Publications

While heavy vehicle loads will provide benefits once they reach their destination, some axle configurations applied to local roads will contribute disproportionately large adverse effects on the road pavements en-route. Unless the affected pavements have been designed to carry the extra loading, heavily loaded axles will inflict an exceptionally large proportion of additional wear on local pavements. The usual result is premature distress to a terminal condition which necessitates full structural rehabilitation or reconstruction rather than periodic resurfacing and maintenance.

Pavement deflection testing is undertaken as the primary means for establishing structural parameters. Part I of this set of two reports discusses “Network Level” deflection testing for asset management purposes. This report, Part II, addresses “Project Level” testing and interpretation for rehabilitation treatment of specific road lengths, or quality control during construction. Typical parameters established from deflection testing on New Zealand roads, including the NZ Transport Agency’s (NZTA) Long Term Pavement Performance (LTPP) benchmark sites, are presented.

This document (Part I of a set of two reports) provides a good practice guide for road controlling authorities about “Network Level” pavement deflection testing for Asset Management. Part II of this report series addresses “Project Level” testing and interpretation for specific road lengths’ rehabilitation treatment, or their quality control during construction. Typical structural parameters from past and on-going studies on New Zealand roads, including the NZ Transport Agency’s (NZTA) Long Term Pavement Performance (LTPP) benchmark sites, are presented.This document (Part I of a set of two reports) provides a good practice guide for road controlling authorities about “Network Level” pavement deflection testing for Asset Management. Part II of this report series addresses “Project Level” testing and interpretation for specific road lengths’ rehabilitation treatment, or their quality control during construction. Typical structural parameters from past and on-going studies on New Zealand roads, including the NZ Transport Agency’s (NZTA) Long Term Pavement Performance (LTPP) benchmark sites, are presented.

The stabilisation of near-surface granular pavement materials is accepted practice in transportation maintenance and capital development projects in Australasia. Stabilisation in this context involves the mechanical introduction of reactive agents, including cement and foamed bitumen, into existing or manufactured granular materials, with or without existing seal inclusion. Present-day design guides characterise stabilised granular materials as either modified or bound, depending primarily on the amount and type of reactive agent used in the stabilising process. Modified materials are modelled as unbound granular materials in a pavement. Bound (cemented) materials are modelled as layers with tensile load-carrying capacity within the pavement. Cracking in the bound pavement layer is governed by ‘fatigue relationships’. This research has shown that stabilisation with smaller reactive agent contents (<3% by dry mass) can deliver materials that should be modelled as lightly bound, delivering cost-effective pavement solutions. This research report describes the collection and interrogation of performance data from New Zealand road pavements that utilised stabilised granular materials. The research, carried out from December 2009 to August 2011, compared actual stabilised pavement performance with the expectations in published design guidelines. A conceptual pavement performance model for near-surface ‘lightly bound’ stabilised granular pavement layers that better matches observed pavement behaviour is proposed.

Premature distress in unbound basecourses has occurred regularly in New Zealand. In 2008, the New Zealand Transport Agency (NZTA) commissioned the assembly of an inventory of problem basecourses and subbases. Study of the inventory found that the long-term degree of saturation of basecourse was highly significant in the case histories of premature distress, ie the pavements failed through shear instability (shoving) in the basecourses. A common feature in basecourses with a high degree of saturation was gap grading in the sand fraction. Existing basecourse specifications limit gap grading through grading shape control requirements but the case histories demonstrate that tighter control is required. The basecourse inventory was used to establish regression equations for predicting the in situ long-term degree of saturation of a basecourse. This approach appears to be very promising. Timely decisions can now be made on acceptance or the need for corrective measures prior to sealing. The above considerations have been used for preparing revised drafts of the NZTA basecourse specification, subbase specification notes as well as a set of recommendations for the compaction specification and the New Zealand supplement to the document, Pavement design – a guide to the structural design of road pavements (Austroads 2004) (Transit NZ 2007a) to implement practical solutions to premature distress in unbound basecourses.

Pavement performance modelling for New Zealand roading networks, currently relies on an adjusted structural number (SNP) which is a single parameter intended to describe the performance of a multilayered pavement structure in terms of its rate of deterioration with respect to all structural distress modes, as well as non-structural modes. This parameter had its origin in the AASHO road test in the late 1950s, before the advent of analytical methods. Hence refinement to keep abreast of current practice in pavement engineering is overdue. This research describes the basis for a new set of structural indices and how these can be used to obtain improved prediction of pavement performance: both at network level and for project level rehabilitation of individual roads. The results are (i) effective use of all the data contained in RAMM, (ii) more reliable assignment of network forward work programmes, (iii) reduced cost through targeting only those sections of each road that require treatment and (iv) more efficient design of pavement rehabilitation through informed appreciation of the relevant distress mechanism that will govern the structural life of each individual treatment length.

The Falling Weight Deflectometer (FWD) which measures pavement deflections was assessed for its ability to predict the life of a newly constructed or rehabilitated pavement. FWD measurements used in the study were from NZ Transport Agency’s test track CAPTIF, roads that have failed and from two performance specified maintenance contracts where the actual life from rutting and roughness measurements could be determined. Three different methods to calculate life from FWD measurements were trialled. The first, a simple Austroads method that uses the central deflection only and was found to either grossly over predict life by a factor of 1000 times more than the actual life or grossly under predict the life. The second two methods trialled were based on Austroads mechanistic pavement design where the life is determined from the vertical compressive strain at the top of the subgrade. For the mechanistic approach the FWD measurements are analysed with specialised software that determines a linear elastic model of the pavement which computes the same surface deflections as those measured by the FWD. From the linear elastic model the subgrade strain is determined and life calculated using the Austroads equation. It was found when using this approach that predictions of life from individual FWD measured points within a project length can range from nearly 0 to over 100 million equivalent standard axles. To cater for this large scatter in results, the 10th percentile value was used as the predicted life of the pavement. In general the Austroads mechanistic approach under-predicted the life, sometimes by a factor of 10 or more. The third approach trialled was adjusting the Austroads mechanistic approach by applying a factor determined from past performance to calibrate the subgrade strain criterion to local conditions. This third approach greatly improved the predictions but it was found that the multiplying factor was not consistent for a geographical area and thus the factor found from one project might not be suitable for another similar project.

Pavement performance modelling for New Zealand (NZ) roading networks currently relies on Adjusted Structural Number (SNP); a single parameter intended to describe the performance of a multi-layered pavement structure in terms of its rate of deterioration with respect to all structural distress modes as well as non-structural modes. This parameter had its origin in the AASHO Road Test in the late 1950’s before the advent of analytical methods. Refinement to keep abreast of current practice in pavement engineering is long over-due.

An advanced model for pavement structural capacity is under development overseas (NCHRP), but it is not in general use because of its complexity (including requirements for destructive test information). The focus of a recent NZ Transport Agency (NZTA) research study has been to set the framework for obtaining the most practical indices for New Zealand pavements based on parameters which are currently stored in RAMM, while at the same time maintaining flexibility for ongoing upgrades that might utilise future developments in the way of pavement data collection. In many pavements, structural distress can be assigned to one or more of at least four discrete categories: rutting, roughness, crack initiation and shear instability (shoving). In this study, data from all NZ’s Long Term Pavement Performance (LTPP) sites have been analysed to explore the benefits of replacing SNP with four separate “Structural Indices” - each being determined mechanistically (considering stresses and strains induced by an Equivalent Standard Axle) from data commonly available in RAMM.

Each Structural Index has been developed to fall within the same range as the traditional SNP, allowing straightforward implementation with minimal additional calibration needed to implement these in existing NZ Pavement Deterioration models and asset management software such as dTIMS. This paper presents the new developments resulting from the ongoing study, and shows the basis for the new set of Structural Indices and how these can be used to obtain improved prediction of pavement performance, both at network level and for project level rehabilitation of individual roads. The results enable: (i) effective use of all the data contained in RAMM, (ii) more reliable assignment of network Forward Work Programmes, (iii) reduced cost by targeting only those sections of each road that require treatment, and (iv) more efficient design of pavement rehabilitation through informed appreciation of the relevant distress mechanism that will govern the structural life of each individual treatment length.

Pavement performance modelling for New Zealand (NZ) roading networks currently relies on Adjusted Structural Number (SNP); a single parameter intended to describe the performance of a multi-layered pavement structure in terms of its rate of deterioration with respect to all structural distress modes as well as non-structural modes. This parameter had its origin in the AASHO Road Test in the late 1950’s before the advent of analytical methods. Refinement to keep abreast of current practice in pavement engineering is long over-due. An advanced model for pavement structural capacity is under development overseas (NCHRP), but it is not in general use because of its complexity (including requirements for destructive test information). The focus of a recent NZ Transport Agency (NZTA) research study has been to set the framework for obtaining the most practical indices for New Zealand pavements based on parameters which are currently stored in RAMM, while at the same time maintaining flexibility for ongoing upgrades that might utilise future developments in the way of pavement data collection. In many pavements, structural distress can be assigned to one or more of at least four discrete categories: rutting, roughness, crack initiation and shear instability (shoving). In this study, data from all NZ’s Long Term Pavement Performance (LTPP) sites have been analysed to explore the benefits of replacing SNP with four separate “Structural Indices” - each being determined mechanistically (considering stresses and strains induced by an Equivalent Standard Axle) from data commonly available in RAMM. Each Structural Index has been developed to fall within the same range as the traditional SNP, allowing straightforward implementation with minimal additional calibration needed to implement these in existing NZ Pavement Deterioration models and asset management software such as dTIMS. This paper presents the new developments resulting from the ongoing study, and shows the basis for the new set of Structural Indices and how these can be used to obtain improved prediction of pavement performance, both at network level and for project level rehabilitation of individual roads. The results enable: (i) effective use of all the data contained in RAMM, (ii) more reliable assignment of network Forward Work Programmes, (iii) reduced cost by targeting only those sections of each road that require treatment, and (iv) more efficient design of pavement rehabilitation through informed appreciation of the relevant distress mechanism that will govern the structural life of each individual treatment length.

A roading authority that has pavement assets to manage needs well developed strategies for network maintenance based on desired levels of service, pavement deterioration and budget constraints supported by options for cost effective pavement rehabilitation once a terminal condition is reached. A purely reactive approach to maintenance or rehabilitation can result in both inefficiencies and unexpected costs. This is being further highlighted by the need to consider impacts of the newly increased VDM limits.

In recent years, a wealth of pavement structural data has been collected and stored in RAMM, but it is only used superficially. The bulk of the data is neglected. The mechanistic approach to pavement management uses Austroads principles to analyse the structural data already collected in RAMM, then calculates layer stiffnesses and stresses and strains induced in the pavement structure by trafficking to rationalise and quantify distress modes. This article addresses the role of structural analysis using a mechanistic approach to reduce the cost of maintenance and rehabilitation by effectively identifying the relevant distress mechanism, and focussing treatment only on subsections where work is essential, rather than using a continuous, and sometimes ineffective, treatment over an assumed treatment length. Often only the carriageway interval or top surface type is used for identification of treatment lengths.

In many cases, the existing raw data already stored in the RAMM database can be reanalysed to minimise any further data collection costs. Significant advancements in the application of pavement structural analysis have taken place in 2008 and 2009 as part of Austroads and NZTA research, and these are summarised in this article.

The following aspects are addressed:

• Part I - Network maintenance and determination of Forward Work Programmes
• Part II - Pavement structural rehabilitation
• Part III - Quality assurance of newly constructed pavements and demonstration that the intended design life will be achieved
• Part IV - Maintenance cost implications of increased vehicle weight limits

The conclusions to each of Parts I , II, III & IV highlight the key benefits that can now be obtained from a mechanistic approach to pavement asset management giving cost-effective design, reduced incidence of premature failure, reduced cost for maintenance, and informed budget estimates for future planning.

A roading authority that has pavement assets to manage needs well developed strategies for network maintenance based on desired levels of service, pavement deterioration and budget constraints supported by options for cost effective pavement rehabilitation once a terminal condition is reached. A purely reactive approach to maintenance or rehabilitation can result in both inefficiencies and unexpected costs. This is being further highlighted by the need to consider impacts of the newly increased VDM limits.

In recent years, a wealth of pavement structural data has been collected and stored in RAMM, but it is only used superficially. The bulk of the data is neglected. The mechanistic approach to pavement management uses Austroads principles to analyse the structural data already collected in RAMM, then calculates layer stiffnesses and stresses and strains induced in the pavement structure by trafficking to rationalise and quantify distress modes. This article addresses the role of structural analysis using a mechanistic approach to reduce the cost of maintenance and rehabilitation by effectively identifying the relevant distress mechanism, and focussing treatment only on subsections where work is essential, rather than using a continuous, and sometimes ineffective, treatment over an assumed treatment length. Often only the carriageway interval or top surface type is used for identification of treatment lengths.

In many cases, the existing raw data already stored in the RAMM database can be reanalysed to minimise any further data collection costs. Significant advancements in the application of pavement structural analysis have taken place in 2008 and 2009 as part of Austroads and NZTA research, and these are summarised in this article. The following aspects are addressed:

    Part I - Network maintenance and determination of Forward Work Programmes
    Part II - Pavement structural rehabilitation
    Part III - Quality assurance of newly constructed pavements and demonstration that the intended design life will be achieved
    Part IV - Maintenance cost implications of increased vehicle weight limits

The conclusions to each of Parts I , II, III & IV highlight the key benefits that can now be obtained from a mechanistic approach to pavement asset management giving cost-effective design, reduced incidence of premature failure, reduced cost for maintenance, and informed budget estimates for future planning.

Pavement maintenance specifications are increasingly calling for “Operational Performance Measures” for rehabilitation treatments. One of these requires that the treatment is constructed so that post-construction materials testing and FWD testing confirm achievement of the design life. Often only post-construction testing is available and the procedures which maintenance contractors should follow to demonstrate achievement, are less well defined. This article is intended as a discussion document, and gives some alternative methods, depending on test data available, with suitable rationale for application. Rehabilitation treatments are addressed primarily. In some cases, new pavement construction can be assessed similarly but with limitations.

Frost damage is a significant factor in the premature demise of unbound granular pavements in Central Otago. During the 1970s and 1980s, extensive lengths of new and rehabilitated pavements were constructed in the Cromwell Gorge and Lindis Pass (where frost penetration to 600 mm was noted). Experience obtained in these projects was documented at the time under research grants addressing both construction in schist and the frost resistance of pavement materials. The work was carried out for the National Roads Board to provide guidelines for the frost resistant design and construction of unbound pavements in similar terrain where schist and glacial deposits are the principal soil types. This update is intended for local practitioners involved with design and/or construction supervision and quality assurance of pavements in frost prone regions.

When pavement structural evaluation is carried out using the Falling Weight Deflectometer, the full displacement-time history for each geophone is measured in the field, but for normal production only the peak deflection data are stored for subsequent processing. Velocity and acceleration information can however be readily captured. A preliminary study has been made to examine whether useful time-dependent parameters could be extracted from the full time history and used by practitioners. Software has been developed for capturing wave speeds and correlating with approximate moduli of the surface layer, to complement conventional back-analyses. Surface wave speeds appear promising indicators to allow rapid distinction of near surface material types and structural condition, at no additional cost or time.

Unbound granular pavement layers show a characteristic increase in moduli from the subgrade to the surface. The ratio of moduli between successive layers provides a direct and effective measure of pavement quality that makes due allowance for site conditions and seasonal effects beyond a contractor’s control. The modular ratio is therefore an effective indicator for performance-based specifications. Modular ratios are readily determined from back-analyses of deflection monitoring. Their applications to new construction quality assurance and performance-based network maintenance contracts are described.

Realistic prediction of pavement performance is a critical component of asset management. Performance in terms of structural capacity is generally measured by the Adjusted Structural Number (SNP), hence a review has been carried out testing alternative methods for deriving this parameter for unbound granular pavements. Simplified methods may work reasonably well when calibrated to typical local conditions, but they are less likely to provide the same reliability in different regions or with different pavement structures. The rigorous methods (which can be readily applied) are recommended wherever deflection bowl information has been recorded. However, it is important to note that the standard equations are based on isotropic moduli for each layer, rather than anisotropic moduli, which the Austroads Pavement Design Guide has adopted for mechanistic analysis. Typical field measured moduli for unbound granular pavements are presented and appropriate calculation procedures suggested.

Calibration or correction of SNP may also be required in a range of circumstances when refining the structural capacity for use in long-term pavement performance prediction. Where representative benchmark sites are set up and actual performance is monitored, appropriate calibration factors, based on specific modes of distress can be developed. These result in effective SNP values for a given network, giving due regard to the local range of materials and construction techniques.

The AUSTROADS Pavement Design Guide has been used for the mechanistic design of pavement rehabilitation in New Zealand for several years. A number of projects have now been completed where pavement testing has been carried out before and after construction, giving the opportunity to compare design prediction with the in situ performance achieved by the rehabilitation. Case histories involving different forms of construction in various parts of New Zealand are presented, quantifying the improvement in performance resulting from the use of AUSTROADS procedures. Data is also presented for typical in situ material properties of unbound granular overlays and cement based stabilised basecourses.

Overlay design for pavement rehabilitation in New Zealand is based on mechanistic procedures given in the AUSTROADS Pavement Design Guide. In July 1997, Transit New Zealand issued a Supplement to the AUSTROADS Guide, promoting some variation to the existing methods to give due regard to the past performance of any road section programmed for rehabilitation.

The New Zealand Supplement overlay design methods have now been applied to many projects on state highways and local authority roads. This article presents the results from a number of case histories for unbound pavements where overlay requirements have been calculated using both the original AUSTROADS design methods and the procedures set out in the New Zealand Supplement. Comparisons show the strengths and limitations of the various methods, providing a guide to the most useful conditions under which each method should be applied to five both cost effective design and assurance of long term performance.

A mechanistic design procedure has been adopted by Transit New Zealand for designing rehabilitation treatments for New Zealand roads. A computer program such as CIRCLY (Wardle 1980) is sued to analyse the reaction of various pavement rehabilitation designs (modelled as multiple layers of linear elastic materials) under a standard wheel load. Other programs such as ELMOD include allowance for nonlinear elastic material. Strains within various critical layers are computed for each rehabilitation design being considered. The designs which are acceptable are those which meet or exceed specific performance criteria for asphalt, cemented bases and subgrade layers. Mechanistic design has the advantage of allowing the design of a range of rehabilitation treatments including: strengthening the existing pavement layers (stabilisation or other means); granular overlay; asphalt overlay; or any combination of these.

Knowledge of the structural condition of a pavement is required to make effective decisions on the type of maintenance or rehabilitation to be carried out on a pavement section. Reliable quantification of the extent of necessary remedial work will ensure the design life is achieved with maximum benefit/cost ratio.

The Falling Weight Deflectometer determines the full dynamic deflection bowl beneath a standard wheel load. Using the associated analysis software, it is possible to quickly and accurately determine the structural condition of the pavement system. The remaining life and appropriate rehabilitation requirements of a pavement section are determined from calculated stresses and strains in each layer. Thus, by using non-linear layered elastic theory, a pavement structure is analysed in the same way as other civil engineering structures. From non-destructive FWD testing, the required overlay or other rehabilitation alternatives are quantified from soundly based principles, producing maximum pavement life at minimum cost.

FWD data are combined with environmental data, layer thicknesses, material response functions and traffic load information to determine remaining life, critical layer, failure mode and required treatment (if any), at each FWD test point. Determination of structurally uniform subsections is based on a statistical evaluation of remaining life for all test points along the design section of pavement.

The performance of unbound granular basecourse surfaced with chip seal has been studied in a field situation using simulated traffic loads. Test variables included subgrade deflection, aggregate type, grading and shape.

Pavement distress due to basecourse instability may be attributed fundamentally to thickness change due to densification and/or lateral shear movement of material beneath a wheel path. Contrary to reports involving laboratory models, the field trials found that lateral movement provided no measurable contribution to rut depth unless either high saturation or excessive pavement deflection had occurred. The most significant factor affecting the stability of basecourses was densification at constant moisture content resulting in a high degree of saturation (over 80%). This produced progressive rutting and lateral shear. The character of highly saturated basecourses was deceptive: excavated specimens were strongly cohesive and appeared quite “dry”. (Testing showed that moisture contents as low as 3.5% give excessive saturation in some basecourses after less than 10% of design traffic loading). The most adequate means of predicting susceptibility of stockpile material to excessive densification is examination of the shape of the particle graduation curve. The sand equivalent test, which reputedly determines the presence of undesirable fine material in aggregates, was found to correlate poorly with basecourse performance. Guidelines are offered for the practitioner wishing to utilise local materials, which do not comply with current NRB requirements for basecourse or subbase.

Unbound granular pavement layers show a characteristic increase in moduli from the subgrade to the surface. The ratio of moduli between successive layers provides a direct and effective measure of pavement quality that makes due allowance for site conditions and seasonal effects beyond a contractor’s control. The modular ratio is therefore an effective indicator for performance-based specifications. Modular ratios are readily determined from backanalysis of Falling Weight Deflectometer bowls recorded during construction monitoring or long-term structural condition studies. Their applications to new construction quality assurance and performance-based network maintenance contracts are described.