AP-T33-04 Austroads 1987 [PDF]

Citation preview

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:



Technical Basis of Austroads Pavement Design Guide 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: [email protected] 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.



Technical Basis of Austroads Pavement Design Guide



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



Technical Basis of Austroads Pavement Design Guide 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 4.2 4.3 4.4 4.5



5.



Broad Issues ......................................................................................................................... 1-19 Elastic Characterisation ........................................................................................................ 1-22 Performance Relationships ................................................................................................... 1-27 Design Traffic ........................................................................................................................ 1-33 Incorporation of Location-Specific Temperature Regime ..................................................... 1-35



DEVELOPMENT OF OVERLAY DESIGN PROCEDURE ........................................................... 1-36 5.1 5.2 5.3 5.4 5.5



Design Deflection Curves...................................................................................................... 1-36 The Curvature Function ........................................................................................................ 1-38 Temperature Correction for Deflection and Curvature.......................................................... 1-38 Reduction in Deflection Parameters due to Overlay Placement .......................................... 1-41 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Although the document records 1979 as its publication date (the Foreword is dated April 1979), it was mid1980 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 quasielastic 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Table 1 Methodology and Rationale for Procedures in IGPTD Procedure F1



F2



F3A



F3B



Method



Rationale



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 condition and estimated traffic



California Highways Department (Porter 1938) TRRL, SRA experience



Modify T for desired terminal roughness



Scala (1977); (AASHO Road Test data)



Check minimum base requirement



SRA experience



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)



Check minimum base requirement



SRA experience



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, Poisson’s Ratio) for each layer



Laboratory, field investigations



Estimate deflection under a Standard Axle



Elastic layer theory



Check estimated deflection against tolerable deflection and, if less, reduce thickness of bound layers



SRA experience; Scala (1973)



Check minimum base requirement



SRA experience



Determine thickness of cover (T) using F1 or F2 (as appropriate)



See procedures F1, F2



Determine individual layer thicknesses (ti) from suggested layer equivalencies (ai) and relationship: T = Σ ai ti Check minimum base requirement F4



SRA experience



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 deflection, dTOL



SRA experience; Scala (1973)



If dREP < dTOL, adopt nominal overlay



–



If dREP > dTOL, design structural overlay



–



For granular overlay, determine thickness on the basis of a 10% deflection reduction for each 45 mm of overlay material



SRA experience



For AC overlay, determine thickness using recommended percentage reduction in deflection per 25 mm for the specific climate zone



SRA experience



AUSTROADS 2004 — 1.5 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



• 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



Extent of Revision



Proposed Format



Person(s) Responsible



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 (ii) subgrade CBR (iii) subgrade modulus (iv) pavement materials (v) traffic (vi) evaluation for rehabilitation



total re-write total re-write total re-write major revision total re-write total re-write



R. Payze R. Payze R. Payze Working Group R. Payze R. Payze



total re-write major revision total re-write



Working Group Working Group Working Group



total re-write



A. Leask supported by E. Haber, J. Cruickshank, R. Elliott



major revision



Working Group



total re-write



P. Lowe



7. Design Procedures for Flexible Pavements F1 F2 F3 8.



Design Procedure for Rigid Pavements



9.



Design Procedures for Overlays



10. Observations



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 CBRTraffic-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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 subbases. 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 inplace 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 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 —



(1)



Technical Basis of Austroads Pavement Design Guide: Part 1



One further piece of information from the IGPTD (Section 2.1.2) which is relevant here is that the CBRthickness-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) T =t.[2R1/(R2 - R1)]0.25



Now, from equation 1, 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: i.e.



log10 (NM/120) = log10 (N/120) * [2/(R2/R1-1)]0.25 log10 NM = [2/(R2/R1-1)]0.25 log10N + [1 - [2/(R2/R1-1)]0.25] * log10120



(3)



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



10 9 8 7



5



10



6 6



10



Ratio of Roughness 5 to Initial Roughness 4



7



10



3 2 1 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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: = 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 Ni 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 stressdependent. 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



• 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 (inplace) 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 CIRCLYestimated 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 oversimplification, 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 straindependency). 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 sublayers, 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



•



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. interpretation of the terms, gives:



Simplifying the relationship, and generalising the



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 Material (MPa) 2,000 5,000 10,000



Value of K 280 200 150



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: Modulus of Cemented Material (MPa) 2,000 3,500 5,000 10,000 15,000



Value of K 440 350 310 260 240



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



AUSTROADS 2004 — 1.30 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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-atbreak 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 Material (MPa) 2000 5000 ≥ 10000



K1



K2



259 244 152



19.9 14.5 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



with values of K as follows: Modulus of Cemented Material (MPa) 2,000 5,000 15,000



Value of K 440 310 240



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 dual-tyred single axle dual-tyred tandem axle dual-tyred triaxle



5.4 t (53.0 kN) 8.2 t (80.4 kN) 13.6 t (133.4 kN) 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 dualtyred 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Table 3 Distribution of Axle Group Type by State/Territory Axle Group Type State/Territory



Single Axle Single Tyres



Single Axle Dual Tyres



Tandem Axle Dual Tyres



Triple Axle 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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



NSW VIC



8



9



10



0.5



5.3



11.8



25.2



50.3



6.4



0.5



0.1



5.7



16.5



17.2



24.5



30.2



5.5



0.2



0.1



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



11



12



13



14



15



16



17



18



19



20



21



22



23



24



25



26



27



28



29



13



14



15



16



17



18



19



20



21



22



23



24



25



26



27



28



29



0.0



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



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



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



0.1



AUSTROADS 2004 — 1.36 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Table 4 (con’t) AG3: Tandem Axle, Dual Tyres Load (t)



1



2



NSW



3



4



5



6



7



8



9



10



11



12



13



14



15



16



17



18



19



20



21



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



22



23



0.0



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



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



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



WA SA 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



24



25



26



27



28



29



27



28



29



NT TOT



AG4: Triple Axle, Dual Tyres Load (t) NSW



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



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



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 0.3



VIC QLD



0.1



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



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



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



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



17.6



11.8



11.8



5.9



5.9



5.9



11.8



5.9



11.8



1.9



WA SA TAS



0.5



ACT



5.9



0.2 0.2



0.2



0.5 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



AUSTROADS 2004 — 1.37 —



0.1



0.1



Technical Basis of Austroads Pavement Design Guide: Part 1



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 (7.14)



Asphalt (5)



Cemented (18)



ESA (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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 highwayspeed 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Figure 3: Predictive ability of deflection parameters (Anderson, 1984a)



AUSTROADS 2004 — 1.40 —



Technical Basis of Austroads Pavement Design Guide: Part 1



• 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 db T



= deflection after overlay, = deflection before overlay, and = 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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. Country Roads Board, Victoria (1960). The Design of Flexible Pavements. Technical Bulletin No. 21. Country Roads Board, Melbourne, Victoria.



AUSTROADS 2004 — 1.43 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Country Roads Board, Victoria (1969). The Design of Flexible Pavements. Technical Bulletin No. 26. Country Roads Board, Melbourne, Victoria. 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 1751, 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 TimeDependent 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 MS11.) 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 Category



Two-Way CV/day



Two-Way CV/20 years



One-Way 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Figure A.7: 1979 NAASRA design chart



AUSTROADS 2004 — 1.55 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 NAASRA Engineer-Secretary to ARRB, May. 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 subbase 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) Let



Soil Moisture Relations Rule 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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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



(c)



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



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



12



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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



W



26-0.195W



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



(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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 Surface Deflection (mm)



Curve 1 1.0



0.8



0.6 Curve 2 0.4



0.2



0 105



2



4



6



8 106



2 4 6 8 107 Design Traffic (ESAs)



2



4



6



8 108



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



"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 1975 CRB Design Deflection (mm)



1979 NAASRA Design Deflection (mm)



Design Traffic One-way ESAs/20 years



1965 Scala Allowable Deflection (mm)



1969 NAASRA Design Deflection (mm)



0-50



8 x 104



1.78



1.78



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



1.80



Design deflections for asphalt-surfaced pavements.



Design deflection (mm)



*



1969 CRB Design Deflection (mm)



Design Traffic Two-way End-of-Life CV/day



2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 1E4



1E5 1E6 1E7 Design Traffic (ESAs)



NAASRA 1979



CRB 1975



NAASRA 1969



Figure C3: Comparison of design deflections



AUSTROADS 2004 — 1.63 —



1E8



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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: where



d t CBR



log(d) = –(0.88+0.0165t) – 0.5 log(CBR) Benkelman Beam maximum deflection (inches), thickness of granular material (inches), and subgrade CBR.



= = =



(C1)



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.



Mean Deflection (mm)



1.8 1.6 1.4 1.2 1 0.8 0.6 2



4



6



200 mm



8 10 12 Subgrade CBR (%) 300 mm



400 mm



14



16



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



where



d t CBR



log(d) = – 0.34(1+0.1t) – 0.7 log(CBR) Benkelman Beam maximum deflection (inches), thickness of granular material (inches), and subgrade CBR



= = =



(C2)



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)



"High" Deflection (mm)



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 2.6 2.2 1.8 1.4 1 0.6 2



4



6



200 mm



8 10 12 Subgrade CBR (%) 300 mm



400 mm



14



16



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



"High" Deflection (mm)



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 Design Traffic Loading (ESAs) 200 mm 500 mm



300 mm AUSTROADS



1E7



400 mm



"High" Deflection (mm)



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 Design Traffic Loading (ESAs) CBR=2 CBR=10



CBR=3 CBR=15



1E7



CBR=5 AUSTROADS



Figure C7: Relationship between "high" Benkelman Beam deflection and design traffic loading for various subgrade CBRs



AUSTROADS 2004 — 1.67 —



Technical Basis of Austroads Pavement Design Guide: Part 1



Mean Deflection (mm)



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.



1.8 1.6 1.4 1.2 1 0.8 0.6 1E4



1E5 1E6 Design Traffic Loading (ESA) 200 mm 500 mm



300 mm AUSTROADS



1E7



400 mm



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 1



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 NAASRA Engineer-Secretary to ARRB, May. 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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



Technical Basis of Austroads Pavement Design Guide 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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



TABLES Page



Table 1: Table 2: Table 3:



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 Faulting criteria for major roads (Packard 1977) ........................................................... 2-14 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 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12



Figure 13 Figure 14 Figure 15 Figure 16 Figure 17



Figure 18



Figure 19



Typical longitudinal section of plain concrete pavement (PCP)....................................... 2-2 Typical longitudinal section of jointed reinforced concrete pavement (JRCP) ................ 2-2 Typical longitudinal section of continuously reinforced concrete pavement (CRCP)............................................................................................................ 2-2 Typical cross-section of Dowelled Plain Concrete Pavement (PCP-D)........................... 2-2 Design model assumption of concrete strength gain with age (PCA 1984) .................... 2-3 Fatigue relationship adopted in design model.. ............................................................... 2-4 Plan of the four most common commercial vehicle axle groups in Australia .................. 2-5 Axle group load distributions for a typical urban site from Appendix I of the Guide...................................................................................................................... 2-5 Axle group load distributions for a typical rural site from Appendix I of the Guide...................................................................................................................... 2-6 The base will curl on a daily cycle provided the base and subbase are debonded ................................................................................................................... 2-8 Equivalent edge stress factor depends on percentage of trucks at pavement edge (PCA 1984) ............................................................................................ 2-9 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 A joint with 100% joint effectiveness will deflect equally across the joint ...................... 2-10 A joint with 0% joint effectiveness will have zero deflection across the joint................. 2-10 Position of axle load group for the critical base flexural stress...................................... 2-12 Position of axle load group for the critical base deflections........................................... 2-13 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 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 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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 groundsupported 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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 & sealed joints



3.5 to 4.5m



Figure 1: Typical longitudinal section of plain concrete pavement (PCP).



Steel mesh



8 to 15 m



Dowels



Sawn & 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 & sealed joints



up to 5 m



Figure 4: Typical cross-section of Dowelled Plain Concrete Pavement (PCP-D)



AUSTROADS 2004 — 2.2 —



Technical Basis of Austroads Pavement Design Guide: Part 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)



Technical Basis of Austroads Pavement Design Guide: Part 2



130% 120%



90-day design



110%



28-day design



100% 90%



10



28 days



100



90



1 yr



1000



3



10000



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.



Stress Ratio



0.9 0.8 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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 Single wheels Dual wheels (SS) (SD)



Dual axles Dual wheels Tandem (TAD)



Three axles Dual wheels Triaxle (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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



45



40



35



30 SS (%)



SD (%)



TAD (%)



TRD (%)



Distribution (%)



25



20



15



10



5



0 10



60



110



160 210 Axle Group Load (kN)



260



310



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



5



0 10



60



110



160 210 Axle Group Load (kN)



260



310



Figure 9: Axle group load distributions for a typical rural site from Appendix I of the Guide



AUSTROADS 2004 — 2.6 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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.



∆L = d/2



Wheel Load



∆u = d/2



Figure 13: A joint with 100% joint effectiveness will deflect equally across the joint



∆L = d



Wheel Load



∆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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



Table 2 Faulting Criteria for Major Roads (Packard 1977) Fault Index (PI)



Average Faulting (mm)



Rating



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



330 310



without shoulder



Base Thickness



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



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 Group Load (kN)



Legal Limit



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



Axle Group



6.4



(kN)



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).



Base Thickness (mm)



330 310



without shoulder



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



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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.



250



Erosion related distress



200



Fatigue related distress



150



Concrete Base Thickness (mm)



300



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.



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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 —



Technical Basis of Austroads Pavement Design Guide: Part 2



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. Austroads (1992). Pavement Design – A Guide to the Structural Design of Road Pavements. Austroads, 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Technical Basis of Austroads Pavement Design Guide 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



TABLES Page



Table 1: Table 2: Table 3: Table 4: Table 5: Table 6: Table 7: Table 8: Table 9: Table 10:



Reliability Traffic Multipliers for Asphalt Surfaced Cemented Treated Base Pavements 2004 Guide Suggested Reliability Factors for Cemented Materials Fatigue Reliability Traffic Multipliers for Full Depth Asphalt Pavements Desired Project Reliability 2004 Guide Suggested Reliability Factors (RF) for Asphalt Fatigue Axle Group Loads Which Cause Equal Damage as a Standard Axle............... Distress Mode Strain Dependency Characteristics of Presumptive Traffic Load Distributions (TLDs)................... Average Project Reliabilities of Outlasting Design Traffic for Flexible Pavements Pavement Type Traffic Multipliers to Outlast the Design Traffic Traffic Multipliers to Outlast the Design Period



3-13 3-14 3-16 3-16 3-17 3-18 3-19 3-21 3-22 3-22



FIGURES Page



Figure 1 Figure 2 Figure 3



Standard Axle loading used in Austroads Pavement Design Model for Mechanistic Design of Flexible pavements .......................................................................... 3-4 Brown’s (1973) loading time equation and 1/V approximation ................................................ 3-10 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 Mr David Potter Mr Geoff Jameson Mr John Hodgkinson Mr Alan Pearson



Department of Main Roads, NSW (Convenor) Australian Road Research Board Country Roads Board, Victoria Cement &Concrete Association of Australia (Federal Office) 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 Mr Mike Butcher Mr Allan Jones Dr Peter Kadar Mr John Lancaster Mr Geoff Jameson Mr Kieran Sharp



ARRB Transport Research (Convenor) Department of Road Transport, South Australia Queensland Department of Main Roads CERTS International Pioneer Road Services ARRB Transport Research 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Also established in late 1993 was an Austroads Working Group on Reliability of Pavement Design. Membership of this Working Group comprised: Mr David Potter Mr Geoff Jameson Mr Geoff Ayton Mr Ian Rickards Mr George Vorobieff



ARRB TR (Convenor) ARRB TR Roads &Traffic Authority, NSW Pioneer Road Services 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 Mr Kieran Sharp Mr Geoff Jameson Mr Chris Mathias Mr Frank Butkus Mr Allan Jones Ms Narelle Dobson Mr Peter Tamsett Mr David Dash Mr Geoff Ayton Mr Greg Arnold Mr David Alabaster Mr Andrew Papacostas Mr Ralph Rallings Mr David Mangan Mr George Vorobieff Mr David Chatwin Mr Scott Matthews Mr Ian Rickards Mr David Potter Mr Geoff Youdale



VicRoads (Convenor) ARRB Transport Research ARRB Transport Research Transport SA Main Roads Western Australia Queensland Department of Main Roads Queensland Department of Main Roads Roads & Traffic Authority, NSW Roads & Traffic Authority, NSW Roads & Traffic Authority, NSW Transit New Zealand Transit New Zealand (replacing Greg Arnold) VicRoads Pitt & Sherry (representing Department of Infrastructure and Environmental Services), Tasmania Australian Asphalt Pavement Association (AAPA) Australian Stabilisation Industry Association C&CAA C&CAA (replacing David Chatwin) Pioneer Road Services Consultant 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 Mr Geoff Jameson Mr Chris Mathias Mr Frank Butkus Mr Allan Jones Mr Peter Tamsett Mr Andrew Papacostas Mr David Alabaster



VicRoads (Convenor) ARRB Transport Research Transport SA Main Roads Western Australia Queensland Department of Main Roads Roads & Traffic Authority, NSW VicRoads Transit New Zealand AUSTROADS 2004 — 3.2 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Mr David Mangan Mr George Vorobieff Mr Scott Matthews Mr Ian Rickards Mr Geoff Youdale



Australian Asphalt Pavement Association (AAPA) Australian Stabilisation Industry Association C&CAA Pioneer Road Services Consultant



Throughout this report this group is referred to as the 2004 Guide Reference Group (RG).



AUSTROADS 2004 — 3.3 —



Technical Basis of Austroads Pavement Design Guide: Part 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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. following changes to the model have resulted in changes in the calculated responses:



However, the



• 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Consequently, the 2004 RG decided that the Austroads design model should be changed to require sublayering 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



•



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



•



to;



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 = asphalt modulus (MPa). EAC 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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: For cemented crushed rock (2) E = 1814 UCS0.88 + 3500 0.88 + 1100 For cemented natural gravel (3) E = 2240 UCS 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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: where



EFLEX UCS k



4.2.9



= = =



(7) EFLEX = k UCS flexural modulus of field beams at 28 days moist curing (MPa); Unconfined Compressive Strength of laboratory specimens at 28 days (MPa); and 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.



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 sublayering. 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 (8) t = 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



Loading Time (secs)



0.07



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 (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 Desired Project Reliability 80% RTMs: average reliability of design process 85% (Jameson & Moffatt , 2001)



0.81



Equivalent RTMs for an average reliability of design process 95%



0.405



85%



1.0 0.5



90%



95%



97.5%



1.3



2.0



3.0



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:



⎡ 6918(0.856 VB + 1.08) ⎤ N = RF ⎢ ⎥ 0.36 S mix µe ⎢⎣ ⎥⎦ where



5



(22)



N



=



allowable number of repetitions of the load;



µε VB Smix RF



= = = =



tensile strain produced by the load (microstrain); percentage by volume of bitumen in the asphalt (%); asphalt mix stiffness (modulus) (MPa); and reliability factor for asphalt fatigue (Table 4). Table 4 2004 Guide Suggested Reliability Factors (RF) for Asphalt Fatigue Desired Project Reliability 80%



85%



90%



95%



97.5%



2.5



2.0



1.5



1.0



0.67



AUSTROADS 2004 — 3.18 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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: ⎡ ⎤ Load on Axle Group No. of Standard axles of damage = ⎢ ⎥ ⎣ Appropriat e Loadfrom Table 1 ⎦



EXP



(23)



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 stressdependent.



AUSTROADS 2004 — 3.20 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table 6 Distress Mode Strain Dependency Distress Mode



Dependence on Strain



Asphalt Fatigue (Section 4.3.3)



⎛ k N = ⎜⎜ ⎝ µε



⎞ ⎟⎟ ⎠



5



Rutting of Unbound Granular Materials and Subgrade (Section 4.3.1)



⎛ k N = ⎜⎜ ⎝ µε



⎞ ⎟⎟ ⎠



7



Cemented Materials Fatigue (Section 4.3.2)



⎛ k N = ⎜⎜ ⎝ µε



⎞ ⎟⎟ ⎠



12



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. • Using a load damage exponent of 7, the number of Standard Axles that produce the same cumulative damage as the design traffic in terms of rutting of granular materials and subgrade is calculated (NSS). • Using a load damage exponent of 12, the number of Standard Axles that produce the same cumulative 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table 7 Characteristics of Presumptive Traffic Load Distributions (TLDs) for Urban and Rural Roads Pavement Type



Damage Type



Damage Index



Presumptive Rural TLD



Presumptive Urban TLD



Granular Pavements with Thin Bituminous Surfacings, designed using Figure 8.4



Overall Damage



ESA/HVAG



0.9 2.5



0.7 1.8



Pavement Containing One or More Bound Layers, Mechanistically Designed



Fatigue of Asphalt



SARa/ESA



1.1



1.1



Rutting and Shape Loss



SARs/ESA



1.6



1.6



Fatigue of Cemented Materials



SARc/ESA



12



12



5.2



ESA/HV



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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-beconstructed) 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 cemented material



90



Thin bituminous seal on cemented material



80



Full depth asphalt



85



Asphalt on granular material



85



Asphalt on granular material on cemented material



85



Asphalt (< 150 mm) on cemented material



85



Asphalt (≥ 150 mm) on cemented material



85



Asphalt (≥ 150 mm) on modified material



85



1



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 material



0.17



0.27



0.49



1.0



2.9



7.4



Thin bituminous seal on 300 mm or more granular material



0.27



0.38



0.58



1.0



2.3



4.4



Thin bituminous seal on granular material on cemented material



0.33



0.44



0.62



1.0



2.1



3.7



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



0.65



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 material



0.18



0.28



0.50



1.0



3.1



8.2



Thin bituminous seal on 300 mm or more granular material



0.28



0.40



0.62



1.1



2.5



4.8



Thin bituminous seal on granular material on cemented material



0.34



0.47



0.67



1.1



2.2



4.1



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



0.69



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



• 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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. 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. 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 WDR98/005 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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:



⎡ EV ⎤ R = ⎢ top of base ⎥ ⎣⎢ E V subgrade ⎦⎥



1



5



(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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



(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 (mm)



Maximum possible modulus of top layer of granular material



25 50 75 100



350 320 290 250



AUSTROADS 2004 — 3.35 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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



12000



Critical microstrain (using modified Potter subdivision approach)



Critical microstrain (using modified Potter subdivision approach)



13000 3



11000 10000 9000



4



8000 5



7000 6000 7



5000 4000



10



3000 15 20



2000 30 1000



12000 Subgrade CBR 11000 3



10000 9000 8000



4



7000 5



6000 5000 7 4000 10



3000 15



2000



20 30



1000



0



0 0



1000



2000



3000



4000



5000



6000



7000



8000



9000



10000



11000



12000



13000



14000



0



1000



2000



3000



4000



Critical microstrain (using original Potter subdivision approach)



Total granular thickness = 100 mm



7000



8000



9000



10000



11000



12000



13000



14000



12000



13000



14000



12000



13000



14000



14000 13000 Critical microstrain (using modified Potter subdivision approach)



13000 Critical microstrain (using modified Potter subdivision approach)



6000



Total granular thickness = 125 mm



14000



2 12000 11000 10000



Subgrade CBR



9000 3 8000 7000 4 6000 5



5000 4000



7



3000



10



2000



15 20 30



1000



12000 11000 2 10000 9000 Subgrade CBR



8000



3



7000 6000 4 5000 5 4000 7



3000 10 2000 30



1000



20



15



0



0 0



1000



2000



3000



4000



5000



6000



7000



8000



9000



10000



11000



12000



13000



0



14000



1000



2000



3000



4000



Total granular thickness = 150 mm 13000 Critical microstrain (using modified Potter subdivision approach)



14000



13000 12000 11000 10000



2



8000 Subgrade CBR



7000 6000



3



5000 4 4000



5



3000



7 10



2000 1000



30



6000



7000



8000



9000



10000



11000



Total granular thickness = 175 mm



14000



9000



5000



Critical microstrain (using original Potter subdivision approach)



Critical microstrain (using original Potter subdivision approach)



Critical microstrain (using modified Potter subdivision approach)



5000



Critical microstrain (using original Potter subdivision approach)



15 20



12000 11000 10000 9000 8000 2 7000 Subgrade CBR



6000



3



5000 4000



4 5



3000 7 2000



10 15 20 30



1000



0



0 0



1000



2000



3000



4000



5000



6000



7000



8000



9000



10000



11000



12000



13000



14000



0



1000



2000



Critical microstrain (using original Potter subdivision approach)



3000



4000



5000



6000



7000



8000



9000



10000



11000



Critical microstrain (using original Potter subdivision approach)



Total granular thickness = 200 mm



Total granular thickness = 225 mm



Figure B.2: Comparison between alternative Potter granular material characterisations



AUSTROADS 2004 — 3.36 —



Technical Basis of Austroads Pavement Design Guide: Part 3



9000 Critical subgrade microstrain (using modified Potter subdivision approach)



Critical subgrade microstrain (using modified Potter subdivision approach)



1800



1600



8000



1400



7000



1200



6000



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



Critical subgrade microstrain (using original Potter subdivision approach)



3000



4000



5000



6000



7000



8000



9000



Critical subgrade microstrain (using original Potter subdivision approach)



Figure B.3: Comparison between alternative Potter granular material characterisations (asphalt surfaced pavements)



Asphalt strain calculated using proposed granular subdivision method (microstrain)



1800



25 100 02 asphalt thickness (mm)



1600



granular thickness (mm)



25 150 02



subgrade CBR (%) XX YYY ZZ



1400 25 100 03



25 200 02



50 100 02



1200



50 150 02



25 150 03 25 100 04



1000



50 200 02



50 100 03



50 150 03 50 100 04 50 200 03 75 100 02 50 150 04 75 150 02 50 02 100 05 75 200 25 150 05 75 100 03 25 200 04 5050 200 0405 150 75 150 03 25 100 07 75 100 04 75 50 200 03 07 100 50 200 05 75 150 04 100 100 02 150 02 75100 100 50 07 25 200 05 100 200 02 05 75150 200150 04 05 75 100 100 03 25 150 07 100 150 03 50 100 10 75 07 20075 05 100 0704 50 200 100 200 03 100 100 75100 150150 07 04 25 100 10 50 100 150 200 10 10004 100 05 150 05 75 100 200 07 10 75 100 100 200 05 25 200 07 50 100 15 50100 200 10 100 75100 150 10 07 150 07 25 150 10 75 200 10 100 200 07 100 100 10 75 100 15 50100 150 50 2015 100 150 10 75 150 15 50 200 15 25 100 15 100 75100 100 20 75 200 15 50 150 20 100 200 10 75 150 20 15 50 200 20 200 100 150 15 25 200 10 100 10075 200 1520 100 100 20 150 200 20 25 25150 10015 20 25 150 20 25 25200 20015 20 25 200 03 25 150 04



800



600



400



200



25 100 05



0 0



100



200



300



400



500



600



700



800



900



1000



Asphalt strain calculated using current (Austroads, 1992) granular subdivision method (microstrain)



Asphalt strain calculated using proposed granular subdivision method (microstrain)



1800



1600



1400



1200



1000



SG AS



800



SG AS



600



SG AS



400 SG AS SG AS AS AS



200



SG AS SG AS SG AS SG AS SG AS SG AS SG AS SG AS SGSG ASAS SG AS AS ASAS AS AS AS AS AS AS ASAS AS AS AS AS AS AS AS AS AS AS AS AS AS AS AS ASAS AS AS AS AS ASAS AS AS AS AS AS AS Indicates that asphalt AS AS AS AS AS AS AS AS ASAS strain criterion AS AS dominates



Indicates that subgrade strain criterion dominates when using current subdivion method, and that asphalt strain dominates when using proposed subdivision method



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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).



⎡ 8511⎤ N=⎢ ⎥ ⎣ µε ⎦



7.14



(B.1)



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



2.



Granular Thicknesses (mm)



100 to 1,000 in steps of 25



Subgrade CBR (%)



2, 3, 4, 5, 7, 10, 15, 20, 30



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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)



5.



6.



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. 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 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).



⎡ 9177 ⎤ N=⎢ ⎥ ⎣ µε ⎦ 7.



7.03



(B.5)



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). (B.6) log10(N) = 27.442 – 6.888 log10( µε ) (B.7) log10(µε) = 3.945 – 0.1392 log10( N ) (B.8) log10(N) = 27.881 – 7.032 log10( µε ) Rearranging, and simplifying the denominator, yields eqn (B.9). ⎡ 9223 ⎤ N=⎢ ⎥ ⎣ µε ⎦



7 .03



where N = allowable repetitions of a full Standard Axle, and µε = maximum compressive strain at the top of the subgrade (microstrain).



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 —



(B.9)



Technical Basis of Austroads Pavement Design Guide: Part 3



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.



⎡ 9300 ⎤ N=⎢ ⎥ ⎣ µε ⎦



7



(B.10)



where N = allowable repetitions of a full Standard Axle, and µε = maximum compressive strain at the top of the subgrade (microstrain). 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.



Number of allowable repetitions of a Standard Axle



1.0E+09



1.0E+08



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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: ⎡ 9300 ⎤ N=⎢ ⎥ ⎣ µε ⎦



7



where N = allowable repetitions of a full Standard Axle, and µε = maximum compressive strain at the top of the subgrade (microstrain).



AUSTROADS 2004 — 3.41 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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.



2. 3. 4. 5.



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) Sublayer thickness t (mm) = T/n (Vertical) modulus of top sublayer is given by: EvTopGR = minimum( revised Table 6.6 entry, 2T/100 x EvSG ) Ratio of (vertical) moduli of adjacent sub-layers is given (as currently) by: R = (EvTopGR /EvSG )1/n 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 thickness (mm)



Maximum modular 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 (ETGR / ESG )max 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0



CURRENT



PROPOSED



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 (mm)



Design ESAs for Tmin (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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 Thickness (mm)



Subgrade CBR



100



2



Critical Strain determined by CIRCLY (µε) 17130



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



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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 0



Pavement Life (ESAs)



0.0E+00



0.0E+00



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table B.2.1 (continued) Results of CIRCLY analysis and life determined from Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence) Granular Thickness (mm)



Subgrade CBR



275



3



Critical Strain determined by CIRCLY (µε) 3460



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



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



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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 1.00E+03



AUSTROADS 2004 — 3.47 —



Pavement Life (ESAs)



1.0E+03



9.0E+04



1.0E+05



Technical Basis of Austroads Pavement Design Guide: Part 3



Table B.2.1 (continued) Results of CIRCLY analysis and life determined from Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence) Granular Thickness (mm)



Subgrade CBR



525



4



Critical Strain determined by CIRCLY (µε) 990.1



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



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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 5.3E+06



AUSTROADS 2004 — 3.48 —



Pavement Life (ESAs)



5.3E+06



9.0E+04



Technical Basis of Austroads Pavement Design Guide: Part 3



Table B.2.1 (continued) Results of CIRCLY analysis and life determined from Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence) Granular Thickness (mm)



Subgrade CBR



775



5



Critical Strain determined by CIRCLY (µε) 435.8



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



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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 6.9E+09



AUSTROADS 2004 — 3.49 —



Pavement Life (ESAs)



6.9E+09



7.0E+04



Technical Basis of Austroads Pavement Design Guide: Part 3



Table B.2.1 (continued) Results of CIRCLY analysis and life determined from Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence) Granular Thickness (mm)



Subgrade CBR



100



10



Critical Strain determined by CIRCLY (µε) 3426



125



10



2929



0



0.0E+00



150



10



2439



1.00E+04



1.0E+04



175



10



2018



6.00E+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



100



15



2283



1.30E+03



125



15



1954



1.20E+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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 0



— 3.50 —



0.0E+00



6.0E+04



1.7E+17



AUSTROADS 2004



Pavement Life (ESAs)



1.7E+17 1.3E+03 1.2E+04



Technical Basis of Austroads Pavement Design Guide: Part 3



Table B.2.1 (continued) Results of CIRCLY analysis and life determined from Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence) Granular Thickness (mm)



Subgrade CBR



350



15



Critical Strain determined by CIRCLY (µε) 581.1



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



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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 8.6E+08



1.00E+04



AUSTROADS 2004 — 3.51 —



Pavement Life (ESAs)



8.6E+08



1.0E+04



Technical Basis of Austroads Pavement Design Guide: Part 3



Table B.2.1 (continued) Results of CIRCLY analysis and life determined from Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence) Granular Thickness (mm)



Subgrade CBR



600



20



Critical Strain determined by CIRCLY (µε) 222.4



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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 1.4E+16



Pavement Life (ESAs)



1.4E+16



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table B.2.1 (continued) Results of CIRCLY analysis and life determined from Austroads (1992) – Figure 8.4 and APRG (1998) – Figure 13.8.2(B) (90% confidence) Granular Thickness (mm)



Subgrade CBR



850



30



Critical Strain determined by CIRCLY (µε) 87.9



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



Life from Life from APRG (1998) Austroads (1992) Figure 13.8.2(B) Figure 8.4 (ESAs) (ESAs) 1.5E+27



AUSTROADS 2004 — 3.53 —



Pavement Life (ESAs)



1.5E+27



Technical Basis of Austroads Pavement Design Guide: Part 3



APPENDIX C GRANULAR MATERIALS MODULI UNDER ASPHALT AND CEMENTED MATERIAL The following text was first published as an internal ARRB Transport Research Working Document WDR98/008 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



50 mm asphalt



350 /350



1



500 /350



1



500 /350



100 mm asphalt



1



210 /150



1



350 /250



5001/350



150 mm asphalt



2101/150



2101/150



3501/250



cemented material



2101/150



2101/150



2101/150



1



1



where material is compacted using Modified compactive effort.



AUSTROADS 2004 — 3.54 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table C.2 Suggested Vertical Modulus (MPa) of Top Sub-Layer of Granular Material (Yeo et al. 1997, Table 6.6) Overlying Material



C.2



1



2000



380 / 270



300 / 220



2401 / 180



50 mm asphalt



5001 / 350



5001 / 350



4601 / 320



3901 / 280



3101 / 230



2601 / 200



2301 / 180



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



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



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



210 / 170



210 / 170



210 / 170



1



210 / 170



1



1



210 / 170



1



1



1



280 / 210



1



1



1



250 mm asphalt



1



1



1



1



125 mm asphalt



1



1



1



1



7000



460 / 330 1



1



6000



500 / 350 1



1



5000



500 / 350 1



1



4000



500 / 350 1



1



3000



=300 mm



150



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 Modulus of Cover1 Material (MPa)



Thickness of Overlying Material



1000



2000



3000



40 mm



500



500



500



4000



480



5000 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



1. Cover material is either asphalt or cemented material or a combination of these materials.



AUSTROADS 2004 — 3.62 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 270MPa



FWD maximum deflections (mm)



350MPa



230MPa



1.3



310MPa



350MPa 350MPa



350MPa



350MPa



350MPa



350MPa



350MPa



1.2 350MPa



1.1 150MPa



170MPa



150MPa



210MPa



1.0



250MPa



290MPa



330MPa



350MPa



350MPa



350MPa



350MPa



350MPa



350MPa 350MPa



0.9



150MPa



190MPa



230MPa 270MPa



310MPa



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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



Relative Granular Modulus



0.9



0.8



0.7



0.6



Based on Dawson (1999) y = 7.40E-08x2 - 4.90E-04x + 1.31



0.5



June 2003 revision Table 6.4 y = -3.804E-04x + 1.377



0.4



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



Relative Granular Moduli



0.9



0.8



0.7



0.6



0.5



0.4 0



500



1000



1500



2000



2500



3000



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 —



3500



Technical Basis of Austroads Pavement Design Guide: Part 3



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 Modulus of Cover Material (MPa)



Thickness of Overlying Material



1000



2000



3000



4000



5000



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



200 mm



220



150



150



150



150



225 mm



180



150



150



150



150



>=250 mm



150



150



150



150



150



150



1. Cover material is either asphalt or cemented material or a combination of these materials.



AUSTROADS 2004 — 3.65 —



Technical Basis of Austroads Pavement Design Guide: Part 3



June 2003 Table 6.4(b) Suggested Vertical Modulus (MPa) of Top Sub-Layer of High-Standard Base Material Modulus of Cover1 Material (MPa)



Thickness of Overlying Material



1000



2000



3000



4000



5000



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



1. Cover material is either asphalt or cemented material or a combination of these materials.



50mm,100mm and 150 mm asphalt on 500mm granular on subgrade CBR=3 1.4 Top Granular Moduli 350MPa



FWD maximum deflections (mm)



1.3 1.2



350MPa



350MPa



350MPa



350MPa



350MPa



350MPa



350MPa



350MPa



350MPa



350MPa



350MPa 350MPa



1.1 1.0 230MPa 250MPa



260MPa



270MPa



280MPa



290MPa



300MPa



310MPa



320MPa



330MPa



280MPa 250MPa



210MPa



0.9



220MPa



340MPa



260MPa



240MPa



200MPa 150MPa



0.8



160MPa



180MPa



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



APPENDIX E



RELATIONSHIP BETWEEN UNCONFINED COMPRESSIVE STRENGTH AND FLEXURAL MODULUS FOR CEMENTED MATERIALS The following text was first published as an internal ARRB Transport Research Working Document WDR98/024 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 = 2240 UCS0.88 + 1100 for cemented natural gravel where E UCS



= =



(E.1) (E.2)



Elastic Modulus (MPa), and 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 UCS = Unconfined Compressive Strength at 28 days (MPa). adjusted R2 = 0.60 Standard Error of Estimate = 2692



20,000 18,000



Flexural Modulus (28 days curing) (MPa)



16,000 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) Lab-Cement Treated Crushed Concrete (VicRoads) Lab-Cement Treated Class 3 Crushed Rock (VicRoads)



Cement Treated Natural Gravel (Otte) Field-Cement Treated Cruched 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



EFLEX = 3,000 UCS where EFLEX UCS



E.5



= =



(E.5)



Flexural Modulus at 28 days moist curing (MPa), and Unconfined Compressive Strength at 28 days (MPa).



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 where EFLEX UCS



= =



for cemented crushed rock and cemented natural gravel



Flexural Modulus at 28 days moist curing (MPa), and Unconfined Compressive Strength at 28 days (MPa).



AUSTROADS 2004 — 3.70 —



Technical Basis of Austroads Pavement Design Guide: Part 3



REFERENCES Austroads (1992). Pavement Design – A Guide to the Structural Design of Road Pavements. (Austroads: 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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) where Eflex UCS



= =



flexural modulus of field beams at 28 days (MPa), and unconfined compressive strength of laboratory specimens at 28 days (MPa).



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) where Eflex = flexural modulus of field beams at 28 days (MPa), and UCS = unconfined compressive strength of laboratory specimens at 7 days (MPa). 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



APPENDIX G



DEVELOPMENT OF RELATIONSHIP TO ADJUST ITT MODULUS FROM TEST LOADING RATE TO OPERATING SPEED INSERVICE 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. log



Ef = 0.365(log(f ) − 1) E10



(G.3)



• 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 ITT Modulus This relationship is plotted in Figure G.1.



AUSTROADS 2004 — 3.75 —



(G.4)



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table G.1 VicRoads Flexural and ITT Modulus Data at 25°C



Mix Type 14H320



14N170



14V320



20T320



20T600



Flex 10Hz 3497 2043 4446



1826 4170



2672



2922 1866 4021



3960 2096 4750



air voids 5.9 6.7 6.5



6 6.3



7.6



6.1 6 7.2



4.8 6.1 7.5



Modulus Adjustment Adj to 7% AV 0.93 0.98 0.97 Mean



Adjusted Flex Modulus 3242 2000 4293 3178



ITT (MPa) 4857 3205 3271



0.93 0.95 Mean



1704 3971 2838



2664 2895



1.04



2792 2792



3551



0.94 0.93 1.01 Mean



2746 1742 4079 2855



3585 3038 3799



0.86 0.94 1.04 Mean



3422 1969 4926 3439



7545 3374 4134



air voids 7.8 8.5 6.6



6.8 7.1



9.1



8.7 7.2 8.1



8.8 7.8 7



Modulus Adjustment Adj to 7% AV 1.06 1.12 0.97 Mean



ITT Adj to 7% AV 5151 3590 3180 3974



0.99 1.01 Mean



2626 2916 2771



0.98



1.18



4178 4178



1.50



1.14 1.01 1.09 Mean



4080 3082 4123 3762



1.32



1.15 1.06 1.00 Mean



8658 3578 4134 5457



1.59



Mean Ratio of ITT/Flex modulus



AUSTROADS 2004 — 3.77 —



ITT/Flex



1.25



1.33



Technical Basis of Austroads Pavement Design Guide: Part 3



1.0



0.9



0.8



y = 0.1714x



0.365



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



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 —



90



100



Technical Basis of Austroads Pavement Design Guide: Part 3



APPENDIX H DEVELOPMENT OF RELATIONSHIP TO ADJUST ITT MODULUS 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 Modulus at 25o C



= exp( −0.08(WMAPT − 25)



(H.1)



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.



Asphalt Stiffness (MPa)



10000



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table H.1 Estimated Asphalt Moduli Used to Develop Austroads Overlay Design Procedures Asphalt Temperature (°C) 12.5 15.0 17.5 20.0 22.5 25 27.5 30.0 32.5 35.0 37.5 42.5



Asphalt Moduli (MPa) 7400 6350 5300 4420 3550 2890 2225 1850 1475 1310 1150 1000



Moduli Relative to 25°C 2.56 2.20 1.83 1.53 1.23 1.00 0.77 0.64 0.51 0.45 0.40 0.35



2.8 2.6 2.4



Moduli relative to moduli at 25C



2.2 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



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 —



40



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



APPENDIX I



DEVELOPMENT OF RELATIONSHIP TO ADJUST ITT MODULUS 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 inservice, 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: (21 - AVIn-Service ) Modulus at In − Service Air Voids = Modulus at Test Air Voids (21 − AVtest )



(I.1)



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.



Modulus at In Service AV / Modulus at Test AV



1.3



1.2



1.1



1.0



0.9



0.8



0.7 2



3



4



5



6



7



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 —



8



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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



1.



Triaxle



Mean and 95% Confidence Limits of SARc/AG



33.1



25.0



5.5 ± 1.1



12.6



30.3



20.4



7.2 ± 2.0



36.0



10.5



30.1



22.5



8.5 ± 8.5



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



Road Functional Class



Number of Sites1



1



45



33.5



7.8



2



29



35.5



3



4



6



Mean Percentage of Axle Group Type Single Axle Single Axle Tandem Single Wheel Dual Wheels Axle



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 heavilytrafficked 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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Class 1 12 ALL DATA



Mean



10



Mean = 6.0 Std Dev = 4.9 n = 46 95% min = 4.5 95% max = 7.4



8 Frequency



EXCLUDING OUTLIER Mean = 5.5 Std Dev = 3.6 n = 45 95% min = 4.4 95% max = 6.6



6



4



2



0 0



5



10



15



20



25



30



SARs/AG (Cemented Materials)



Figure J.1



Class 2 12



Mean



10



Frequency



8



Excluding Outliers Mean = 7.2 Std Dev = 5.4 n = 29 95% min = 5.2 95% max = 9.3



ALL DATA Mean = 15.1 Std Dev = 28.2 n = 32 95% min = 4.9 95% max = 25.2



6



4



2



0 0



10



20



30



40



50



60



70



80



SARs/AG (Cemented Materials)



Figure J.2



AUSTROADS 2004 — 3.85 —



90



100



110



120



130



Technical Basis of Austroads Pavement Design Guide: Part 3



Class 3 2



Frequency



Mean



All Data Mean = 15.0 Std Dev = 15.4 n=5 95% min = 0 95% max = 34



Excluding Outlier Mean = 8.5 Std Dev = 5.7 n=4 95% min = 0 95% max = 17.6



1



0 0



5



10



15



20



25



30



35



40



45



50



SARs/AG (Cemented Materials)



Figure J.3



Class 6 3 ALL DATA



Mean



Frequency



2



EXCLUDING OUTLIERS



Mean = 43.6 Std Dev = 80.9 n = 20 95% min = 5.8 95% max = 81.5



Mean = 20.9 Std Dev = 20.3 n = 18 95% min =10.8 95% max = 31.0



1



0 0



50



100



150



200



250



SARs/AG (Cemented Materials)



Figure J.4



AUSTROADS 2004 — 3.86 —



300



350



400



Technical Basis of Austroads Pavement Design Guide: Part 3



Class 7 3 ALL DATA



Mean



Mean = 53.1 Std Dev = 65.6 n = 11 95% min = 7.9 95% max = 98.4



Frequency



2



1



0 0



10



20



30



40



50



60



70



80



90



100



110



120



130



140



150



160



170



180



SARs/AG (Cemented Materials)



Figure J.5



Class 8 2



Frequency



Mean



All Data Mean = 12.7 Std Dev = 16.8 n=4 t =3.18 95% min = 0 95% max = 39.5



Excluding Outlier Mean = 4.4 Std Dev = 2.6 n= 3 t =4.3 95% min = 0 95% max = 10.8



1



0 0



5



10



15



20



25



SARs/AG (Cemented Materials)



Figure J.6



AUSTROADS 2004 — 3.87 —



30



35



40



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



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 —



Technical Basis of Austroads Pavement Design Guide: Part 3



Table J.2: Presumptive Traffic Load Distribution for Rural Roads Axle Group Load (kN) 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 Total



SAST % 0.08 3.97 8.20 11.34 31.51 33.87 9.85 1.10 0.08 0.01



Axle Group Load SADT TAST % % 1.25 7.30 17.41 18.44 17.80 12.70 9.57 6.78 4.45 2.74 0.96 0.37 0.15 0.05 0.01 0.00



0.01 0.09 0.29 0.97 4.05 10.30 15.68 15.92 14.34 15.11 10.94 6.29 3.20 1.71 0.62 0.22 0.10 0.09 0.05 0.03



TADT %



TRDT %



0.15 0.56 0.97 1.97 3.73 4.98 5.66 5.85 5.87 6.53 6.38 6.89 7.44 8.31 10.37 9.13 7.05 4.18 2.08 1.07 0.44 0.21 0.11 0.05 0.03 0.01 0.00 0.00 0.00



0.03 0.15 0.52 1.57 3.30 3.83 4.32 4.41 4.16 4.15 3.43 3.34 3.52 3.92 4.97 5.23 6.26 7.44 8.40 9.00 6.63 4.76 2.95 1.66 0.97 0.44 0.24 0.13 0.08 0.07 0.05 0.02 0.02 0.01 0.00 0.00 0.00



100.00



100.00



100.00



100.00



100.00



Proportion of 0.344 Each Axle Group



0.098



0.007



0.320



0.230



AUSTROADS 2004 — 3.90 —



NHVAGs =



Material Type Component Granular Pavements (4th Power)



2.9



SARs/ AG



1.0



Asphalt (5th Power)



1.2



Subgrade (7th Power)



1.7



Cemented mater (12th Power)



7.1



Technical Basis of Austroads Pavement Design Guide: Part 3



Table J.3: Presumptive Traffic Load Distribution for Urban Roads



Axle Group Load (kN) 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 Total



Axle Group Type SAST % 0.26 13.53 18.02 19.99 25.74 17.11 4.37 0.77 0.12 0.06 0.01 0.01 0.01 0.01



SADT % 2.18 10.23 20.67 17.99 13.42 8.30 6.27 7.68 6.37 3.58 1.68 0.92 0.44 0.19 0.05 0.02 0.00 0.01 0.00



TAST % 0.10 0.96 1.26 1.33 4.52 13.66 17.95 17.36 13.23 9.92 9.77 4.66 2.33 1.19 0.87 0.33 0.31 0.13 0.10 0.03



TADT % 0.10 0.68 1.41 3.76 7.73 10.32 10.22 8.56 6.76 5.34 4.38 4.15 4.29 4.71 6.15 5.71 4.97 3.40 2.64 1.70 1.19 0.83 0.42 0.21 0.16 0.08 0.08 0.01 0.01 0.00 0.01 0.00



TRDT % 0.00 0.11 0.25 1.04 4.92 9.44 9.79 8.62 6.53 4.35 3.12 2.70 2.47 2.65 3.09 3.42 3.81 4.94 6.24 7.22 5.24 3.70 2.02 1.45 0.90 0.60 0.62 0.31 0.20 0.16 0.04 0.03 0.03 0.02 0.00 0.00 0.00 0.00



100.00



100.00



100.00



100.00



100.00



Proportion of 0.380 Each Axle Group



0.217



0.022



0.259



0.122



AUSTROADS 2004 — 3.91 —



NHVAGs =



Material Type Component Granular Pavements (4th Power)



2.6



SARs/ AG



0.80



Asphalt (5th Power)



0.96



Subgrade (7th Power)



1.7



Cemented materials (12th Power)



19



Technical Basis of Austroads Pavement Design Guide: Part 3



Table J.4: Representative Distribution of Loads on Axle Groups for Use on Lightly-Trafficked Urban Roads



p Axle Group Load (kN) 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 Total



g



y



Axle Group Type SAST % 14.90 31.46 17.82 10.95 5.115 12.44 5.515 1.319 0.2903 0.1452 0.0484



SADT % 20.47 24.74 9.148 7.730 6.623 8.403 3.022 4.191 5.508 4.177 3.080 2.200 0.5231 0.1916



TADT % 0.00 1.460 8.957 4.380 1.202 1.804 1.202 7.644 11.77 12.11 12.37 14.69 5.926 4.724 1.460 1.804 0.8589 1.460 2.663 1.202 0.6012 1.202 0.0000 0.2577 0.2577



TRDT % 0.00 0.00 4.790 0.00 0.00 0.00 0.00 0.00 4.790 9.481 0.00 4.790 14.27 18.96 9.481 4.790 0.00 0.00 0.00 9.481 0.00 0.00 4.790 0.00 0.00 0.00 4.790 0.00 4.790 0.00 0.00 4.790



100.00



100.00



100.00



100.00



Proportion of 0.4909 Each Axle Group



0.4594



0.0445



0.0028



AUSTROADS 2004 — 3.92 —



NHVAGs =



2.0



ESA/HVAG =



0.6



ESA/HV =



1.2



Damage Type



SARs/ ESA



Asphalt Fatigue (5th Power)



1.3



Rutting and Shape Loss (7th Power)



2.3



Cemented Materials Fatigue (12th Power)



21



Technical Basis of Austroads Pavement Design Guide: Part 3



Table J.5: Representative Distribution of Loads on Axle Groups for Use on Urban Roads Axle Group Load (kN) 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 Total



Axle Group Type SAST % 0.27 7.83 15.46 15.71 29.94 23.29 6.50 0.79 0.11 0.04 0.02 0.02 0.02



SADT % 3.47 8.70 23.46 21.93 16.80 9.61 6.50 4.62 2.97 1.39 0.41 0.12 0.02



TAST % 0.03 0.24 0.28 0.58 2.89 10.27 16.81 16.61 15.95 14.42 9.77 5.90 2.94 1.54 0.84 0.43 0.23 0.14 0.07 0.06



TADT % 0.15 0.58 0.62 1.98 6.50 9.51 10.94 9.77 7.61 7.24 6.27 5.95 5.88 6.53 8.03 5.72 3.55 1.86 0.85 0.33 0.08 0.03 0.02



TRDT % 0.01 0.13 0.33 1.32 4.17 7.42 9.78 8.34 6.15 5.03 3.70 3.30 3.15 3.36 4.01 4.11 4.82 6.10 7.73 8.43 5.14 2.34 0.78 0.25 0.09 0.01



100.00



100.00



100.00



100.00



100.00



Proportion of 0.393 Each Axle Group



0.191



0.009



0.259



0.148



AUSTROADS 2004 — 3.93 —



NHVAGs =



2.5



ESA/HVAG =



0.7



ESA/HV =



1.8



Damage Type



SARs/ ESA



Asphalt Fatigue (5th Power)



1.1



Rutting and Shape Loss (7th Power)



1.6



Cemented Materials Fatigue (12th Power)



12



Technical Basis of Austroads Pavement Design Guide: Part 3



Table J.6: Representative Distribution of Loads on Axle Groups for Use on Rural Roads



p Axle Group Load (kN) 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 Total



SAST % 0.50 6.74 9.24 15.30 28.74 32.19 5.65 1.21 0.25 0.08 0.06 0.03



Axle Group Load SADT TAST % % 3.03 0.01 4.97 0.08 6.86 1.29 8.48 1.99 10.23 5.12 14.16 11.48 13.58 15.24 13.29 16.19 11.17 14.59 7.84 15.37 4.13 11.13 2.00 6.40 0.18 0.63 0.07 0.22 0.10 0.09 0.05 0.02



TADT % 0.38 0.57 2.36 4.01 4.26 5.17 4.87 4.85 5.30 6.24 9.31 9.17 10.55 12.92 10.35 3.77 2.57 1.60 0.87 0.43 0.23 0.13 0.06 0.03



TRDT % 0.01 0.12 1.88 3.86 5.21 4.36 4.08 3.84 3.79 3.84 4.12 3.57 6.69 7.89 8.36 8.13 7.38 6.08 4.98 4.42 2.83 1.93 1.14 0.63 0.35 0.22 0.13 0.08 0.04 0.03 0.01



99.99



99.99



100.00



100.00



100.00



Proportion of 0.344 Each Axle Group



0.098



0.007



0.320



0.231



AUSTROADS 2004 — 3.94 —



NHVAGs =



2.8



ESA/HVAG =



0.9



ESA/HV =



2.5



Damage Type



SARs/ ESA



Asphalt Fatigue (5th Power)



1.1



Rutting and Shape Loss (7th Power)



1.6



Cemented Materials Fatigue (12th Power)



12



Technical Basis of Austroads Pavement Design Guide: Part 4



Technical Basis of Austroads Pavement Design Guide 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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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



A select material zone under the subbase assists in providing uniform support to the pavement, especially in cut and fill transition areas



4-3



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



Figure 5 Figure 6



AUSTROADS 2004 — 4.iii —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 & sealed joints



3.5 to 4.5 m



Figure 1: Typical longitudinal-section of plain concrete pavement (PCP) Steel fibre reinforced concrete is sometimes used for PCP



Steel mesh



8 to 15m



Dowels



Sawn & 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 & sealed joints



up to 5 m



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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



330



Base Thickness (mm)



310



without shoulder



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 Single wheels Dual wheels Dual wheels Tandem (SS) (SD) (TAD)



Three axles Dual wheels Triaxle (TRD)



Figure 7: Plan of the four most common heavy vehicle axle groups in Australia



Single axle Single wheels (SAST)



Single axle Dual wheels (SADT)



Two axles Single wheels (TAST)



Three axles Dual wheels Triaxle (TRDT)



Two axles Dual wheels Tandem (TADT) Four axles Dual wheels Quad (QADT)



Figure 8: Plan of the six heavy vehicle axle groups used in the Guide



AUSTROADS 2004 — 4.5 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 Design 1.0 1.1 1.2



Adjusted LSF for Roundabouts 1.3 1.4 1.5



AUSTROADS 2004 — 4.6 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 ⎤ when Sr > 0.55 ⎣ 0.0828 ⎥⎦



log (Nf) = ⎢



⎡ 4.258 ⎤ Nf = ⎢ ⎥ ⎣ S r − 0.4325⎦ where



Sr Se f’cf P LSF F1



⎡ P.LSF ⎤ Se = ⎢ ⎥ 0.944 f ' cf ⎣ 4.45F1 ⎦ = = = = = = = = = = =



(1)



3.268



when 0.45 ≤ Sr ≤ 0.55



(2)



0.94



(3)



equivalent stress (MPa); design characteristic flexural strength (MPa); axle group load (kN); load safety factor; and 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:



⎛ Se ⎞ ⎛ P ⎞ Sr = ⎜ ⎟ x⎜ ⎟ ⎝ Mr ⎠ ⎝ 18 ⎠



0.94



AUSTROADS 2004 — 4.7 —



(4)



Technical Basis of Austroads Pavement Design Guide: Part 4



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:



⎡⎛ P.L ⎞ 2 10 F3 ⎤ SF ⎟⎟ log (F2 Ne) = 14.524 - 6.777 ⎢⎜⎜ − 9 .0 ⎥ ⎢⎣⎝ 4.45 F1 ⎠ 41.354 ⎥⎦ where



0.103



(5)



P, LSF and F1 are similar to previous definitions with the exception of F1 for QUAD axle groups; = adjustment for slab edge effects; F2 = 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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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.



Number of Allowable Repetitions



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



SAST



SADT



1.0E+00 0



100



200 Axle Group



300



400



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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 pullout 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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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 —



Technical Basis of Austroads Pavement Design Guide: Part 4



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