites
in urban and rural locations in NSW and Victoria. ARRB Transport Research notes that this data is the axle
loads with less than a 5% dynamic component. The designer should always strive to derive the load
distribution for the site based on historical data.
Using the representative load distributions in Appendix I, example design charts were produced as shown in
Figures 9.7 to 9.10. Cautionary notes on any wider application of the axle distributions and example design
charts are given in Chapter 9.
Load safety factors (LSF ) have been used in concrete pavement design in Australia since the interim
NAASRA Guide (NAASRA 1979). The basic design axle loads are multiplied by a factor generally in the
range 1.0 to 1.2 before inputting into the design procedure. It can be demonstrated that the magnitude of
axle loads rather than numbers of overall traffic has a major influence on the determination of the concrete
pavement thickness, and it becomes clear by inspection from the Example design charts in Chapter 9.
AUSTROADS 2004
— 2.5 —
45
40
35
30
25 SS (%) SD (%) TAD (%) TRD (%)

Distribution (%)
20
15
10
0
10 60 110 160 210 260 310
Axle Group Load (kN)
Figure 8: Axle group load distributions for a typical urban site from Appendix I of the Guide
40
35
30
25
SS (%) SD (%) TAD (%) TRD (%)
20
15
10
0
10 60 110 160 210 260 310
Axle Group Load (kN)
Figure 9: Axle group load distributions for a typical rural site from Appendix I of the Guide
AUSTROADS 2004
— 2.6 —
For this reason the LSF is applied to axle load tonnage rather than traffic volume repetitions as in flexible
pavement design.
LSF is a design input safety factor against possible changes in traffic patterns for a particular road during the
period for which traffic is estimated, or occasional unpredictable heavy loads whether legal or not. Table 1
shows the legal axle limits, the ratio of the maximum axle load within each group to the legal limit and the
percentage of vehicles over the legal limit. This data indicates that CULWAY and WIM static axle load data
with load safety factors is a conservative approach to the design model.
In the use of traffic data the designer should take into consideration the possible development of a new
secondary industry complex adjacent to a road being designed and thereby generating an increasing in the
assumed load distribution, say 10 to 15 years into the traffic analysis period. Also, the designer should
ensure that traffic growth of commercial vehicles rather than AADT is considered in the analysis.
Guidance on the distribution of traffic within lanes is not contained in the Guide and the designer should
seek input from the road authority. Typically the lane use factor for commercial vehicles in the left lane or
heavy truck lane will be in the range 85-95%. It is possible to complete designs giving two thicknesses such
that a thicker base is constructed in the left lane and a thinner base in the other lanes. As noted in this
commentary, concrete base thickness design is more sensitive to the magnitude of axle loads than it is to
traffic volumes and accordingly such a two-tiered thickness design will produce differences in thickness of
probably only 20 to 30 mm. When considered in the context of an overall 10 m carriageway (two lanes and
a concrete shoulder), a 'tapered' pavement cross-section has a lower base transverse gradient of about 0.2%.
This is generally considered impracticable for underlying subgrade/subbase layer level tolerance control and
therefore, the typical approach is to design for the heavy truck lane and carry the thickness across the whole
carriageway.
3.4 Foundation Support

The principal contribution of the subbase to base thickness design is to increase the foundation CBR. It also
has other non-thickness roles including providing an erosion resistant construction working platform under
the base. The thickness design of a concrete road pavement is not particularly sensitive to modest variations
in CBR values.
Table 1
Legal Axle Loads and Extent of Overloads for the Typical Rural and Urban Road Axle
Group Distributions in Appendix I of the 1992 Guide
Road Type Axle Group

SS SD TAD TRD
6.0 t 9.0 t 16.5 t 20 t
Ratio of Maximum Axle Load Group to Legal Limit
Rural road 1.87 1.59 1.61 1.68
Urban road 1.70 1.47 1.55 1.68
Percentage of Axle Loads Within Groups Above the
‘Static’ Axle Legal Limit
Rural road 20% 14% 15% 22%
Urban road 9% 3% 5% 11%
The correlation of Elastic Modulus to CBR as discussed in Chapter 5 " Subgrade Evaluation " can be
reasonably applied to concrete pavement design.
Almost all international road engineering agencies now include bound subbases in varying forms in concrete
pavement design guides and catalogues relevant to highway and similar road classifications.
Figure 9.1 provides guidance on the minimum subbase requirements and only bound or lean-mix concrete
subbases are noted.
AUSTROADS 2004
— 2.7 —
The genesis of current subbase design information in Chapter 9 can be attributed to work by leading
international engineers Michel Ray from SETRA in France (Ray 1981) and the late Prof. Eldon Yoder of
Purdue University in the USA (Yoder 1978). Until the work of Ray on subbases in the late 1970s and early
1980s, concrete slab thickness design had a simplistic approach. The solution was simply to make the slab
thicker using a relatively thin unbound subbase. One of the forms of distress which was not being analysed
was erosion occurring under joints as the result of plastic foundation soils becoming wet and being ejected
upwards and out of joints by joint/crack deflections caused by large numbers of truck loads i.e. "pumping".
Experience had shown that simply making a slab thicker did not always address the issues of drainage and
erosion.
Ray (1981) drew the link between load transfer (dowels or not dowels), varying traffic load intensity and
erosion under the slab. This also include discussion on internal pavement drainage and the selection of
subbase materials which would offer resistance to erosion based on the above three factors. This led to the
progressive development of practical design information in Australia now contained in Figure 9.1 of Chapter
9.
Yoder presented information in an invited paper to the ARRB Conference in 1978 showing, among a range
of other design issues, the benefit to design, expressed in terms of an increase in design CBR by placing a
subbase of any type. For unbound subbases the benefit was very small. As concrete thickness design is
relatively insensitive to variations in CBR compared with a flexible pavement it was decided to not assign
any increase in CBR for an unbound subbase.9
The increase in the CBR value from the top of the subgrade to that at the top of the subbase is shown in
Figure 9.2 in Chapter 9 and is termed the "effective" CBR. This chart was prepared from the PCA Manual
(PCA 1984) based on the AASHO Road Test research results. It is the value assigned to the foundation
support input factor used in the thickness design procedure. Whilst the horizontal axis of Figure 9.2 assigns
input CBR values ranging from 2 to 35%, the designer needs to assure that the subgrade CBR is achievable
in the field when sections of the pavement are constructed on fill. It is not prudent to specify a 150 mm layer
with a CBR of 35% on a layer of consisting of CBR at 5%. The Guide does not provide suitable guidelines
with respect to multi-layered subgrades and the engineer should seek geotechnical advice.
The subbase thickness and material type are general recommendations which link traffic loads, susceptibility
to erosion (a low CBR subgrade will be more prone to erosion than a high one), and joint load transfer
(dowels or no dowels). A guide such as Austroads cannot attempt to provide detailed advice on all
conditions around Australia and presents the preferred general approach. As pointed out in Chapter 9 if good
local research or experience shows good performance with a less demanding solution such as good quality
unbound materials then the Guide does not restrict local experience. However, adoption of overseas
technology of subbase construction should be carefully examined as many practices perform well under the
appropriate climatic conditions, for example permeable subbases for frozen subgrades.
The thickness design of the subbase is arbitrary and supported by experience. For highway construction in
the early 1990s the minimum thickness which was believed to be capable of construction was 100 mm.
Many engineers have found that a more appropriate minimum construction thickness for the subbase is
125 mm.
9
Whilst not mentioned in the Guide it is implicit by its exclusion.
AUSTROADS 2004
— 2.8 —
The design of a concrete pavement base slab is relatively insensitive to modest variations in CBR, such that
an increase in subbase thickness above 150 mm will prove to be uneconomical for a given design traffic
load. For example, an increase in subbase thickness from 150 mm to 200 mm could be structurally matched
by an increase in base thickness of about 5 mm. Therefore, no recommendations are given for subbase
thicknesses greater than 150 mm.
Slab moves up
Tension
Daytime curling of base

Slab moves up at edges
Compression
Night time curling of base
Figure 10: The base will curl on a daily cycle provided the base and subbase are debonded
3.5 Base and Subbase Debonding

The thickness design approach in the Guide is based on the assumption that the base and subbase are two
individual layers such that the base slab can “freely” curl according to daily changes in temperature (Fig.
10). The Guide cannot be used for conditions where the base is bonded to the subbase.
3.6 Concrete Shoulders

Critical wheel induced stresses occur at the outside edge of a slab and therefore, positioning of the
commercial vehicles to minimise edge loadings and slab thickness is essential to the design procedure. A
number of field observations in the USA and analytical studies, principally by Zollinger and Barenberg
(1989) in the USA have shown two characteristics which are built into the procedure, namely:
• Field studies (PCA 1984) showed that typical outer wheel paths for trucks are about 600 mm in from the
outside edge of the main heavy truck lane and about 6% of all axles in the outer wheel path are at the
edge or very close to it.
• For the stress analysis, fatigue was computed at mm increments inward from the slab edge for different
truck wheel positions. This analysis showed that the equivalent edge-stress factors, as shown in Fig. 11,
could be generated such that a statistical analysis could derive a conservative equivalent stress factor to
allow for 6% of the truck wheels travelling at the edge. These edge stress ratios for the assumed truck
wander was incorporated in the design model.
AUSTROADS 2004
— 2.9 —
Ratio to Edge Stress

for Same Fatigue
0.95
0.90
0.85
0.80
0.75
0.70
0 1 2 3 4 5 6 7 8
Percent Trucks at Edge
Figure 11: Equivalent edge stress factor depends on percentage of trucks at pavement edge
(PCA 1984)
The concept of using concrete shoulders is to allow the base to be designed to interior loading and hence a
thinner base thickness. Concrete shoulders also assist in the shedding of water away from the subbase and
subgrade in the main carriageway area. Field performance studies in the USA were carried out using 10 foot
wide shoulders and much of the analysis to date has been based on 10 feet (i.e. ≈3 m) wide tied concrete
shoulders. The Guide allows the use of 1.5 m tied shoulders and to date there has been no research to
indicate that this is unsatisfactory.
Although outside the scope of Chapter 9, a common Australian design detail is to widen the distance
between the two longitudinal joints forming the heavy truck lane, without changing an overall multiple-lane
carriageway width and paint an edge line about 500 to 600 mm in from the outside one of these joints. This
is referred to as a " widened truck lane " and is discussed in Ayton (1993). As trucks will generally respond
to linemarking, the outer wheel path is further shifted from the edge of the truck lane slab and adds further
conservatism. This pavement detail is widely used in Europe, particularly in France.
3.7 Load Transfer at Joints

Many long-term performance problems of concrete pavements in Australia and internationally causing slab
cracking are not always simply due to lack of slab thickness, but are due to voids under the slab from the
erosion of subgrade material. The voids are formed by large joint deflections and resulting "pumping" of the
underlying material upwards through joints or cracks (Fig. 12). The slab is no longer in the assumed state of
being fully supported.
Approach slab Leave slab
Accumulated fines
Figure 12: The void under the slab allows water to push the fines from the subbase/subgrade material, and in
some cases water has been seen to spray out of transverse joints
AUSTROADS 2004
— 2.10 —
Experience over the last 60 years has shown that erosion of the subgrade is principally a heavy
truck-trafficked highway condition and it is not usually found in residential streets (lightly trafficked).
Load transfer is a measure of the vertical shear load transferred across the joint when the load is positioned
adjacent to the joint as shown in Fig. 13. When there is no load transfer the differential deflection or faulting
across the joint is at maximum (Fig. 14) and conversely the differential deflection is zero when there is 100%
load transfer across the joint. Faulting across the joint is sometimes referred to as stepping.
Wheel
Load
∆L = d/2 ∆u = d/2
Figure 13: A joint with 100% joint effectiveness will deflect equally across the joint
Wheel
Load
∆L = d ∆u = 0
Figure 14: A joint with 0% joint effectiveness will have zero deflection across the joint
Load transfer across transverse and longitudinal joints is typically achieved by aggregate interlock,
corrugated formed side faces, mechanical dowels and shear keys for lightly-trafficked roads. For CRCP load
transfer at transverse cracks is achieved by aggregate interlock and therefore, it is crucial to have narrow
cracks formed in the base. The reinforcement does not provide dowel action but holds the cracks together
from repetitive opening and closing of cracks due to temperature variations.
The erosion analysis is concerned with the power of deflections at joints and cracks applied to the
foundation. The deflections are influenced both by traffic speed and slab thickness. The analysis is also
influenced by the type of load transfer at a joint i.e. whether it has dowels or no dowels. The finite analysis
leading to the design procedure models assumes that in a dowelled joint the dowel provides effectively all
the load transfer and is modelled as a series of small elastic beams, one per dowel. An aggregate interlock
type joint is modelled as a spring.
In general terms a thin slab will feel a greater 'punch' from a moving load than a thick slab. Although
deflections in a concrete pavement are small, dowelled, joint will deflect relatively less than an undowelled
joint for a particular load. For design purposes and with the amount of longitudinal reinforcement in CRCP,
CRCP thickness design follows the same criteria as for a dowelled jointed pavement. A concrete shoulder
will also reduce joint deflections and will influence the erosion analysis.
AUSTROADS 2004
— 2.11 —
4. DESIGN PROCEDURE METHODOLOGY

4.1 General
The fatigue life of the slab has been based on the trucks causing fatigue damage The core design
methodology adopted by Austroads is taken from a published USA procedure known as the "1984 PCA
Method" (PCA 1984). The Canadian Edition is used for convenience as it is in metric units but in all other
respects is identical to the USA edition. This analytical design procedure to determine subbase and base
thickness for various pavement types is based on the premise that the pavement will reach some lower level
of satisfactory service by either the onset of fatigue cracking or erosion of the subbase or subgrade at joints
or cracks.
Other published Highway Authority documents, particularly those in the UK and Europe, are frequently
issued in catalogue form and do not present or discuss the design criteria in detail making it difficult to assess
their relevance or compatibility with Australian conditions. They are often based on statutory axle loads that
are much higher than in Australia and include provision for very harsh winter conditions not usually
experienced in Australia.
The AASHTO method developed in the USA was empirically derived from the performance of a series of
test tracks using pavement compositions not used in Australia and under some climatic conditions not
considered relevant to Australia. This procedure was not considered for this reason. A discussion on
selected overseas design guides and comparative thicknesses is discussed in Hodgkinson (Hodgkinson 1993).
Much of the Austroads procedure incorporates information derived from finite element analysis models and
software developed in the USA in the late 1970s and early 1980s (Packard and Ray 1986; Packard and
Tayabli 1985; Heinrichs et al. 1988). In this regard the 1992 Austroads Guide marks a departure from the
previous Guides based on equations developed by Westergaard for calculating stresses in slabs for single
wheel loads and the "Pickett and Ray" influence charts which assess the influence of multiple wheel
assemblies at edges or within slabs.
The use of CIRCLY for rigid pavements has not been considered technically relevant as it assumes infinite
layers of homogeneous and isotropic properties around the area of load application. It cannot take into
account discontinuities, such as joints/cracks, or the positioning of wheel loads near an outside edge or at a
joint, both of which have a direct bearing on design of rigid pavements.
4.2 Flexural Fatigue Analysis

The fatigue analysis has two steps; calculating the "equivalent stress" at an outside edge caused by one
application of a particular load; determining the fatigue 'damage' caused by the numbers of loads in the
design load distribution. In terms of establishing the equivalent stress the design model assumes that the
critical position of the vehicle axle group is at mid-slab and adjacent to the edge as shown in Fig. 15. The
equivalent stresses in Tables 9.2 and 9.3 of the Guide have allowed for the truck wander as indicated in Fig.
11.
Based on a trial base thickness and effective design CBR an equivalent stress for a pavement with or without
a concrete shoulder is obtained from Table 9.2 in Chapter 9 for each axle type. The reduction in equivalent
stress with increasing foundation support, trial thickness and the inclusion of a shoulder can be seen by
inspection. The equivalent stress is divided by the design concrete flexural strength to obtain the "stress ratio
factor " for each axle type. The stress ratio factor forms the basis for the fatigue analysis.
AUSTROADS 2004
— 2.12 —
Traffic
Lane
Pavement Edge
Figure 15: Position of axle load group for the critical base flexural stress
The fatigue analysis aims to keep flexural stresses within safe limits by limiting the 'fatigue damage' to
100%. This is determined by summing the individual fatigue damage from each axle type/load in the design
load distribution. This is based on the "Miner hypothesis' approach which simply says that the balance of
fatigue damage after considering one load is available for all other loads.
If the stress ratio is kept small and given the general response of a concrete element to loads then it is highly
unlikely that slab rupture will occur for a particular load. This leads to the use of the term 'unlimited' load
repetitions. Based on work in the USA dating back to the 1960s and until the 1987 Austroads Guide, the
unlimited load classification applied to a stress ratio less than 0.5. At a stress ratio of 0.5, the allowable
repetitions were 500,000 and this value progressively declined as the stress ratio increased. During the
1980s it became clear in the USA (PCA 1984) that with increasing truck traffic, design axle load repetitions
were now often approaching and exceeding one hundred million compared with the 1970s when designs
were in more often in tens of millions. However, the same thickness results often emerged because of the 0.5
stress ratio and this was considered to be 'unrealistic'. Accordingly, the 1984 PCA Procedure and 1992
Austroads Guide have extended the unlimited repetition stress ratio down to 0.45 (Fig. 6). This has led to
more conservative designs at the higher end of design traffic volumes than previously.
For design inputs yielding a thickness of about 150 mm the changed fatigue design from 1987 Austroads
Guide has negligible effect. For a design yielding a thickness of about 230 to 250 mm, and all other factors
remaining constant the base thickness has increased by about 10 mm arising from the changed stress ratio
conditions alone.
Using the load per wheel for a particular axle group and the stress ratio factor for that axle group the
allowable repetitions for different loads are determined from the nomograph in Figure 9.4 of the 1992 Guide.
The inclusion or otherwise of a concrete shoulder has already been taken into account within the equivalent
stress.
For each axle load the expected load are divided by the allowable loads to calculate the fatigue damage for
that load. The individual fatigue damages are summed to calculate the overall fatigue damage.
The manual design proforma in Appendix I of the Guide allows this to be tabulated.
4.3 Erosion Analysis

Many repetitions of heavy axle loads at slab corners and edges cause pumping resulting in the erosion of
subgrade, subbase, and shoulder materials leading to voids under and adjacent to the slab. The voids create
faulting of pavement joints and most notable in pavements with undowelled transverse joints. Finite element
analysis shows that the critical axle group location for erosion distress is at the edge and adjacent to the joint
as shown in Fig. 16.
AUSTROADS 2004
— 2.13 —
Traffic
Lane
Concrete
Shoulder
Pavement Edge
Figure 16: Position of axle load group for the critical base deflections
The PCA Design Guide (PCA 1984) notes that:

“Correlations of deflections computed from the finite element analysis with AASHO Road Test
performance data were not completely satisfactory for design purposes. (The principal mode of
failure of concrete pavements at the AASHO Road Test was pumping or erosion of the granular
subbase from under the slabs.) It was found that to be able to predict the AASHO Road Test
performance, different values of deflection criteria would have to be applied to different s ' lab
thicknesses, and to a small extent, different foundation moduli (k values).
More useful correlation was obtained by multiplying the computed corner deflection values (w)
by computed pressure values (p) at the slab-foundation interface. Power, or rate of work, with
which an axle load deflects the slab is the parameter used for the erosion criterion-for a unit
area, the product of pressure and deflection divided by a measure of the length of the deflection
basin (l-radius of relative stiffness, in millimeters). The concept is that a thin pavement with its
shorter deflection basin receives a faster load punch than a thicker slab. That is, at equal pw’s
and equal truck speed, the thinner slab is subjected to a faster rate of work or power (watts). A
successful correlation with road test performance was obtained with this parameter.
The development of the erosion criterion was also generally related to studies on joint faulting.
These studies included pavements in Wisconsin, Minnesota, North Dakota, Georgia, and
California, and included a range of variables not found at the AASHO Road Test, such as a
greater number of trucks, undowelled pavements, a wide range of years of pavement service,
and stabilized subbases.”
The erosion analysis follows a similar procedure to that of the fatigue analysis and is based on a suitable base
thickness which will keep joint/crack deflections within safe limits by limiting the 'erosion' damage' to 100%.
This is determined by summing the individual erosion damage from each axle group load in the design load
distribution. The 100% damage limit has been based on performance studies with limits placed on the
serviceability index which correlates to limiting the faulting at transverse joints to the range of 3 to 6 mm at
terminal conditions (Packard and Tayabji 1985). Table 2 lists work by Packard (Packard 1977) to relate
average faulting across joints to driver comfort conditions.
AUSTROADS 2004
— 2.14 —
Table 2
Faulting Criteria for Major Roads (Packard 1977)
Fault Index (PI) Average Rating

Faulting (mm)
0 to 5.0 0-0.8 Excellent
5.1 to 10.0 0.8–1.6 Very Good
10.1 to 15.0 1.6–2.4 Good
15.1 to 20.0 2.4–3.2 Fair
20.1 to 25.0 3.2–4.0 Poor
25.0 4.0 Very poor
The erosion criterion was suggested for use as a guideline and it was always the researchers intention that it
could be modified according to local experience since climate, drainage, local factors, and new design
innovations may have an influence. To the authors knowledge there have been no known cases in Australia
where design engineers have amended the 100% erosion damage limit for a specific project.
AUSTROADS 2004
— 2.15 —
5. REINFORCEMENT FOR CRCP

Section 9.5 of the Austroads Guide details equations to assess the minimum requirement for reinforcement in
JRCP and CRCP. The equations to estimate the amount of longitudinal and transverse reinforcement has
been based on theoretical modelling with limits derived from experience (Haber and Cruickshank 1981).
In the USA, AASHTO has reinforcement guidelines in the form of nomographs derived from algorithms
based on performance data. These results or algorithms were not used in the design procedure adopted by
Austroads.
The design procedure and limits are based on the production of evenly spaced “tight” transverse cracks based
on the premise that the concrete will shrink and expand, there is interlayer friction and the reinforcement will
hold the cracks together. The two equations used to derive those in the Guide are based on balancing the
bond force between the concrete and steel with the tensile strength of the concrete, that is:
fb Lcr ∑πd = fct Ac (1)
where fb = is the average bond strength,
fct = concrete tensile strength,
Lcr = theoretical critical spacing of cracks,
d = diameter of reinforcement, and
Ac = area of concrete related to reinforcement.
The second important equation is:
r = (∑πd2) / 4Ac (2)
In this equation “r” is the ratio of reinforcement area to the area of concrete.
Based on experience in Australia, the minimum amount of longitudinal reinforcement to achieve a suitable
crack pattern was set at 0.6% in the early 1990s. This minimum limit has been raised in recent years based
on further pavement performance studies at the RTA, NSW.
Probably one of the shortcomings of the reinforcement procedure in the Guide is the lack of attention to
placing an upper bound to concrete strength, such that the reinforcement will not yield and the crack widths
become larger than desired.
AUSTROADS 2004
— 2.16 —
6. OVERVIEW OF DESIGN RESULTS

6.1 General
Arising from the changed methodology and criteria in the 1992 Austroads Design Guide, this commentary
would be incomplete without some consideration of relative results for different inputs and design options. A
comparison of thicknesses for a typical major Australian highway traffic load using a variety of International
design guides and 1992 Austroads is given in Hodgkinson (1993) with interpretations as noted.
Within the 1992 Austroads Guide and because of the additional features of designing for shoulders and joint
types and the addition of the erosion analysis some clear trends emerge from the criteria used for design. It
should be noted that Austroads has no recommended preference for pavement type. Each pavement type
should reach and exceed its design traffic life when designed, detailed and constructed in accordance with
best practice. The following comments provide a basis for informed consideration by designers.
The example design charts in Chapter 9 (refer to Figures 9.7 to 9.10) use a flexural strength of 4.25 MPa that
approximates to a 28-day characteristic compressive strength of 32 MPa. However, this does not constitute a
recommendation by Austroads for concrete strength. The design procedure can accommodate various
concrete strengths and the optimum strength should be selected by the designer based on considerations of
variations in thickness, durability and costs associated with varying strengths.
6.2 Minimum Thickness Limit

For practical purposes and within the scope of Chapter 9 the tables and nomographs have been based on
traffic volumes (CVAG) in the range 106 to 3x108 and thickness in the range of 150 to 350 mm. These limits
should do not be interpreted as the limits for concrete pavements but practical limits for Australian traffic
conditions. Constructing road pavements below 150 mm are possible but require good quality control by the
contractor. For further information on the thickness design for base thickness less than 150 mm refer to
Austroads Lightly Traffic Roads Guide (APRG 1997).
6.3 Maximum Wheel Load

The upper limit on wheel load for the nomographs is 65 kN and there are no sighted references to justify this
limit. Table 3 shows the static load equivalence for various axle groups for a load safety factor of 1.2 and it
is apparent that the upper load limit is sufficient for Australian legal axle load limits. There is no guidance in
Australia or overseas literature on using the thickness design method for wheel loads exceeding 65 kN.
The use of the design method for forklift trucks and other industrial vehicles is not recommended.
Across the range of input factors the presence of a concrete shoulder will reduce base thickness from about
25 mm at lower traffic levels to about 40 to 50 mm at the higher traffic levels (Fig. 17). The 30 years of
experience constructing pavements with concrete shoulders in Australia has shown that concrete shoulders
improves pavement performance compared to flexible pavement shoulders (Zollinger and Barenberg 1989).
AUSTROADS 2004
— 2.17 —
330
310
290
without shoulder
Base Thickness
270
250
230
with shoulder
210
190
170
150
1.0E+06 2.1E+07 4.1E+07 6.1E+07 8.1E+07
CVAG
Figure 17: Base thickness for varying CVAGs for plain concrete pavement with and without
dowels. Note: Chart for PCP, LSF = 1.1, undowelled joints, effective CBR15% and rural traffic
distribution
Table 3
Static Load Equivalence for Various Commercial Vehicle Axle Groups with a
Maximum Wheel Load of 65 kN and a LSF of 1.2
Static Axle Legal Limit

Axle Group Group Load
(kN) (kN)
Single axle, single wheel (SS) 108 60
Single axle, dual wheel (SD) 217 90
Dual axle, dual wheel (TAD) 433 165
Triaxle, dual wheel (TRD) 650 200
6.4 Influence of Concrete Shoulders

Across the range of input factors the presence of a concrete shoulder will reduce base thickness from about
25 mm at lower traffic levels to about 40 to 50 mm at the higher traffic levels (Fig. 18). The 30 years of
experience constructing pavements with concrete shoulders in Australia has shown that concrete shoulders
improves pavement performance compared to flexible pavement shoulders (Zollinger and Barnberg 1989).
330
310
Base Thickness (mm)
290 without shoulder

270
250
230
210 with shoulder
190
170
150
1.0E+06 2.1E+07 4.1E+07 6.1E+07 8.1E+07
CVAGs
Figure 18: Base thickness for varying CVAGs for plain concrete pavement with
and without dowels. Note: Chart for PCP, LSF = 1.1, undowelled joints, effective
CBR of 15% and Rural axle load distribution
AUSTROADS 2004
— 2.18 —
6.5 Relative Influence of Traffic Volumes and Axle Loads

By inspection from the example design charts in the Guide (Figures 9.7 to 9.10) it can be seen that the
influence of axle loads on base thickness is more crucial than that of traffic volumes. For example, a PCP
with a rural-traffic distribution and with undowelled transverse joints the increase base thickness is
approximately equal to the increase in axle loads (i.e. a LSF jump from 1.1 to 1.2). For the same pavement
type with an increase of pavement thickness of about 10%, the traffic volume capacity is doubled.
Also, by inspection of the worked example in Appendix I, only the upper 'tail' of axle loads influences the
design (i.e. lower axle loads typically have unlimited repetitions to failure).
Some judgement needs to be exercised in relation to a literal interpretation of the 100% damage levels. For
practical purposes pavement bases are likely to be specified in thickness increments of 5 or 10 mm. It is a
characteristic of the design procedure that as a trial thickness approaches the design thickness, small changes
in trial thickness can have a considerable impact on fatigue and erosion damage. Most designers would
specify the base thickness to the nearest 5 mm and therefore, the cumulative damage is likely to be less than
95%.
Fig. 19 shows that at a lower volume of traffic the erosion analysis does not have a significant influence on
design base thickness. As the numbers of vehicles and the severity of the loading regime increase three
trends emerge, namely:
• The 100% damage limit will occur for fatigue or erosion damage much more quickly than the other. It is
unlikely that 100% damage can be optimised for both.
• For dowelled jointed pavements the fatigue analysis tends to control the joint. This is explained by the
relatively smaller joint faulting at a dowelled joint than at a joint relying on aggregate interlock and
therefore, the lower likelihood of erosion controlling base thickness.
For undowelled jointed pavements the erosion analysis tends to control the base thickness. This is due to the
relatively larger deflections at transverse joints relying on aggregate interlock for load transfer and therefore,
the lower likelihood of flexural fatigue controlling the base thickness.
300
Concrete Base Thickness (mm)
Fatigue related distress Erosion related distress

250
200
150
1.00E+6 3.00E+6 1.00E+7 3.00E+7 1.00E+8

Commercial Vehicle Axle Groups
Figure 19: Typical rigid pavement thickness design curve for a specific effective CBR
and concrete flexural strength demonstrating that erosion of the transverse joints is
generally the dominant distress mechanism for high-volume vehicle traffic
AUSTROADS 2004
— 2.19 —
6.6 Influence of Pavement Type

The base thickness for a PCP-D, CRCP and JRCP are considered to be the same pavement thickness and this
has been based on experience in the field. A PCP will always be significantly higher than a dowelled joint,
but the additional cost of concrete is offset by the additional expense and difficulty to place dowels when
slipforming concrete.
In arriving at the optimum economic design there is justification in considering the combination of pavement
type and subbase before arriving at the preferred design.
AUSTROADS 2004
— 2.20 —
REFERENCES
APRG (1997). A Guide to the Design of New Pavements for Light Traffic. Austroads Pavement Research
Group. Published by ARRB Transport Research, Vermont South, Australian.
Sydney.
Ayton, G.P. (1993). Concrete Highway Pavements in Australia. Proc. 8th Int. Conf. on Concrete Pavements,
Purdue University, USA.
Haber, E. and Cruickshank, J. (1981). Design Procedure for CRCP Based On Theoretical Considerations
and Service Behaviour. Proc. 2nd International Conference on Concrete Pavement Design, Purdue
University (USA).
Heinrichs, K. et al. (1988). Rigid Pavement Design and Analysis. Federal Highway Administration (USA)
Report No. FHWA-RD-88-068.
Hodgkinson, J. (1993). Contemporary Design Methods For Concrete Highway Pavements. Proc. 16th
Biennial Conf., Concrete Inst. of Australia.
NAASRA (1979). Interim Guide to Pavement Thickness Design. National Assoc. of Australian State Road
Authorities, Sydney.
Packard, R.G. (1977). Design Considerations for Control of Joint Faulting of Undowelled Pavements. Proc.
1st International Conference on Concrete Pavement Design, Purdue University, USA.
Packard, R. and Ray, G.K. (1986). Update of Portland Cement Concrete Pavement Design. American
Society of Civil Engineers Highway Division, May.
Packard, R. and Tayabji, S. (1985). New PCA Thickness Design Procedure For Concrete Highway and
Street Pavements. Proc 3rd Int. Conf. on Concrete Pavements, Purdue University, USA.
PCA (1984). Thickness Design for Concrete Highway and Street Pavement. Portland Cement Association
(USA), EBA 209.01P.
Ray, M.A. (1981). European Synthesis on Drainage Subbase Erodibility and Load Transfer in Concrete
Pavements. Proc. 2nd Int. Conf. on Concrete Pavement Design, Purdue University, USA.
RTA (1996). Concrete Pavement Manual–- Design and Construction. Roads and Traffic Authority, NSW
Edition 2, 5 June.
Vorobieff, G. (1998). Australian Concrete Road Training Manual. Head to Head International Edition 1.4,
Sydney.
Yoder, E.J. (1978). Design Principles and Practices - Concrete Pavements. Proc. 9th ARRB Conf.
Zollinger, D.G. and Barenberg, E.J. (1989). A Mechanistic Based Design Procedure for Jointed Concrete
Pavements. Proc 4th Int. Conf. on Concrete Pavements, Purdue University, USA.
AUSTROADS 2004
— 2.21 —

Part 3: 2004 Guide Procedures for Design of Flexible Pavements
Geoff Jameson
December 2003
SUMMARY
This report records the work undertaken to revise the Austroads Pavement Design – A Guide to the
Structural Design of Road Pavements for publication in 2004. The Guide was initially published by the
National Association of Australian State Road Authorities (NAASRA) in 1987 and subsequently revised and
re-issued by Austroads in 1992.
The work undertaken to develop previous edition of the Guide plus its predecessor, the Interim Guide to
Pavement Thickness Design, has been well-documented by Potter (Part 1 of this report).
This report records the work undertaken to revise the guidelines for the design of flexible pavements for the
2004 edition of the Austroads Guide (Austroads, 2004).
The major changes to the Austroads guidelines are:
• changes to the pavement response model, including the use of full Standard Axle rather than a half axle
loading to calculate critical strains;
• improved procedures for estimating the design moduli of selected subgrade materials and unbound
granular materials;
• improved methods of estimating asphalt design moduli, including the derivation of design moduli from
measured moduli obtained using the indirect tensile test;
• revision of the subgrade strain criterion for application with the full Standard Axle load;
• procedures to enable design to a desired reliability of the constructed pavement outlasting the design
traffic;
• modification to the cemented materials fatigue and asphalt fatigue relationships to include variable factors
to enable design to a desired project reliability; and
• significant revision to the procedures used to calculate the design traffic, including a database of traffic
load distributions obtained at over 100 WIM sites throughout Australia.
This report does not address the changes made to Chapter 9 of the 1992 Guide – the Design of Rigid
Pavements – which is the subject of another report in this document (Part 4).
AUSTROADS 2004
— 3.i —
TABLE OF CONTENTS
Page
1. INTRODUCTION............................................................................................................................. 3-1
2. ESTABLISHMENT OF WORKING GROUPS TO REVISE THE 1987 GUIDE .............................. 3-1
3. GRANULAR PAVEMENTS WITH THIN BITUMINOUS SURFACINGS........................................ 3-2
4. DEVELOPMENT OF THE MECHANISTIC PROCEDURE FOR FLEXIBLE
PAVEMENTS .................................................................................................................................. 3-3
4.1 Response Model ..................................................................................................................... 3-3
4.2 Elastic Characterisation .......................................................................................................... 3-4
4.3 Performance Relationships ................................................................................................... 3-11
5. DESIGN TRAFFIC ........................................................................................................................ 3-17

5.1 Units of Loading .................................................................................................................... 3-17
5.2 Traffic Load Distributions ...................................................................................................... 3-19
6. PROJECT RELIABILITY .............................................................................................................. 3-20

6.1 Definition................................................................................................................................ 3-20
6.2 Procedures in 1992 Guide .................................................................................................... 3-20
6.3 Development of Procedures in draft 2001 Guide.................................................................. 3-21
6.4 Development of Procedures in 2004 Guide .......................................................................... 3-23
7. SUMMARY .................................................................................................................................... 3-23

REFERENCES ........................................................................................................................................ 3-24
APPENDIX A ORIGINS OF UNBOUND GRANULAR THICKNESS CHART ...................................... 3-26
APPENDIX B CHARACTERISATION OF GRANULAR MATERIALS AND
DEVELOPMENT OF A SUBGRADE STRAIN CRITERION .................................................................. 3-27
APPENDIX C GRANULAR MATERIALS MODULI UNDER ASPHALT AND
CEMENTED MATERIAL (APRIL 1998) ................................................................................................. 3-47
APPENDIX D GRANULAR MATERIALS MODULI UNDER ASPHALT AND
CEMENTED MATERIAL (JUNE 2003)................................................................................................... 3-55
APPENDIX E RELATIONSHIP BETWEEN UNCONFINED COMPRESSIVE
STRENGTH AND FLEXURAL MODULUS FOR CEMENTED MATERIALS ........................................ 3-60
APPENDIX F DISCUSSION NOTE ON ESTIMATING CEMENTED MATERIALS
MODULUS FROM UCS (JUNE 2000) .................................................................................................... 3-63
APPENDIX G DEVELOPMENT OF RELATIONSHIP TO ADJUST ITT MODULUS
FROM TEST LOADING RATE TO OPERATING SPEED IN-SERVICE................................................ 3-65
APPENDIX H DEVELOPMENT OF RELATIONSHIP TO ADJUST ITT MODULUS
FROM MEASUREMENT TEMPERATURE TO WMAPT ....................................................................... 3-68
APPENDIX I DEVELOPMENT OF RELATIONSHIP TO ADJUST ITT MODULUS
FROM TEST SPECIMEN AIR VOIDS TO IN-SERVICE AIR VOIDS ..................................................... 3-70
APPENDIX J DEVELOPMENT OF PRESUMPTIVE TRAFFIC LOAD DISTRIBUTIONS.................... 3-71
AUSTROADS 2004
— 3. ii —
TABLES
Page
Table 1: Reliability Traffic Multipliers for Asphalt Surfaced Cemented Treated Base Pavements 3-13
Table 2: 2004 Guide Suggested Reliability Factors for Cemented Materials Fatigue 3-14
Table 3: Reliability Traffic Multipliers for Full Depth Asphalt Pavements Desired
Project Reliability 3-16
Table 4: 2004 Guide Suggested Reliability Factors (RF) for Asphalt Fatigue 3-16
Table 5: Axle Group Loads Which Cause Equal Damage as a Standard Axle............... 3-17
Table 6: Distress Mode Strain Dependency 3-18
Table 7: Characteristics of Presumptive Traffic Load Distributions (TLDs)................... 3-19
Table 8: Average Project Reliabilities of Outlasting Design Traffic for Flexible
Pavements Pavement Type 3-21
Table 9: Traffic Multipliers to Outlast the Design Traffic 3-22
Table 10: Traffic Multipliers to Outlast the Design Period 3-22
FIGURES
Page
Figure 1 Standard Axle loading used in Austroads Pavement Design Model
for Mechanistic Design of Flexible pavements .......................................................................... 3-4
Figure 2 Brown’s (1973) loading time equation and 1/V approximation ................................................ 3-10
Figure 3 Comparison of allowable loadings in terms of asphalt fatigue of 1987 Guide
and draft 2001 Guide with and without use of a shift factor..................................................... 3-15
AUSTROADS 2004
— 3. iii —
1. INTRODUCTION
This report records the work undertaken to revise the Pavement Design – A Guide to the Structural Design of
Road Pavements for publication in 2004. The Guide was initially published by the National Association of
Australian State Road Authorities (NAASRA) in 1987 and subsequently revised and re-issued by Austroads
in 1992.
The work undertaken to develop the 1992 edition of the Guide, plus its predecessor, the Interim Guide to
Pavement Thickness Design (NAASRA 1979), has been well-documented by Potter (refer Part 1 of this
report).
The work undertaken to develop the guidelines for design of new rigid pavements in the 2004 Guide is
described by Vorobieff (refer Part 4 of this report).
2. ESTABLISHMENT OF WORKING GROUPS TO REVISE

THE 1987 GUIDE
In 1989, the (then) Austroads Pavement Research Group (APRG) established a Working Group to revise the
rigid pavement design guidelines of the 1987 Guide. The 1987 Guide had been published without major
revision to the IGPTD procedures that had been based on the 1966 US Portland Cement Association (PCA)
method for thickness design. This review was warranted in view of PCA’s revision of their design method in
1984.
The Working Group for the Design of Rigid Pavements comprised:

Mr Ed Haber Department of Main Roads, NSW (Convenor)
Mr David Potter Australian Road Research Board
Mr Geoff Jameson Country Roads Board, Victoria
Mr John Hodgkinson Cement &Concrete Association of Australia (Federal Office)
Mr Alan Pearson Cement &Concrete Association of Australia (NSW Office)
The revised procedures for rigid pavement design were published in the 1992 Austroads Guide. The
development of these procedures is discussed by Vorobieff (refer Part 4 of this report).
In late 1993, APRG established a Working Group on Asphalt Characterisation with the task of revising the
text of Chapter 6 of the 1992 Guide. Membership of this Working Group comprised:
Mr David Potter ARRB Transport Research (Convenor)
Mr Mike Butcher Department of Road Transport, South Australia
Mr Allan Jones Queensland Department of Main Roads
Dr Peter Kadar CERTS International
Mr John Lancaster Pioneer Road Services
Mr Geoff Jameson ARRB Transport Research
Mr Kieran Sharp ARRB Transport Research
In 1996, this Working Group produced APRG Document 96/03 (DA) titled Pavement Materials –Asphalt:
Draft Revision of the Austroads Pavement Design Guide.
AUSTROADS 2004
— 3.1 —
Also established in late 1993 was an Austroads Working Group on Reliability of Pavement Design.
Membership of this Working Group comprised:
Mr David Potter ARRB TR (Convenor)
Mr Geoff Jameson ARRB TR
Mr Geoff Ayton Roads &Traffic Authority, NSW
Mr Ian Rickards Pioneer Road Services
Mr George Vorobieff Cement and Concrete Association of Australia (C&CAA)
Under the direction of this Working Group, ARRB TR produced two substantial reports (Potter et al. 1996;
Moffatt et al. 1998) which formed the basis of the reliability procedures adopted in the 2004 Austroads
Guide.
In 1998 APRG established a Reference Group for the revision of 1992 Austroads Guide. Members of the
Reference Group comprised:
Mr Lance Midgley VicRoads (Convenor)
Mr Kieran Sharp ARRB Transport Research
Mr Chris Mathias Transport SA
Mr Frank Butkus Main Roads Western Australia
Ms Narelle Dobson Queensland Department of Main Roads
Mr Peter Tamsett Roads & Traffic Authority, NSW
Mr David Dash Roads & Traffic Authority, NSW
Mr Geoff Ayton Roads & Traffic Authority, NSW
Mr Greg Arnold Transit New Zealand
Mr David Alabaster Transit New Zealand (replacing Greg Arnold)
Mr Andrew Papacostas VicRoads
Mr Ralph Rallings Pitt & Sherry (representing Department of Infrastructure and
Environmental Services), Tasmania
Mr David Mangan Australian Asphalt Pavement Association (AAPA)
Mr George Vorobieff Australian Stabilisation Industry Association
Mr David Chatwin C&CAA
Mr Scott Matthews C&CAA (replacing David Chatwin)
Mr David Potter Consultant
Mr Geoff Youdale Consultant
Throughout this report this group is referred to as the 2001 Guide Reference Group (RG).
This Reference Group produced the 2001 Austroads Pavement Design Guide (Final Draft) in November
2001, which was issued as a draft for evaluation by users.
In late 2002 APRG re-established the Reference Group to review comments on the draft 2001 Guide and
finalise the text of the 2004 Guide. Member of the 2004 Guide Reference Group comprised:
Mr Steve Brown VicRoads (Convenor)
Mr Chris Mathias Transport SA
Mr Frank Butkus Main Roads Western Australia
Mr Peter Tamsett Roads & Traffic Authority, NSW
Mr Andrew Papacostas VicRoads
Mr David Alabaster Transit New Zealand
AUSTROADS 2004
— 3.2 —
Mr David Mangan Australian Asphalt Pavement Association (AAPA)

Mr George Vorobieff Australian Stabilisation Industry Association
Mr Scott Matthews C&CAA
Mr Geoff Youdale Consultant
Throughout this report this group is referred to as the 2004 Guide Reference Group (RG).
AUSTROADS 2004
— 3.3 —
3. GRANULAR PAVEMENTS WITH THIN BITUMINOUS

SURFACINGS
The 2004 Guide RG decided to retain the CBR-Thickness-Traffic chart (Figure 8.4 of the Guide) for the
design of granular pavements with thin bituminous surfacings because it was the consensus view of
pavement designers that the field performance of pavements designed in accordance with it was consistent
with design expectations.
Jameson (1996) reviewed the origins of Figure 8.4, and, as an extract of this report is included Appendix A
of Part 1 of this document, it is not repeated here. Further details of interest are also contained in Part 1.
Briefly, this design chart evolved from the 1940s Californian State Highway design thickness curves. These
curves were then refined and extended by the UK Road Research Laboratory. With minor modifications,
these design curves have been widely used by Australian State Road Authorities for over 30 years. Whilst
there has been no research undertaken to verify these design curves, a limited survey of experienced
engineers indicated that pavements designed using these charts and constructed in accordance with State
Road Authority specifications had about a 90% chance of exceeding the design traffic loading. This survey
was largely based on the perceived performance of pavements in other than wet tropical areas of Australian.
Due to the moisture-sensitivity nature of unbound granular pavements, the performance of this type
pavement in heavily-trafficked, wet tropical areas is variable.
In the draft 2001 Guide it had been proposed to provide Reliability Traffic Multipliers (RTM) to adjust the
design traffic to enable design to a selected reliability of the project outlasting the design traffic. However,
comments received on the draft 2001 Guide expressed concern about the changes in thickness of unbound
granular pavements that would result for the use of the proposed RTMs. The 2004 Guide RG considered that
appropriate levels of reliability are built into this empirical chart across the range of design traffic levels
covered in the Guide rather than the 90% reliability assumed in the development of RTM. Consequently, the
2004 Guide RG decided to delete the RTMs for unbound granular pavements.
Implicit in the design procedure for granular pavements with thin bituminous surfacings (Figure 8.4), is a
terminal condition which is considered to be unacceptable and hence signifies the end of life of the
pavement. According to Potter (refer Part 1 of this report), the terminal conditions may be considered to be:
• an average rut depth of about 20 mm, and
• a terminal roughness about three times the initial roughness.
AUSTROADS 2004
— 3.4 —
4. DEVELOPMENT OF THE MECHANISTIC PROCEDURE FOR

FLEXIBLE PAVEMENTS
4.1 Response Model

The Austroads mechanistic pavement design model predicts the elastic response of a selected pavement
under a Standard Axle load. This predicted response is then used to estimate the allowable loading which is
determined using performance relationships (refer Section 4.3)
Potter (refer Part 1 of this report) describes the basis for the linear elastic model adopted in the 1987 and
2004 Guides for the mechanistic design of flexible pavements.
Given the above, the 2004 Guide RG decided to retain the CIRCLY response model. However, the
following changes to the model have resulted in changes in the calculated responses:
• In both the 1987 and 2004 Guides responses are calculated under a Standard Axle load. In the 1987
Guide, responses were calculated using only one of the two sets of wheels of a Standard Axle, as the
influence of the other set of wheels was presumed to be insignificant. Since the 1987 Guide was
published there have been significant increases in both the thickness and stiffness of pavements. As such,
it can no longer be assumed that the effect of the other set of wheels is insignificant for all pavement
configurations. Hence, full Standard Axle modelling as described in Fig. 1 has been adopted for all
flexible pavements, consistent with the procedures used for rigid pavements.
• In the 1987 Guide, the tyre-surface contact stress was 550 kPa, a typical value for cross-ply tyres
commonly used in Australia in the past. Radial tyres are now commonly used and inflation pressures of
500-1000 kPa have been measured in the field (Chowdhury and Rallings 1994). In recognition of the
more damaging effects of radial tyres an amendment to the 1992 Guide issued in 1997 (Austroads 1997)
changed the Standard Axle tyre pressure to 750 kPa.
• In the late 1990s, improvements were made to the linear elastic model CIRCLY. These improvements
significantly altered the calculated responses within thin asphalt layers.
Figure 1: Standard Axle loading used in Austroads Pavement Design Model

for Mechanistic Design of Flexible Pavements
AUSTROADS 2004
— 3.5 —
4.2 Elastic Characterisation
4.2.1 Isotropic or Anisotropic Characterisation?

This issue relates to the elastic characterisation of pavement and subgrade materials associated with the use
of CIRCLY. Isotropic materials have the same properties in all directions, whereas anisotropic materials do
not. In the Austroads Guide, bound material are characterised as isotropic whereas granular materials and
subgrades are characterised as cross-anisotropic, with vertical modulus being twice the horizontal modulus.
Potter (refer Part 1 of this report) provides a detailed description of the reasons behind this method of
characterisation. In summary:
• A literature review indicated that there was little evidence of anisotropic behaviour in the cases of asphalt
and cemented materials, while there was definite evidence of anisotropic behaviour in the cases of
granular and subgrade materials.
• A response-to-load analyses using finite element modelling (FEM) suggested that the adoption of
anisotropic characterisation of granular materials (with Eν/Eh > 1) within the CIRCLY model would be ‘a
step in the right direction’ towards encompassing their known stress-dependency.
• The use of anisotropic characterisation for both granular material and subgrades resulted in narrower
deflection bowls being predicted which were more in line with the shape of measured deflection bowls.
Consequently, cross-anisotropic characterisation of granular and subgrade materials is adopted in the
Austroads Guide to at least partly compensate for the absence of a lateral stress dependent mechanism for
elastic modulus in the linear elastic model (CIRCLY) adopted. As already discussed, for asphalt and
cemented materials, isotropic characterisation was considered to be adequate and, hence, was adopted.
4.2.2 Values for Poisson’s Ratios

Potter (refer Part 1 of this report) details the origin of the Poisson’s Ratios adopted in 1987 Guide. No
change was made to these values in the 2004 Guide.
4.2.3 Relationship between Subgrade Modulus and CBR

The linear elastic response model requires that the design subgrade modulus be estimated. As subgrades are
usually characterised in the CBR test in terms of strength rather than resilient modulus, a procedure was
required to estimate subgrade design modulus from CBR test results.
Potter (refer Part 1 of this report) describes in detail the origins of the E = 10.CBR relationship adopted in
the 1987 Guide.
It should be noted that the relationship E = 10.CBR was derived assuming isotropic characterisation of the
subgrade and hence some designers have questioned its suitability given the use of anisotropic subgrade
characterisation in the Guide. As discussed in Appendix B (Moffatt and Jameson 1998a), it was estimated
that E = 9.1 x CBR1.03 was an equivalent relationship using anisotropic characterisation. However, the use of
this more complex relationship was not recommended as it did not produce significantly different allowable
loadings than the simpler E = 10.CBR relationship.
Modulus of Selected Subgrade Materials

In using the 1987 Guide, it was common practice to characterise the elastic properties of selected subgrade
materials in the same manner as for the underlying insitu subgrade.
The 2004 RG were concerned that this practice often resulted in the moduli of selected subgrade material
being higher than those of granular materials placed in the same layer. This practice did not recognise that
the modulus of a select material depends on the modulus of the layer on which it is compacted: the entire
layer thickness of select material was assumed to have a vertical modulus of 10 times it CBR regardless of
the strength of the underlying material.
AUSTROADS 2004
— 3.6 —
Consequently, the 2004 RG decided that the Austroads design model should be changed to require sub-
layering of selected subgrade material in a similar manner as for granular materials (refer Section 4.2.6).
For granular materials, the moduli are calculated for each of five sub-layers assuming that, for a 125 mm
sub-layer thickness, the modulus is double the modulus of the underlying material. For selected subgrade
material it was considered that a greater thickness of material was required to double the modulus as this
would reflect the lower bearing capacity of selected subgrade materials compared to unbound granular
materials. Consequently, the 2004 RG decided to base the sub-layering rules on the assumption that the
modulus is two times the modulus of the underlying material for every 150 mm of select material.
This led to the following design rules for selected subgrade materials:
a) Divide the total depth of all selected subgrade materials into 5 equi-thick sub-layers.
b) The vertical modulus of the top sub-layer of selected subgrade is the minimum of 10 times the design
CBR of the selected subgrade material and that determined using:
E V top sublayer = E V insitu subgrade × 2(total selected subgrade thickness 150)

c) The ratio (R) of the moduli (E) of the adjacent sub-layers is given by:
1
⎡ E top selected subgrade sublayer ⎤ 5
R = ⎢ ⎥
⎣⎢ E insitu subgrade ⎦⎥
d) The modulus of each sub-layer may then be calculated from the modulus of the adjacent underlying
sub-layer, beginning with the insitu subgrade, the modulus of which is known.
Where the trial pavement configuration includes more than one type of selected subgrade material, a check
needs to be made that the vertical modulus calculated for each sub-layer (step (d)) does not exceed 10 times
the design CBR of each selected subgrade material within the sub-layer. If this condition is not met, an
alternative trial subgrade configuration needs to be selected.
4.2.5 Sub-Layering of Granular Materials and Assignment of Moduli

The typical modulus values for unbound granular materials listed in Tables 6.3 of the 2004 Guide are based
on the results of triaxial testing undertaken by Youdale (1978) on a range of materials. The values for base
materials are those achieved in the upper portion of the material when covered by a thin bituminous
surfacing.
As stated by Potter (refer Part 1 of this report):
“With the knowledge that:
• the modulus of a granular material depends on the stress conditions it is subjected to;
• these stress conditions depend on the stiffness(es) of the underlying material(s);
• the stress levels decrease within a pavement structure as the depth below the surface
increases; and
• in the response model CIRCLY, layers are constrained to having a fixed modulus
value;
then it was obvious that the simulation within CIRCLY of actual stress-strain conditions would
be improved for many pavement configurations by considering the granular material to be
comprised of more than one layer, each with a distinct modulus value.
The rules for determining when the sub-layering is necessary and, if so, the appropriate number
of sub-layers, together with appropriate modulus for each sub-layer, were presented in Section
8.2.2 of the 1987 Guide.”
Potter (refer Part 1 of this report) details the origin of the 1987 Guide sub-layering rules.
AUSTROADS 2004
— 3.7 —
In reviewing the design method for the Guide, the 2004 Guide RG considered that the 1987 rules could be
inappropriately used by less experienced designers as they allowed the maximum modulus of a granular
material to be developed regardless of the thickness of the granular material or the strength of the subgrade.
As detailed in Appendix B, the revised rules adopted by the 2004 Guide RG were developed such that the
maximum modulus of a granular material depends on the both the granular material thickness and subgrade
strength. These rules are presented in Section 8.2.3 of the Guide.
These new rules significantly changed the calculated asphalt strains for some pavement configurations and
this needed to be taken into account in the developing the reliability procedures (see Section 6).
4.2.6 Modulus of Top Sublayer of Granular Material

The entries in Table 6.6 of the 1987 Guide were derived from results of a study by Youdale (1983) wherein a
finite element package (PAVAN1) was used to determine modulus values in a granular material which was
surfaced with asphalt and subjected to surface loading representing one side of a Standard Axle. Potter (refer
Part 1 of this report) details aspects of this analysis.
Table 6.6 of the 1987 Guide listed the variation in maximum modulus of the granular material for various
asphalt temperatures. With the wide variety of asphalt mixes currently used, it was considered more
appropriate to modify Table 6.6 for the 2004 Guide so that it gave the variation in maximum modulus with
the modulus of the overlaying material (asphalt and/or cemented material).
Accordingly, based on work detailed in Appendix C, revised granular moduli were developed for the draft
2001 Guide.
For some pavement configurations these draft 2001 Guide granular moduli were significantly less than those
in the 1987 Guide. Users of the draft Guide expressed concern that these moduli resulted in low asphalt
fatigue lives for thin asphalt-surfaced pavements. It was concluded that considerably more research was
required – possibly involving finite element modelling – before significant changes to the granular moduli in
the 1987 Guide could be made.
Consequently, the 2004 Guide RG requested that granular moduli be developed which were reasonably
consistent with the values in the 1987 Guide. Using the granular moduli in Table 6.6 of the 1987 Guide, and
assuming the modulus of asphalt at 20°C, 25°C and 30°C was 5,200 MPa, 3,500 MPa and 2,300 MPa
respectively, the following relationship was developed for granular moduli under various thicknesses and
stiffnesses of asphalt relative to the value under a sprayed seal surface (see Appendix D):
RGM = min(1, 1.377 − 0.00003804 * ASTH 3 E AC ) (1)
where RGM = relative granular moduli, granular moduli under asphalt divided by the granular
moduli under a sprayed seal surface, maximum value 1.0 and a minimum of 0.429
(= 150/350) for Normal Standard granular material and 0.42 (=210/500) for High Standard
granular material;
ASTH = thickness of asphalt overlying granular material (mm); and
EAC = asphalt modulus (MPa).
Similar procedures were developed to demonstrate that there was no significant difference in the maximum
top-layer moduli if the material overlying the granular material was cemented rather than asphalt provided
the moduli were the same (Appendix C). Hence the revised Table 6.4 in the 2004 Guide gives the top-layer
granular material for various thicknesses and moduli of asphalt and or cemented material.
AUSTROADS 2004
— 3.8 —
4.2.7 Modulus of Granular Material Overlying a Cemented Layer

Potter (refer Part 1 of this report) reports:
“Section 8.2.2(1) of the Guide states, “For granular materials placed directly on a stiff cemented
sub-base, no sub-layering is required”. This statement is based on results of a finite element
analysis undertaken by Youdale. The modelling and analysis was as outlined above (Section
4.2.7). Six pavements were modelled – granular thicknesses of 150, 200 and 250 mm on
cemented material having stiffnesses of 2,000 and 5,000 MPa. The study is reported in
Youdale (1984b).”
No change was made to these moduli in the 2004 Guide.
4.2.8 Modulus of Cemented Materials

The typical modulus values for cemented material listed in Table 6.6 of the 1987 Guide were based on
limited testing undertaken by the Queensland Department of Main Roads (QDMR), supplemented by test
results reported in the literature. The values were revised by the 2004 RG to reflect current typical design
moduli.
Two relationships were presented in Section 6.3.2.3 of the 1987 Guide for estimating the modulus of
cemented material from its UCS value, one for cemented crushed rock and one for cemented natural gravel:
E = 1814 UCS0.88 + 3500 For cemented crushed rock (2)
0.88
E = 2240 UCS + 1100 For cemented natural gravel (3)
Potter (refer Part 1 of this report ) presents the origin of these equations.
These relationships were reviewed by Yeo (1997). It was concluded that Otte’s data, from which the above
relationships were derived, was too variable to assign different relationships for the two types of cemented
material. In addition, the draft 2001 Guide RG considered that a simple linear relationship between UCS and
modulus was warranted given this variability. Moffatt and Yeo (1998) subsequently recommended the
following relationship (refer Appendix E):
E flex = 3,000 UCS (4)
In late 1999, the draft 2001 Guide RG decided against adopting eqn (4) as the moduli estimated were
considerably higher than the presumptive moduli used by State Road Authorities.
In a subsequent note prepared for the draft 2001 Guide RG (Appendix F), results reported by Alderson
(1999) were analysed. Alderson reported the following relationship between the flexural modulus of
laboratory beams at 28 days and the UCS of laboratory specimens at 28 days:
E flex lab = 2,460 UCS (5)
In this limited study, Alderson also reported that the laboratory flexural moduli were about 2.5 times higher
than those of the field beams. However, the field beams were about 3% lower in density than the laboratory
beams. Correcting for this difference in air voids, the laboratory beam moduli were about 1.6 times the field
beam values. Applying this factor to Alderson’s relationship, the following relationship was derived to
estimate the field 28-day flexural modulus from the UCS tests of laboratory specimens at 28 days:
E flexfield = 1,500 UCS (6)
This relationship was adopted in the draft 2001 Guide.

Comments received on the draft 2001 Guide expressed concern that the moduli predicted using eqn (6) were
higher than common road agency practices. For example:
• RTA NSW use a design modulus of 5,000 MPa for 28-day UCS of 4MPa; and
• QDMR use design moduli of 2,000 MPa and 5,000 MPa for 7-day UCS values of 2 and 3 MPa
respectively. For GP cement these are equivalent to 28-day UCS values of about 2.9 MPa and 4.3 MPa
respectively.
AUSTROADS 2004
— 3.9 —
Whereas eqn (6) allows the mean modulus estimated from a UCS value obtained by testing a range of
materials to be calculated, current practice is to use a conservative estimate of modulus which reflects the
uncertainty associated with estimating the design modulus of a given material from its UCS. The
relationship between UCS and modulus varies with Road Agency laboratory testing practices and
construction specifications for cemented materials. Consequently, the 2004 Guide RG decided to adopt the
following relationship, applicable to UCS values up to 5 MPa:
EFLEX = k UCS (7)
where EFLEX = flexural modulus of field beams at 28 days moist curing (MPa);
UCS = Unconfined Compressive Strength of laboratory specimens
at 28 days (MPa); and
k = values of 1,000 to 1,250 are typically used for General Purpose cements, the value
depending on laboratory testing practices and construction specifications for
cemented materials.
4.2.9 Characterisation of Cracked Cemented Materials

As stated by Potter (refer Part 1 of this report ):
“Section 8.5 of the 1987 Guide discusses the possibility of significant life remaining in a
pavement subsequent to the cracking of cemented material, indicating that it is appropriate to
characterise the cracked material as granular material. The intention of this latter statement is
that the cracked material be considered as a granular material in all respects (2:1 anisotropy,
Poisson’s Ratio = 0.35, sub-layering as per Section 8.2.2). Due to an oversight, no guidance
was offered on which of the two sets of entries in Table 6.6 (for Modified and Standard
compactive efforts) was appropriate. The intention was that materials whose initial stiffnesses
were 5,000 MPa or more be assigned values corresponding to Modified compactive effort, and
materials with lower stiffnesses be assigned values corresponding to Standard compactive
effort.”
Since the 1987 Guide was published, VicRoads and QDMR have characterised cracked cement-treated
crushed rock using a vertical modulus of 500 MPa, a horizontal modulus of 250 MPa and without sub-
layering. These Authorities considered a well-graded cracked cemented material had moduli superior to
those of the material in the unbound state.
Hence, in the 2004 Guide, cracked cemented materials produced from a well-graded granular material are
assumed to have a presumptive vertical modulus of 500 MPa and a Poisson’s Ratio of 0.35. The layer is not
sub-layered and is considered to be cross-anisotropic, with a degree of anisotropy of 2.
Characterisation of Asphalt
1987 Guide
Section 6.4.3 of the 1987 Guide presented a procedure for measuring asphalt moduli and for estimating
modulus using the Shell nomographs (Shell 1978).
At the time the 1987 Guide was published, equipment and test procedures for routinely measuring resilient
modulus were not available. Consequently, most Road Authorities calculated their design moduli using the
Shell nomographs and:
• average values of local mix properties;
• local Weighted Mean Annual Pavement Temperature (WMAPT); and
• a loading time obtained using the relationship = 1/V given in Section 6.4.2.5 of the 1987 Guide
The WMAPT was generally used to characterise the pavement damage due to loads applied at various
operating temperatures, even though the 1987 Guide provided two methods: WMAPT and Pavement Life
Multipliers.
AUSTROADS 2004
— 3.10 —
The 1987 WG used procedures adapted from the Shell Pavement Design Manual (1978) to develop a
procedure for calculating WMAPTs. The procedure used is included in Appendix 6.1 of the 2001 Guide.
The 1987 WG recognised that there were two significant deficiencies in the use of WMAPTs to characterise
asphalt operating temperature, namely:
• it was assumed that loads were applied uniformly throughout the day, with no allowance made for
situations where the hourly loading varied with hourly asphalt temperature; and
• the weighting factors used to account for the damage at each operating temperature were applicable to
thick (>150 mm) asphalt layers rather than to thin asphalt layers.
Accordingly, Youdale (1984d) suggested an alternative approach, viz. the use of Pavement Life Multipliers
(PLM) for the design of asphalt-surfaced granular pavements.
To adjust asphalt moduli for the rate of loading, the 1987 WG suggested the following relationship between
loading time of a step-shaped pulse (t, seconds) and vehicle speed (V, km/h):
1
t = (8)
V
Eqn (8) seems to be approximation of the following equation developed by Brown (1973):
log t = 0.0005T − 0.2 − 0.94 log V (9)
where t = loading time (s),
V = vehicle speed (km/h), and
T = asphalt thickness (mm).
Eqn (8) is compared in Fig. 2 with eqn (9) for various thicknesses of asphalt.
It is apparent from Fig. 2 that the commonly used 1/V formula is a reasonable approximation to Brown’s
relationship for asphalt thicknesses of 200 mm, except for low speeds, where it overestimates loading time.
0.10
0.09
0.08
0.07
Loading Time (secs)
0.06
0.05
0.04
0.03
0.02
0.01
0.00
0 10 20 30 40 50 60 70 80 90 100
Vehicle Speed (km/h)
T=100mm T=150 mm T=200mm 1/V
Figure 2: Comparison of Brown’s (1973) loading time equation and 1/V approximation
AUSTROADS 2004
— 3.11 —
2004 Guide
The 2004 RG decided to include the following two methods of estimating asphalt design moduli:
• the resilient modulus measured using the standard indirect tensile test (ITT) adjusted to the in-service
temperature (WMAPT) and for the rate of loading in the road-bed; or
• use the Shell nomographs to estimate the bitumen properties and mix volumetrics and the in-service
temperature (WMAPT) and rate of loading in the road-bed.
The 2004 RG decided to use the WMAPT as the means of characterising the damage to the pavement over
its range of operating temperatures as most Road Authorities used WMAPT values. The 2004 Guide RG
also decided to delete the PLM from the 2004 Guide as they had been seldom used.
Jameson and Hopman (2000) developed an approach for adjusting the modulus for the rate of loading based
on modelling the viscoelastic characteristics of asphalt. The RG considered this approach but decided to
retain the 1/V formula for use with the Shell nomographs. The main reasons for this were as follows:
• Although Jameson and Hopman (2000) had developed a procedure for the variation in loading time with
asphalt depth, only the relationship for a 100 mm thick layer was applicable as the temperature
characterisation (WMAPT) was calculated only for a 100 mm depth of asphalt rather than allowing for
the variation in temperature with asphalt depth. If the more complex approach to loading time provided
by Jameson and Hopman were to be adopted then a more complex approach to temperature was also
required.
• The Jameson and Hopman loading time relationship for 100 mm depth was similar to the simpler 1/V
formula except for slow (<40 km/h) traffic speeds.
If the ITT results are to be used to estimate the design moduli, then procedures are required to adjust the
measured ITT moduli at 40 ms rise time to the in-service value at the design vehicle speed. As the ITT uses
neither a step-shaped pulse nor a sinusoidal pulse, the simple t = 1/V formula is not appropriate.
Accordingly, the following relationship was derived and adopted in the 2004 Guide.
Modulus at Speed V
= 0.17V 0.365 (10)
ITT Modulus
The derivation of this procedure is given in Appendix G. A procedure was also required to adjust the ITT
results from the measurement temperature (25°C) to the operating temperature (WMAPT). Appendix H
details the method used to develop the following relationship, which was also adopted in the 2004 Guide:
Modulus at WMAPT
= exp( −0.08(WMAPT − 25) (11)
Modulus at 25o C
Appendix I details the origins of the procedure adopted to adjust the moduli from the air voids of the
measured specimen to the in-service air voids.
The typical moduli values for asphalt listed in Table 6.10 of the Guide are based on extensive indirect tensile
(MATTA) testing conducted by VicRoads, RTA NSW and QDMR of laboratory-manufactured mixes
conforming to their specifications. The data includes specimens compacted using the Marshall drop hammer
and Gyratory compactor. Results obtained on specimens that had a test air void level other than 5% were
corrected to 5% using the relationship given in Equation 6.5 of the Guide.
AUSTROADS 2004
— 3.12 —
4.3 Performance Relationships
4.3.1 Subgrade Strain Criterion

The 1987 WG decided to follow the approach adopted by Shell (1978) and select the maximum vertical
strain generated at the top of the subgrade (by a Standard Axle) as the pavement response best suited for the
prediction of allowable traffic before permanent deformation in the granular layers and subgrade reached an
unacceptable level.
Potter (refer Part 1 of this report ) details the origin of the 1987 Guide subgrade strain criterion.
Due to changes in the 2004 Guide to the granular materials characterisation, the subgrade strain criterion
needed to be revised. Using a method similar to that used by Youdale (1984c), Moffatt and Jameson (1998a)
derived the following relationship:
N = (9300/µ∈)7 (14)
where N = the allowable number of Standard Axles of loading before an unacceptable level of
rutting and roughness; and
µε = the vertical compressive strain under a Standard Axle (microstrain) predicted using
the Austroads procedures (see also Appendix B).
Note that Figure 8.4 of the 1992 Guide, from which this relationship was derived, expresses the allowable
loading in terms of Equivalent Standard Axles (ESA). In the mechanistic design of new pavements, strains
are calculated under a Standard Axle and hence it is more appropriate to also express the allowable loading
in terms of Standard Axles of loading rather than ESAs (refer Section 4.4.1). Consequently, it was necessary
to convert the Figure 8.4 allowable loading from ESAs to Standard Axles. A factor of 1.1 was used by
Moffatt and Jameson (1998a), in line with Section 7.5 of the 1987 Guide.
Note that the subgrade strain criterion is based on the correlation of strains calculated using the Austroads
procedure with the observed allowable loading to a terminal amount of rutting and roughness (see Section 3)
due to deformation of the granular layers and subgrade.
In the draft 2001 Guide it had been proposed that Reliability Traffic Multipliers (RTM) be provided to adjust
the design traffic to enable design to a selected reliability of the project outlasting the design traffic (refer
Section 6.3). Comments received on the draft 2001 Guide expressed concern about the changes in thickness
of unbound granular pavements that would result from the use of the proposed RTMs. The 2004 Guide RG
considered that appropriate levels of reliability were built into the Figure 8.4 empirical chart across the range
of design traffic levels covered in the Guide rather than the 90% reliability assumed in the development of
the RTM.
Moreover, the 2004 RG considered that a simpler approach to reliability would be to delete the RTM and
incorporate variable factors for reliability in the performance relationships. As the subgrade strain
relationship was derived from Figure 8.4, and as appropriate levels of reliability were incorporated into this
empirical chart, there was no need to provide a Reliability Factor for this performance relationship.
However, as discussed in Sections 4.3.2 and 4.3.3, Reliability Factors were required for the cemented
materials fatigue and asphalt fatigue relationships.
AUSTROADS 2004
— 3.13 —
4.3.2 Fatigue Relationship for Cemented Materials

The performance relationship in the 1987 Guide was:
N = (K/µ∈)18 (15)
where N is the number of repetitions of tensile strain at the bottom of the cemented layer before fatigue
failure occurs, i.e. when the level of this strain is µ∈ microstrain. The numerator K depends on the stiffness
of the material as follows:
Material (MPa)
2,000 280
5,000 200
10,000 150
Potter (refer Part 1 of this report) details the development of these relationships.
In November 1997, a revision to the Guide was issued by Austroads (1997) which replaced eqn (15) with the
following:
N = (K/µ∈)12 (16)
and the numerator K depended on the stiffness of the material as follows:
Material (MPa)
2,000 440
3,500 350
5,000 310
10,000 260
15,000 240
As stated by Potter (refer Part 1 of this report):

“Subsequent to the 1987 WG’s adoption of the relationships with exponent 18, Angell (1988), in
the course of development of a pavement design manual for Main Roads, undertook a further
review of the literature and reported fatigue exponents of 32, 9, 12.7, 12.2 and 12 for
relationships developed by workers in four countries.
In light of this, together with his proposition that, if the true exponent were 18, then cement-
treated pavements in Queensland would be failing very early in their design life because of
vehicle overloading, Angell opted for an exponent of 12 and derived numerator (K1) values such
that the revised relationships (in his words) “allow approximately the same levels of strain as the
relationships previously used”. Angell developed the following relationships:
N = (K/µ∈)12 (17)
with values of K as follows:

Material (MPa)
2,000 440
5,000 310
15,000 240
The WG was apprised of MRDQ’s intention to adopt these revised relationships while the
original (1987 version) Guide was in press.”
AUSTROADS 2004
— 3.14 —
These relationships were adopted in the Austroads Pavement Design Guide Revision Note (1997), together
with additional relationships for materials with moduli of 3,500 and 10,000 MPa. The additional
relationships were determined from the following general relationship derived (Jameson et al. (1995)) from
Angell’s (1988) three relationships:
(
⎡ 112,664/ E0.804 + 190.7 ⎤
N =⎢ ⎥
) 12
(18)
⎣ µε ⎦
The issue of the Revision Note (Austroads 1997) was prompted by a literature review conducted by Jameson
(1995) and by the findings of an ALF trial on cemented materials (Jameson et al. 1995).
For the draft 2001 Guide, eqn (18) was simplified as follows:
N = ⎢
(
⎡ 113,000 / E 0.804 + 191 ⎤ ) 12
(19)
⎥
⎣ µε ⎦
where N = allowable number of repetitions of the load;
µε = tensile strain produced by the load (microstrain); and
E = modulus of cemented material modulus (MPa).
In the draft 2001 Guide, to design to a desired project reliability, the design traffic was adjusted using
Reliability Traffic Multipliers (RTM). Comments received on the draft 2001 Guide, indicated that the
cemented material fatigue lives were significantly under-predicted and modification to the procedures were
required.
The 2004 RG considered that a simpler approach to reliability would be to delete the RTM and incorporate a
variable factor for reliability in the cemented materials fatigue relationship. It was considered that eqn (18)
resulted in 95% reliability of a project outlasting the design traffic. To provide adjustment factors to enable
design to other reliabilities, the variability of performance between projects was required. Shown in Table 1
are the RTM given by Jameson and Moffatt (2001) (refer Table 5 below) for a pavement comprising up to
150 mm of asphalt on cemented materials. For this pavement type the dominant distress mode is cemented
materials fatigue.
Table 1
Reliability Traffic Multipliers for Asphalt Surfaced Cemented Treated Base Pavements
Desired Project Reliability

80% 85% 90% 95% 97.5%
2001 draft Guide RTMs: average reliability 0.70 1.0 1.6 3.3 6.5
of design process 85%
Equivalent RTMs for an average reliability 0.21 0.30 0.48 1.0 2.0
These RTM assume that the average reliability of the design process is 85% (RTM=1 for a desired project
reliability of 85%). As stated above, the 2004 Guide considered the above reliability of the design process
was 95% rather than 85%. Consequently, Table 1 shows the adjusted RTMs for an average reliability of
95% obtained by dividing the 85% RTM values by 3.3.
AUSTROADS 2004
— 3.15 —
Rather than provide for project reliability by adjusting the design traffic using these 95% RTM values, the
2004 Guide RG simplified the process by providing factors to the allowable loading predicted using the
cemented materials fatigue relationship. These factors are the inverse of the 95% RTM values given in Table
1. Hence the following cemented materials fatigue relationship was adopted in the 2004 Guide:
12
⎡ ⎛113,000/E 0.804 + 191⎞ ⎤
⎢ ⎜ ⎟
N = RF⎢ ⎝ ⎠⎥ (20)
⎥
⎢ µε ⎥
⎣ ⎦
µε = tensile strain produced by the load (microstrain);
E = modulus of cemented material modulus (MPa); and
RF = reliability factor for cemented materials fatigue (Table 2).
Table 2
2004 Guide Suggested Reliability Factors for
Cemented Materials Fatigue

80% 85% 90% 95% 97.5%
4.7 3.3 2.0 1.0 0.5
4.3.3 Fatigue Relationship for Asphalt

After due consideration, detailed by Potter (refer Part 1 of this report ), the 1987 WG adopted the Shell
relationship for recommended use in the Guide. However, as noted by Potter :
“It is now well understood that the Shell relationship is for fatigue failure in the laboratory and
not in the field. This fact was not appreciated by the WG at the time. Commenting on the Shell
relationship, Anderson stated that it is “supposedly directly applicable to design, allowing for
crack propagation through the asphalt.” In discussing desirable attributes of a relationship,
Anderson states:
“Consideration should be given to crack propagation and tolerable cracking as well as the
possibility of crack healing occurring under intermittent loading. Shell (1978) report field
to laboratory fatigue life correction factors of 10-20.”
Further, it has been noted above that the Shell relationship, as adopted, was considered by the
WG to provide allowance for crack propagation and to tolerate some cracking in the field.
Hence, while it was the desire and the intention of the WG that the adopted relationship
incorporate laboratory-to-field corrections, the relationship ultimately adopted contains no such
correction.
A detailed review of the development of the Shell laboratory fatigue relationships (including
materials tested, test conditions, etc.), together with Shell’s use of correction factors in its
design procedure, is provided in Potter (1997).”
As part of the development of the draft 2001 Guide, Jameson (1999) reviewed existing information on the
relationship between asphalt fatigue observed in the field (mainly from accelerated loading trials) and
performance predicted using the Austroads design procedures, including the Shell laboratory fatigue
relationship. It was noted that the observed asphalt fatigue lives generally exceeded the predicted lives and
Jameson concluded that the use of a laboratory-field correction factor, or shift factor, to the Shell laboratory
fatigue relationship needed to be considered for adoption with the project reliability procedures adopted in
the draft 2001 Guide.
AUSTROADS 2004
— 3.16 —
The draft 2001 Guide RG considered the effects on asphalt fatigue life of the following changes between the
1987 and draft 2001 Guides:
• granular materials characterisation, and
• full Standard Axle modelling, including an increase in contact tyre-surface stress to 750 kPa.
These changes reduced the predicted asphalt fatigue life, with the effect varying markedly with pavement
configuration as shown in Figure 3. Note that the draft 2001 Guide predictions are presented with and
without the use of an asphalt fatigue shift factor of five. The changes to the design method had most effect
on the allowable loading of thin asphalt-surfaced granular pavements and the least effect on full-depth
asphalt pavements.
50 mm asphalt on unbound granular 100 mm asphalt on unbound granular

subgrade CBR=5%, asphalt modulus 3000 MPa subgrade CBR=5% asphalt modulus 3000 MPa
550 500
500
450
450
400
Granular thickness (mm)
400

350 350
300
300
250
250
200
150 200
100
1.00E+03 1.00E+04 1.00E+05 1.00E+06 1.00E+07 150
Allowable loading (ESAs) 1.00E+04 1.00E+05 1.00E+06 1.00E+07
Allowable loading (ESAs)
2001 Guide w/o SF 2001 Guide with SF=5 1987 Guide
150 mm asphalt on unbound granular 200 mm asphalt on granular

subgrade CBR=5%, asphalt modulus 3000 MPa subgrade CBR=5%, asphalt modulus 3000 MPa
450
600
400
500 350
300
400
250
300 200
150
200
100
50
100
0
1.00E+06 1.00E+07 1.00E+08
0 Allowable loading (ESAs)
1.00E+05 1.00E+06 1.00E+07 2001 Guide w/o SF 2001 Guide with SF=5 1987 Guide
Allowable loading (ESAs)
Figure 3: Comparison of allowable loadings in terms of asphalt fatigue of 1987 Guide

and draft 2001 Guide with and without use of a shift factor
After considering these changes, the proposed changes to the reliability procedures and the comparison of
observed and predicted performance (Jameson 1999), the draft 2001 Guide RG decided to adopt a shift factor
of five to the Shell laboratory fatigue relationship as follows:
5
⎡ 6918(0.856 VB + 1.08) ⎤
N = 5⎢ 0.36 ⎥ (21)
⎢⎣ Smix µε ⎥⎦
VB = percentage by volume of bitumen in the asphalt (%); and
Smix = asphalt mix stiffness (modulus) (MPa).
Eqn (21) was provided in the draft 2001 Guide.
AUSTROADS 2004
— 3.17 —
The 2004 Guide RG reviewed comments received on the draft 2001 Guide. Concern was expressed of
significant reductions in thickness of thick (>150mm) asphalt surfaced pavements resulting from the draft
2001 Guide design process. Hence the 2004 RG reconsidered the shift factor of five and the use of RTM to
enable design to a desired project reliability.
The 2004 Guide RG considered that the use of the Shell laboratory fatigue relationship with a shift factor of
one resulted on average in a 95% probability of a project outlasting the design traffic. This average
reliability was largely based on the performance of dense-graded hot mix asphalt which has been widely
used in Australia. To provide adjustment factors to enable design to other reliabilities, the variability of
performance between projects was required. The RTM values given in Jameson and Moffatt (2001) (refer
Table 9 below) for full depth asphalt pavements are given in Table 3. For this pavement type the dominant
distress mode is asphalt fatigue. These RTM are based on the assumption that the average reliability of the
design process is 85%.
Table 3
Reliability Traffic Multipliers for Full Depth Asphalt Pavements

80% 85% 90% 95% 97.5%
RTMs: average reliability of design process 0.81 1.0 1.3 2.0 3.0
85% (Jameson & Moffatt , 2001)
Equivalent RTMs for an average reliability 0.405 0.5 0.65 1.0 1.5
The RTMs reported by Jameson and Moffatt (2001) assume that the average reliability of the design process
is 85% (RTM=1 for a desired project reliability of 85%). As stated above, the 2004 Guide RG considered
the average reliability of the design process was 95% rather than 85%. Consequently, Table 3 shows the
adjusted RTMs for an average reliability of 95% obtained by dividing the RTMs based on an average
reliability of 85% by a factor of 2.
Rather than provide for project reliability by adjusting the design traffic using these RTM values, the 2004
Guide RG simplified the process by providing factors to the allowable loading predicted using the asphalt
fatigue relationship. These reliability factors are the inverse of the RTM values for an average reliability of
95% given in Table 3. Hence the following asphalt fatigue relationship was adopted in the 2004 Guide:
5
⎡ 6918(0.856 VB + 1.08) ⎤
N = RF ⎢ ⎥ (22)
S mix µe
0.36
⎢⎣ ⎥⎦
VB = percentage by volume of bitumen in the asphalt (%);
Smix = asphalt mix stiffness (modulus) (MPa); and
RF = reliability factor for asphalt fatigue (Table 4).
Table 4
2004 Guide Suggested Reliability Factors (RF)
for Asphalt Fatigue

80% 85% 90% 95% 97.5%
2.5 2.0 1.5 1.0 0.67
AUSTROADS 2004
— 3.18 —
Note that these reliability factors are transfer functions that relate a mean laboratory fatigue life, as
determined by the relationship above (Shell 1978), to the in-service fatigue life predicted using this Guide at
a desired project reliability. In effect they comprise two components:
• a shift factor relating mean laboratory fatigue life to a mean in-service fatigue life, taking account of the
differences between the laboratory test conditions and the conditions applying to the in-service pavement;
and
• a reliability factor relating mean in-service fatigue life to the in-service life predicted using this Guide at a
desired project reliability, taking into account those factors (e.g. construction variability, environment,
traffic loading).
AUSTROADS 2004
— 3.19 —
5. DESIGN TRAFFIC
5.1 Units of Loading
The design procedures for flexible pavements are based on assessing the response under a Standard Axle
load, which is defined in the 2004 Guide as a single axle with two sets of dual wheels that carries a load of
80 kN.
The loads on other axle groups that cause the same damage as a Standard Axle are given in Table 5. Potter
(refer Part 1 of this report ) describes the origin of the Table 5 axle loads, except for quad axle loading, the
origins of which are discussed by Vuong et al. (2002).
Table 5
Axle Group Loads Which Cause Equal Damage as a Standard Axle
Axle Configuration Load (kN)

Single Axle Single Wheels 53
Single axle Dual Wheels 80
Tandem Axle Single Wheels 90
Tandem Axle Dual Wheels 135
Triaxle 181
Quad axle 221
If axle group loads other than those in Table 5, then the damage caused is expressed as the number of
Standard Axles which produce the same damage and is calculated as follows:
EXP
⎡ Load on Axle Group ⎤ (23)
No. of Standard axles of damage = ⎢ ⎥
⎣ Appropriat e Loadfrom Table 1 ⎦
where the exponent EXP varies with the distress type. During the AASHO Road Test, it was concluded that
damage to the test pavements was related to the 4th power of the axle load. Hence a value of 4 is commonly
adopted for the exponent, in this case the number of Standard Axles of damage is termed the number of
Equivalent Standard Axles (ESAs).
The empirical thickness design chart (Figure 8.4) for thin bituminous-surfaced unbound granular pavements
presents the allowable loading in terms of equivalent standard axles (ESA). Hence to use this design chart,
the design traffic loading needs to be calculated in terms of ESAs.
In the Austroads mechanistic design procedures, the allowable loading varies with strain level as shown in
Table 6.
In the Austroads Guide, the exponent of strain dependency for each distress mode is adopted as the exponent
of load dependency in eqn (23). This is based on the simplifying assumption that all materials are linear
elastic, even though it is acknowledged that the moduli of granular and subgrade materials are stress-
dependent.
AUSTROADS 2004
— 3.20 —
Table 6
Distress Mode Strain Dependency
Distress Mode Dependence on Strain

Asphalt Fatigue (Section 4.3.3) ⎛ k ⎞
5
N = ⎜⎜ ⎟⎟
⎝ µε ⎠
Rutting of Unbound Granular Materials and ⎛ k ⎞
7
Subgrade (Section 4.3.1) N = ⎜⎜ ⎟⎟

⎝ µε ⎠
Cemented Materials Fatigue (Section 4.3.2) 12
⎛ k ⎞
N = ⎜⎜ ⎟⎟
⎝ µε ⎠
Consequently, for the mechanistic design of flexible pavements, design traffic is expressed in terms of the
following Standard Axles of loading rather than ESA (4th power) of loading:
• Using a load damage exponent of 5, the number of Standard Axles that produce the same cumulative
damage as the design traffic in terms of asphalt fatigue (NSA) is calculated.
damage as the design traffic in terms of rutting of granular materials and subgrade is calculated (NSS).
damage as the design traffic in terms of cemented material fatigue (NSC) is calculated.
In the 1987 Guide, the most common practice of calculating these Standard Axles (see Appendix E of the
1987 Guide) was as follows:
• First, the cumulative number of heavy vehicles over the design period was calculated.
• Based on Table E5, or other relevant data, the number ESA of loading per heavy vehicle and hence the
cumulative ESA of loading (NE) were calculated.
• The Standard Axles of loading per ESA for each distress type were then calculated based on the following
presumptive values or other relevant information:
NSA = 1.1 NE
NSS = 1.1 NE
NSC = 20 NE
• Using these factors, the number of Standard Axles of loading for each distress type were calculated.
For the design of rigid pavements, the cumulative number of heavy vehicles per axle groups (HVAGs) was
calculated. For the 2004 Guide, the process was simplified, with the cumulative number of HVAGs for all
pavements calculated.
To calculate Standard Axles of loading from the cumulative HVAGs of loading, the factors are calculated for
the average number of Equivalent Standard Axles (ESA) of damage per HVAG and the Standard Axle
Repetitions (SARs) per ESA for each distress type based on the traffic load distribution.
In the absence of traffic load distribution data, the presumptive values of ESA/HVAGs and SAR/ESA given
in Table 7 are given based on the presumptive traffic load distributions discussed in Section 5.2.
AUSTROADS 2004
— 3.21 —
Table 7
Characteristics of Presumptive Traffic Load Distributions (TLDs)
for Urban and Rural Roads
Pavement Type Damage Type Damage Index Presumptive Presumptive

Rural TLD Urban TLD
Granular Pavements with Overall Damage ESA/HVAG 0.9 0.7
Thin Bituminous 2.5 1.8
Surfacings, designed using ESA/HV
Figure 8.4
Pavement Containing One Fatigue of Asphalt SARa/ESA 1.1 1.1
or More Bound Layers, Rutting and Shape Loss SARs/ESA 1.6 1.6
Mechanistically Designed
Fatigue of Cemented Materials SARc/ESA 12 12
5.2 Traffic Load Distributions

In the early to mid-1980s when Chapter 7 of the 1987 Guide was written there was limited axle load data.
Given the increased amount of weigh-in-motion (WIM) data now available, and the increase in axle loads
over the last 15-20 years, it was considered essential that this element of the design procedure be revised for
the 2004 Guide.
Koniditsiotis (1996, 1997 and 1998) collated WIM data collected by State Road Authorities at over 100 sites
and measured WIM data on 12 lightly-trafficked roads in the Melbourne metropolitan area.
After validation and quality assurance of the traffic load data, Koniditsiotis (1998) used clustering analysis to
group similar traffic load distributions (TLDs). Based on this analysis, 25 presumptive TLDs were
developed, comprising:
• nine TLDs for urban roads, with presumptive TLDs varying with type of heavy vehicle use (e.g. General
Freight, Industrial, etc.) and Austroads Road Functional Class 6, 7 and 8; and
• 16 TLDs for rural roads, with presumptive TLD varying with type of heavy vehicle use (eg General
Freight, Industrial, etc) and Austroads Road Functional Class 1, 2 and 3.
A Working Group lead by David Dash (RTA, NSW) reviewed these 25 TLDs and considered the data could
be further reduced to 12 presumptive TLDs. Based on this analysis, the text of Chapter 7 was prepared by
the Working Group and discussed at the June 2000 meeting of the 2001 Guide RG.
In this draft of the Guide, Table 7.5 provided presumptive traffic load distribution (TLDs) for urban roads
and Table 7.6 provided presumptive traffic load distribution (TLDs) for rural roads. Again, the appropriate
TLD depended on the type of heavy traffic use and the Austroads Road Functional Class.
The draft 2001 Guide RG requested that a review be undertaken of the data, as some of the values in the
Tables appeared to be counter-intuitive.
Following the June 2000 RG meeting, the SRAs were asked to confirm the Road Functional classes assumed
in the analysis undertaken to derive Tables 7.5 and 7.6. The SRAs were forwarded the June 2000 draft
Appendix 7.2, which summarised the data for 119 weigh-in-motion sites through Australia. This review
resulted in the following changes to Appendix 7.2:
• based on SRA advice, five of the South Australian sites, two of the Victorian sites and one of the Western
Australian sites were deleted due to concerns about the validity of the data; and
• the Road Functional Class of 32 of the remaining 111 sites was changed.
AUSTROADS 2004
— 3.22 —
Due to these significant changes to road functional class, it was apparent that new presumptive TLDs would
need to be developed from the WIM data reported by Koniditsiotis (1996). Given the limited resources
available for this additional development work, the data was analysed to only assess whether presumptive
TLDs for each road functional class were appropriate, rather than repeat the more detailed clustering analysis
previously undertaken by Koniditsiotis (1998).
Based on this analysis (see Appendix J), the 2004 Guide RG adopted the following presumptive TLDs:
“urban” and “rural”.
In addition to these presumptive TLDs, Appendix 7.2 of the Guide lists the TLDs measured at over 100 sites
throughout Australia. Designers may select a TLD from this database.
AUSTROADS 2004
— 3.23 —
6. PROJECT RELIABILITY
6.1 Definition
An integral part of the pavement design process is an assessment by the designer of how well the outcome of
the design – the constructed pavement – will perform. No pavement design process can ever guarantee with
certainty that a subsequently constructed pavement will perform to design expectations. The reasons for this
are as follows:
• No design process perfectly models how a specific pavement will perform in a controlled environment
with a specified traffic loading, let alone in its allotted environment when subjected to its actual traffic.
• The design values chosen for material properties are, at best, gross simplifications of the complex and
variable properties of pavement and subgrade materials.
• No construction process can produce a pavement in complete conformance with a design configuration,
both in terms of layer thicknesses and (simplistic) material properties.
Because of this lack of certainty, an appropriate measure of the anticipated performance of the (yet-to-be-
constructed) pavement is its Project Reliability which is defined as follows:
The Project Reliability is the probability that the pavement when constructed to the chosen design will
outlast its Design Traffic before major rehabilitation is required. In regard to these reliability procedures,
a project is defined as a portion from a uniformly designed and (nominally) uniformly constructed road
pavement which is subsequently rehabilitated as an entity.
A more detailed definition and description of project reliability is contained in Potter et al. (1996).
6.2 Procedures in 1992 Guide

Section 7.9 of the 1987 Guide presented a method for modifying the design traffic to improve the reliability
of the design. For flexible pavements, the number of load repetitions was increased above the number
actually anticipated over the design period.
During the development of the revised design procedures for rigid pavements in the early 1990s concern was
expressed regarding the different methods used to allow for project reliability in rigid pavement and flexible
pavement design.
The existing flexible pavement procedures contained in the 1987 Guide allow for variability by increasing
the number of load repetitions by a factor of up to 4. The existing rigid pavement design procedures, on the
other hand, allow for variability by using ‘load safety factors’ (LSF) (e.g. 1.0, 1.1 and 1.2), to adjust the axle
loads used in the design. Neither of these procedures enables pavements to be designed to a selected level of
reliability.
6.3 Development of Procedures in Draft 2001 Guide

In 1996 Austroads commissioned ARRB TR to develop reliability guidelines that were consistent across all
new pavement types. An Austroads Working Group was established to guide and assist in this task.
Reports by Potter et al. (1996), Moffatt et al. (1998) and Jameson and Moffatt (2001) detail the development
of these reliability guidelines and the effect of the adoption on these guidelines on generic pavement designs.
The method used to derive these procedures requires knowledge of: (i) the variability between road projects
of input parameters to the design processes, and (ii) the average reliability of projects designed using the
current design processes and constructed and maintained to current SRA standards and specifications.
AUSTROADS 2004
— 3.24 —
As discussed by Jameson and Moffatt (2001), given the paucity of pavement performance data, there was
considerable difficulty in reaching agreement on these average reliability levels. Table 8 lists the average
reliabilities that were used in the development of the draft 2001 Guide reliability procedures.
Table 8
Average Project Reliabilities of Outlasting
Design Traffic for Flexible Pavements (draft 2001 Guide)
Pavement Type Reliability (%) of the

Project Withstanding its
Design Traffic
Thin bituminous seal on granular material 90
Thin bituminous seal on granular material on 90
cemented material
Thin bituminous seal on cemented material 80
Full depth asphalt 85
Asphalt on granular material 85
Asphalt on granular material on cemented 85
material
Asphalt (< 150 mm) on cemented material 85
Asphalt (≥ 150 mm) on cemented material 85
Asphalt (≥ 150 mm) on modified material 1
85
1
Material modified with a quantity and type of binder that ensures the material
will perform as a granular material.
For a given project the desired project reliability may differ from the average values shown in Table 8. To
design a project to a reliability level other than the values given in Table 8 the distribution of allowable
loadings of projects identically designed and (nominally) identically constructed was required. These
distributions were generated using Monte Carlo simulations and estimations of the variability between road
projects of input parameters to the design processes. For each pavement type, these distributions were used
to estimate how to adjust the design traffic using a Traffic Multiplier to design to a project reliability other
than the values given in Table 8 (Jameson and Moffatt 2001).
In the draft 2001 Guide procedures for the design of new flexible pavements, project reliability was allowed
for by altering the number of Standard Axle loads by so-called Reliability Traffic Multipliers. Reliability
Traffic Multipliers (RTM) were estimated for a range of flexible pavement types to enable pavements to be
designed to project reliabilities of 75%, 80%, 85%, 90%, 95% and 97.5%. RTMs were developed both for
the reliability of outlasting the design traffic expected during the design period and of outlasting the design
period. Table 9 shows the RTMs for outlasting the design traffic, while Table 10 shows the RTMs for
outlasting the design period. Note that the RTMs are greater for outlasting the Design Period than the
Design Traffic as the former includes the additional uncertainty associated with estimating the design traffic
during the design period. As the values in Tables 9 and 10 were similar, the draft 2001 Guide RG decided to
simply adopt the Design Traffic RTM (Table 9) in the 2001 draft Guide.
AUSTROADS 2004
— 3.25 —
Table 9
Traffic Multipliers to Outlast the Design Traffic
Pavement Type Desired Project Reliability

75% 80% 85% 90% 95% 97.5%
Thin bituminous seal on less than 300 mm granular 0.17 0.27 0.49 1.0 2.9 7.4
material
Thin bituminous seal on 300 mm or more granular 0.27 0.38 0.58 1.0 2.3 4.4
material
Thin bituminous seal on granular material on 0.33 0.44 0.62 1.0 2.1 3.7
cemented material
Thin bituminous seal on cemented material 0.62 1.0 1.8 3.9 14 42
Full depth asphalt 0.68 0.81 1.0 1.3 2.0 3.0
Asphalt on granular material 0.70 0.82 1.0 1.3 1.9 2.7
Asphalt on granular material on cemented material 0.68 0.81 1.0 1.3 2.0 3.1
Asphalt (<150 mm) on cemented material 0.52 0.70 1.0 1.6 3.3 6.5
Asphalt (>=150 mm) on cemented material 0.65 0.79 1.0 1.4 2.3 4.3
Asphalt (>=150 mm) on modified1 material 0.71 0.83 1.0 1.3 1.9 2.6
1. Material modified with a quantity and type of binder that ensures the layer will not fail by fatigue cracking.
Table 10
Traffic Multipliers to Outlast the Design Period
Pavement Type Desired Project Reliability

75% 80% 85% 90% 95% 97.5%
Thin bituminous seal on less than 300 mm granular 0.18 0.28 0.50 1.0 3.1 8.2
material
Thin bituminous seal on 300 mm or more granular 0.28 0.40 0.62 1.1 2.5 4.8
material
Thin bituminous seal on granular material on 0.34 0.47 0.67 1.1 2.2 4.1
cemented material
Thin bituminous seal on cemented material 0.64 1.03 1.9 4.1 15 44
Full depth asphalt 0.74 0.89 1.1 1.5 2.4 3.6
Asphalt on granular material 0.78 0.92 1.1 1.5 2.3 3.3
Asphalt on granular material on cemented material 0.74 0.90 1.1 1.5 2.4 3.7
Asphalt (<150 mm) on cemented material 0.54 0.74 1.1 1.7 3.6 7.3
Asphalt (>=150 mm) on cemented material 0.69 0.86 1.1 1.5 2.6 4.3
Asphalt (>=150 mm) on modified1 material 0.78 0.92 1.1 1.5 2.2 3.3
1. Material modified with a quantity and type of binder that ensures the layer will not fail by fatigue cracking.
6.4 Development of Procedures in 2004 Guide

The 2004 Guide RG viewed comments on the draft 2001 Guide reliability procedures. The following issues
were raised:
• Given the lack of data about the average reliabilities of the design procedures, it was considered that the
provision of RTM for nine separate pavements was unwarranted. In addition, for some pavement types
there was a wide scatter of RTMs.
• There was concern that the average reliabilities may be misinterpreted as reflecting the intrinsic reliability
of pavement types themselves rather the reliabilities of the design processes for pavement types.
AUSTROADS 2004
— 3.26 —
• There was concern that RTM values provided for unbound granular pavements with thin bituminous
surfacings would result in changes in the thickness of these pavements. These RTMs were based on the
assumption that the average reliability of pavements designed using the empirical thickness design charts
(Figure 8.4) was 90%. As the performance of pavements designed with the empirical chart has been
adequate across a wide range of traffic loadings it was concluded that Figure 8.4 already included
appropriate levels of reliability across a range of traffic loadings and that this was more appropriate than
adopting an average reliability of 90%.
The 2004 Guide RG decided that the process to enable design to a desired project reliability should be
simplified. The approach adopted was to adjust the predicted allowable loading for each distress mode rather
than the applying RTM to change the design traffic loadings.
Sections 4.3.2 and 4.3.3 of this report explain how Reliability Factors were developed to enable design to a
desired project reliability in terms of cemented materials fatigue and asphalt fatigue respectively. As
discussed in Section 4.3.1, the subgrade strain relationship was derived by back-analysis of empirical chart
for unbound granular pavements (Figure 8.4). The RG considered this chart to implicitly include appropriate
levels of reliability across a range of traffic loadings. As such, there was no need to provide Reliability
Factors for use with the subgrade strain criterion.
Note that the Reliability Factors provided in the 2004 Guide are applicable to the design processes in the
Guide and to current SRA construction and maintenance specifications. In the event of changes to these
procedures these factors should be reviewed.
As the design of rigid pavements is more sensitive to load magnitude rather than to axle load, project
reliability is allowed for by altering the project axle load distribution using Load Safety Factors (LSFs).
LSFs were estimated for reliability of outlasting both the design traffic and the design period.
AUSTROADS 2004
— 3.27 —
7. SUMMARY
This report has recorded the work undertaken to revise the guidelines for the design of flexible pavements in
Pavement Design – A Guide to the Structural Design of Road Pavements for publication in 2004. The Guide
was initially published by the National Association of Australian State Road Authorities (NAASRA) in 1987
and subsequently revised and re-issued by Austroads in 1992.
The work undertaken to develop the previous editions of the Guide plus its predecessor, the Interim Guide to
Pavement Thickness Design (NAASRA 1979), has been well-documented by Potter (refer to Part 1 of this
report ).
The major changes to the Austroads guidelines for the design of flexible pavements are:
• changes to the pavement response model, including the use of full Standard Axle loading rather than a
half Standard Axle loading to calculate critical strains;
• improved procedures for estimating the design moduli of selected subgrade materials and unbound
granular materials;
• improved methods of estimating asphalt design moduli, including the derivation of design moduli from
measured moduli obtained using the indirect tensile test;
• revision of the subgrade strain criterion for application with the full Standard Axle load;
• procedures to enable design be to be conducted to a desired reliability of the constructed pavement
outlasting the design traffic;
• modification to the cemented materials fatigue and asphalt fatigue relationships to include variable factors
to enable design to a desired project reliability; and
• significant revision to the procedures used to calculate the design traffic, including a database of traffic
load distributions obtained at over 100 WIM sites throughout Australia.
AUSTROADS 2004
— 3.28 —
REFERENCES
Alderson, A.J. (1999). Summary of VicRoads Research into Cement-Treated Materials. ARRB Transport
Research Ltd Contract Report RC 90216.
Angell, D. (1988). Technical Basis for the Pavement Design Guide. Report RP 1265. Pavements Branch.
Department of Main Roads, Queensland.
Sydney.
Austroads (1997). Revision Note for Pavement Design – A Guide to the Structural Design of Road
Pavements. APRG, November.
Austroads (2004). Pavement Design – A Guide to the Structural Design of Road Pavements. Pub. No.
AP-G17/04. Austroads, Sydney.
Brown, S.F. (1973). Determination of Young’s Modulus for Bituminous Materials in Pavement Design.
Transp. Res. Rec., pp. 38-49.
Chowdhury, F. and Rallings, R.A. (1994). A Survey of Truck Tyre Pressures in Tasmania. Road and
Transport Research 3(3), September.
Jameson, G.W. (1995). Response of Cementitious Pavement Materials to Repeated Loadings. ARRB
Transport Research, Contract Report RI 949, March.
Jameson, G.W. (1996). Origins of AUSTROADS Design Procedures for Granular Pavements. ARRB
Transport Research, Research Report ARR No. 292.
Jameson, G.W. (1999). An Assessment of the Need to Incorporate Shift Factors for Predicting the Fatigue
Life of Asphalt. APRG Document 99/43, December.
Jameson, G.W. and Hopman, P.C. (2000). Austroads Pavement Design Guide Chapter 6: Development of
Relationships Between Laboratory Loading Rate and Traffic Speed. APRG Document 00/16, June.
Jameson, G.W. and Moffatt, M.A. (2001). Development of Austroads Pavement Design Reliability
Guidelines. APRG Document 00/17, June.
Jameson, G.W., Sharp, K.G. and Yeo, R. (1992). Cement-Treated Crushed Rock Pavement Fatigue Under
Accelerated Loading: the Mulgrave (Victoria) ALF trial, 1989/1991. Australian Road Research Board.
Research Report ARR No. 229.
Jameson, G.W., Dash, D.M., Tharan, Y. and Vertessy, N.J. (1995). The Performance of Deep-Lift In-Situ
Pavement Recycling Under Accelerated Loading: the Cooma ALF Trial 1994. APRG Report No. 11.
Kenis, W.J., Sherwood, J.A., and McMahon, T.F. (1981). Verification and Application of the VESYS
Structural Subsystem. Proc. 5th Int. Conf. on Structural Design of Asphalt Pavements, Ann Arbor, Michigan.
Koniditsiotis, C. (1996). Update of Traffic Design Chapter in the Austroads Pavement Design Guide – Status
Report. ARRB TR Working Document WD TI96/024.
Koniditsiotis, C. (1997). Update of Austroads Pavement Design Guide – Traffic Design Chapter. Report on
Consultation with Stakeholders and Outline (Draft) of Proposed Traffic Design Chapter. ARRB TR
Working Document WD R97/019.
Koniditsiotis, C. (1998). Update of the Austroads Pavement Design Guide – Traffic Design Chapter. Final
Draft of New Traffic Design Chapter. ARRB TR Working Document WD R98/030.
AUSTROADS 2004
— 3.29 —
Moffatt, M.A. and Jameson, G.W. (1997). Development of a Subgrade Strain Criterion for Full Width
Standard Axle Loading. ARRB TR Working Document WD-R97/068, October.
Moffatt, M.A. and Jameson, G.W. (1998). Granular Materials Moduli Under Asphalt and Cemented
Material. ARRB TR Working Document WD-R98/008, April.
Moffatt, M.A. and Jameson, G.W. (1998). Characterisation of Granular Material and Development of a
Subgrade Strain Criterion. ARRB TR Working Document WD-R98/005, March.
Moffatt, M.A. and Yeo, R. (1998). Relationships Between Unconfined Compressive Strength and Flexural
Modulus for Cemented Materials. ARRB TR Working Document WD-R98/024, June.
Moffatt, M.A, Jameson, G.W, Cropley, S. and Ramsey, E. (1998). Traffic Multipliers for Incorporating
Design Reliability into the Austroads Pavement Design Guide. ARRB TR Working Document WD-R98/026,
November.
NAASRA (1979). Interim Guide to Pavement Thickness Design. National Association of Australian State
Road Authorities (NAASRA), Sydney.
NAASRA (1987). Pavement Design – A Guide to the Structural Design of Road Pavements. National
Association of Australian State Road Authorities, Sydney.
Potter, D.W. (1997). Appropriate Laboratory Fatigue Testing of Asphalt for Australia and its Role in
Australian Pavement Design. ARRB TR Working Document R97/021, April.
Potter, D.W., Jameson, G.W., Makarov, A., Moffatt, M.A. and Cropley, S.M. (1996). A Basis for
Incorporating Reliability in the Austroads Pavement Design Procedures. ARRB TR Working Document
TI96/014, June.
Rallings, R.A. (1997). APRG Workshop on Structural Behaviour of Unbound Granular Pavements. Report
APRG 97/03(DA).
Shell (1978). Pavement Design Manual.
Vuong, B.T. (1994). Prediction Versus Performance of a Granular Pavement Tested with the Accelerated
Loading Facility (ALF). Proc. Symp. on Prediction Versus Performance in Geotechnical Engineering,
Bangkok. A.A. Balkema, Rotterdam.
Wardle, L.J. (1977). Program CIRCLY User’s Manual. CSIRO Division of Applied Geomechanics.
Yeo, R.E.Y. (1997). Basis for Revision of Modulus Correlations for Cemented Materials. ARRB TR
Working Document WD-R97/072, December.
Yandell, W.O. (1981). Applications of the Mechano-Lattice Analysis in Materials Engineering. Proc. 2nd
Aust. Conf. on Engineering Materials, Sydney, pp. 401-19. University of New South Wales.
Youdale, G.P. (1978). Repeated Load Triaxial Tests on Granular Pavement Materials. Materials Research
Laboratory Test Report No. RS21 PTII. DMR, July 26.
Youdale, G.P. (1983). Investigation of the Variation of Stiffness With Depth of a Granular Layer Under
Variable Thickness of Asphaltic Concrete. Report to WG, July 8, 8 pp.; October 25, 5 pp.
Youdale, G.P. (1984a). Investigation of the Effects of and the Interaction Between the Stress Dependency of
Moduli and the Anisotropy of Granular Pavement Materials on the Results of Pavement Analysis Using
CIRCLY. Report to WG, April 13, 7 pp.
Youdale, G.P. (1984b). Investigation of the Variation of Stiffness With Depth of a Granular Layer Over a
Bound Subbase Layer. Reports to WG, March 9, 9 pp.
AUSTROADS 2004
— 3.30 —
Youdale, G.P. (1984c). Review of Limiting Subgrade Strain Criteria. Submission to NAASRA Working
Group on the Revision of NAASRA Interim Guide to Pavement Thickness Design (IGPTD), April 13, 7 pp.
Youdale, G.P. (1984d). The Design of Asphalt Pavements for Particular Temperature Environments. Proc.
12th ARRB Conf. 12(3), pp. 78-88.
Road Research Laboratory (1952). Soil Mechanics for Road Engineers. (HMSO: London.)
Road Research Laboratory (1955). Construction of Housing-Estate Roads Using Granular Base and
Subbase Materials. Road Note 20. (HMSO: London.)
AUSTROADS 2004
— 3.31 —
APPENDIX A
ORIGINS OF UNBOUND GRANULAR THICKNESS CHART

This excerpt is presented in Appendix A of Part 1 of this report and is not repeated here.
AUSTROADS 2004
— 3.32 —
APPENDIX B
CHARACTERISATION OF GRANULAR MATERIALS AND

DEVELOPMENT OF A SUBGRADE STRAIN CRITERION
The following text was first published as an internal ARRB Transport Research Working Document WD-
R98/005 by M. Moffatt and G. Jameson, March 1998.
B.1 INTRODUCTION
Moffatt and Jameson (1997) recommend the adoption of full width Standard Axle modelling in CIRCLY for
the design of new pavements in the next edition of the Austroads Pavement Design Guide (APDG)
(Austroads 1992). At the inaugural meeting of the Austroads Pavement Design Guide Reference Group
(APDGRG), held in January 1998, it was decided to adopt this recommendation, and use the geometry
shown here in Figure B.1. Moffatt and Jameson describe the process whereby a new relationship between
allowable Standard Axle repetitions and maximum compressive strain at the top of the subgrade was
developed, and also recommends a new relationship developed by such a process using full width Standard
Axle modelling.
330 mm 330 mm
1800 mm
Figure B.1: Geometry of a full Standard Axle
It was also decided at the APDGRG meeting that the current “rules” in the Design Guide (Section 8.2.2) for
subdividing granular materials could be inappropriately used by less experienced designers. The current
“rules” allow the maximum modulus of a granular material to be developed regardless of the thickness of the
granular layer or the strength of the subgrade. The meeting decided that the subdividing process be altered to
prevent such practices. As a result of adopting a new procedure for sub-layering granular materials the
process used to develop a revised subgrade strain performance relationship needed to be repeated.
This report recommends a revised procedure for the sub-layering of granular materials and also proposes a
revised subgrade strain performance relationship.
AUSTROADS 2004
— 3.33 —
B.2 UNBOUND GRANULAR MATERIALS

CHARACTERISATION
B.2.1 Background
Jameson and Moffatt (1998) conduced a review of the current “rules” for sub-layering granular materials,
and also investigated at procedures used by Main Roads Queensland (Angell 1988), the United States Army
Corps of Engineers (USACE 1989) and the Shell Pavement Design Manual (1978). Subsequent to this
report an alternative procedure was suggested (in two slightly different forms) by David Potter (Potter 1998).
This procedure is reproduced here in Attachment B1, as it is not published elsewhere.
Members of the APDGRG were invited to review the alternative procedures and reach a decision as to which
procedure should be adopted for use in the APDG. The general consensus was to adopt Potter’s modified
approach (whereby granular material is always divided into 5 equi-think layers). It was also thought that the
use of a value of 125 in Step 3 of this procedure was appropriate. Some concerns were raised as to whether
numerical instability in CIRCLY may result from the small sublayer thicknesses used for thin granular
layers. A quick exercise, discussed below, confirms that excessive numerical instability does not occur.
The following section contains the recommended procedure for sub-layering granular material, and written
so as to form a direct replacement for the current Section 8.2.2 of the APDG.
B.2.2 Proposed Procedure of Elastic Characterisation of Granular Materials

The moduli of granular materials are dependent not only on the intrinsic characteristics of these materials,
but also on the stress level at which they operate and the stiffness of underlying layers. As a result, the
moduli of pavement materials subjected to vertical loading will decrease with depth to an extent influenced
by the stiffness of the subgrade.
Iterative analyses with a finite element model would permit allowance to be taken of the stress dependant
nature of the modulus of granular material, but would not make allowance for the degree of support provided
by underlying layers. In addition, as such models are not readily available to pavement designers, the
procedure in this Guide utilises the elastic model CIRCLY with the granular layers partitioned into several
sub-layers, with each assigned a modulus value according to the following:
(1) For granular materials placed directly on a stiff cemented sub-base, no sub-layering is required. The
modulus is determined using the procedures which are discussed in Chapter 6 of the 1992 Guide.
(2) For granular materials placed directly on the subgrade, sub-layering is required and should be conducted
as follows:
(a) Divide the granular materials into 5 equi-thick sub-layers.
(b) The vertical modulus of the top sublayer is the minimum of the value indicated in Table 6.6 and
that determined using:
E V top of base = E V subgrade × 2 ( total granular material thickness 125 )
(c) The ratio of moduli of adjacent sub-layers is given by:

1
⎡ EV ⎤ 5
R = ⎢ top of base ⎥
⎣⎢ E V subgrade ⎦⎥
(d) The modulus of each layer may then be calculated from the modulus of the adjacent underlying
layer, beginning with the subgrade whose modulus is known.
AUSTROADS 2004
— 3.34 —
(3) For all granular materials, the other stiffness parameters required for each sublayer may be calculated
from the following relationships:
Eh = 0.5 Ev – refer to Section 6.1
Fv = Ev / (1 + νv) – refer to CIRCLY Manual
B.2.3 Check for Numerical Instability in CIRCLY for Low Sub-Layer Thicknesses
As noted above, some concerns were raised when deciding upon the most appropriate sub-layering procedure
as to whether use of very thin (around 20 mm) sub-layers would lead to numerical instability in CIRCLY.
The recommended procedure described above can produce thin sub-layers for low total material thicknesses.
After discussions with Leigh Wardle of Mincad Systems it was felt that a suitable means of determining if
any numerical instability had occurred for a specific analysis was to examine the CIRCLY output file
(*.CLO) and check for an asterisk (*) in the first column for each coordinate analysed. The asterisk indicates
CIRCLY has failed to converge on a solution.
A series of granular pavements with compositions shown in Table B.1 were analysed (using CIRCLY 3.0,
update 56) using both Potter’s original suggested procedure and his modified procedure as shown in Section
B.2.2.
Table B.1
Granular Material Thicknesses and Subgrade (Vertical) Moduli used in CIRCLY Analysis
Granular Thicknesses (mm) 100 to 1000 in steps of 25

Subgrade CBR (%) 2, 3, 4, 5, 7, 10, 15, 20, 30
This comparison did not show an significant difference in the number times CIRCLY reported, via the
asterisk, that instability had occurred in determining a solution.
The results of these analyses also shows that the maximum compressive strain at the top of the subgrade
(which could occur anywhere along the full axle) did not significantly vary between each of the Potter
approaches. For pavement thicknesses greater than 225 mm there was no difference, as both procedures use
five sub-layers for granular layers greater than or equal to 250 mm. For thicknesses of 225 mm or below, the
differences between strains determined using both Potter granular characterisations can be seen in Figure
B.2.
A second series of CIRCLY analyses, this time of pavement with a thin asphalt surfacing, was conducted
using both Potter’s original suggested procedure and his modified procedure. Table B.2 shows the pavement
compositions used for the analysis. Figure B.3 shows that there is no significant difference between the
asphalt or subgrade critical strains determined using the two Potter approaches.
Table B.2
Granular Material Thicknesses and Subgrade (Vertical) Moduli used in CIRCLY Analysis
Asphalt Modulus (MPa) 3,000

Asphalt Thicknesses (mm) 25, 50, 75, 100
Granular Thicknesses (mm) 100, 150, 200
Subgrade CBR (%) 2, 3, 4, 5, 7, 10, 15, 20, 30
Thickness of overlying asphalt Maximum possible modulus of top
(mm) layer of granular material
25 350
50 320
75 290
100 250
AUSTROADS 2004
— 3.35 —
A comparison of the critical asphalt strains determined from the modified Potter approach and the current
(Austroads 1992) subdivision approach, shown in Figure B.4, indicates that there is a substantial difference
in strain levels for thin asphalt layers. The current Austroads approach yields considerably lower strain
levels than the proposed modified Potter method. Figure B.4 shows that as pavement thicknesses increase
the agreement in strain levels between the two methods improves. However, this difference in strain is most
important when pavement life is dominated by asphalt fatigue rather than subgrade (and pavement) rutting.
Figure B.4 indicates that the difference in strains are considerably less in these cases.
In all of the exercise relating to asphalt-surfaced pavements, it was found that CIRCLY did not encounter
any more numerical instability using the modified Potter approach than with the original method proposed
by Potter.
14000 14000
Subgrade CBR
13000 13000
Critical microstrain (using modified Potter subdivision approach)

12000 3 12000
Subgrade CBR
11000 11000
10000 10000 3
9000 4 9000
8000 8000
4
5
7000 7000
5
6000 6000
7
5000 5000
7
4000 4000
10
10
3000 3000
15
15
2000 20 2000 20
30 30
1000 1000
0 0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000
Critical microstrain (using original Potter subdivision approach) Critical microstrain (using original Potter subdivision approach)
Total granular thickness = 100 mm Total granular thickness = 125 mm

14000 14000
13000 13000
2
12000 12000
11000 11000
2
10000 10000
Subgrade CBR
9000 9000
3
8000 8000 Subgrade CBR
7000 7000 3
4
6000 6000
5 4
5000 5000
5
4000 7 4000
7
3000 3000
10
10
2000 15 2000
20 15
20
30 30
1000 1000
0 0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000

14000 14000
13000 13000
12000 12000
11000 11000
10000 10000
9000 9000
2
8000 8000
2
7000 Subgrade CBR 7000
Subgrade CBR
6000 3 6000
5000 5000 3
4
4000 4000 4
5
5
3000 3000
7
7
2000 10 2000
10
15 15
20 20
1000 30 1000 30
0 0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000
Figure B.2: Comparison between alternative Potter granular material characterisations
AUSTROADS 2004
— 3.36 —
1800 9000
Critical subgrade microstrain (using modified Potter subdivision
Critical subgrade microstrain (using modified Potter subdivision

1600 8000
1400 7000
1200 6000
approach)
approach)
1000 5000
800 4000
600 3000
400 2000
200 1000
0 0
0 200 400 600 800 1000 1200 1400 1600 1800 0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Critical subgrade microstrain (using original Potter subdivision approach) Critical subgrade microstrain (using original Potter subdivision approach)
Figure B.3: Comparison between alternative Potter granular material characterisations

(asphalt surfaced pavements)
1800
25 100 02
Asphalt strain calculated using proposed granular subdivision method
asphalt thickness (mm)
1600 granular thickness (mm)

25 150 02
subgrade CBR (%)
XX YYY ZZ
1400
25 200 02 25 100 03
50 100 02
1200 50 150 02
25 150 03
25 100 04 50 200 02
50 100 03
(microstrain)
1000 50 150 03
25 200 03
25 150 04 25 100 05 50 100 04
50 200 03 75 100 02
50 150 04 75 150 02
50 02
75 200 100 05
800 25 150 05 75 100 03
25 200 04 5050
2001500405
75 150 03
25 100 07 75 50
200 75 100 04
03 07
100
50 200 05 75 150 04 100 100 02
50 75100
100150 02
02 05
600
25 200 05 75150 07
100
200150
75 04 05200
100 100 03
25 150 07 50 200 50
75 07 100
20075 100
0510
100 150 03
0704
100 200
100 03
100
25 100 10 75100
150150
07 04
50 100
150 200
1010004100 05
75 100
200
75 150
07 10
100 05
25 200 07 100 200 05
5010075100
50
200 10 100
100
150
150 10 07
15
07
400 25 150 10 100
75 75
100
100
50100 200
200
150 100
1510
07 10
75
5050
100
150
200 150
152015
15 10
25 100 15 10075100
75
50
100
75
50 100
200
150
150
200200
20
1520
15
20
2010
25 200 10 100 10075
100
100
200
150
200
100
150
200 1520
15
20
20
25
25150
10015
20
25
25 150
25200 20
2001520
200
0
0 100 200 300 400 500 600 700 800 900 1000
Asphalt strain calculated using current (Austroads, 1992) granular subdivision method (microstrain)
1800
Asphalt strain calculated using proposed granular subdivision method
1600
1400
1200
(microstrain)
1000 SG AS
SG AS
SG AS SG AS
SG AS SG AS
SG AS
SG AS
800
SGSG
ASAS SG AS
SG AS
ASAS AS
AS AS AS AS
SG AS AS
600 ASAS AS
AS
AS AS
AS ASAS AS Indicates that subgrade
AS AS
ASAS
AS AS
AS AS
ASAS
strain criterion dominates when using
SG AS AS
AS
AS
ASAS AS current subdivion method, and that
400 AS
AS
ASAS asphalt strain dominates when using
ASAS
AS
AS Indicates that asphalt
ASAS
AS
AS
AS proposed subdivision method
SG AS ASAS
AS
AS
AS strain criterion
SG AS dominates
AS
AS
200
0
0 100 200 300 400 500 600 700 800 900 1000
Asphalt strain calculated using current (Austroads, 1992) granular subdivision method (microstrain)
Figure B.4: Comparison between Critical Asphalt Strains produced using proposed Granular Material
Characterisation with those produced using the current Austroads Characterisation
AUSTROADS 2004
— 3.37 —
B.3 DEVELOPMENT OF A REVISED SUBGRADE STRAIN

PERFORMANCE RELATIONSHIP
The current equation, eqn (B.1), relating the number of allowable strain repetitions to the subgrade strain
level, was developed using a half Standard Axle model. Jameson (1996) describes how Youdale (1984)
derived the relationship from back-analyses of 25 pavements selected from the NAASRA granular thickness
chart (now Figure 8.4 of Austroads (1992)). For each pavement CIRCLY was used to calculate the
maximum compressive vertical strain at the top of the subgrade between the dual wheels of a Standard Axle
(550 kPa tyre pressure was assumed). A regression of these strains versus the allowable traffic loading
indicated from the design chart yielded eqn (B.1).
7.14
⎡ 8511⎤
N=⎢ ⎥ (B.1)
⎣ µε ⎦
where N = allowable repetitions of a half Standard Axle, and
µε = maximum compressive strain at the top of the subgrade (microstrain).
With the adoption of full width modelling of the Standard Axle, and an increased tyre pressure of 750 kPa
(Austroads 1997), a new relationship needed to be determined. This was done and is reported in Moffatt and
Jameson (1997). As noted above, the Austroads Pavement Design Guide Reference Group (APDGRG)
decided that a modification of the procedure for sub-layering granular materials was needed. As such a
modification would mean that the analysis described in Moffatt and Jameson would need to be repeated, the
APDGRG decided that a further revision of the subgrade strain performance relationship should be
conducted using the new sub-layering technique.
Using the sub-layering procedure described in Section B.2.2, the following procedure was followed in order
to determine the revised relationship:
1. A series of CIRCLY Version 3 runs was conducted for unbound granular pavements, with a range of
granular thicknesses and subgrade CBRs. These runs were conducted using modelling parameters
considered to be applicable to the conditions relevant to the pavements considered in the
development of Figure 8.4 of Austroads (1992):
• 550 kPa tyre pressure
• Traffic distribution shown in Table 8.3 of Austroads (1992)
• Maximum vertical modulus of top sublayer of granular material, as per Table 6.6 of Austroads
(1992) – 350 MPa (value for standard compactive effort).
• Full width Standard Axle load geometry, as described in Section 1.
The pavement compositions modelled are shown in Table B.3.
Table B.3
Granular Material Thicknesses and Subgrade (Vertical) Moduli Used in CIRCLY Analysis
Granular Thicknesses (mm) 100 to 1,000 in steps of 25

Subgrade CBR (%) 2, 3, 4, 5, 7, 10, 15, 20, 30
2. The maximum compressive vertical strain at the surface of the subgrade of these pavements was
determined. These values are shown in Attachment B2.
AUSTROADS 2004
— 3.38 —
3. The allowable loading (expressed in terms of ESAs) of the pavement configurations was calculated
using Figure 8.4 of Austroads (1992)10. If, for a specific CBR value, the pavement composition was
such that the granular thickness was less than the minimum value shown in Figure 8.4, then the 90%
confidence chart (Figure 13.8.2(B)) in APRG (1998) was used to determine the allowable loading in
ESAs. The allowable loadings in ESAs were converted to allowable repetitions of a Standard Axle
by multiplying by a factor of 1.1 as per Section 7.5 of Austroads (1992) for the traffic distribution
used. These allowable loadings are also shown in Attachment B2.
4. A linear regression was fitted through the log-log data to predict the allowable number of repetitions
of load from the critical strain level. Eqn (B.2) shows the equation determined.
log10(N) = 27.427 – 6.888 log10(µε) (B.2)
where N = allowable repetitions of a Standard Axle,

µε = maximum compressive strain at the top of the subgrade (microstrain), and
r2 for regression = 0.96.
5. A linear regression was fitted through the log-log data to predict the critical strain level from the
number of allowable repetitions of load. Eqn (B.3) shows the equation determined.
log10(µε) = 3.9431 – 0.1392 log10( N ) (B.3)
r2 for regression = 0.96
6. As only one relationship should be used regardless of which parameter is being predicted, the most
appropriate relationship would be the following equation which bisects eqns (B.2) and (B.3).
log10(N) = 27.865 – 7.032 log10( µε ) (B.4)
Rearranging eqn (B.4) yields the more familiar equation form shown in eqn (B.5).
7.03
⎡ 9177 ⎤
N=⎢ ⎥ (B.5)
⎣ µε ⎦
7. In the above calculations the assumption was made that, based on the selected traffic distribution
(Table 8.3 of Austroads (1992)) and eqn (B.1) the factor to convert numbers of ESA repetitions to a
Standard Axles count was 1.1. Using eqn (B.5), which has a strain exponent of 7.03 instead of 7.14
in eqn (B.1), and the procedure detailed in Moffatt (1997), a revised figure of 1.14 was calculated.
Repeating steps 4, 5 and 6, eqns (B.1) and (B.7) were determined. The bisector of this equation is
shown in eqn (B.8).
log10(N) = 27.442 – 6.888 log10( µε ) (B.6)
log10(µε) = 3.945 – 0.1392 log10( N ) (B.7)
log10(N) = 27.881 – 7.032 log10( µε ) (B.8)
Rearranging, and simplifying the denominator, yields eqn (B.9).
7 .03
⎡ 9223 ⎤
N=⎢ ⎥ (B.9)
⎣ µε ⎦
where N = allowable repetitions of a full Standard Axle, and
10
For ease of use the equational form of the Figure was used:
Thickness = [219 – 211(log(CBR)+58(log CBR)2]log (N/120)
AUSTROADS 2004
— 3.39 —
8. As the power of eqn (B.9) is not different to that of eqn (B.5) no subsequent change occurred in the
factor to convert between Standard Axle repetitions and ESA count.
9. Eqn (B.9)is, therefore, appropriate to use to determine the number of allowable repetitions of a
Standard Axle, before deformation failure of the subgrade and overlying material. In order to ensure
that a sense of (unjustified) precision is not conveyed by the relationship, an analysis was undertaken
to determine a relationship containing less significant figures which matches well with that shown in
eqn (B.9). It was found that eqn (B.10) matched very well with eqn (B.9), and so it is proposed that
eqn (B.10) be the form of the subgrade strain relationship adopted in the APDG.
7
⎡ 9300 ⎤
N=⎢ ⎥ (B.10)
⎣ µε ⎦
Figure B.5 compares the allowable repetitions of a Standard Axle determined from subgrade strain using the
current relationship (eqn (B.1)) and that shown in eqn (B.10). It can be seen from this Figure that there is a
difference in the two relationships, particularly for high strains. The higher strains correspond to thin
pavements, where the effects of the changes in granular material subdividing will be apparent. The
difference between the two curves is, therefore, to be expected.
1.0E+09
1.0E+08
Number of allowable repetitions of a Standard Axle
1.0E+07
1.0E+06
1.0E+05
1.0E+04
1.0E+03
0 500 1000 1500 2000 2500 3000 3500 4000
Maximum compressive vertical strain at top of subgrade (microstrain)
N = (8511/micostrain)^7.14 N = (9300/micostrain)^7
Figure B.5: Comparison of Current (eqn B.1) with Proposed (eqn B.10) Relationship
between Subgrade Strain and Allowable Standard Axle Repetitions
AUSTROADS 2004
— 3.40 —
B.4 SUMMARY AND RECOMMENDATIONS

The following recommendations are made in the light of the work discussed above:
• that Austroads replace the existing procedure for sub-layering granular material for CIRCLY modelling
by that shown here in Section B.2.2; and
• that Austroads adopt the following equation for determining the number of allowable repetitions of a
Standard Axle before deformation failure of the subgrade and overlying material:
7
⎡ 9300 ⎤
N=⎢ ⎥
⎣ µε ⎦
AUSTROADS 2004
— 3.41 —
REFERENCES
Angell, D.J. (1988). Technical Basis for the Pavement Design Manual. Main Roads Department,
Queensland. RP-1265.
Austroads (1992). Pavement Design – A Guide to the Structural Design of Road Pavements (Austroads:
Sydney.)
Austroads (1997). October 1997 Updates to: Pavement Design – A Guide to the Structural Design of Road
Pavements. (Austroads: Sydney.)
Austroads Pavement Research Group (1998). A Guide to the Design of New Pavements for Light Traffic – A
Supplement to Austroads Pavement Design. APRG Report No. 21. ARRB Transport Research, Vermont
South.
Jameson, G.W. (1996). Origins of Austroads Design Procedures for Granular Pavements. Research Report
ARR 292. ARRB Transport Research Ltd, Vermont South, Victoria.
Jameson, G.W. and Moffatt, M.A. (1998). Review of Granular Moduli for Use in Pavement Design. ARRB
Transport Research Working Document WD-R98/002, February 1998.
Mincad Systems (1996). CIRCLY – User Manual. MINCAD Systems, Richmond, Victoria.
Moffatt, M.A. (1997). Example of the Effect of Traffic Distribution on Pavement Design Life. ARRB
Transport Research Working Document WD-R97/014, March.
Moffatt, M.A. and Jameson, G.W. (1997). Development of a Subgrade Strain Criterion for Full Width
Standard Axle Modelling. ARRB Transport Research Working Document WD-R97/068, October 1997.
SHELL (1978). Shell Pavement Design Manual – Asphalt Pavements and Overlays For Road Traffic. (Shell
International Petroleum Company Limited: London.)
US Army Corps of Engineers (1989). Flexible Pavement Design for Airfields (Elastic Layered Method).
Departments of the Army and the Airforce Technical Manual 5-825-2-1/AFM 88-6.
AUSTROADS 2004
— 3.42 —
Attachment B1
Potter’s Proposed Procedures for Granular Subdivision And Modulus Assignment
Covering Letter to Geoff Jameson
Geoff,
Attached is my suggestion for a set of design rules for assignment of modulus to granular material placed on
the subgrade. The change from the current rules – although subtle – is significant. Before any set of rules is
adopted, I believe that the effect of the change of rules on levels of critical strains needs to be investigated
and considered.
Subsequent to writing the attachment, I now believe that the process of modulus assignment can be further
improved, and at the same time simplified, by replacing my Rules 1 and 2 by the following:
Amended Design Rules
1. Divide the granular material into 5 equi-thick sub-layers
2. Sublayer thickness t (mm) = T / 5
The effect of these amendments is to remove the discontinuities that occur in the plot of level of critical
strain versus total granular thickness as one changes from n sub-layers to n+1 sub-layers. I have spoken to
Leigh Wardle re minimum layer thicknesses allowable in CIRCLY 3 (or 4). He has advised me not to go
below 20 mm at the surface. This can be reduced somewhat – perhaps to 10 mm – within the pavement.
Based on this advise, the rules as amended would be OK for (total) granular thickness down to 50 mm when
the granular material is well covered, and down to 100 mm when there is minimal cover. These constraints
don’t appear to me to be restrictive.
DAVID POTTER
Consulting Engineer
AUSTROADS 2004
— 3.43 —
David Potter’s Initial Proposed Design Rules For Granular Subdivision / Modulus
Assignment
(see covering note for extra refinement to these rules)
Suggested Design Rules For Assignment Of Modulus To Granular Material Placed On The
Subgrade
Situation
We are placing T mm of granular material on a subgrade with (vertical) modulus EvSG.
Design Rules
1. Divide the granular material into n equi-thick sub-layers, where:
n = minimum(5, integer(T/50))
(This ensures that the minimum sublayer thickness is 50 mm, while maximum number of subgrade is
5)
2. Sublayer thickness t (mm) = T/n
3. (Vertical) modulus of top sublayer is given by:
EvTopGR = minimum( revised Table 6.6 entry, 2T/100 x EvSG )
4. Ratio of (vertical) moduli of adjacent sub-layers is given (as currently) by:
R = (EvTopGR /EvSG )1/n
5. Determine (vertical) moduli for each sublayer in the current manner.
Notes
1. In design rule (1), the values 5 and 50 were (somewhat arbitrarily) selected on the grounds of
practicality. Either or both may be changed without upsetting the approach. The rule can be
presented as a simplified version of the current (1992 Guide) Table 8.2.
2. The additional constrain in design rule (3) simply ensures that the modular ratio, R, stays within
reasonable bounds when the sublayer thickness is small. Its effect is illustrated in the following
Table.
Sublayer Maximum
thickness modular
(mm) ratio (R)
50 1.4 ( = 20.5 )
100 2
150 2.8 ( = 21.5 )
Corresponding current values are 2, 4 and 8 respectively. See the attached plot. If one wishes to
“tighten the screws more”, one simply increases the value of 100 in the design rule to 125 or 150 or
whatever. For example, a change to 150 changes the above 3 values to 1.26, 1.59 and 2 respectively.
AUSTROADS 2004
— 3.44 —
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
(ETGR / ESG )max
16
15 CURRENT
14
13
12
11
10
9
PROPOSED
8
7
6
5
4
3
2
1
0
0 100 200 300 400 500
T (mm)
3. As the accompanying Table shows, the above rules imply that, for sprayed sealed granular
pavements designed according to Figure 8.4, a value of 500 MPa can be achieved for the top
(sublayer) of the granular material when Design ESAs > approx 2x105. In my view, this is a
reasonable result.
Minimum Design ESAs For Sprayed-Sealed Granular Pavements such that the
Top (Sublayer) of the Granular Layer CAN be Assigned a Value of 500 MPa
Under the Above Rules
CBR ETGR / ESG Tmin Design ESAs for Tmin

(mm) (Fig. 8.4)
2 25 464 < 105
3 16.7 406 1.6 x 105
4 12.5 364 105
5 10 332 2.5 x 105
7 7.1 283 2.5 x 105
10 5 232 4 x 105
15 3.3 174 2 x 105
20 2.5 132 105
30 1.7 74 < 105
AUSTROADS 2004
— 3.45 —
Attachment B2
Results of CIRCLY Analysis for Determination of Subgrade Strain Performance
Relationship
Table B.2.1
Results of CIRCLY analysis and life determined from
Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence)
Granular Subgrade Critical Strain Life from Life from Pavement Life
Thickness CBR determined by APRG (1998) Austroads (1992) (ESAs)
(mm) CIRCLY Figure 13.8.2(B) Figure 8.4
(µε) (ESAs) (ESAs)
100 2 17130 0 0.0E+00
125 2 14690 0 0.0E+00
150 2 12180 0 0.0E+00
175 2 10110 0 0.0E+00
200 2 8372 0 0.0E+00
225 2 7056 0 0.0E+00
250 2 6044 0 0.0E+00
275 2 5204 0 0.0E+00
300 2 4577 0 0.0E+00
325 2 4057 0 0.0E+00
350 2 3674 0 0.0E+00
375 2 3297 0 0.0E+00
400 2 2934 0 0.0E+00
425 2 2613 0 0.0E+00
450 2 2326 0 0.0E+00
475 2 2064 1.1E+05 1.1E+05
500 2 1841 1.5E+05 1.5E+05
525 2 1675 2.2E+05 2.2E+05
550 2 1553 3.2E+05 3.2E+05
575 2 1444 4.5E+05 4.5E+05
600 2 1345 6.5E+05 6.5E+05
625 2 1256 9.3E+05 9.3E+05
650 2 1176 1.3E+06 1.3E+06
675 2 1103 1.9E+06 1.9E+06
700 2 1037 2.7E+06 2.7E+06
725 2 976.3 3.9E+06 3.9E+06
750 2 921.2 5.6E+06 5.6E+06
775 2 870.9 8.0E+06 8.0E+06
800 2 824.8 1.1E+07 1.1E+07
825 2 782.5 1.6E+07 1.6E+07
850 2 743.5 2.3E+07 2.3E+07
875 2 707.6 3.3E+07 3.3E+07
900 2 674.4 4.8E+07 4.8E+07
925 2 643.8 6.8E+07 6.8E+07
950 2 615.3 9.8E+07 9.8E+07
975 2 588.9 1.4E+08 1.4E+08
1000 2 564.4 2.0E+08 2.0E+08
100 3 11440 0 0.0E+00
125 3 9780 0 0.0E+00
150 3 8163 0 0.0E+00
175 3 6737 0 0.0E+00
200 3 5610 0 0.0E+00
225 3 4705 0 0.0E+00
250 3 3999 0 0.0E+00
AUSTROADS 2004
— 3.46 —
Table B.2.1 (continued)

275 3 3460 1.00E+03 1.0E+03
300 3 3044 5.00E+03 5.0E+03
325 3 2710 1.02E+04 1.0E+04
350 3 2440 3.50E+04 3.5E+05
375 3 2193 9.00E+04 9.0E+04
400 3 1948 1.3E+05 1.3E+05
425 3 1740 2.0E+05 2.0E+05
450 3 1564 3.2E+05 3.2E+05
475 3 1442 4.9E+05 4.9E+05
500 3 1332 7.6E+05 7.6E+05
525 3 1234 1.2E+06 1.2E+06
550 3 1145 1.8E+06 1.8E+06
575 3 1065 2.8E+06 2.8E+06
600 3 992.7 4.4E+06 4.4E+06
625 3 927.2 6.8E+06 6.8E+06
650 3 867.9 1.0E+07 1.0E+07
675 3 814.1 1.6E+07 1.6E+07
700 3 765.1 2.5E+07 2.5E+07
725 3 720.4 3.9E+07 3.9E+07
750 3 679.6 6.0E+07 6.0E+07
775 3 642.2 9.4E+07 9.4E+07
800 3 608 1.5E+08 1.5E+08
825 3 576.5 2.2E+08 2.2E+08
850 3 547.5 3.5E+08 3.5E+08
875 3 520.8 5.4E+08 5.4E+08
900 3 496.1 8.3E+08 8.3E+08
925 3 473.3 1.3E+09 1.3E+09
950 3 452.1 2.0E+09 2.0E+09
975 3 432.5 3.1E+09 3.1E+09
1000 3 414.2 4.8E+09 4.8E+09
100 4 8561 0 0.0E+00
125 4 7324 0 0.0E+00
150 4 6094 0 0.0E+00
175 4 5020 0 0.0E+00
200 4 4197 0 0.0E+00
225 4 3539 1.00E+03 1.0E+03
250 4 3006 3.50E+03 3.5E+03
275 4 2608 1.20E+04 1.2E+04
300 4 2284 3.50E+04 3.5E+04
325 4 2037 1.00E+05 1.0E+05
350 4 1829 1.5E+05 1.5E+05
375 4 1638 2.5E+05 2.5E+05
400 4 1476 4.2E+05 4.2E+05
425 4 1358 6.9E+05 6.9E+05
450 4 1251 1.2E+06 1.2E+06
475 4 1155 1.9E+06 1.9E+06
500 4 1068 3.2E+06 3.2E+06
AUSTROADS 2004
— 3.47 —

525 4 990.1 5.3E+06 5.3E+06
550 4 919.4 8.8E+06 8.8E+06
575 4 855.5 1.5E+07 1.5E+07
600 4 797.7 2.5E+07 2.5E+07
625 4 745.3 4.1E+07 4.1E+07
650 4 697.7 6.8E+07 6.8E+07
675 4 654.5 1.1E+08 1.1E+08
700 4 615 1.9E+08 1.9E+08
725 4 579.1 3.1E+08 3.1E+08
750 4 546.2 5.2E+08 5.2E+08
775 4 516.1 8.7E+08 8.7E+08
800 4 488.4 1.4E+09 1.4E+09
825 4 463 2.4E+09 2.4E+09
850 4 439.6 4.0E+09 4.0E+09
875 4 418 6.7E+09 6.7E+09
900 4 398.1 1.1E+10 1.1E+10
925 4 379.6 1.8E+10 1.8E+10
950 4 362.5 3.1E+10 3.1E+10
975 4 346.6 5.1E+10 5.1E+10
1000 4 331.8 8.5E+10 8.5E+10
100 5 6850 0 0.0E+00
125 5 5871 0 0.0E+00
150 5 4870 0 0.0E+00
175 5 4030 0 0.0E+00
200 5 3359 0 0.0E+00
225 5 2823 6.00E+03 6.0E+03
250 5 2404 2.30E+04 2.3E+04
275 5 2081 9.00E+04 9.0E+04
300 5 1826 1.2E+05 1.2E+05
325 5 1628 2.2E+05 2.2E+05
350 5 1462 3.8E+05 3.8E+05
375 5 1345 6.8E+05 6.8E+05
400 5 1238 1.2E+06 1.2E+06
425 5 1141 2.2E+06 2.2E+06
450 5 1053 3.9E+06 3.9E+06
475 5 972.7 6.9E+06 6.9E+06
500 5 900.3 1.2E+07 1.2E+07
525 5 834.9 2.2E+07 2.2E+07
550 5 775.7 3.9E+07 3.9E+07
575 5 722.1 6.9E+07 6.9E+07
600 5 673.5 1.2E+08 1.2E+08
625 5 629.4 2.2E+08 2.2E+08
650 5 589.3 3.9E+08 3.9E+08
675 5 552.8 6.9E+08 6.9E+08
700 5 519.5 1.2E+09 1.2E+09
725 5 489.2 2.2E+09 2.2E+09
750 5 461.3 3.9E+09 3.9E+09
AUSTROADS 2004
— 3.48 —

775 5 435.8 6.9E+09 6.9E+09
800 5 412.4 1.2E+10 1.2E+10
825 5 390.9 2.2E+10 2.2E+10
850 5 371.1 3.9E+10 3.9E+10
875 5 352.8 7.0E+10 7.0E+10
900 5 335.8 1.2E+11 1.2E+11
925 5 320.2 2.2E+11 2.2E+11
950 5 305.6 3.9E+11 3.9E+11
975 5 292.2 7.0E+11 7.0E+11
1000 5 279.6 1.2E+12 1.2E+12
100 7 4897 0 0.0E+00
125 7 4188 0 0.0E+00
150 7 3479 0 0.0E+00
175 7 2880 3.00E+03 3.0E+03
200 7 2395 5.00E+04 5.0E+04
225 7 2012 7.00E+04 7.0E+04
250 7 1720 1.3E+05 1.3E+05
275 7 1487 2.7E+05 2.7E+05
300 7 1317 5.4E+05 5.4E+05
325 7 1200 1.1E+06 1.1E+06
350 7 1107 2.2E+06 2.2E+06
375 7 1022 4.4E+06 4.4E+06
400 7 943.6 8.9E+06 8.9E+06
425 7 871.4 1.8E+07 1.8E+07
450 7 805.4 3.6E+07 3.6E+07
475 7 745.4 7.3E+07 7.3E+07
500 7 690.8 1.5E+08 1.5E+08
525 7 641.3 3.0E+08 3.0E+08
550 7 596.3 6.0E+08 6.0E+08
575 7 555.5 1.2E+09 1.2E+09
600 7 518.4 2.4E+09 2.4E+09
625 7 484.6 4.9E+09 4.9E+09
650 7 453.9 9.9E+09 9.9E+09
675 7 425.9 2.0E+10 2.0E+10
700 7 400.3 4.0E+10 4.0E+10
725 7 376.9 8.1E+10 8.1E+10
750 7 355.4 1.6E+11 1.6E+11
775 7 335.7 3.3E+11 3.3E+11
800 7 317.7 6.6E+11 6.6E+11
825 7 301 1.3E+12 1.3E+12
850 7 285.7 2.7E+12 2.7E+12
875 7 271.5 5.4E+12 5.4E+12
900 7 258.4 1.1E+13 1.1E+13
925 7 246.2 2.2E+13 2.2E+13
950 7 235 4.5E+13 4.5E+13
975 7 224.5 9.0E+13 9.0E+13
1000 7 214.8 1.8E+14 1.8E+14
AUSTROADS 2004
— 3.49 —

100 10 3426 0 0.0E+00
125 10 2929 0 0.0E+00
150 10 2439 1.00E+04 1.0E+04
175 10 2018 6.00E+04 6.0E+04
200 10 1676 1.3E+05 1.3E+05
225 10 1412 3.1E+05 3.1E+05
250 10 1229 7.4E+05 7.4E+05
275 10 1088 1.8E+06 1.8E+06
300 10 977.2 4.2E+06 4.2E+06
325 10 890.8 1.0E+07 1.0E+07
350 10 823.4 2.4E+07 2.4E+07
375 10 762.7 5.8E+07 5.8E+07
400 10 706 1.4E+08 1.4E+08
425 10 653.5 3.3E+08 3.3E+08
450 10 605.2 7.9E+08 7.9E+08
475 10 561 1.9E+09 1.9E+09
500 10 520.7 4.5E+09 4.5E+09
525 10 483.9 1.1E+10 1.1E+10
550 10 450.4 2.6E+10 2.6E+10
575 10 419.9 6.2E+10 6.2E+10
600 10 392.1 1.5E+11 1.5E+11
625 10 366.8 3.5E+11 3.5E+11
650 10 343.7 8.5E+11 8.5E+11
675 10 322.5 2.0E+12 2.0E+12
700 10 303.2 4.8E+12 4.8E+12
725 10 285.5 1.2E+13 1.2E+13
750 10 269.3 2.8E+13 2.8E+13
775 10 254.3 6.6E+13 6.6E+13
800 10 240.6 1.6E+14 1.6E+14
825 10 228 3.8E+14 3.8E+14
850 10 216.3 9.1E+14 9.1E+14
875 10 205.5 2.2E+15 2.2E+15
900 10 195.5 5.2E+15 5.2E+15
925 10 186.3 1.2E+16 1.2E+16
950 10 177.7 3.0E+16 3.0E+16
975 10 169.7 7.1E+16 7.1E+16
1000 10 162.3 1.7E+17 1.7E+17
100 15 2283 1.30E+03 1.3E+03
125 15 1954 1.20E+04 1.2E+04
150 15 1625 1.0E+05 1.0E+05
175 15 1367 3.2E+05 3.2E+05
200 15 1161 9.9E+05 9.9E+05
225 15 997.4 3.1E+06 3.1E+06
250 15 869 9.4E+06 9.4E+06
275 15 768.3 2.9E+07 2.9E+07
300 15 689.5 9.0E+07 9.0E+07
325 15 628.2 2.8E+08 2.8E+08
AUSTROADS 2004
— 3.50 —

350 15 581.1 8.6E+08 8.6E+08
375 15 540.2 2.6E+09 2.6E+09
400 15 501.5 8.2E+09 8.2E+09
425 15 465.4 2.5E+10 2.5E+10
450 15 431.9 7.8E+10 7.8E+10
475 15 401.1 2.4E+11 2.4E+11
500 15 372.8 7.4E+11 7.4E+11
525 15 347 2.3E+12 2.3E+12
550 15 323.3 7.1E+12 7.1E+12
575 15 301.7 2.2E+13 2.2E+13
600 15 281.9 6.7E+13 6.7E+13
625 15 263.9 2.1E+14 2.1E+14
650 15 247.4 6.4E+14 6.4E+14
675 15 232.3 2.0E+15 2.0E+15
700 15 218.4 6.1E+15 6.1E+15
725 15 205.7 1.9E+16 1.9E+16
750 15 194 5.8E+16 5.8E+16
775 15 183.3 1.8E+17 1.8E+17
800 15 173.4 5.5E+17 5.5E+17
825 15 164.2 1.7E+18 1.7E+18
850 15 155.8 5.3E+18 5.3E+18
875 15 148 1.6E+19 1.6E+19
900 15 140.8 5.0E+19 5.0E+19
925 15 134.1 1.6E+20 1.6E+20
950 15 127.9 4.8E+20 4.8E+20
975 15 122.1 1.5E+21 1.5E+21
1000 15 116.7 4.6E+21 4.6E+21
100 20 1714 1.00E+04 1.0E+04
125 20 1485 1.0E+05 1.0E+05
150 20 1260 3.9E+05 3.9E+05
175 20 1065 1.5E+06 1.5E+06
200 20 905.9 5.9E+06 5.9E+06
225 20 778.4 2.3E+07 2.3E+07
250 20 677.7 8.7E+07 8.7E+07
275 20 598.8 3.4E+08 3.4E+08
300 20 537.1 1.3E+09 1.3E+09
325 20 489 5.0E+09 5.0E+09
350 20 452.3 1.9E+10 1.9E+10
375 20 421.5 7.4E+10 7.4E+10
400 20 392.2 2.9E+11 2.9E+11
425 20 364.6 1.1E+12 1.1E+12
450 20 338.9 4.2E+12 4.2E+12
475 20 315.1 1.6E+13 1.6E+13
500 20 293.2 6.3E+13 6.3E+13
525 20 273.1 2.4E+14 2.4E+14
550 20 254.7 9.4E+14 9.4E+14
575 20 237.8 3.6E+15 3.6E+15
AUSTROADS 2004
— 3.51 —

600 20 222.4 1.4E+16 1.4E+16
625 20 208.2 5.4E+16 5.4E+16
650 20 195.3 2.1E+17 2.1E+17
675 20 183.4 8.0E+17 8.0E+17
700 20 172.5 3.1E+18 3.1E+18
725 20 162.5 1.2E+19 1.2E+19
750 20 153.3 4.6E+19 4.6E+19
775 20 144.8 1.8E+20 1.8E+20
800 20 137 6.8E+20 6.8E+20
825 20 129.8 2.6E+21 2.6E+21
850 20 123.1 1.0E+22 1.0E+22
875 20 116.9 3.9E+22 3.9E+22
900 20 111.2 1.5E+23 1.5E+23
925 20 105.9 5.8E+23 5.8E+23
950 20 101 2.2E+24 2.2E+24
975 20 96.39 8.6E+24 8.6E+24
1000 20 92.11 3.3E+25 3.3E+25
100 30 1169 1.1E+05 1.1E+05
125 30 1028 5.9E+05 5.9E+05
150 30 879.9 3.2E+06 3.2E+06
175 30 746.9 1.8E+07 1.8E+07
200 30 635.9 9.6E+07 9.6E+07
225 30 546.4 5.3E+08 5.3E+08
250 30 475.5 2.9E+09 2.9E+09
275 30 419.7 1.6E+10 1.6E+10
300 30 376.1 8.6E+10 8.6E+10
325 30 342.1 4.7E+11 4.7E+11
350 30 316.2 2.6E+12 2.6E+12
375 30 295.4 1.4E+13 1.4E+13
400 30 275.7 7.7E+13 7.7E+13
425 30 256.9 4.2E+14 4.2E+14
450 30 239.2 2.3E+15 2.3E+15
475 30 222.8 1.3E+16 1.3E+16
500 30 207.7 6.9E+16 6.9E+16
525 30 193.7 3.8E+17 3.8E+17
550 30 180.9 2.1E+18 2.1E+18
575 30 169.1 1.1E+19 1.1E+19
600 30 158.2 6.2E+19 6.2E+19
625 30 148.3 3.4E+20 3.4E+20
650 30 139.2 1.8E+21 1.8E+21
675 30 130.8 1.0E+22 1.0E+22
700 30 123.1 5.5E+22 5.5E+22
725 30 116 3.0E+23 3.0E+23
750 30 109.4 1.7E+24 1.7E+24
775 30 103.4 9.0E+24 9.0E+24
800 30 97.81 4.9E+25 4.9E+25
825 30 92.66 2.7E+26 2.7E+26
AUSTROADS 2004
— 3.52 —

850 30 87.9 1.5E+27 1.5E+27
875 30 83.49 8.1E+27 8.1E+27
900 30 79.39 4.4E+28 4.4E+28
925 30 75.59 2.4E+29 2.4E+29
950 30 72.05 1.3E+30 1.3E+30
975 30 68.76 7.2E+30 7.2E+30
1000 30 65.69 4.0E+31 4.0E+31
AUSTROADS 2004
— 3.53 —
APPENDIX C
GRANULAR MATERIALS MODULI UNDER ASPHALT AND CEMENTED

MATERIAL
R98/008 by M. Moffatt and G. Jameson, April 1998.
C.1 INTRODUCTION
Moffatt and Jameson (1998) contains a proposed procedure for the characterisation of modulus of granular
materials for use in the revised Austroads Pavement Design Guide. The current procedures detailed in
Austroads (1992) allow the maximum modulus of a granular material to be developed regardless of the
thickness of the granular layer or the strength of the subgrade. The procedure detailed in Moffatt and
Jameson alters characterisation procedure, so that the maximum modulus of granular material is dependent
upon both granular material thickness and subgrade strength.
Table 6.6 of Austroads (1992), reproduced here as Table C.1, provides suggested values of the vertical
modulus of the top-sublayer of granular material when covered by asphalt (and cemented material) of
different thickness, and at different temperatures. With the wide variety of alternative asphalt mixes, the use
of temperature in this chart is no longer considered appropriate. A far more appropriate parameter is the
actual modulus of the overlying material.
Yeo et al. (1997) proposed an alternative form of this table, reproduced here as Table C.2, showing the
suggested vertical modulus of granular material as a function of overlying asphalt thickness and modulus. At
the inaugural meeting of the Austroads Pavement Design Guide Reference Group, it was decided that the
form of the revised table was appropriate for inclusion in the next edition of the Design Guide, but that it
contained internal inconsistencies which needed to be rectified before being adopted.
This report proposes a new version of this table, and also briefly documents the procedure used to develop
the values contained in the Table.
Table C.1
Suggested Vertical Modulus (MPa) of Top Sub-Layer of Granular Material
(Austroads 1992, Table 6.6)
Overlying Material Weighted Mean Annual Pavement

Temperature
20°C 25°C 30°C
less than 25 mm asphalt 5001/350 5001/350 5001/350
1 1 1
50 mm asphalt 350 /350 500 /350 500 /350
1 1
100 mm asphalt 210 /150 350 /250 5001/350
150 mm asphalt 2101/150 2101/150 3501/250
cemented material 2101/150 2101/150 2101/150
1
where material is compacted using Modified compactive effort.
AUSTROADS 2004
— 3.54 —
Table C.2
Suggested Vertical Modulus (MPa) of Top Sub-Layer of Granular Material
(Yeo et al. 1997, Table 6.6)
Overlying Modulus of Asphalt (MPa)

Material 1000 2000 3000 4000 5000 6000 7000
1 1 1 1 1 1
<25 mm asphalt 500 / 350 500 / 350 500 / 350 460 / 330 380 / 270 300 / 220 2401 / 180
50 mm asphalt 5001 / 350 5001 / 350 4601 / 320 3901 / 280 3101 / 230 2601 / 200 2301 / 180
1 1 1 1 1 1
75 mm asphalt 500 / 350 500 / 350 410 / 290 320 / 230 230 / 180 210 / 170 2101 / 170
100 mm asphalt 5001 / 350 4501 / 320 3401 / 250 2601 / 200 2201 / 170 2101 / 170 2101 / 170
1 1 1 1 1 1
125 mm asphalt 500 / 350 400 / 280 270 / 200 210 / 170 210 / 170 210 / 170 2101 / 170
150 mm asphalt 5001 / 350 3301 / 240 2401 / 180 2101 / 170 2101 / 170 2101 / 170 2101 / 170
200 mm asphalt 4301 / 310 2301 / 180 2101 / 170 2101 / 170 2101 / 170 2101 / 170 2101 / 170
1 1 1 1 1 1
250 mm asphalt 280 / 210 210 / 170 210 / 170 210 / 170 210 / 170 210 / 170 2101 / 170
300 mm asphalt 2101 / 170 2101 / 170 2101 / 170 2101 / 170 2101 / 170 2101 / 170 2101 / 170
1 1 1 1 1 1
cemented material 210 / 170 210 / 170 210 / 170 210 / 170 210 / 170 210 / 170 2101 / 170
1
where material is compacted using modified compaction
C.2 DERIVATION OF TABLE 6.6 IN AUSTROADS (1992)

The moduli values in Table 6.6 of Austroads (1992), reproduced here as Table C.1, were derived from a
study by Geoff Youdale (Youdale 1983a and b). Youdale used a finite element package, PAVAN1, to
determine modulus values at various locations within a granular material which was covered with asphalt.
The granular material model was based on a octahedral stress-dependency model of a Hornsby Breccia
modified with 2% hydrated lime, representing a good quality crushed rock base material, well compacted at
roughly 2% below optimum.
Contour plots of granular modulus were made, and from these plots the entries in Table C.1 for material
compacted using Standard compactive effort, covered by asphalt at 25°C of thicknesses 25, 50, 100 and
150 mm and covered by asphalt at 30°C of thicknesses 100 and 150 mm. From these values all other values
in Table C.1 were extrapolated.
As part of this report, Youdale’s reports were reviewed. The contour plots developed do not seem to bear
out the conclusions drawn from them. It would appear that a value of 350 MPa had already been identified
as the maximum modulus achievable by granular materials, compacted using Standard compactive effort,
and that the significantly lower values shown in the contour plots were shifted upwards. Such a shift would
have been conducted in order that a match occur between a value of 350 MPa for the top sub-layer with a
spray seal surface, and the value indicated by the finite element analysis and subsequently derived contour
plot.
It worth noting that a large number of the values in Table C.1 are based on extrapolations of a limited
number of analyses. It would be inappropriate to further expand upon these extrapolations, with a view to
expanding the Table’s range of both thickness and modulus of cover, as has been conducted in Yeo et al.
(1997), and shown here in Table C.2.
With a view to providing an expanded version of this table, additional modelling was conducted, and is
described below.
AUSTROADS 2004
— 3.55 —
C.3 PROCEDURE USED TO DEVELOP NEW TABLE

A series of CIRCLY (Mincad 1996) analyses were conducted in order to determine the stresses in the top
sublayer of granular material under covers of different combinations of asphalt thickness and modulus.
Two different qualities of granular material were examined consistent with Austroads (1992). These
different qualities were characterised by the maximum and minimum modulus values the material was
considered to be able to achieve. The first quality, 350/150, was considered to be able to achieve a
maximum modulus of 350 MPa, and a minimum modulus of 150 MPa. This range of modulus values
matches with the standard compactive effort modulus values in Table C.1. The second quality was 500/210,
which corresponds with Modified compactive effort material in Table C.1.
The maximum modulus of the granular material was assumed to occur in the top-sublayer under a cover of
25 mm of 1,000 MPa asphalt, which could be considered to be roughly equivalent to a spray seal. The
minimum modulus was assumed to be achieved under a cover of 150 mm of 3,000 MPa asphalt, similar to
the Austroads (1992) design procedures.
A series of granular pavements of thicknesses (300, 400 and 500 mm) were modelled with subgrade vertical
moduli of 70, 100, 150 and 200 MPa. Weaker pavements were not modelled as these pavements may not
have been able to develop the maximum modulus. For each pavement the granular material was divided into
five sub-layers, as per the procedure in Moffatt and Jameson (1998). Each pavement was analysed with a
cover of 25 mm of 1,000 MPa asphalt, and the mean principal stress11 was determined at the mid-layer point
of the top granular sublayer. The pavements were loaded with a half-standard axle load (with a tyre pressure
of 750 kPa) and the principal stresses determined beneath one of the two wheels.
Similarly, each pavement was modelled with a cover of 150 mm of 3,000 MPa asphalt, and the mean
principal stress determined at the mid-layer depth of the top granular sublayer.
Pavements whose top sublayer of granular material did not achieve its maximum modulus, due to inadequate
thickness of granular material for the specific subgrade strength, under 25 mm of 1,000 MPa asphalt were
excluded from any further analysis.
These two analyses provided mean principal stresses associated with the upper and lower bounds of the top
granular sublayer modulus. By taking the base-10 logarithm of both the stresses and the modulus, and
assuming that a linear relationship existed between them, these data then enabled the interpolation of top-
sublayer modulus for a given mean principal stress. Figure C.1 schematically shows the derivation of the
straight line relationship. These relationships were developed for each of the two material qualities.
Subsequent modelling of the pavements with a variety of combinations of asphalt thickness and modulus was
then conducted. This modelling determined the mean principal stress under a specific combination of asphalt
thickness and modulus. Using the linear relationship between the logarithm of stress and the logarithm of
modulus, for the specific combination of granular material thickness and subgrade strength, to determine a
modulus of that material. This modulus was then used to seed a subsequent iteration of the analysis, which
produced another mean principal stress, and hence estimation of modulus. This iteration stage was
conducted three times.
11
In determining the mean principal stress, principal tensile stresses were set to zero.
AUSTROADS 2004
— 3.56 —
1000 MPa 25 mm
350 MPa
Log( Etop )
Log( σmean )
3000 MPa 150 mm
150 MPa
Figure C.1: Schematic of process used to derive relationship between

mean principal stress and vertical modulus
C.4 RESULTS OF ANALYSIS

For each combination of asphalt thickness and modulus and granular material quality, the modulus values for
the range of pavements with varying subgrade strengths and granular thicknesses were averaged, and results
rounded to the nearest 5 MPa to produce Tables C.3-C.4.
Table C.3
Suggested Vertical Modulus (MPa) of Top Sub-Layer of Granular Material for
Standard-Quality Granular Materials Under Asphalt Cover

Material 1000 2000 3000 4000 5000 6000 7000 8000
25 mm 350 335 330 325 320 315 310 305
50 mm 325 310 300 290 285 280 275 270
35 mm 295 275 265 255 245 240 235 230
75 mm 255 235 220 210 200 195 190 185
100 mm 225 205 190 180 170 165 160 160
125 mm 205 180 165 160 150 150 150 150
150 mm 185 160 150 150 150 150 150 150
175 mm 165 150 150 150 150 150 150 150
200 mm 155 150 150 150 150 150 150 150
225 mm 150 150 150 150 150 150 150 150
250 mm 150 150 150 150 150 150 150 150
275 mm 150 150 150 150 150 150 150 150
300 mm 150 150 150 150 150 150 150 150
AUSTROADS 2004
— 3.57 —
C.5 BOUND CEMENT MATERIAL COVER

Table C.1 contains suggested values for the top sublayer modulus of granular material when covered by
cement treated material. In order to determine whether the values contained in Tables C.3-C.4 are equally
applicable to bound cement material cover as they are to asphalt cover, the above analysis procedure was
repeated for cover moduli of 2,000 and 5,000 MPa, but changing the Poisson’s Ratio from 0.40 (asphalt) to
0.2 (bound cement material). The results, shown in Tables C.5-C.6, indicate show very little difference from
the corresponding values in Tables C.3-C.4. Therefore, Tables C.3-C.4 can be re-written to express the
modulus of the top sublayer of granular material as a function of covering bound material thickness and
modulus, and not just asphalt thickness and modulus as was the case in Tables C.3-C.4. This has been done
in Tables C.7-C.8.
Table C.4
High-Quality Granular Materials Under Asphalt Cover

Material 1000 2000 3000 4000 5000 6000 7000 8000
25 mm 500 475 465 460 450 445 440 435
35 mm 460 440 425 415 410 400 395 385
50 mm 420 395 380 365 350 345 335 330
75 mm 370 335 315 300 285 275 270 265
100 mm 325 290 270 255 245 235 225 220
125 mm 290 255 235 220 210 210 210 210
150 mm 260 230 210 210 210 210 210 210
175 mm 235 210 210 210 210 210 210 210
200 mm 215 210 210 210 210 210 210 210
225 mm 210 210 210 210 210 210 210 210
250 mm 210 210 210 210 210 210 210 210
275 mm 210 210 210 210 210 210 210 210
300 mm 210 210 210 210 210 210 210 210
Table C.5
Suggested Vertical Modulus (MPa) of Top Sub-Layer of Granular Material for Standard-
Quality Granular Materials Under Bound Cement Material Cover (c.f. Table C.3)
Overlying Modulus of Cemented

Material Material (MPa)
2000 5000
25 mm 335 315
35 mm 305 280
50 mm 275 245
75 mm 230 200
100 mm 200 170
125 mm 180 150
150 mm 160 150
175 mm 150 150
200 mm 150 150
225 mm 150 150
250 mm 150 150
275 mm 150 150
300 mm 150 150
AUSTROADS 2004
— 3.58 —
Table C.6
High-Quality Granular Materials under Bound Cement Material Cover (c.f. Table C.4)
Overlying Modulus of Cemented

Material Material (MPa)
2000 5000
25 mm 470 440
35 mm 435 400
50 mm 390 345
75 mm 330 285
100 mm 285 245
125 mm 255 210
150 mm 230 210
175 mm 210 210
200 mm 210 210
225 mm 210 210
250 mm 210 210
275 mm 210 210
300 mm 210 210
Table C.7
Suggested Vertical Modulus (MPa) of Top Sublayer of
Granular Material for Standard-Quality Granular Materials
Overlying Modulus of Cover Material (MPa)

Material 1000 2000 3000 4000 5000 6000 7000 8000
25 mm 350 335 330 325 320 315 310 305
50 mm 325 310 300 290 285 280 275 270
35 mm 295 275 265 255 245 240 235 230
75 mm 255 235 220 210 200 195 190 185
100 mm 225 205 190 180 170 165 160 160
125 mm 205 180 165 160 150 150 150 150
150 mm 185 160 150 150 150 150 150 150
175 mm 165 150 150 150 150 150 150 150
200 mm 155 150 150 150 150 150 150 150
225 mm 150 150 150 150 150 150 150 150
250 mm 150 150 150 150 150 150 150 150
275 mm 150 150 150 150 150 150 150 150
300 mm 150 150 150 150 150 150 150 150
AUSTROADS 2004
— 3.59 —
Table C.8
Suggested Vertical Modulus (MPa) of Top Sublayer of
Granular Material for High Quality Granular Materials
Overlying Modulus of Cover Material (MPa)

Material 1000 2000 3000 4000 5000 6000 7000 8000
25 mm 500 475 465 460 450 445 440 435
35 mm 460 440 425 415 410 400 395 385
50 mm 420 395 380 365 350 345 335 330
75 mm 370 335 315 300 285 275 270 265
100 mm 325 290 270 255 245 235 225 220
125 mm 290 255 235 220 210 210 210 210
150 mm 260 230 210 210 210 210 210 210
175 mm 235 210 210 210 210 210 210 210
200 mm 215 210 210 210 210 210 210 210
225 mm 210 210 210 210 210 210 210 210
250 mm 210 210 210 210 210 210 210 210
275 mm 210 210 210 210 210 210 210 210
300 mm 210 210 210 210 210 210 210 210
C.6 COVER COMPRISED OF MULTIPLE BOUND MATERIALS

Some pavements comprise layers of different moduli of bound material (asphalt or cemented material)
overlying granular layers. It is proposed that for such pavements an equivalent modulus (Ee) of the bound
material be calculated using the following formula:
3
⎛ n
⎛⎜ t E 13 ⎞⎟ ⎞⎟
∑ ⎝ i( i) ⎠⎟
⎜
E e = ⎜⎜ i =1
⎟
T
⎜⎜ ⎟⎟
⎝ ⎠
where
E e = equivalent modulus of total thickness of bound material (MPa)
E i = modulus of layer i (MPa)
t i = thickness of layer i (mm)
T = total thickness of overlying bound materials (mm)
Based on values of Ee and T so calculated, the vertical modulus of the top sublayer of granular materials are
then calculated from Tables C.7-C.8.
C.7 SUMMARY AND RECOMMENDATIONS

It is recommended that Tables C.7-C.8 replace the existing Table 6.6 in the Austroads Pavement Design
Guide. These tables provide suggested values for the vertical modulus of the top sublayer of granular
materials as a function of thickness and modulus of covering material. Furthermore, it is also recommended
that the procedure described in Section C.6 above be incorporated into the Guide to provide guidance on
estimating the effective combined modulus of a cover comprised of multiple materials.
AUSTROADS 2004
— 3.60 —
REFERENCES
Sydney).
Moffatt, M.A. and Jameson, G.W. (1998). Characterisation of Granular Material and Development of a
Subgrade Strain Criterion. ARRB TR Working Document No. WD-R98/005.
Yeo, R.E.Y, Jameson, G.W. and Moffatt, M.A. (1997). Pavement Materials – Granular Materials/Cemented
Materials – Draft Revision of the Austroads Pavement Design Guide. ARRB TR Working Document No.
WD-R97/071, November 1997.
Youdale, G.P. (1983a). Investigation of the Variation of Stiffness With Depth of a Granular Layer Under
Variable Thickness of Asphaltic Concrete. Report to Working Group, 8 July, 8 pp.
Youdale, G.P. (1983b). Investigation of the Variation of Stiffness With Depth Of A Granular Layer Under
Variable Thickness of Asphaltic Concrete. Report to Working Group, October 25, 5 pp.
AUSTROADS 2004
— 3.61 —
APPENDIX D
GRANULAR MATERIALS MODULI UNDER ASPHALT AND CEMENTED

MATERIAL
2004 Austroads Pavement Design Guide Reference Group
Discussion Note Table 6.4: Granular Moduli Under Asphalt
June 2003
Currently, the 2004 Austroads Pavement Rehabilitation Guide overlay design procedures are being revised
due to the changes that have recently been to the procedures in the Austroads Pavement Design Guide.
These changes included changes to Table 6.4 of the Pavement Design Guide which gives granular moduli
under various thicknesses and moduli of overlaying asphalt. The current Tables (May 2003 Austroads Guide
draft) are given below. The values in the Tables were derived from the moduli in Table 6.6 of the 1992
Guide.
May 2003 Table 6.4(a)
Suggested Vertical Modulus (MPa) of Top Sub-Layer of Normal Standard Base Material
Thickness of Modulus of Cover1 Material (MPa)

Overlying 1000 2000 3000 4000 5000
Material
40 mm 350 350 350 340 280
75 mm 350 350 340 270 210
100 mm 350 350 290 220 160
125 mm 350 340 240 170 150
150 mm 350 290 190 150 150
175 mm 350 240 150 150 150
200 mm 350 190 150 150 150
250 mm 260 150 150 150 150
>=300 mm 150 150 150 150 150
1. Cover material is either asphalt or cemented material or a combination of these materials.
May 2003 Table 6.4(b)

Suggested Vertical Modulus (MPa) of Top Sub-Layer of High Standard Base Material

Overlying 1000 2000 3000 4000 5000
Material
40 mm 500 500 500 480 400
75 mm 500 500 480 380 300
100 mm 500 500 410 310 230
125 mm 500 490 340 240 210
150 mm 500 410 270 210 210
175 mm 500 340 210 210 210
200 mm 500 270 210 210 210
250 mm 380 210 210 210 210
>=300 mm 210 210 210 210 210
AUSTROADS 2004
— 3.62 —
Recently the moduli estimated from these tables have been used to predict the variation in FWD deflections
with asphalt temperature, as part of the development of procedures to adjust measured deflections from the
measurement temperature to the WMAPT.
As seen from the examples given in Figure D.1, the May 2003 Table 6.4 results in a counter-intuitive
variation in FWD deflections with temperature for some pavement configurations.
50mm,100mm and 150 mm asphalt on 500mm granular on subgrade CBR=3
1.4
190MPa Top Granular Moduli
350MPa
230MPa 350MPa
1.3 350MPa 350MPa
270MPa 350MPa
310MPa 350MPa
FWD maximum deflections (mm)
350MPa 350MPa
1.2
350MPa
1.1
350MPa 350MPa
150MPa 170MPa
150MPa 210MPa 250MPa 290MPa 330MPa 350MPa 350MPa
1.0
350MPa
350MPa 350MPa
0.9 230MPa 270MPa 310MPa
150MPa 190MPa
150MPa
150MPa
0.8 150MPa
150MPa
0.7
0.6
10 15 20 25 30 35 40
Asphalt Temperature (C)
50mm 100mm 150mm
Figure D.1: Examples of predicted FWD maximum deflections (566 kPa)

variation with temperature using May 2003 Table 6.4
To shed further light on the issue, the literature was briefly examined. Of particular interest was a paper by
Andrew Dawson of the University of Nottingham summarising research undertaken by Lois Plaistow and
himself. The authors used an FEM package called FENLAP to calculate the granular moduli variation with
depth under asphalt thicknesses of 50 mm, 100 mm, 150 mm and 250 mm and asphalt moduli of 2,000, 5,000
and 8,000 MPa. The data was reported for only one type of granular material, a hard limestone with a high
modulus.
Reading from the graphs in the paper, the granular moduli 50 mm below the bottom of the asphalt were
estimated. These moduli were assumed to be similar to the Austroads top granular modulus. Given the
granular material tested was atypical of commonly-used Australian materials, the use of the Dawson data
was limited to an examination of the relative effects of asphalt thickness and modulus on granular moduli as
follows:
• Firstly, the data was used to estimate the moduli under 50 mm of asphalt with an asphalt modulus of
3,000 MPa. The granular modulus was about 560 MPa.
• The ratio of the granular modulus at each asphalt thickness and asphalt modulus to the reference modulus
of 560 MPa, was then calculated. These were called relative granular moduli.
• Based on the assumption that the stiffness of an asphalt layer is related to its moment of inertia (I)
multiplied by its modulus (E) (Odemark’s method of equivalent thickness), the following stiffness
parameter was calculated for each combination of asphalt thickness and asphalt modulus:
= asphalt thickness x 3 Asphalt Modulus
AUSTROADS 2004
— 3.63 —
The relative granular moduli variation with the stiffness parameter (asphalt thickness x Eac0.3333) estimated
from Dawson’s data is given in Figure D.2. As anticipated granular modulus is closely related to the
stiffness parameter as the stiffness of the overlaying asphalt influences the load-induced stresses in
underlying granular materials.
1.0
0.9
0.8
Relative Granular Modulus
0.7
Based on Dawson (1999)

0.6 y = 7.40E-08x2 - 4.90E-04x + 1.31
0.5
0.4 June 2003 revision Table 6.4

y = -3.804E-04x + 1.377
0.3
500 1000 1500 2000 2500 3000 3500
Asphalt thickness*Eac^0.3333
Revised Table 6.4 Dawson 1999
Figure D.2: Comparison of relative granular moduli variation with the stiffness parameter
estimated from Dawson’s data
In contrast to Dawson’s results, the May 2003 Table 6.4 moduli do not vary consistently with the stiffness
parameter as seen from Figure D.3.
Based on May 2003 Table 6.4
1.0
0.9
Relative Granular Moduli
0.8
0.7
0.6
0.5
0.4
0 500 1000 1500 2000 2500 3000 3500
Asphalt thickness X E^0.3333
40 mm 75mm 100mm 125mm 150mm 175mm 200mm
Figure D.3: Comparison of relative granular modulus and asphalt thickness

(May 2003, Table 6.6 data)
AUSTROADS 2004
— 3.64 —
Based on the above it was concluded that the May 2003 Table 6.4 need to be revised. Two possible options
are identified:
1. Revise Table 6.4 completely used Dawson’s and other published data. Although this option would
provide improved granular material moduli, it has the disadvantage that it requires the recently revised
Example Design Charts to be recalculated. Given the current project timings this option is not
considered viable in the short-term, but is recommended for future investigation.
2. Retain the May 2003 Table 6.4 granular moduli for asphalt modulus of 3,000 MPa as these granular
moduli were recently used to revise the Example Charts. The granular moduli for other asphalt moduli
can then be estimated assuming the granular moduli are related to the stiffness parameter.
Consequently in this discussion note, the granular moduli using Option (2) were calculated.
The first step was to use the May 2003 granular moduli for 3,000 MPa to develop a relationship between
relative granular modulus to the stiffness parameter. The results are plotted in Figure D.2, marked “June
2003 Table 6.4”.
Using the “June 2003” relationship given in Figure D.2, granular moduli for all thickness and moduli of
overlying asphalt were then estimated. These revised granular moduli are given below:
Using these June 2003 granular moduli, the variation of FWD deflection with temperature was predicted
(Figure D.4). The revised granular moduli result in more logical variations in deflection with temperature.
June 2003 Table 6.4(a)

Suggested Vertical Modulus (MPa) of Top Sub-Layer of Normal-Standard Base Material
1
Thickness of Modulus of Cover Material (MPa)
Overlying 1000 2000 3000 4000 5000
Material
40 mm 350 350 350 350 350
75 mm 350 350 340 320 310
100 mm 350 310 290 270 250
125 mm 320 270 240 220 200
150 mm 280 230 190 160 150
175 mm 250 190 150 150 150
200 mm 220 150 150 150 150
225 mm 180 150 150 150 150
>=250 mm 150 150 150 150 150
AUSTROADS 2004
— 3.65 —
June 2003 Table 6.4(b)

Suggested Vertical Modulus (MPa) of Top Sub-Layer of High-Standard Base Material

Overlying 1000 2000 3000 4000 5000
Material
40 mm 500 500 500 500 500
75 mm 500 500 480 460 440
100 mm 500 450 410 390 360
125 mm 450 390 350 310 280
150 mm 400 330 280 240 210
175 mm 360 270 210 210 210
200 mm 310 210 210 210 210
225 mm 260 210 210 210 210
>=250 mm 210 210 210 210 210
50mm,100mm and 150 mm asphalt on 500mm granular on subgrade CBR=3
1.4
Top Granular Moduli
350MPa
1.3 350MPa
350MPa
350MPa 350MPa
FWD maximum deflections (mm)
350MPa 350MPa
350MPa
1.2 350MPa
350MPa
350MPa
350MPa 350MPa
1.1
340MPa
320MPa 330MPa
310MPa 280MPa
290MPa 300MPa
1.0 270MPa 280MPa
260MPa 250MPa 260MPa
230MPa 250MPa
220MPa 240MPa
0.9 210MPa
200MPa
180MPa
150MPa 160MPa
0.8 150MPa
150MPa
0.7
0.6
10 15 20 25 30 35 40
Asphalt Temperature (C)
50mm 100mm 150mm
Figure D.4: Examples of variation in predicted FWD deflection with temperature

(using June 2003 Table 6.4)
Note that, compared to the May 2003 Table 6.4 moduli, the June 2003 granular moduli are:
• lower for WMAPTs > 30°C,
• the same for WMAPTs 25°C-30°C, and
• higher for WMAPTs < 25°C.
Based on the assumption that additional funds/time will not be provided to revise the Example Design
Charts, it is recommended that the revised Table 6.4 given in this Discussion Note be adopted in the 2004
Austroads Pavement Design Guide.
Geoff Jameson
Principal Research Scientist
AUSTROADS 2004
— 3.66 —
REFERENCE
Dawson, A. (1999). Implications of Granular Material Characteristics on the Response Of Different
Pavement Constructions. Published in Unbound Granular Materials: Laboratory Testing, In-situ Testing and
Modelling, Edited by A. Gomes Correia.
AUSTROADS 2004
— 3.67 —
APPENDIX E
RELATIONSHIP BETWEEN UNCONFINED COMPRESSIVE

STRENGTH AND FLEXURAL MODULUS FOR CEMENTED
MATERIALS
R98/024 by M. Moffatt and R. Yeo, August 1998.
E.1 BACKGROUND
This document details the basis for a change in the relationship between flexural modulus and Unconfined
Compressive Strength (UCS) used for estimation of the elastic properties of cemented granular materials in
the Austroads Pavement Design Guide (Austroads, 1992). This work has been undertaken as part of the
Austroads National Strategic Research Program Project NT&E9617B titled “Revision of the Austroads
Pavement Design Guide”.
E.2 EXISTING RELATIONSHIPS BETWEEN UCS AND FLEXURAL

MODULUS OF CEMENTED MATERIALS
Austroads (1992) currently provides two equations relating UCS to modulus. These relationships were
included in the Guide to provide a means for estimation of design modulus where actual moduli values were
not available. The two relationships are reproduced below as eqns (E.1) and (E.2).
E = 1814 UCS0.88 + 3500 for cemented crushed rock (E.1)
E = 2240 UCS0.88 + 1100 for cemented natural gravel (E.2)
where E = Elastic Modulus (MPa), and
UCS = Unconfined Compressive Strength (MPa).
A previous project report (Yeo et al, 1997) contained draft revised sections of Chapter 6 of the Austroads
Guide. This report suggested a more suitable relationship would be that reproduced here as eqn (E.3).
EFLEX = 2966 UCS0.83 for cemented crushed rock and cemented natural gravel (E.3)
where EFLEX = Flexural modulus at 28 days moist curing (MPa), and
UCS = Unconfined Compressive Strength at 28 days (MPa).
Yeo (1997) detailed the origin of eqns (E.1) and (E.2), and also documented the procedure used to derive eqn
(E.3). Yeo noted that the equations are based on the work by Otte (1978) in South Africa on data collected
from flexural beam samples of field-placed and cured cemented material.
Yeo (1997) examined Otte’s data set, and other possible sources of new, and in particular Australian, data,
and concluded that, at the current time, Otte’s data represented the best available.
The data collected by Otte included values of flexural modulus in the range of 1,000 MPa to nearly 40,000
MPa (corresponding to a UCS of about 34 MPa). As this data far exceeded that required for the pavement
design incorporating cemented materials in Australia, Yeo limited the data set to values of flexural modulus
below 20,000 MPa and determined a new relationship between UCS and flexural modulus (eqn (E.3)).
AUSTROADS 2004
— 3.68 —
Copies of the Yeo at al. (1997) and Yeo (1997) reports were circulated to members of the Austroads
Pavement Design Guide Reference Group for comment. The Group decided that the form of eqn (E.3)
implied a degree of precision unwarranted by the amount of scatter in the data used to determine the
regression relationship. The Group determined that a straight line relationship would provide an equally
valid fit to the data, and would not imply such a high degree of precision. This report documents the
procedure used to derive such a simplified regression equation.
E.3 DATA SET USED TO DERIVE NEW RELATIONSHIP

The data set used to derive a revised equation was that collected by Otte (1978), limited to values of UCS
below 5 MPa.
In addition to the Otte data, VicRoads was able to supply a very limited amount of preliminary data from a
current research project. This data included both UCS and flexural modulus for samples which were
manufactured in the laboratory, as well as data from field-placed and cured material. All the VicRoads data
was determined at 28 days. The VicRoads project is continuing, and will be comprehensively documented
elsewhere.
E.4 REGRESSION ANALYSIS

Figure E.1 shows the data, as well as the straight regression line (eqn (E.4)) which was found to best fit the
data.
EFLEX = 3,013 UCS (E.4)
where EFLEX = Flexural Modulus at 28 days moist curing (MPa), and
adjusted R2 = 0.60
Standard Error of Estimate = 2692
20,000
18,000
16,000
Flexural Modulus (28 days curing) (MPa)
14,000
12,000
10,000
8,000
6,000
4,000
2,000
-
- 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
Unconfined Compressive Strength (UCS) (28 days curing) (MPa)
Cement Treated Crusher Run (Otte) Cement Treated Natural Gravel (Otte)
Lab-Cement Treated Crushed Concrete (VicRoads) Field-Cement Treated Cruched Rock (VicRoads)
Lab-Cement Treated Class 3 Crushed Rock (VicRoads) E = 3013 x UCS
Figure E.1: Relationship between UCS and Flexural Modulus at 28 days
Eqn (E.4) is a simpler relationship than that suggested by Yeo (1997). Furthermore it was found that
simplifying the multiplier of UCS to 3000 (eqn (E.5)) yielded a line which fitted the data equally well as that
shown in eqn (E.4):
AUSTROADS 2004
— 3.69 —
EFLEX = 3,000 UCS (E.5)

E.5 RECOMMENDATION
It is recommended that the suggested relationship for relating flexural modulus to UCS for cemented
materials in the Austroads Pavement Design Guide be revised to the following:
EFLEX = 3000 UCS for cemented crushed rock and cemented natural gravel
AUSTROADS 2004
— 3.70 —
REFERENCES
Sydney.)
Otte, E. (1978). A Structural Design Procedure for Cement-Treated Layers in Pavements. Thesis
(D.Sc.(Eng)) University of Pretoria.
Yeo, R.E.Y. (1997). Basis for Revision of Modulus Correlations for Cemented Materials. ARRB TR
Working Document No. WD-R97/072, December.
Yeo, R.E.Y., Jameson, G.W. and Moffatt, M.A. (1997). Pavement Materials – Granular Materials/Cemented
Materials – Draft Revision of the Austroads Pavement Design Guide. ARRB TR Working Document No.
WD-R97/071.
AUSTROADS 2004
— 3.71 —
APPENDIX F
DISCUSSION NOTE ON ESTIMATING CEMENTED MATERIALS

MODULUS FROM UCS
Prepared for 2001 Guide Reference Group (June 2000)
At December 1999 meeting of the Austroads Pavement Design Guide Reference Group, concern was
expressed about the adoption of the following relationship between UCS and flexural modulus (Moffatt and
Yeo 1998):
E flex = 3000 UCS (F.1)
where Eflex = flexural modulus of field beams after 28 days moist curing (MPa), and
UCS = unconfined compressive strength of field specimens at 28 days (MPa)
A material with a 7-day UCS of 2 MPa has a 28-day UCS of field core of about 2.5 MPa. Using eqn (F1),
the 28-day flexural modulus is 7,500 MPa. This value is very high compared to QDMR specifications and
design procedures where a design modulus of 2,000 MPa is associated with a crushed rock with a 7-day UCS
of laboratory specimens of 2 MPa.
More specifically the QDMR relationships are as follows:
A design modulus of 5,000 MPa is used for Category 1 cemented materials
Crushed rocks covering a wide range of source rocks but excluding materials that are likely to break
down further in service. Materials are required to have a minimum 7-day UCS of 3 MPa.
A design modulus of 2,000 MPa is used for Category 2 cemented materials
Type A materials: crushed rocks covering a wide range of source rocks but excluding materials which
are likely to break down further in service. Materials are required to have a minimum 7-day UCS of 2
MPa.
Type B materials: loams, gravels with excess fines or coarse-grained decomposed rock. The minimum
7-day UCS of these materials is 3 MPa.
These QDMR design modulus values were based on a laboratory tested program together with moduli
back calculated from pavement deflection testing (Litwinowicz 1986).
It is considered that at least part of the reason eqn (F.1) produces higher modulus values is that the
relationship predicts the relationship between the UCS of field cores and the flexural modulus of field beam
rather than the UCS of laboratory specimens and the flexural modulus of field beams.
A study conducted for VicRoads (Alderson 1999) suggests that the laboratory UCS values are on average 1.4
times the UCS values of field cores when tested at 28-days.
Using this factor and eqn (F.1), the following relationship is obtained:
E flex = 2140 UCS (F.2)
where Eflex = flexural modulus of field beams at 28 days (MPa), and
UCS = unconfined compressive strength of laboratory specimens at 28 days (MPa)
Using eqn (F.2), a 7-day UCS of 2 MPa (28-day 2.5 MPa) results in a design modulus of 5,350 MPa
compared to the QDMR value of 2,000 MPa.
Similarly, a 7-day UCS of 3 MPa (28 day 3.9 MPa) results in a design modulus of 8,350 MPa compared to
the QDMR value of 5,000 MPa.
AUSTROADS 2004
— 3.72 —
The values using eqn (F.2) are still well in excess of current QDMR practice.
Since the production of the Moffatt and Yeo (1998) report, VicRoads has completed a project assessing
relationships between various cemented material test parameters (Alderson 1999).
Alderson reported the following relationship between flexural modulus and UCS of laboratory specimens:
E flex = 2460 UCS (F.3)
where Eflex = flexural modulus of laboratory beams at 28 days (MPa), and
UCS = unconfined compressive strength of laboratory specimens at 28 days (MPa).
Alderson reported that the laboratory flexural moduli were 2.5 times the flexural moduli of field beams.
However, the field and laboratory beams were about 3% different in density on average. Correcting for this
difference in modulus, it is estimated that the flexural modulus of the laboratory beams was about 1.6 times
the flexural modulus of the field beams. Using this factor of 1.6 and eqn (F.3), and rounding off, the
following relationship was derived:
E flex = 1,500 UCS (F.4)
Note that eqn (F.4) still estimates flexural moduli in excess of the current QDMR procedures. For example,
for a 7-day UCS of 2 MPa (28-day UCS of 2.9 MPa) the flexural modulus is 4,300 MPa, compared to a
value of 2,000 MPa used by QDMR.
During the laboratory testing that formed the basis of the current QDMR relationships (Litwinowicz 1986), it
was observed that:
• for a 7-day UCS of 2 MPa (Category 2 material), the 28-day laboratory flexural modulus was about 4,500
MPa; and
• for a 7-day UCS of 3 MPa (Category 1 material), the 28-day laboratory flexural modulus was about 8,500
MPa.
However, lower design moduli were adopted because the laboratory moduli were not being achieved in the
field and premature pavement distress had been observed on occasions.
A relationship better fitting current QDMR practice has been calculated as follows:
For Category 1 materials and Category 2 Type A materials:
E flex = 420 UCS2.26 (F.5)
For Category 2 Type B materials only the modulus of 2,000 MPa for a 7-day UCS of 3 MPa was used;
therefore there was no need for a prediction equation.
The draft text of Chapter 6 of the 2001 Austroads Guide has been changed to include eqn (F.4). However,
given the considerable scatter in the relationship between UCS and modulus and QDMR practice, it is
recommended that the Reference Group consider:
• whether the Guide should include a relationship, and if so:
• simply adopt a relationship based on QDMR practice, viz. eqn (F.5);
• alternatively, undertake a detailed review of the data concerning this relationship.
The Reference Group should also consider adopting the Indirect Tensile Modulus rather than the UCS as the
specified material parameter.
AUSTROADS 2004
— 3.73 —
REFERENCES
Alderson, A.J. (1999). Summary of VicRoads Research into Cement-Treated Materials. ARRB Transport
Research Ltd, Contract Report RC 90216.
Litwinowicz, A. (1986). Characterisation of Cement Stabilised Crushed Rock. M. Eng. Thesis, University of
Queensland.
AUSTROADS 2004
— 3.74 —
APPENDIX G
DEVELOPMENT OF RELATIONSHIP TO ADJUST ITT MODULUS

FROM TEST LOADING RATE TO OPERATING SPEED IN-
SERVICE
A procedure was required in the 2004 Austroads Pavement Design Guide to adjust the indirect tensile test
(ITT) asphalt moduli from the test value obtained at a rise time of 40 ms to an in-service value at a design
vehicle speed.
As the ITT uses neither a step-shaped pulse nor a sinusoidal pulse, the following relationships – currently
used in the Austroads Guide to predict asphalt modulus using the Shell nomographs – are not appropriate:
loading time (t, seconds) = 1/Speed (V, km/h) (G.1)
frequency (f, Hz) = Speed (V, km/h)/ 2π (G.2)
The following procedure was used to develop the relationship adopted in the 2004 Austroads Guide:
• Using VicRoads results at 25°C of flexural modulus (10 Hz) and ITT on a wide range of approved
Victorian asphalt mixes, it was calculated that the ratio of the ITT modulus at 40 ms to the flexural
modulus at 10 Hz to was 1.33. These results are summarised in Table G.1.
• Using the following relationship for the variation in flexural modulus with test frequency of a typical
Victorian Class 320 binder dense-graded asphalt (Jameson and Hopman 2000), it was calculated that a
frequency of loading in the flexural beam test of 20 Hz would be required to increase the modulus to the
same modulus as the ITT with a rise time of 40 ms.
Ef
log = 0.365(log(f ) − 1) (G.3)
E10
• Using eqn (G.3), the sinusoidal frequencies for a range of vehicle speeds were calculated and the ratios of
the moduli to the values at 20 Hz calculated. Note that, as the flexural modulus at 20 Hz was assumed to
be equivalent to an ITT modulus at 40 ms, these moduli ratios were equivalent to the ratios of the
modulus at the design traffic speed to the ITT modulus at 40 ms. A regression analysis of this data
yielded the following relationship:
Modulus at Speed V
= 0.17 V 0.365 (G.4)
ITT Modulus
This relationship is plotted in Figure G.1.
AUSTROADS 2004
— 3.75 —
REFERENCE
Jameson, G.W. and Hopman, P.C (2000). Austroads Pavement Design Guide Chapter 6: Development of
Relationship between Laboratory Loading Rates and Traffic Speed. APRG Document 00/16(DA).
AUSTROADS 2004
— 3.76 —
Table G.1
VicRoads Flexural and ITT Modulus Data at 25°C
Mix Flex air Modulus Adjustment Adjusted Flex ITT air Modulus Adjustment ITT ITT/Flex
Type 10Hz voids Adj to 7% AV Modulus (MPa) voids Adj to 7% AV Adj to 7% AV
14H320 3497 5.9 0.93 3242 4857 7.8 1.06 5151
2043 6.7 0.98 2000 3205 8.5 1.12 3590
4446 6.5 0.97 4293 3271 6.6 0.97 3180
Mean 3178 Mean 3974 1.25
14N170 1826 6 0.93 1704 2664 6.8 0.99 2626

4170 6.3 0.95 3971 2895 7.1 1.01 2916
Mean 2838 Mean 2771 0.98
14V320 2672 7.6 1.04 2792 3551 9.1 1.18 4178

2792 4178 1.50
20T320 2922 6.1 0.94 2746 3585 8.7 1.14 4080

1866 6 0.93 1742 3038 7.2 1.01 3082
4021 7.2 1.01 4079 3799 8.1 1.09 4123
Mean 2855 Mean 3762 1.32
20T600 3960 4.8 0.86 3422 7545 8.8 1.15 8658

2096 6.1 0.94 1969 3374 7.8 1.06 3578
4750 7.5 1.04 4926 4134 7 1.00 4134
Mean 3439 Mean 5457 1.59
Mean Ratio of ITT/Flex modulus 1.33
AUSTROADS 2004
— 3.77 —
1.0
0.9
0.8
0.365
y = 0.1714x
0.7
Ev / E ITT(40ms) 0.6
0.5
0.4
0.3
0.2
0 10 20 30 40 50 60 70 80 90 100
Design Speed (km/h)
Figure G.1: Relationship to estimate modulus a design speed from

ITT modulus using 40 ms rise time
AUSTROADS 2004
— 3.78 —
APPENDIX H

FROM MEASUREMENT TEMPERATURE TO WMAPT
In Section 6.4 of the 2004 Guide, the following relationship is provided to adjust ITT moduli from the
standard test temperature of 25°C to the operating temperature in-service (WMAPT):
Modulus at WMAPT
= exp( −0.08(WMAPT − 25) (H.1)
Modulus at 25o C
This relationship is only applicable for WMAPT values between 10°C and 42°C; all listed WMAPTs in the
2004 Austroads Guide are within this range.
The origins of this relationship are as follows:
• Asphalt moduli were back-calculated form Falling Weight Deflectometer (FWD) deflections measured on
a full-depth asphalt pavement as part of the Mulgrave ALF trial (Jameson, Sharp and Vertessy 1992).
The results are given in Figure H.1.
10000
Asphalt Stiffness (MPa)
1000
12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Asphalt Temperature (Deg. C)
Figure H.1: Variation in back-calculated moduli with temperature based on analysis

of Mulgrave ALF trial data (Jameson, Sharp, Vertessy 1992)
• Based on the Figure H.1 values, and assuming the asphalt moduli asymptote to about 1,000 MPa for
temperature exceeding 40°C, the moduli given in Table H.1 were adopted in the development of the
Austroads mechanistic overlay design procedures (Potter et al. 1994).
• Using the relative moduli in Table H.1, eqn (H.1) was derived using regression analysis.
Figure H.2 compares the temperature correction adjustment obtained using eqn (H.1) with temperature
adjustment factors obtained using the Shell nomograph for: PI = –0.7, T800pen = 58, and loading times of 0.1
seconds (10 km/h), 0.0167 seconds (60 km/h) and 0.00833 seconds (120 km/h).
It is apparent that the 2004 Austroads Guide relationship is similar to that obtained using the Shell
nomographs for other than low traffic speeds.
AUSTROADS 2004
— 3.79 —
Table H.1
Estimated Asphalt Moduli Used to Develop
Austroads Overlay Design Procedures
Asphalt Asphalt Moduli (MPa) Moduli Relative to

Temperature (°C) 25°C
12.5 7400 2.56
15.0 6350 2.20
17.5 5300 1.83
20.0 4420 1.53
22.5 3550 1.23
25 2890 1.00
27.5 2225 0.77
30.0 1850 0.64
32.5 1475 0.51
35.0 1310 0.45
37.5 1150 0.40
42.5 1000 0.35
2.8
2.6
2.4
2.2
Moduli relative to moduli at 25C
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
15 20 25 30 35 40
Asphalt temperature (C)
Shell 60km/h Shell 10km/h Austroads 2001 Shell 120km/h
Figure H.2: Comparison of variations of asphalt moduli with temperature
AUSTROADS 2004
— 3.80 —
REFERENCES
Jameson, G.W., Sharp, K.G. and Vertessy, N.J. (1992). Full-Depth Pavement Fatigue Under Accelerated
Loading: The Mulgrave (Victoria) ALF Trial, 1989/1991. ARRB TR Research Report, ARR No. 224.
Potter, D.W., Jameson, G.W., Vuong, B.T., Moffatt, M.A., Yeo, R., Makarov, A., and Armstrong, P.W.
(1994). The Development of the Australian Mechanistic Approach to Overlay Design. Proc. 17th ARRB
Conf., 17(2), pp. 265-97.
AUSTROADS 2004
— 3.81 —
APPENDIX I

FROM TEST SPECIMEN AIR VOIDS TO IN-SERVICE AIR VOIDS
As the air voids of mixes tested in the ITT test may be different from the characteristics of the pavement in-
service, it was necessary in the Austroads Guide to adjust the measured ITT modulus to values applicable to
in-service voids.
Oliver et al. (1995) reported the variation in ITT moduli with air voids of two Class 320 dense-graded
asphalt mixes.
Using the variation in the average ITT modulus with air voids, the following equation was obtained:
Modulus at In − Service Air Voids (21 - AVIn-Service )
= (I.1)
Modulus at Test Air Voids (21 − AVtest )
Eqn (I.1) is plotted in Figure I.1 for test air voids of 5%, together with the adjustment equation provided in
VicRoads Technical Bulletin No. 37 (TB 37). The TB 37 relationship was obtained by regression analysis of
moduli calculated using the Shell nomographs. It is apparent there is reasonable agreement between the two
methods provided that the test air voids are within 2-3% of the in-service air voids.
1.3
Modulus at In Service AV / Modulus at Test AV
1.2
1.1
1.0
0.9
0.8
0.7
2 3 4 5 6 7 8
In Service Air Voids (%)
Oliver et al 1995 VicRoads TB 37
Figure I.1: Comparison of variations of asphalt moduli with air voids
AUSTROADS 2004
— 3.82 —
REFERENCES
Oliver, J.W.H., Alderson, A.J., Tredrea, P.F. and Karim, M.R. (1995). Results of the Laboratory Program
Associated With the ALF Asphalt Deformation Trial. ARRB TR Research Report, ARR No. 272.
VicRoads (1993). VicRoads Guide to Pavement Design. Technical Bulletin 37. VicRoads, Kew, Victoria,
Australia.
AUSTROADS 2004
— 3.83 —
APPENDIX J
DEVELOPMENT OF PRESUMPTIVE TRAFFIC LOAD

DISTRIBUTIONS
As discussed in Section 4.4.2 of this report, the traffic load distribution (TLD) data originally obtained by
Koniditsiotis (1996) was analysed to obtain presumptive TLDs for the draft 2001 Guide. This Appendix
summarises the procedure used to derive the presumptive TLD in the 2004 Guide.
Given the limited resources available for this additional development work, the data was analysed to only
assess whether presumptive TLDs for each road functional class were appropriate rather than repeating the
more detailed clustering analysis previously undertaken by Koniditsiotis (1998).
To assess the consistency of the TLDs of each functional class, the calculated Standard Axle Repetitions of
cemented materials fatigue damage (SARc) per axle group (AG) were calculated and compared. For each
Road Functional Class the frequency of occurrence of SARc/AG and various statistical parameters are plotted
in Figures J.1-J.7 and summarised in Table J.1.
Table J.1
Comparison of statistics of Traffic Load Distributions
of Various Road Classes
Road Number of Mean Percentage of Axle Group Type Mean and 95%
Functional Sites1 Single Axle Single Axle Tandem Triaxle
Confidence
Class Single Wheel Limits of
Dual Wheels Axle SARc/AG
1 45 33.5 7.8 33.1 25.0 5.5 ± 1.1

2 29 35.5 12.6 30.3 20.4 7.2 ± 2.0
3 4 36.0 10.5 30.1 22.5 8.5 ± 8.5
6 18 39.1 22.7 24.3 12.2 20.9 ± 10.0
7 11 38.3 26.7 23.6 8.7 53.1 ± 45.3
7 (reduced) 5 34.1 18.2 31.7 12.4 11.8 ± 6.7
8 3 49.1 45.9 4.4 0.0 4.4 ± 6.4
1. Number of sites after deletion of outliers
Comparing the spread of SAR/AG for each road functional class it was observed that:
• As expected, the three urban road classes (Classes 6, 7 and 8) had lower percentages of tandem and triaxle
groups and higher percentage of dual axles with dual wheels than the rural road classes. The Class 6 and
7 urban road classes also had higher SARc/AG than the rural road distributions.
• The Class 7 SARc/AG results were highly variable considering the sample size and were unexpectedly
high compared to the more heavily-trafficked Class 6 urban roads.
• As expected, Class 8 roads – their main function being to provide access to abutting property – had very
low percentages of tandem axles and triaxles and a significantly lower SARc/AG than more heavily-
trafficked urban roads.
• As there was considerable overlap of the SARc/AG of the three rural road classes (Classes 1, 2 and 3), it
was concluded that their results could be pooled to derive a presumptive “Rural” TLD.
AUSTROADS 2004
— 3.84 —
Class 1
12
ALL DATA EXCLUDING

10 OUTLIER
Mean
Mean = 6.0 Mean = 5.5
Std Dev = 4.9 Std Dev = 3.6
n = 46 n = 45
8 95% min = 4.5 95% min = 4.4
95% max = 7.4 95% max = 6.6
Frequency
0
0 5 10 15 20 25 30
SARs/AG (Cemented Materials)
Figure J.1
Class 2
12
ALL DATA Excluding Outliers

Mean = 15.1 Mean = 7.2
10 Std Dev = 28.2 Std Dev = 5.4
n = 32 n = 29
Mean
95% min = 4.9 95% min = 5.2

95% max = 25.2 95% max = 9.3
8
Frequency
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Figure J.2
AUSTROADS 2004
— 3.85 —
Class 3
2
All Data
Mean = 15.0
Std Dev = 15.4
n=5
Mean
95% min = 0
95% max = 34
Frequency
Excluding Outlier
Mean = 8.5
1 Std Dev = 5.7
n=4
95% min = 0
95% max = 17.6
0
0 5 10 15 20 25 30 35 40 45 50
Figure J.3
Class 6
3
ALL DATA EXCLUDING

OUTLIERS
Mean = 43.6
Std Dev = 80.9 Mean = 20.9
n = 20 Std Dev = 20.3
95% min = 5.8 n = 18
Mean
2 95% max = 81.5 95% min =10.8

95% max = 31.0
Frequency
0
0 50 100 150 200 250 300 350 400
Figure J.4
AUSTROADS 2004
— 3.86 —
Class 7
3
ALL DATA
Mean = 53.1
Std Dev = 65.6
Mean
n = 11
95% min = 7.9
95% max = 98.4
2
Frequency
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180
Figure J.5
Class 8
2
All Data Excluding Outlier

Mean = 12.7 Mean = 4.4
Std Dev = 16.8 Std Dev = 2.6
Mean
n=4 n= 3
t =3.18 t =4.3
95% min = 0 95% min = 0
95% max = 39.5 95% max = 10.8
Frequency
0
0 5 10 15 20 25 30 35 40
Figure J.6
AUSTROADS 2004
— 3.87 —
These results were discussed at the February 2001 RG meeting. The results for the Class 7 roads were
reviewed and it was decided to delete the results of six Class 7 Victorian sites from the analysis. The 2001
Guide RG requested the revised data be analysed to develop presumptive TLDs.
As the remaining five Class 7 sites only represented data from two roads, the data was considered
insufficient to recommend a separate TLD for Class 7 roads. Hence, the Class 6 and Class 7 (reduced set)
data was pooled. For each TLD, the percentage occurrence of each axle group type and each axle group load
was calculated. The average of these percentages for all Class 6 roads and the Class 7 reduced data set was
then calculated. The resulting “Urban” TLD is given in Table J.2.
As mentioned above, considerable overlap of the distributions of SARc/AG was observed for the three rural
road classes (Classes 1, 2 and 3). Hence the results were pooled to derive a presumptive “Rural” TLD (Table
J.3).
Generally Class 8 roads would have design traffic loading less than 105 ESA. These roads are designed in
accordance with A Guide to the Design of New Pavements for Light Traffic (APRG 21) rather than the
Austroads Pavement Design Guide. The average Class 8 TLD is given in Table J.4.
Based on this analysis it was decided that:
• Table J.2 be included in the draft 2001 Guide as the presumptive of “urban” traffic load distribution; and
• Table J.3 be included in the draft 2001 Guide as the presumptive of “rural” traffic load distribution.
The 2004 RG reviewed re-examined the data at the December 2002 meeting. The results for the Class 7 road
were reviewed and it was decided to delete the results from the calculation of the presumptive axle
distribution due to very limited amount of data on which they were based. Hence only the Class 6 data was
used to generate the presumptive urban distribution. For each TLD the percentage occurrence of each axle
group type and each axle group load was calculated. The average of these percentages for all Class 6 roads
was then calculated. The resulting “Urban” TLD is given in Table J.5.
In terms of the Table J.3 presumptive “Rural” distribution, some 2004 RG members were concerned that the
distribution resulted in less damage in fatigue of cemented materials and concrete pavements than Table J.5
presumptive ‘Urban” TLD. This was due to the small number of very high single axle single tyre axle loads
in the “Urban” TLD. Consequently in June 2003, RTA NSW used their WIM data on rural roads to produce
a presumptive “Rural” TLD (Table J.6) more consistent with the presumptive (Table J.5).
Tables J.5 and J.6 were adopted in the 2004 Austroads Pavement Design Guide.
AUSTROADS 2004
— 3.88 —
REFERENCES
Koniditsiotis, C. (1996). Update of Traffic Design Chapter in the Austroads Pavement Design Guide – Status
Report. ARRB TR Working Document WD TI96/024.
Koniditsiotis, C. (1998). Update of the Austroads Pavement Design Guide – Traffic Design Chapter. Final
Draft of New Traffic Design Chapter. ARRB TR Working Document WD R98/030.
AUSTROADS 2004
— 3.89 —
Table J.2: Presumptive Traffic Load Distribution for Rural Roads
Axle Group Axle Group Load

Load SAST SADT TAST TADT TRDT
(kN) % % % % %
10 0.08 1.25 0.01 0.15 0.03

20 3.97 7.30 0.09 0.56 0.15 NHVAGs = 2.9
30 8.20 17.41 0.29 0.97 0.52
40 11.34 18.44 0.97 1.97 1.57
50 31.51 17.80 4.05 3.73 3.30
60 33.87 12.70 10.30 4.98 3.83
70 9.85 9.57 15.68 5.66 4.32
80 1.10 6.78 15.92 5.85 4.41 Material Type
90 0.08 4.45 14.34 5.87 4.16 Component SARs/ AG
100 0.01 2.74 15.11 6.53 4.15
110 0.96 10.94 6.38 3.43 Granular
120 0.37 6.29 6.89 3.34 Pavements
130 0.15 3.20 7.44 3.52 (4th Power) 1.0
140 0.05 1.71 8.31 3.92
150 0.01 0.62 10.37 4.97 Asphalt 1.2
160 0.00 0.22 9.13 5.23 (5th Power)
170 0.10 7.05 6.26
180 0.09 4.18 7.44 Subgrade 1.7
190 0.05 2.08 8.40 (7th Power)
200 0.03 1.07 9.00
210 0.44 6.63 Cemented mater 7.1
220 0.21 4.76 (12th Power)
230 0.11 2.95
240 0.05 1.66
250 0.03 0.97
260 0.01 0.44
270 0.00 0.24
280 0.00 0.13
290 0.00 0.08
300 0.07
310 0.05
320 0.02
330 0.02
340 0.01
350 0.00
360 0.00
370 0.00
380
390
400
Total 100.00 100.00 100.00 100.00 100.00
Proportion of 0.344 0.098 0.007 0.320 0.230

Each Axle Group
AUSTROADS 2004
— 3.90 —
Table J.3: Presumptive Traffic Load Distribution for Urban Roads
Axle Group Axle Group Type

Load
(kN) SAST SADT TAST TADT TRDT
% % % % %
10 0.26 2.18 0.10 0.10 0.00
20 13.53 10.23 0.96 0.68 0.11
30 18.02 20.67 1.26 1.41 0.25 NHVAGs = 2.6
40 19.99 17.99 1.33 3.76 1.04
50 25.74 13.42 4.52 7.73 4.92
60 17.11 8.30 13.66 10.32 9.44
70 4.37 6.27 17.95 10.22 9.79
80 0.77 7.68 17.36 8.56 8.62 Material Type
90 0.12 6.37 13.23 6.76 6.53 Component SARs/ AG
100 0.06 3.58 9.92 5.34 4.35
110 0.01 1.68 9.77 4.38 3.12 Granular
120 0.01 0.92 4.66 4.15 2.70 Pavements
130 0.01 0.44 2.33 4.29 2.47 (4th Power) 0.80
140 0.01 0.19 1.19 4.71 2.65
150 0.05 0.87 6.15 3.09 Asphalt 0.96
160 0.02 0.33 5.71 3.42 (5th Power)
170 0.00 0.31 4.97 3.81
180 0.01 0.13 3.40 4.94 Subgrade 1.7
190 0.00 0.10 2.64 6.24 (7th Power)
200 0.03 1.70 7.22
210 1.19 5.24 Cemented materials 19
220 0.83 3.70 (12th Power)
230 0.42 2.02
240 0.21 1.45
250 0.16 0.90
260 0.08 0.60
270 0.08 0.62
280 0.01 0.31
290 0.01 0.20
300 0.00 0.16
310 0.01 0.04
320 0.00 0.03
330 0.03
340 0.02
350 0.00
360 0.00
370 0.00
380 0.00
390
400
Total 100.00 100.00 100.00 100.00 100.00
Proportion of 0.380 0.217 0.022 0.259 0.122

Each Axle Group
AUSTROADS 2004
— 3.91 —
Table J.4: Representative Distribution of Loads on Axle Groups for Use on

Lightly-Trafficked Urban Roads
p g y
Load
(kN) SAST SADT TADT TRDT
% % % %
10 14.90 20.47 0.00 0.00
20 31.46 24.74 1.460 0.00
30 17.82 9.148 8.957 4.790
40 10.95 7.730 4.380 0.00
50 5.115 6.623 1.202 0.00
60 12.44 8.403 1.804 0.00 NHVAGs = 2.0
70 5.515 3.022 1.202 0.00
80 1.319 4.191 7.644 0.00 ESA/HVAG = 0.6
90 0.2903 5.508 11.77 4.790
100 0.1452 4.177 12.11 9.481 ESA/HV = 1.2
110 0.0484 3.080 12.37 0.00
120 2.200 14.69 4.790
130 0.5231 5.926 14.27 Damage Type SARs/ ESA
140 0.1916 4.724 18.96
150 1.460 9.481 Asphalt Fatigue 1.3
160 1.804 4.790 (5th Power)
170 0.8589 0.00
180 1.460 0.00 Rutting and Shape Loss 2.3
190 2.663 0.00 (7th Power)
200 1.202 9.481
210 0.6012 0.00 Cemented Materials 21
220 1.202 0.00 Fatigue (12th Power)
230 0.0000 4.790
240 0.2577 0.00
250 0.2577 0.00
260 0.00
270 4.790
280 0.00
290 4.790
300 0.00
310 0.00
320 4.790
330
340
350
360
370
380
390
400
Total 100.00 100.00 100.00 100.00
Proportion of 0.4909 0.4594 0.0445 0.0028

Each Axle Group
AUSTROADS 2004
— 3.92 —

Urban Roads

Load
(kN) SAST SADT TAST TADT TRDT
% % % % %
10 0.27 3.47 0.03 0.15 0.01
20 7.83 8.70 0.24 0.58 0.13
30 15.46 23.46 0.28 0.62 0.33
40 15.71 21.93 0.58 1.98 1.32
50 29.94 16.80 2.89 6.50 4.17
60 23.29 9.61 10.27 9.51 7.42 NHVAGs = 2.5
70 6.50 6.50 16.81 10.94 9.78
80 0.79 4.62 16.61 9.77 8.34 ESA/HVAG = 0.7
90 0.11 2.97 15.95 7.61 6.15
100 0.04 1.39 14.42 7.24 5.03 ESA/HV = 1.8
110 0.02 0.41 9.77 6.27 3.70
120 0.02 0.12 5.90 5.95 3.30
130 0.02 0.02 2.94 5.88 3.15 Damage Type SARs/ ESA
140 1.54 6.53 3.36
150 0.84 8.03 4.01 Asphalt Fatigue 1.1
160 0.43 5.72 4.11 (5th Power)
170 0.23 3.55 4.82
180 0.14 1.86 6.10 Rutting and Shape Loss 1.6
190 0.07 0.85 7.73 (7th Power)
200 0.06 0.33 8.43
230 0.02 0.78
240 0.25
250 0.09
260 0.01
270
280
290
300
310
320
330
340
350
360
370
380
390
400
Total 100.00 100.00 100.00 100.00 100.00
Proportion of 0.393 0.191 0.009 0.259 0.148

Each Axle Group
AUSTROADS 2004
— 3.93 —

Rural Roads
p
Axle Group Axle Group Load
Load SAST SADT TAST TADT TRDT
(kN) % % % % %
10 0.50 3.03 0.01 0.38 0.01
20 6.74 4.97 0.08 0.57 0.12
30 9.24 6.86 1.29 2.36 1.88
40 15.30 8.48 1.99 4.01 3.86
50 28.74 10.23 5.12 4.26 5.21 NHVAGs = 2.8
60 32.19 14.16 11.48 5.17 4.36
70 5.65 13.58 15.24 4.87 4.08 ESA/HVAG = 0.9
80 1.21 13.29 16.19 4.85 3.84
90 0.25 11.17 14.59 5.30 3.79 ESA/HV = 2.5
100 0.08 7.84 15.37 6.24 3.84
110 0.06 4.13 11.13 9.31 4.12
120 0.03 2.00 6.40 9.17 3.57 Damage Type SARs/ ESA
130 0.18 0.63 10.55 6.69
140 0.07 0.22 12.92 7.89 Asphalt Fatigue 1.1
150 0.10 10.35 8.36 (5th Power)
160 0.09 3.77 8.13
170 0.05 2.57 7.38 Rutting and Shape Loss 1.6
180 0.02 1.60 6.08 (7th Power)
190 0.87 4.98
220 0.13 1.93
230 0.06 1.14
240 0.03 0.63
250 0.35
260 0.22
270 0.13
280 0.08
290 0.04
300 0.03
310 0.01
320
330
340
350
360
370
380
390
400
Total 99.99 99.99 100.00 100.00 100.00
Proportion of 0.344 0.098 0.007 0.320 0.231

Each Axle Group
AUSTROADS 2004
— 3.94 —

Part 4: 2004 Guide Procedures for Design of Rigid Pavements
George Vorobieff
December 2003
SUMMARY
This report records the work undertaken in the revision of Chapter 9 (Design of New Rigid Pavements) of
the 1992 edition of Pavement Design – A Guide to the Structural Design of Road Pavements, published by
Austroads in 1992.
The 2004 revision of Chapter 9 of the Guide includes improvements to the thickness design procedures
based on ten years of concrete pavement design and construction experience in Australia, primarily in NSW.
The Guide provides guidance on the structural design of pavements, specifically the establishment of
appropriate thicknesses of the pavement layers to withstand the design traffic at a specified project design
reliability using pavement materials which meet specified mechanical properties. The information in
Chapter 9 the 2004 edition of the Guide is for new pavements with a traffic volume exceeding 1 million
heavy vehicle axle groups (HVAGs). The revision to this Chapter of the Guide also took account, where
appropriate, of revisions to other Chapters of the Guide.
AUSTROADS 2004
— 4.i —
TABLE OF CONTENTS
Page
1. GENERAL ....................................................................................................................................... 4-1

2. PAVEMENT TYPES ...................................................................................................................... 4-2
2.1 Base Concrete......................................................................................................................... 4-2
2.2 Subbase Types ....................................................................................................................... 4-3
3. THICKNESS DETERMINATION .................................................................................................... 4-4

3.1 Subgrade and Subbase Stiffness............................................................................................ 4-4
3.2 Design Traffic .......................................................................................................................... 4-4
3.3 Concrete Shoulders................................................................................................................. 4-6
3.4 Load Safety Factors ................................................................................................................ 4-6
4. BASE THICKNESS DESIGN PROCEDURE .................................................................................. 4-7
4.1 General.................................................................................................................................... 4-7
4.2 Dowel and Tie Bars ................................................................................................................. 4-9
5. REINFORCEMENT DESIGN PROCEDURES ............................................................................. 4-11

REFERENCES ........................................................................................................................................ 4-12
AUSTROADS 2004
— 4.ii —
TABLES
Page
Table 1: Load safety factors (LSF) for Rigid Pavement Types 4-5
Table 2: Recommended load safety factors (LSF) for Roundabouts 4-5
FIGURES
Page
Figure 1 Typical longitudinal section of plain concrete pavement (PCP).
Steel reinforced concrete is sometimes used for PCP 4-2
Figure 2 Typical longitudinal-section of jointed reinforced concrete pavement (JRCP) 4-2
Figure 3 Typical longitudinal-section of continuously reinforced concrete pavement 4-2
Figure 4 Typical cross-section of dowelled plain concrete pavement (PCP-D).
Steel fibre reinforced concrete is sometimes used for PCP-D 4-2
Figure 5 A select material zone under the subbase assists in providing uniform
support to the pavement, especially in cut and fill transition areas 4-3
Figure 6 Concrete base thickness versus traffic volume for a PCP supported
on a 150 mm lean mix concrete subbase with and without concrete shoulders 4-4
Figure 7 Plan of the four most common heavy vehicle axle groups in Australia 4-4
Figure 8 Plan of the six heavy vehicle axle groups used in the Guide 4-4
Figure 9 Allowable repetitions to failure for fatigue from SAST and SADT axle group
loads – the two vertical dashed lines represent the 65 kN limit for each axle
group for a LSF = 1.2 4-8
AUSTROADS 2004
— 4.iii —
1. GENERAL
The 2004 revision of Chapter 9 of the Austroads Pavement Design Guide includes improvements to the
thickness design procedures based on ten years of concrete pavement design and construction experience in
Australia, primarily in NSW. The Guide provides guidance on the structural design of pavements,
specifically the establishment of appropriate thicknesses of the pavement layers to withstand the design
traffic at a specified project design reliability using pavement materials which meet specified mechanical
properties. The information in Chapter 9 the 2004 edition of the Guide is for new pavements with a traffic
volume exceeding 1 million heavy vehicle axle groups (HVAGs). The revision to this Chapter of the Guide
also took account, where appropriate, of revisions to other Chapters of the Guide.
The Chapter in the Austroads Pavement Design Guide addressing the design of new rigid pavements has
traditionally dealt with thickness design but not structural design detailing. The Working Group updating
Chapter 9 decided to ensure that the revision adopted a similar approach to the Chapter in the Guide dealing
with the design of flexible pavements. The revised Chapter provides guidance on the:
• minimum subbase type and thickness;
• thickness determination of the basic four pavement types;
• load safety factors according to pavement type and project reliability;
• minimum base thickness for increasing traffic volume;
• quantity of longitudinal and transverse reinforcement for CRCP;
• provision of tie bars in longitudinal joints;
• provision of dowels in dowelled transverse joints; and
• requirements for base anchors.
For guidance on the structural detailing of joints and surface details the designer should seek information
from various technical publications, such as the Concrete Pavement Manual – Design and Construction
Roads (RTA 1996), Concrete Roundabout Pavements – A Guide to their Design and Construction (RTA
2003a), Treatment of Moisture in Cuttings (RTA 1999) and RTA standard drawings (RTA 2003b).
As detailed in this report, the major change to Chapter 9 was the replacement of the nomographs and tables
listing design coefficients with a series of algorithms so that the trial-and-error thickness procedure can be
carried out using a computer program or spreadsheet macros. This will enable the designer to develop a
more efficient design procedure and reduce the potential for human errors associated with the use of the
nomographs and coefficients.
Several terms were revised in this Chapter as follows:
• ‘commercial vehicle axle groups (CVAG)’ has been changed to ‘heavy vehicle axle groups (HVAG)’;
• ‘effective subgrade strength’ has been changed to ‘effective subgrade stiffness’;
• the term ‘wearing surface’ has been introduced to describe the use of thin layers of asphalt above the
concrete base; and
• abbreviations for the heavy vehicle axle groups have been introduced.
For a full list of current pavement definitions refer to the Glossary of Terms (Standards Australia 2002).
AUSTROADS 2004
— 4-1 —
2. PAVEMENT TYPES
2.1 Base Concrete
The Working Group considered that the information about various base types in the 1992 Guide was too
brief if designers were to clearly distinguish between the appropriate base types. This has been improved
without detailing a comprehensive list of advantages and disadvantages for each type, and Figs 1 to 3 show
basic details of the four base types outlined in the Guide. Whilst not specifically in the current Guide, Fig. 4
shows a PCP with dowelled joints. More detailed information about the use of the four base types can be
found in the various Road Authority manuals and technical directions.
Induced &
3.5 to 4.5 m
sealed joints
Figure 1: Typical longitudinal-section of plain concrete pavement (PCP)

Steel fibre reinforced concrete is sometimes used for PCP
Steel mesh Dowels Sawn &

8 to 15m sealed
joints
Figure 2: Typical longitudinal-section of jointed reinforced concrete pavement (JRCP)
0.5 to 1.5 m
Steel bars
Figure 3: Typical longitudinal-section of continuously reinforced concrete pavement
Dowels Induced &

up to 5 m
sealed joints
Figure 4: Typical cross-section of dowelled plain concrete pavement (PCP-D)

Steel fibre reinforced concrete is sometimes used for PCP-D
AUSTROADS 2004
— 4.2 —
2.2 Subbase Types

The 2004 edition of the Guide emphasises the use of lean mix concrete (LMC) subbases and this
recommendation is detailed in Section 9.1 of the Guide. However, the text in the Guide does not preclude
the use of other types of subbases.
The RTA NSW experience with the use of LMC subbase has been very successful at preventing erosion
distress at the subbase level to the extent that little to no erosion distress has been detected for concrete
pavements with this type of subbase.
In addition, as the experience in NSW has shown that the construction of subbase layers 100 mm thick has
been difficult to achieve by contractors, the Working Group considered that the minimum subbase thickness
should be 125 mm for LMC and bound materials. This is now reflected in Table 9.1 and Figure 9.1 of the
Guide.
AUSTROADS 2004
— 4.3 —
3. THICKNESS DETERMINATION
3.1 Subgrade and Subbase Stiffness
The design thickness of the base is a function of the traffic loading, material properties and the cumulative
stiffness of the subbase and subgrade. Many concrete pavement failures have been attributed to uneven
support conditions that may occur over large underground services, culverts or at the transition of the cut and
fill zones. Hence, the text in this section brings the designer’s attention to the fact that the concrete base
layer should be longitudinally and laterally uniformly supported by the subbase and subgrade layers.
At cut and fill transitions with predominantly rock at the cut zone, constructing a pavement onto the rock is a
sound approach and it will provide substantial support to the traffic loading. Unfortunately, the strong
support offered by rock formations cannot be carried though the fill region, and it makes sense that the
selected material from the cut zone is used to form a continuous “bedding” for the concrete subbase as shown
in Figure 5. This selected material may be stabilised with lime or cement to assist with the long-term
stability of the material and ensure a strong working platform for the delivery of the concrete to the concrete
slipformer.
Rock
Select material
Figure 5: A select material zone under the subbase assists in providing uniform
support to the pavement, especially in cut and fill transition areas
This section of the Chapter emphasises the need for long-term uniform and volumetric stable material12 near
the top of the subgrade. Where several subgrade layers are used to achieve this condition, Chapter 5
provides some guidance on how the designer could derived the design subgrade strength at the top of the
subgrade. It is known from experience that, in order to minimise the impact of vertical movement on the
subbase from potentially expansive subgrades, a minimum layer thickness of 600 mm is required for the
select material zone.
3.2 Design Traffic

It is well known that the thickness of concrete pavements is sensitive to traffic loading. A reduction of
10 mm in thickness can represent some 24 million HVAGs of pavement traffic life, as shown in Fig. 6.
With Governments coming under increasing pressure from various industry groups to increase the legal axle
load limits, the estimation of traffic volume and loading over the next ten to forty years has become
increasingly difficult for designers. The new representative statistical data of traffic loading in Chapter 7 of
the Guide represents recent information obtained from weigh-in-motion (WIM) sites and replaces the
previous rural and urban traffic distribution tables generated in the late 1970s (refer to Appendix I of the
1992 edition). This data should provide more reliable estimates of traffic loading distribution for designers.
12
A volumetric stable material may be defined as a material which will not significantly change volume, e.g. swell, shrink with
changes in moisture content.
AUSTROADS 2004
— 4.4 —
330
310
without shoulder
Base Thickness (mm)
290
270
250
230
with shoulder
210
190
170
150
1.0E+06 2.1E+07 4.1E+07 6.1E+07 8.1E+07
HVAGs
Figure 6: Concrete base thickness versus traffic volume for a PCP supported on a 150 mm
lean mix concrete subbase with and without concrete shoulders
The current Guide also introduces two new axle groups and a revision to the abbreviations used to describe
the heavy vehicle axle groups, as shown in Figures 7 and 8.
Single axle Single axle Dual axles Three axles

Single wheels Dual wheels Dual wheels Dual wheels
(SS) (SD) Tandem Triaxle
(TAD) (TRD)
Figure 7: Plan of the four most common heavy vehicle axle groups in Australia
Single axle Single axle Two axles Two axles

Single wheels Dual wheels Single wheels Dual wheels
(SAST) (SADT) (TAST) Tandem
(TADT)
Three axles Four axles
Dual wheels Dual wheels
Triaxle Quad
(TRDT) (QADT)
Figure 8: Plan of the six heavy vehicle axle groups used in the Guide
AUSTROADS 2004
— 4.5 —
3.3 Concrete Shoulders

The structural definition of concrete shoulders was reviewed by the Working Group and no new material was
found which would justify changing the requirements for concrete shoulders. US studies (e.g. Heinrichs et
al. 1988) confirmed Australian experience that the provision of concrete shoulders resulted in a significant
improvement in the performance of the concrete pavement.
It is noted in the Guide that the designer should also be aware that the width of the concrete shoulder should
take into consideration safety requirements to allow drivers to park clear of the fast-moving traffic in the
outer lanes of freeways.
3.4 Load Safety Factors

One of the criticisms of the definition of load safety factors (LSF) given in the 1992 Guide was the loose
description assigned to the factors. The Working Group considered that traffic volume would also assist the
designer to select an appropriate LSF. For instance, LSF = 1.2 may be used for major freeways and other
multi-lane projects carrying uninterrupted flows of high volumes of heavy vehicles (i.e. greater than 5 x 107
HVAGs) and LSF = 1.1 may be used for freeways, highways and arterial road projects with moderate
volumes of heavy vehicles (i.e. greater than 1 x 106 HVAGs).
The purpose of the introduction of project reliability in Chapter 2 of the Guide was to improve the
designation of traffic safety factors for both flexible and rigid pavements. For the design of flexible
pavements the material fatigue life is adjusted according to the assigned project design reliability (refer to
Table 2.1 of the Guide) whereas, for rigid pavements, the assigned project design reliability assigns a load
safety factor according to pavement type (Table 1). This approach addresses the criticisms levelled at the
1992 Guide and provides the designer with the ability to determine equivalent flexible or rigid pavement
configurations for a given project design reliability.
Table 1
Load Safety Factors (LSF) for Rigid Pavement Types
Pavement Type Project Design Reliability

80% 85% 90% 95% 97.5%
PCP 1.15 1.15 1.20 1.30 1.35
Dowelled & CRCP 1.05 1.05 1.10 1.20 1.25
In addition, the Chapter provides LSF values for roundabouts based on the work by Ayton (RTA 2003) and
these are listed in Table 2.
Table 2
Recommended Load Safety Factors (LSF) for Roundabouts
LSF Selected for Adjusted LSF for
Design Roundabouts
1.0 1.3
1.1 1.4
1.2 1.5
AUSTROADS 2004
— 4.6 —
4. BASE THICKNESS DESIGN PROCEDURE

4.1 General
The overall design procedure in the 2004 Guide fundamentally follows the approach adopted in the 1992
edition of the Guide. However, the following changes have been made:
• the nomographs for the determination of the allowable load repetitions have been replaced by algorithms
as discussed earlier;
• the ‘equivalent stress’ and ‘erosion factor’ tables have been replaced with algorithms;
• average concrete strength has been adjusted to characteristic concrete strength;
• two new axle groups – tandem axle with single wheels and quad axles with dual wheels – have been
introduced; and
• the recommended minimum base thickness has been increased.
Vorobieff (1996) published the algorithms that formed the nomographs as Figures 9.4 to 9.6 in the 1992
edition of the Guide. These algorithms were metric conversions from equations presented by Packard and
Tayabji (1985) and correspondence with the Portland Cement Association (Packard, 1994).
The fatigue distress equation is based on the determination of the allowable load repetitions (Nf) for a given
axle load is:
⎡ 0.9719 − S r ⎤
log (Nf) = ⎢ when Sr > 0.55 (1)
⎣ 0.0828 ⎥⎦
3.268
⎡ 4.258 ⎤
Nf = ⎢ ⎥ when 0.45 ≤ Sr ≤ 0.55 (2)
⎣ S r − 0.4325⎦
0.94
Se ⎡ P.LSF ⎤
where Sr = ⎢ ⎥ (3)
0.944 f ' cf ⎣ 4.45F1 ⎦
Se = equivalent stress (MPa);
f’cf = design characteristic flexural strength (MPa);
P = axle group load (kN);
LSF = load safety factor; and
F1 = load adjustment due to axle group
= 9 for single axle with single wheel (referred to as SAST axle group)
= 18 for single axle with dual wheel (referred to as SADT axle group)
= 18 for tandem axle with single wheel (referred to as TAST axle group)
= 36 for tandem axle with dual wheel (referred to as TADT axle group)
= 54 for triaxle with dual wheel (referred to as TRDT axle group)
= 72 for quad axle with dual wheel (referred to as QADT axle group).
Eqns (1) and (2), being dimensionless, required no conversion from imperial to metric units. However, the
stress ratio was derived from the following equation:
0.94
⎛ Se ⎞ ⎛ P ⎞
Sr = ⎜ ⎟ x⎜ ⎟ (4)
⎝ Mr ⎠ ⎝ 18 ⎠
AUSTROADS 2004
— 4.7 —
where Mr = modulus of rupture of the concrete base slab, and

P = axle load (kips).
In eqn (4), both parts of the right-hand side of the equation are dimensionless if the correct imperial or metric
units are input. The value 18 for the denominator is for single axle groups, and this becomes 36 for tandem
axle groups. Therefore, Vorobieff modified eqn (4) to eqn (3) using the following changes:
• Mr became the design characteristic flexural strength (f’cf) and 0.944 was included to allow for the
difference between the average and characteristic flexural strength specified in Australia and the
difference between the PCA’s assumption in the design model and Australian experience of concrete
strength gain beyond 28-days.
• The value of 18 was replaced by F1 to allow various axle groups to be included in the equation.
• LSF was included to allow for the load safety factors in the calculations.
• A value of 4.45 was used to convert kips to kN and allow for four wheels for the single axle group in eqn
(4).
• F1 values for TAST and QADT were determined through numerical analysis by comparing stresses and
cumulative distress for similar axle configurations.
Whilst it is obvious, the Guide notes that values of Nf are infinite or commonly referred to as unlimited when
Sr is less than 0.45.
In terms of the erosion distress mode, the work by Packard and Tayabji (1985) found that the allowable load
repetitions (Ne) for a given axle load were:
0.103
⎡⎛ P.L ⎞ 2 10 F3 ⎤
log (F2 Ne) = 14.524 - 6.777 ⎢⎜⎜ SF
⎟⎟ − 9 .0 ⎥ (5)
⎢⎣⎝ 4.45 F1 ⎠ 41.354 ⎥⎦
where P, LSF and F1 are similar to previous definitions with the exception of F1 for
QUAD axle groups;
F2 = adjustment for slab edge effects;
= 0.06 for base with no shoulder;
= 0.94 for base with shoulder; and
F3 = erosion factor.
Only one area of eqn (5) required conversion from imperial to metric units, namely the P2 section. As noted
for the fatigue distress equations, Vorobieff introduced the load safety factor and the conversion from kips to
kN. In using eqn (5) the value derived in the square brackets ([ … ]) must be greater than zero to ensure that
a real number (i.e. versus imaginary number) is derived. Therefore, it is noted in the Guide that values less
than zero (or where the deflection of the base over the transverse joint is low) there are unlimited allowable
repetitions for the axle load within the axle load group.
Whilst these algorithms provided some acceleration in the design calculations they still relied on the
numerous equivalent stresses and erosion factors in Tables 9.2 and 9.3 of the 1992 Guide. In 1997, ARRB
Transport Research were commissioned (Moffatt 1997) to replace the equivalent stress and erosion factors in
Tables 9.2 and 9.3 to make way for the use of algorithms in the current Guide. At first glance the algorithm
selected by the researchers looks complex. However, it was derived after a series of trials with various
alternatives to ensure that the difference between the coefficients in the previous Tables did not exceed the
values generated by the algorithm by 5%.
AUSTROADS 2004
— 4.8 —
The Austroads Working Group reviewed the application of the concrete flexural strength in the use of the
tables in the 1992 Guide and the use of the algorithms. It was concluded that the original nomographs used
required the designer to use the average concrete flexural strength rather than the characteristic concrete
strength (PCA 1984). Due to the practice in Australia and New Zealand of specifying the characteristic
concrete strength, the Working Group therefore made an allowance of 15% in the fatigue distress algorithm.
This adjustment appears within the denominator of 0.944 in the definition of Sr in eqn (9.2) of the Guide.
The 2004 Guide also introduces two new axle groups and these are noted in the revisions in the traffic
Chapter. Neither the existing guide nor the PCA design method provide any guidance on the selection of an
F1 value (see eqn (3)) for the new axle groups, and the Working Group had little research data available to it
to enable new coefficients for the axle groups to be generated. Therefore, a conservative approach was taken
to the selection of the coefficients for the axle groups based on experience with using the design procedure.
Over the last ten years, experience has shown that heavy-duty pavements in Australia are being subject to
numerous overloaded trucks with axle loads exceeding the legal limit. Unusual forms of pavement distress
are also being observed that appear to be mainly related to environmental loading. Whilst the Working
Group was satisfied with the thickness design model, there have been instances where relatively thin plain
concrete pavements have been built using this model with heavy traffic loading – defined by the RTA NSW
as HVAG exceeding 1 x 107 during the first 20 years of operation (also refer to Section 3.4). As the
thickness design procedure does not directly consider environmental loading parameters, and it would be
inappropriate to increase the load safety factor, the Working Group developed, using their experience, a
series of recommended minimum design base thicknesses for various base types, and for specific ranges of
traffic volumes as listed in Table 9.7 of the Guide.
One of the interesting issues related to the use of the algorithms in the 1992 Guide was the implication of the
65 kN load per tyre limit set in Table 9.1 (Step 10). Numerous attempts by the author to correspond with the
author of the PCA (1984) method failed to identify why such a limit existed. It is noted that, with some
WIM data, the load per tyre value can be exceeded for single axle load groups, especially when high LSF (ie.
> 1.2) values are used.
Upon examination of the behaviour of the curves, as shown in Figure 9, the algorithms do not “misbehave”
above 65 kN. As raising or eliminating the limit was not addressed by the Working Group, it was decided to
continue to caution designers when (4.5PLsf/F1) exceeded 65 kN. It is emphasised that the design procedure
was prepared assuming normal vehicular traffic loadings on multi-lane roads. It is not prudent, therefore, to
use the design procedure for industrial pavements where the spacing between wheels and axles, and axle
loads, are typically much higher.
1.0E+10
Number of Allowable
SAST SADT
Repetitions
1.0E+00
0 100 200 300 400
Axle Group
d
Figure 9: Allowable repetitions to failure for fatigue from SAST and SADT axle group loads – the
two vertical dashed lines represent the 65 kN limit for each axle group for a LSF = 1.2
AUSTROADS 2004
— 4.9 —
4.2 Dowel and Tie Bars

The determination of the size and quantity of dowel and tie bar has not changed in the revision to the
Chapter.
Australian reinforcing bar companies have new high strength steels (i.e. 500 MPa) and most of the work and
experience of using dowels and tie bars is based on the traditional 400Y grade bar. No research could be
located by the Working Group which showed that the increase in yield strength improved tie bar
performance or crack spacing generation for CRCP. One of the key design elements of tie bars is their pull-
out strength, and should this be a weak link and the tie bar fails in its operation, the pavement either side of
the contraction joint is likely to shows signs of fatigue distress.
AUSTROADS 2004
— 4.10 —
5. REINFORCEMENT DESIGN PROCEDURES

Only minor changes to the design procedures have been made in the 2004 Guide. A summary of these
changes follows.
• A clearer definition of L in eqn (9.5) of the Guide as the distance to untied joints or edges of the base.
• An update of the reference for the new Australian Standard for reinforcing bars.
• Updated guidance on the indicative values of the coefficient of friction based on current construction
practices using various curing compounds and debonding materials (Ayton and Haber 1997).
• The minimum reinforcement requirement for CRC has been increased from 0.60 to 0.65 based on
experience with regard to the desired crack spacing.
• The optimum crack spacing is now between 0.5 to 2.5 m to accommodate the higher concrete strengths
specified and the provision of the interlayer between base and subbase.
As noted in the previous section, the equations presented in the Guide were modelled against a quantitative
assessment of the performance of these pavements subjected to environment and traffic loadings. The use of
high strength steel may result in less longitudinal bars per metre width of pavement, but may not produce a
suitable crack pattern conducive to long-term performance.
AUSTROADS 2004
— 4.11 —
REFERENCES
Austroads (2004). Pavement Design – A Guide to the Structural Design of Road Pavements. Pub. No.
AP-G17/04. Austroads, Sydney.
Ayton, G.P. and Haber, E.W. (1997). Curing and Interlayer Debonding. Proc. 6th Int. Conf. on Concrete
Pavement Design and Materials for High Performance, Indianapolis, USA, November.
Heinrichs, K. et al. (1988). Rigid Pavement Design and Analysis.. Federal Highway Administration (USA)
Report No. FHWA-RD-88-068.
Moffatt, M.A. (1997). Regression Equations For Determination of Equivalent Stresses and Erosion Factors
for Rigid Pavement Design. Austroads Pavement Research Group Report No. RE7110, September.
Packard, R. and Tayabji, S. (1985). New PCA Thickness Design Procedure For Concrete Highway and
Street Pavements. Proc. 3rd Int. Conf. on Concrete Pavement Design, Purdue University, USA.
Packard, R. (1994). Private Communication to George Vorobieff. Cement & Concrete Association of
Australia, Sydney, 14th June.
PCA (1984). Thickness Design for Concrete Highway and Street Pavements. Portland Cement Association
(USA), EBA 209.01P.
RTA NSW (1996). Concrete Pavement Manual – Design and Construction. Roads and Traffic Authority,
NSW Edition 2,5 June.
RTA NSW (1999). Treatment of Moisture in Cuttings. RTA Technical Direction 99/7.
RTA NSW (2003a). Concrete Roundabout Pavements – A Guide to their Design and Construction. Roads
and Traffic Authority, NSW Edition 3, June.
RTA, NSW (2003b). Standard Drawings List. Volume 1: Continuously Reinforced Concrete Pavements
(Drawing MD.R84.CC.A). Volume 2: Plain Concrete Pavements (Drawing MD.R83.CP.A). Volume 3:
Jointed Reinforced Concrete Pavements (Drawing MD.R83.CJ.A). Volume 4: Steel Fibre Reinforced
Concrete Pavements (Drawing MD.R83.CF.A).
Standards Australia (2002). Road and Traffic Engineering – Glossary of Terms. AS 1348.
Vorobieff, G. (1996). Rigid Pavement Design Using Spreadsheets. Proc. ‘Roads 96’ Conference, ARRB
Transport Research, Christchurch, NZ.
AUSTROADS 2004
— 4.12 —
INFORMATION RETRIEVAL
Austroads (2004), Technical Basis of Austroads Pavement Design Guide, Sydney,

A4, 221pp, AP-T33/04
______________________________________________________________
KEYWORDS:
Pavement design, flexible pavement, rigid pavement, mechanistic design, pavement

materials, design moduli, subgrade, anisotropic, performance relationships, asphalt
fatigue, cemented materials fatigue, subgrade strain, design reliability, axle load
ABSTRACT:
The Austroads publication “Pavement Design – A Guide to the Structural Design of
Road Pavements” is intended to assist those required to plan and design new
pavements. It was originally produced in 1987 as a result of review of the NAASRA
“Interim Guide to Pavement Thickness Design” (1979). In 1992, the Guide was
revised to include an updated procedure for the design of rigid pavements and also
relevant portions of Chapter 6 (Pavement Materials) and Chapter 7 (Design Traffic).
An essential element in the use of the Guide is a thorough understanding of the origins
of the design procedures, their scope and limitations. Accordingly, this report contains
the following four technical reports detailing the technical basis of both the 1992 and
2004 editions of the Guide:
• Part 1: 1992 Guide procedures for the design of flexible pavements
• Part 2: 1992 Guide procedures for the rigid of flexible pavements
• Part 3: 2004 Guide procedures for the design of flexible pavements
• Part 4: 2004 Guide procedures for the design of rigid pavements

AP-T33-04 Austroads 1987

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

AP-T33-04 Austroads 1987

Uploaded by

Copyright:

Available Formats

AP-T33

TECHNICAL BASIS OF AUSTROADS

© Austroads Inc. 2004

National Library of Australia

Technical Basis of Austroads Pavement Design Guide

Austroads Project No. PUB.PT.C.007

Austroads Publication No. AP-T33/04

Published by Austroads Incorporated

Technical Basis of Austroads Pavement Design Guide

2. SETTING THE SCENE ................................................................................................................... 1-2

3. GRANULAR PAVEMENTS WITH THIN BITUMINOUS SURFACINGS ...................................... 1-11

4. DEVELOPMENT OF THE MECHANISTIC PROCEDURE........................................................... 1-19

5. DEVELOPMENT OF OVERLAY DESIGN PROCEDURE ........................................................... 1-36

REFERENCES CITED IN THE TEXT ..................................................................................................... 1-43

2. SETTING THE SCENE

2.1.1 Format of the IGPTD

Procedure Method Rationale

2.1.2 Release of the IGPTD

2.2 Establishment of a Working Group to Revise the IGPTD

Proposed Format Extent of Person(s) Responsible

2.2.1 A Guide – not a Manual

3. GRANULAR PAVEMENTS WITH THIN BITUMINOUS SURFACINGS

3.1.1 Quality of Pavement Material and its Cover Requirements

3.1.2 Concluding Comments

3.2 Terminal Condition

3.2.1 Modification of Terminal Condition

3.2.2 Implicit Model for Roughness Progression

4. DEVELOPMENT OF THE MECHANISTIC PROCEDURE

reaches a value of 1, where:

4.2 Elastic Characterisation

4.2.1 Isotropic or Anisotropic Characterisation?

It was found that similar effects on FEM-estimated responses were obtained by :

4.2.2 Values for Poisson’s Ratios

4.2.3 Relationship between Subgrade Modulus and CBR

4.2.4 Typical Modulus Values

4.2.5 Stress Regimes for Triaxial Testing of Granular Materials

4.2.6 Sub-Layering of Granular Materials and Assignment of Moduli

4.2.7 Modulus of Top Sub-Layer of Granular Material

4.2.8 Modulus of Granular Material Overlying a Cemented Layer

4.2.9 Relationships between Modulus and UCS for Cemented Materials.

4.2.10 Characterisation of Cracked Cemented Materials

4.3 Performance Relationships

4.3.1 Subgrade Strain Criterion

• Hence, the relationship can be re-written as follows:

4.3.2 Fatigue Cracking of Cemented Material

The development of these two sets of relationships can be summarised as follows.

with values of K as follows:

4.3.3 Fatigue Relationship for Asphalt

4.4 Design Traffic

4.4.2 Derivation of “Standard Axle (or Traffic) Factors”

Derivation of Tables 8.3(a) and 8.3(b)

Axle Group Type

State/Territory Single Axle Single Axle Tandem Axle Triple Axle

Derivation of “Standard Axle Factors”

AG2: Single Axle, Dual Tyres

AG4: Triple Axle, Dual Tyres

Distress Mode Number of Standard Axles per Axle

4.4.3 Cumulative Growth Factor for Estimation of Design Traffic

4.5 Incorporation of Location-Specific Temperature Regime

5. DEVELOPMENT OF OVERLAY DESIGN PROCEDURE

5.1 Design Deflection Curves

5.1.1 Adoption of IGPTD Curve 1

5.1.2 Adoption of IGPTD Curve 4